Wiley Cytometry Web Site - ISAC 2000 International Congress
ISAC Archive
ISAC Meeting Program and Abstracts Archive

ISAC logo ISAC XX International Congress

Plenary and Parallel Session Abstracts
        - Go to Program
        - Go to Poster Abstracts


Saturday, May 20, 2000


8:00 am – 10:00 am


Scientific Tutorial I. Flow Analysis: Principles and Instruments/Highspeed Sorting in Theory and Practice

Tutors: David M. Coder and Ger van den Engh

Flow cells, filter selection, PMT operation and options, etc. Sorting rare events, high speed sorting, sorting tumor cells for molecular analyses, including PCR and arrays

 


Scientific Tutorial II. Cell Cycle Proteins

Tutors: James Jacobberger and Frank N. Traganos

Principles of the cell cycle engine; cyclin/cdk/cdki; detection by flow cytometry; subcellular location by imaging, immunoprecipitation, protein complexes, post-translational modification, inhibitors; problems with antibody specificity and controls.


10:30 am – 12:30 pm


Scientific Tutorial III. Multicolor FISH and Comparative Genomic Hybridization

Tutor: Yuval Garini

Multicolor FISH and Comparative Genomic Hybridization (CGH) represents important applications of advanced imaging techniques in the life sciences. Unlike single color (or up to three colors) imaging methods, multicolor (or spectral) information that exist in images, cannot be fully analyzed by the human eye, because of limited color resolution. New developments in imaging and spectral techniques enable researchers to acquire and analyze multiple color images. This opens the opportunity to detect in a single image multiple genetic (or other) probes with large intensity variations among them (high dynamic range). When combined with optimized fluorochromes and molecular labeling methods, a large variety of applications opens up.

  The tutorial will include explanation of the fundamentals of multicolor techniques and CGH, specifically:

  1.Optical instrumentation for multicolor FISH data acquisition.

  2. Image analysis methods for CGH.

  3. Signal and image processing algorithms for analyzing multicolor images.

  4. Methods for displaying the multi-parameter results found in multicolor FISH and CGH.

  5. Formalism for analyzing the signal to noise (S/N) of multicolor measurements.

  6. The future of multicolor FISH techniques.

 


Scientific Tutorial IV. Live Cell Cytometry: Intracellular Environments, Functional Assays, and Cell Signaling

Tutors: Paul J. Smith and Rachel J. Errington

Review of the use of probes for subcellular compartment analysis, detection of cell signaling, dynamics of probe-target interactions, and the tracking of biological responses in rapid and protracted time frames.

 


Scientific Tutorial V. Multiparameter DNA Measurement of Clinical Samples

Tutors: W. E. Corver, Alberto Orfao de Matos, Michael G. Ormerod

The tutorial will cover:

  1. Methods of samples preparation (for biopsies, fine needle aspirates, paraffin blocks).

  2. Signal processing.

  3. Use of light scatter.

  4. Using antibodies to define tumor cell populations.

  5. Practical clinical applications.


1:30 pm – 3:30 pm


Scientific Tutorial VI. High Density Array Technologies

Tutors: Donna Albertson and Dan Pinkel

This tutorial will present an introduction to high density array methods for measurement of mRNA expression and DNA copy number. Since this is a young, rapidly developing field, the presentations are intended to introduce the fundamental concepts, and to provide an overview of several of the main approaches. Specific topics to be covered include: fundamental aspects of single and multicolor methods, choice of DNA sequences for array spots, critical features of robotic printing systems, binding of DNA to array substrates, fluorescence imaging systems and analytical software, and data interpretation software. Contact information for the various manufacturers currently offering instrumentation and reagents for microarray analysis will also be provided.

 


Scientific Tutorial VII. Cell Death

Tutors: Zbigniew Darzynkiewicz, Michael G. Ormerod, Patrice X. Petit

Introduction. Why is apoptosis (programmed cell death) important? A brief description of the properties of apoptotic cells, including morphology and DNA gel electrophoresis. How flow cytometry can be used to (i) study components the apoptotic cascade and (ii) to quantify apoptotic cells.

Functional assays for studying apoptosis (including the role of mitochondria in the apoptotic process). Mitochondrial membrane potential, pH, calcium ions, oxidative species.

Methods for quantification of apoptotic cells in cell cultures. Annexin V binding. Assays of DNA breakdown. Simultaneous measurement of cell proliferation and apoptosis. Membrane permeability assays. Light scatter.

Clinical studies. Monitoring treatment in leukemias.

 


Scientific Tutorial VIII. Multiparameter Flow Cytometry of Clinical Samples-Including 4-color Immunofluorescence

Tutors: Carleton C. Stewart and Sigrid J. Stewart

This tutorial will describe the procedures for evaluating antibody quality, antibody titer, verifying antibody combinations and processing cells. Instrumentation quality assurance procedures, electronic or software compensation procedures for up to 4 colors will be covered. The unique problems and their solutions for using PE-CY5 conjugated antibodies will be described. The participant should expect to come away with an understanding of the proper methods for performing multiparameter immunophenotyping and the scientific basis for performing each process.


3:30 pm – 5:30 pm


Scientific Tutorial IX. Genotyping of Single Nucleotide Polymorphisms

Tutor: John P. Nolan

The human genome project is poised to provide the first complete human DNA sequence. Because differences in DNA sequence confer phenotypic differences among individuals, genetic variation will be the subject of intense research in the coming decades. The most common type of genetic variation is the single nucleotide polymorphism (or SNP). Biomedical science in the 21st century will use SNPs to identify disease-related genes, diagnose diseases, and target treatments. In addition, SNPs will have wide applications in agriculture, microbiology, and forensics. Multiplexed SNP analysis using flow cytometry offers a rapid and cost effective means to perform SNP-based genotyping in the research or clinical lab on many different sample and throughput scales. This tutorial will cover the nature and uses of SNPs, general features of genomic analysis using flow cytometry, and the development and validation of microsphere-based flow cytometric genotyping assays.

 


Scientific Tutorial X. Practical Confocal Microscopy

Tutors: Robert M. Zucker and Nicholas Terry

The laser scanning confocal microscope (CLSM) has considerable ability to visualize structures at depth within cells, tissues and organisms. Many applications complement analytical flow cytometry by providing spatial information. This tutorial will discuss the basics of confocal microscopy, optical design, fluorescence, filters, sample preparation, and data archiving and presentation in 2D and 3D. The new Leica TCS-SP and Zeiss 510 confocal microscopy systems will be used to illustrate the features of modern CLSM systems. The system variables that affect the accuracy and precision of CLSM image capture will be discussed. Various test methods to insure the instrument is delivering optimum performance will be presented. For optimal images it is important to control the variables of laser power, PMT tube voltage, signal averaging and photo-bleaching. The relation hips between PMT voltage, laser power, and signal averaging, determined with a 10 micron Spherotech test bead will be demonstrated.

Specific applications of CLSM in the areas of morphometry, apoptosis and cell kinetics in both tissue culture cells and thick tissues (murine embryos, human colo-rectal crypts), will be presented to illustrate the power and flexibility of the technique. Attempts to quantify fluorescence from thick tissue (fetal limb, fetal eye, embryo) will also be discussed.

 


Scientific Tutorial XI. Bugs in the Beam-Cytometric Measurement of Microorganisms

Tutors: Susann Müller, Philippe Lebaron, Gerhard Nebe-von Caron

Following the growing popularity of microbial cytometry and the interest for the subject at the last conference this tutorial will cover technical requirements and difficulties as well as practical applications. It is aimed at cytometrists who would like to:

Open up new fields of research (and funding) or be environmentally challenged or push their instrument towards its limit - and push the microbiologists out of their Petri-dish. Technical development in microbial cytometry has gone a long way from the first manual image and flow cytometers by Leeuwenhook and Tyndall. Improvements in instrument sensitivity and fluorochrome intensity put the measurements of microorganisms into everyone’s reach. Apart from a technical background, the tutorial will cover methods for functional characterization of bacteria and discuss non-culturability and its problems and misconceptions.

The practical part will include methods for bacterial process control which are now well established in industry and give special considerations to methods applied in the field of environmental microbiology with its particular problems.

As at the last conference, there will be a number of people present with experience in the field of microbial cytometry and a variety of instruments, therefore giving interested people the opportunity for informal but informative discussion and networking.


5:30 pm – 7:30 pm


Scientific Tutorial XII. Green Fluorescent Protein

Tutors: David Galbraith and Mike Anderson

This tutorial will cover the use of GFP in both flow and image cytometry. Specific recent advances in the use of the Green Fluorescent Protein in mammalian cells will be discussed. These will include:

Simultaneous detection of three distinct GFPs in mammalian cells.

The next generation of dual GFP reporters that are distinct in fluorescence, hybridization and immunodetection.

Recent progress in the development of intrinsically labelled antibodies.

The use of IRES elements linked to GFP for the evolution of proteins through random mutagenesis and FACS selection.

In addition, future directions in new forms of energy transfer will be explored.

Flow cytometric analysis of GFP expression will also be discussed for higher plants and a variety of other life-forms. Recent advances in the use of GFP in image cytometry will include: methods for acquisition and manipulation of confocal images; production of movies; multiphoton applications; subcellular targeting of GFP.

 


Scientific Tutorial XIII. Detection of Minimal Residual Disease

Tutor: James F. Leary

This tutorial will address theoretical and methodological issues regarding detection of minimal residual disease (MRD) in patients with acute leukemia. Theoretical issues will include definition of MRD, rationale for studying MRD, potential clinical applications of MRD studies. Methodological issues will include: specific advantages and disadvantages of molecular and immunological methods for MRD detection, identification of leukemia-associated markers for detecting MRD by flow cytometry at diagnosis, combinations of surface and intracellular antigens that identify leukemic cells, gating and data analysis procedures, and controls. Finally, issues pertinent to the organization of clinical studies of MRD will also be discussed, including optimal time of sampling for MRD studies, use of bone marrow versus peripheral blood, and update of results of clinical studies.

 


Scientific Tutorial XIV. Cytometric Resources

Tutor: David M. Coder

Brief history of cytometry, different systems currently available, cytometry information sources, typical facilities, major facilities, ISAC, affiliated societies, collaborative groups, etc.

This presentation will be free. It is aimed at newcomers to the field. It will be supported during the meeting by a booth at which participants can gain information about ISAC, affiliated societies, bulletin boards, etc.

 


Sunday, May 21, 2000

8:00 am – 9:00 am

Frontiers in Plant Biology


Visualizing Clock Gene Function — Understanding What Makes Us Tick

Steve A. Kay, The Scripps Research Institute

Circadian clocks are likely to have evolved in most organisms to provide an adaptive advantage for life under light/dark cycles. The range of biological processes that have been found to be under circadian control is striking: from photosynthesis in plants and bacteria to cognitive functions in humans. There has recently been intense interest in identifying the molecules responsible for generating timing information, for transducing temporal signals to downstream biological events, and for providing environmental sensing of light signals. We have been interested in identifying these molecules in Drosophila and Arabidopsis using molecular genetic techniques.

In Drosophila we have developed in vivo reporter gene technology that allows us to monitor gene expression and behavior in the same live animal. We have shown that two bHLH-PAS proteins, clock and cycle heterodimerize and activate period and timeless expression via an E-box target sequence in both promoters. Consequently, the PER/TIM heterodimer feeds back into the clock and inhibits the positive activation. We are currently interested in the process by which this inhibition occurs. Light signals for entrainment act through the putative circadian photoreceptor cryptochrome. We have recently demonstrated that CRY can interact with TIM or the PER/TIM heterodimer in a light dependent fashion and inhibit TIM function. We also find CRY colocalizes with the PER/TIM heterodimers in transfected cells. CRY therefore represents a circadian photoreceptor molecule that interacts directly with core clock components. We wish to know the precise photochemical mechanism by which CRY exhibits light-dependent protein complex formation. We are currently investigating output of the circadian clock in Drosophila. By constructing an enhancer trap vector that uses luciferase as a sniffer, we are isolating circadian-regulated genes in the fly. In vivo imaging of the luciferase or GFP expression patterns then provide a key step to elucidating the physiological functions of clock regulated genes in the whole animal.

Finally, we wish to use the information generated from model systems such as Drosophila to investigate clock regulation in humans. As many behavioral pathologies appear to have a circadian component (such as sleep disorder, bipolar disorder and epilepsy), we are developing cell-based assays for circadian regulation from patients. The assays can then be used as a basis for diagnosis, or even for the development of drug screens.


9:00 am – 10:00 am

Frontiers in Marine Science


Flow Cytometry in Oceanography: Two Decades of Discoveries and Progress

Louis Legendre, Laval University

During the last two decades, cell-by-cell analysis contributed to fundamentally change our understanding of marine ecosystems. These are now seen as complex systems of interacting organisms, whose sizes range from <0.1 m viruses to 30 m baleen whales. New groups of increasingly smaller and very abundant marine organisms have been discovered, often using flow cytometry, e.g. free viruses, prochlorophytes, single-celled cyanobacteria, bacterial-sized eucaryotes. New models were proposed for the functioning of pelagic food webs. In parallel, the biogeochemical cycling of carbon became central to biological oceanography because of the role played by oceans in the regulation of atmospheric carbon dioxide and thus global climate.


10:15 am – 11:15 am

Frontiers in Cell Cycle Regulation


On Understanding the Mechanisms by which Extracellular Signals Govern the Choice between Proliferation and Differentiation

Andrew Koff, Memorial Sloan-Kettering Cancer Center

Once every cell cycle, a cell must determine if environmental conditions are appropriate for proliferation, and if so, it must commit irreversibly to S-phase, replicate its DNA and eventually divide to give two equivalent daughter cells. Additionally, the cell must be able to respond to differentiation-inducing signals, which prevent entry into S-phase and allow cells to withdraw from the cell cycle during G1 phase and enter a state wherein they become competent for the transcriptional changes necessitated by acquisition of the differentiation phenotype. The interpretation of these signals at the level of the cell cycle, the choice between proliferation and differentiation, has important ramifications in the development of all tissues and in many diseases, most notably cancer and related diseases of inappropriate cell proliferation and differentiation. In this talk, Dr. Koff will discuss how the regulation of p27Kip1, a major mediator of anti-mitogenic signals to the G1 cyclin-dependent kinases, impacts on this decision in S-phase and G2 phase cells.


1:00 pm – 2:00 pm

Frontiers in Cytometry


Single Molecule Flow Cytometry and Cell Sorting in "Femtofluidic" Chips

Stephen Quake, California Institute of Technology

Single molecule flow cytometry offers the possibility of rapid, ultra-sensitive detection and analysis of biological molecules. We have developed a microfabricated flow cytometry chip as a replacement for analytical pulsed field gel electrophoresis. Assays with these chips are two orders of magnitude faster than pulsed field gels and use a million times less material. Because they are detecting single molecules, their sensitivity is comparable to PCR based techniques. We have also developed a microfabricated fluorescence activated cell sorter and demonstrated its use in screening bacterial cells. In the course of this research we have developed novel valve and pump components for on-chip fluidic manipulation that will be useful in fabricating more complex chip designs.


2:00 pm – 3:00 pm

Frontiers in Oncology


Tyrosine Kinase Display — Discovery and Application in Cancer

Hsing-Jien Kung, University of California-Davis Cancer Center

While a minor class of protein kinases, tyrosine kinases represent a major class of oncogenes. They participate in various cellular functions such as growth, differentiation, apoptosis, migration and chemotaxis. Aberrant expressions or activations of tyrosine kinases are often the underlying causes for the abnormal properties of tumor cells including anchorage and serum-independent growth, angiogenesis, metastasis and resistance to chemo- and radio- therapy. Increasing evidence suggests that tyrosine kinases do not act alone, but cross talk with one another to diversify the signals. The fate of a cell is determined by the integrated signals summoned from the entire network of tyrosine kinases. To have a full appreciation of tumor phenotypes and to have the ability to predict the behaviors of tumor cells, it is important to obtain the entire expression profile of tyrosine kinases of a tumor cell. We have developed an effective approach, tyrosine kinase display, which can describe all or nearly all expressed tyrosine kinases in a single gel and by a single RT-PCR reaction. This restriction-based approach does not require cloning and sequencing and allows all tyrosine kinases be "read" from the gel by their characteristic sizes. Differentially expressed kinases can be easily identified and unknown kinases readily spotted.

Human genome encodes approximately 100 tyrosine kinases and each cell type expresses about 40 such kinases, a number large enough to give characteristic expression pattern, but small enough to be analyzed in a single gel. The catalytic domain of all tyrosine kinases share common motifs which can be used to anchor degenerate primers to amplify the great majority, if not all, tyrosine kinase transcripts in a single PCR RT-PCR reaction. Upon digestions with restriction enzymes, different tyrosine kinase amplicons would segregate by their characteristic sizes. To reduce complexity and to increase sensitivity, the sense primer can be radio-labeled at the 5˘ end, so that only the 5˘ end fragment will be visualized and each kinase is represented by one band, and the intensity reflects the molar concentration of the transcript. We have successfully applied this approach to a variety of cancers and obtained comprehensive profiles of them. As a result, tumor-specific or grade-specific tyrosine kinases have been identified and novel kinases that respond to hormone or drugs uncovered. Enormous amount of information concerning the signal pathways is also obtained. This approach is generally highly quantitative and can pick up variation of expression within two to three folds. In the presentation, we will discuss several useful prostate cancer markers uncovered by this approach. This approach can also be extended to other gene families of which degenerate primers are available. Applications to phosphatases will also be described.


3:15 pm – 5:15 pm

Parallel Session I:

Emerging Technologies in Cytometry and Advanced Microscopy

Chair: Quentin Hanley


6040

HT-PS and Plug Flow Cytometry For Multi-well Plate Drug Screening

Bruce S. Edwards, Frederick W. Kucuck, Eric R. Prossnitz, Larry A. Sklar, University of New Mexico, Cancer Research and Treatment Center; Alex Okun, John T. Ransom, Axiom Biotechnologies, Inc.

Because of the information-rich data sets generated by flow cytometry (FCM) analysis, high throughput FCM offers great potential for efficiently screening and evaluating complex compound libraries for drug discovery. We have developed a "plug flow" interface for an Axiom HT-PS (High Throughput Pharmacology System) that enables automated sampling from 96-well plates, exposure to drugs, and delivery of discrete 5 microliter volumes (plugs) of treated cells to an FCM for analysis. In one mode of compound characterization, HT-PS exposes cells to a programmed gradient of compound (linear or exponential spanning 3.5 logs) yielding a complete dose-response profile in 2 minutes. In initial system validation experiments using fluorescent beads, a HT-PS-generated linear gradient of beads was accurately delivered and detected by flow cytometry at a sampling rate of 9 plugs per minute. To verify mixing and autosampling of samples prior to FCM analysis, separate populations of particles were delivered over time and mixed with a linear gradient of distinct particles. Each population behaved independently. Calcium responses of U937 cells transfected with the formyl peptide receptor were efficiently characterized in response to step, linear and exponential gradientsof formyl peptide. Simultaneous multiparametric analyses of formyl peptide binding (fluorescein) and calcium response (Fura Red) were achieved with formyl peptide gradients and samples drawn from multi-well plates. Cell calcium responses were essentially all-or-none and initiated by receptor occupancy on the order of 1% of the expressed receptors. Sampling from multi-well plates in a single-dose endpoint assay format was achieved at rates of 6 to 7 sample wells per minute. These data demonstrate the utility of HT-PS plug flow cytometry in primary screening and compound profiling. The exploration of the relationship between receptor occupancy and response illustrates just one of the novel pharmacological approaches made available by the technology.


6105

A Compact Flow Cytometer for DNA Fragment Size Analysis

Robert Habbersett, Xiaomei Yan, James Jett, Los Alamos National Laboratory, Bioscience Division

DNA fragment sizing technology has been refined and improved recently resulting in an instrument, with enhanced performance, that is smaller and more robust. Exploiting a new DNA staining dye, it is capable of sizing DNA fragments ranging in length from0.125 to greater than 300 kilobase pairs with high accuracy (~2% uncertainty in absolute size). The excitation light source is a frequency doubled, solid state, Nd:YAG laser that emits at 532 nm. The laser beam passes through a polarizer and a polarization-sensitive beam splitter to provide power adjustment. Typically, 5 mw or less of 532 nm light is passed through crossed cylindrical lenses that focus the light to a 50 x 11 µm spot. The sample (2-200nL are analyzed, depending on fragment concentration) is delivered under pressure, controlled by an electronic pressure regulator, through a 40 µm ID capillary to the center of a 250 µm square bore flow cell. Fluorescence photons, emitted as a burst when a fragment passes through the laser beam, are detected by a photon-counting avalanche photodiode detector (EG&G Canada). The photon count rate history is recorded by operating a data acquisition card (National Instruments PCI-MIO-16E) in a multichannel-scaling mode. A LabView program, residing in a personal computer, controls data acquisition and display functions. Both the laser and the detector are powered by a single high-efficiency power supply (5V, 15A) which, along with all pneumatic and electronic components, including the laser controller, thepressure regulator, and a digital pressure sensor, fit easily in a 12" x 12" x 6" chassis. The apparatus itself occupies less than two square feet on an optical breadboard. Software has been developed to extensively explore the relationship between burst amplitude, width and area using combinations of 1D, 2D and 3D graphs. Editing and highlighting functions, analogous to gating capabilities of conventional cytometric software, are essential to understanding the data presented by a heterogeneous mix ofDNA fragments. This work supported by the Department of Energy Office of Nonproliferation and National Security (NN-20) and the NIH National Center for Research Resources Grant RR-01315 for the National Flow Cytometry Resource.

 


6303

A Micro-Electronic Flow Cytometer

Lydia L. Sohn, Princeton University, Dept. of Physics; Andrew J. Beavis, Daniel A. Notterman, Princeton University, Dept. of Molecular Biology

Flow Cytometry has become a very powerful tool for the diagnosis and management of hematological malignancies. Large numbers of cells can be analyzed rapidly for quantitation of DNA content in specific subpopulations of cells. However, this technologyis expensive requiring sophisticated instrumentation and skilled operators. Furthermore, sample preparation and staining with specific fluorochromes adds additional time and expense to experiments. We have developed a novel integrated microfluidic device which can measure and record directly the unique dielectric properties of material flowing through a micron-sized channel. Our device can detect eucaryotic DNA in vivo and differentiate cells on the basis of DNA content. An aliquot of mouse myeloma cells (SP2/0) was analyzed with our device. Cells in G2/M phase of the cell cycle resulted in twice the change in capacitance measured across the device as compared to those cells in G0/G1 phase, as they have twice the DNA content. Results were compared to a separate aliquot of cells stained with propidium iodide and analyzed by flow cytometry. The percentage of cells in each phase of the cell cycle were consistent with that obtained by our device. In addition to mouse myeloma cells, avian and mammalian red blood cells were also interrogated. Whereas avian red blood cells possess 2N DNA, mammalian red blood cells, which are approximately the same diameter as avian cells, contain no DNA. We measured capacitance changes when avian cells were flowing through our device. No significant changes were detected when interrogating mamalian red blood cells, confirming that we are indeed measuring DNA content rather than some other property of the cell with our device. Since the detection scheme of our device is nonoptical and direct, sample preparation is minimal. Equally important, because integrated circuit technology is used for device fabrication, the entire system can be miniature, robust, and portable. This type of device provdes for an inexpensive but rapid screening of samples based upon DNA content and could be used to pre-screen samples prior to flow cytometric analysis. Furthermore, this compact, portable device would allow for analysis of samples in remote or hazardous areas where flow cytometers cannot be used.

 


6512

The ACIS: Emerging Technology for Cellular Imaging

William Decker, Carl Priddy, Kenneth Bauer, Jose Torre-Bueno, Douglas Harrington, ChromaVision Medical Systems, Inc.

The ACIS (Automated Cellular Imaging System) is a system for improving diagnostic certainty for pathologists. It features two core technologies: Color Space Transformations and Morphological Sorting. The Color Space Transformation turns full color images into black and white images for further processing. These black and white images are then evaluated for their morphological characteristics. Multiple colors can be defined and evaluated, yielding parameters meaningful to the pathologist, such as the nucleus-to-cytoplasm ratio used in tumor cell evaluation. The core technologies have been packaged into technology platforms that can be put together easily into various applications. The four technology platforms that can be assembled into applications are the following: (1) Tissue evaluation, (2) Scoring, (3) Rare Event Detection, (4) Rapid Counting. Using object-oriented software and modular design techniques, these platforms form applications for tissue, blood, bone marrow, and cytopathology specimen evaluation. The ACIS has been applied to several clinical applications. Evaluations with the ACIS of tissue sections that might over-express HER-2/neu - using the Tissue Evaluation and Scoring technology platforms - have been shown to correlate well with manual microscopy. In fact, in a study conducted with 50 cases and numerous pathologists, whenever the pathologists agreed with each other with regard to the HER-2/neu score, the ACIS score also agreed. Reproducibility of the scoring was at a level far exceeding human evaluation - 97% versus 60%-75%. Similar results were achieved when applying the ACIS Rare Event and Rapid Counting technologies to detecting cytokeratin positive cells in bone marrow. The ACIS was shown to reproducibly find rare events; in a study of bone marrow samples, 44% of the cases called negative by a pathologist were found to actually be positive when the pathologist used the ACIS to detect the rare events. The ACIS features many more applications, such as ER, PR, p53, Ki-67,CMV antigenemia, and micrometasteses detection in lymph node sections, and others. It also offers the opportunity for researchers to define their own applications. An application design interface allows customers to adjust well-tested applications to suit their own needs and variations. Thus, the core technologies of Color Space Transformation and Morphological Sorting, and the technology platforms of Tissue Evaluation, Scoring, Rare Event Detection, and Rapid Counting become tools in researching new ways to provide diagnostic certainty to new clinical applications.

 


6712

The Violet Laser Diode as a Light Source for Cytometry

Howard Shapiro, M.D., West Newton, MA

Violet laser diodes developed by Nakamura and colleagues at Nichia Corporation in Japan have recently become commercially available. These devices now emit 5 mW at a nominal wavelength of 405 nm (range 395-415 nm); powers in the 50-100 mW range have been achieved in the laboratory. Several companies now offer violet laser diodes with regulated power supplies, in housings incorporating thermoelectric coolers, which are necessary to maintain stable output. Such systems now cost several thousand dollars, but should drop substantially in price as violet diodes are mass produced for use in high-density DVD recorders. Lifetimes are estimated to be thousands of hours, and short- and long-term output stability should be excellent, with RMS noise well below 1%. Previously, mercury arc lamps, with a relatively weak 405 nm line, and large frame krypton ion lasers, with several lines in the 406-422 nm range, were the only violet sources available for cytometry, and the high cost and power and cooling requirements of the latter prevented them from being widely used. The violet diode can be used for excitation of a number of widely used fluorescent probes, including some now used with UV sources. Devices selected for emission at 400 nm and below provide about 25% maximum excitation for DAPI, and can also be used with selected Hoechst dyes for stoichiometric staining of DNA in viable cells, as well as in fixed or permeabilized cells or nuclei. In addition, violet diodes can excite mithramycin and chromomycin A3, among DNA stains, and Cascade Blue and Cascade Yellow, covalent labels made by Molecular Probes (Eugene, OR). Both the native form of green fluorescent protein and newer blue and cyan variants are excited efficiently. Although the violet diode is not usable with indo-1, the most popular emission ratiometric calcium probe, it is likely that a calcium probe which would absorb near 400 nm could be made by modifiying the organic ring structure of indo-1. The 400 nm wavelength region is also useful for studies of porphyrin absorption and fluorescence. The violet laser diode is efficient, compact, and relatively inexpensive, and should become a useful light source for both flow and scanning cytometry in the near future.

This work was supported by NIH Grant RR13252.

 


6835

Three-dimensional arrangement of GFP-tagged Upstream Binding Factor (UBF) during the cell cycle of living cells.

Marie-Françoise O’Donohue, France; Hervé Kaplan, France; Hervé Le Moal, France; Thierry Cheutin, France; Dominique Ploton, UFR de Médecine, Biologie Cellulaire, Reims, France

The nucleolus is the site of synthesis and maturation of ribosomal RNA (rRNA). The tandemly repeated rDNA genes are transcribed by RNA polymerase I and its co-factors (Upstream Binding Factor (UBF), TBP, TAFs). Since rRNA synthesis is a major cellular function, most of the RNA polymerase I transcriptional machinery remains associated to rDNA genes even during mitosis. It is thus possible to follow these proteins throughout the cell cycle. Previously, we have taken advantage of this particularity to observe UBF in fixed cells using specific antibodies. In the present work, we have developed chimeric proteins using GFP in order to :

(i) observe either UBF1 or UBF2, since these two isoforms cannot be discriminated by antibodies and appreciate their relative distribution compared to that of RNA polymerase I ;

(ii) evaluate the precise order of the different steps of mitosis and their relative length ;

(iii) study the effect of inhibitors in real time.

The cDNAs coding for UBF1 and UBF2 were amplified by PCR and cloned into pEGFP (Clontech). Cells grown on coverslips were transfected and fixed for 5 min with 3% paraformaldehyde sixteen to eighteen hours after transfection. After permeabilization with 0.1% Triton X-100, immunolabelings with antibodies directed against various nucleolar proteins (UBF, RNA polymerase I, nucleolin, etc ) were performed to check their relative positioning with that of GFP-tagged UBF1 or UBF2. For study on living cells, the coverslip was mounted on a device allowingthe observation of the transfected cells at constant temperature (37°C), with a perfusion system to renew the culture medium and/or add inhibitors (Bioptechs). The cells were observed with a Biorad ES1024 confocal microscope equipped with an inverted microscope (Olympus). Thirty to fifty optical sections were recorded every three minutes for several hours within the whole volume of each cell. After processing, volumic rendering methods were realised for each time point to obtain high quality and interactive three-dimensional visualization of the labeling. A wide range of 3D visualization methods and tomography (ray tracing method, transparency views, rotations, stereopairs, split-views) was performed and films of these cellular events were done to get a complete picture of UBF reorganization throughout the cell cycle. This study reveals in real time the modifications of UBF organization, the equal partitioning of this transcription factor between the two daughter-cells during anaphase, as well as the recruitement of UBF at the onset of transcription.

 


6840

The Optical Sectioning Programmable Array Microscope (PAM)

Quentin Hanley, Donna Arndt-Jovin, Thomas Jovin, Max Planck Institute for Biophysical Chemistry, Göttingen, Germany

The distinctive feature of a PAM is the placement of a spatial light modulator (SLM) in an image plane to create structured illumination and/or detection. We have developed two families of fluorescence PAMs. One uses the SLM to create an optical sectioning condition, while the other exploits the SLM for spatial encoding. These instruments have used both digital micromirror devices (DMDs) and liquid crystal displays.

The core of our optical sectioning PAM consists of a DMD to provide structured illumination and conjugate detection. This arrangement allows the thickness of the optical section to be tuned under programmable control to optimally match the sectioning strategy to a particular sample and/or experimental need. Regions of interest may be defined, the light intensity presented to a particular region can be varied on a pixel by pixel basis, and the illumination light can serve to initiate photochemical reactions.

Additional components can be added to this core capability. For example, we have incorporated an acousto-optic modulator to provide frequency modulated excitation and a modulated image intensifier for the measurement of frequency domain fluorescent lifetimes. The result of this modification is a powerful combination of optical sectioning and fluorescent lifetime measurements.

We will present images allowing the comparison of optical sectioning strategies in individual specimens, images taken at high frame rates, as well as data obtained using add-on components.We will also discuss some of the unique challenges presented by the PAMs in the areas of image acquisition, real-time processing, analysis, display, and simulation of acquired data sets.

Selected Optical Sectioning PAM Publications

Verveer, P. J., Q. S. Hanley, P. W. Verbeek, L. J. van Vliet, and T. M. Jovin. 1998. Theory of confocal fluorescence imaging in the Programmable Array Microscope (PAM). J. Microsc. 189:192-198.

Verveer, P. J., and T. M. Jovin. 1998. Improved restoration from multiple images of a single object with application to fluorescence microscopy. Appl. Optics 37:6240-6246.

Hanley, Q. S., P. J. Verveer, M. J. Gemkow, and T. M. Jovin. 1999. An optical sectioning programmable array microscope implemented with a digital micromirror device. J. Microsc., in press, Dec.


3:15 pm – 5:15 pm

Parallel Session I:

Basic Cancer Biology

Chair: Peter Rabinovitch


6489

Pancolonic Chromosomal Instability Precedes Dysplasia and Cancer in Ulcerative Colitis

Peter Rabinovitch, Shawna Dziadon, Teresa Brentnall, Mary Emond, David Crispin, Rodger Haggitt, Mary Bronner, University of Washington, Seattle, Washington

Patients with long-standing Ulcerative Colitis (UC) are at increased risk for colon cancer. These cancers are thought to arise from preexisting dysplasia in a field of abnormal cells that often exhibits aneuploidy and p53 abnormalities. Using dual color fluorescence in situ hybridization (FISH) with centromere probes and locus specific arm probes for chromosomes 8,11,17, and 18 we demonstrate that chromosomal instability (CIN) is present throughout the colon of UC patients with high grade dysplasia or cancer. In rectal biopsies that were negative for dysplasia, abnormalities in chromosomal arm markers, especially losses, were most common, whereas centromere gains were most common in dysplasia and cancer. The frequency and type of abnormalities varied between chromosomes examined; chromosome 8 was least affected, and 17p loss was found to be an early and frequent event. Chromosomal instability was generally not present in biopsies from chronic UC patients that had not progressed to dysplasia or cancer.There was great heterogeneity in the FISH abnormalities seen in any single biopsy, bringing into question the degree of clonality of the abnormal cell populations. To examine this, we microdissected single colon crypts and placed them on slides for FISH. In cypts from histologically negative biopsies similar FISH abnormalities were found in small groups of 2 to 8 cells, and crypts showed as much heterogeneity in kinds of FISH abnormalities as did whole biopsies. These results imply that the earliest CINin UC is microclonal, as clones of FISH-abnormal cells have not expanded to comprise a substantial portion of even one crypt. Chromosomal arm instability showed 100% sensitivity and specificity for distinguishing individual control biopsies from histologically negative rectal biopsies from these UC patients, raising the possibility that a screen for chromosomal instability might detect the subset of UC patients that are at greatest risk for development of dysplasia and cancer. These results suggest that dysplasia and cancer in UC arise from a process of chromosomal instability that affects the entire colon; this may provide the mutator phenotype that predisposes to loss of tumor suppressor genes and evolution towards cancer.

 


6491

Nuclear Chromatin Texture Analysis Of Histologically Normal Colorectal Mucosa In Patients With Adenomatous Polyps And Colorectal Cancer

Peter Hamilton, Katrina Dillon, James Diamond, Queen’s University of Belfast, Department of Pathology, Belfast, N Ireland; Peter Bartels, Deborah Thompson, University of Arizona, Optical Sciences Centre, Tucson, AZ; David Alberts, Arizona Cancer Center, Tucson, AZ

It has been shown previously that histologically normal tissue in tumor bearing organs demonstrate subtle abnormalities in nuclear chromatin arrangement as measured by quantitative texture analysis. These alterations have often been referred to as "malignancy associated changes" or MACs. In this study, we examined nuclear chromatin texture at various stages of colon cancer progression in order to assess potential differences in karyometric features found in individual nuclei within these samples. The samples included rectal mucosal biopsies from subjects with adenomatous polyps and the matching adenomatous polyp (n=21) and 20 colorectal cancers with matching adjacent mucosa (within 1 cm) as well as distant mucosa (at approximately 10 cm or the furthest margin of resection). Five micron thick tissue sections were taken from each sample and stained using a modified haematoxylin and eosin stain. Digital images of nuclei were recorded at x100 objective magnification using a calibrated photometer system. Individual nuclei were segmented using KS400 software (Carl Zeiss) and analyzed for chromatin texture using purpose built software. At least thirty nuclei were measured in each sample and a total of 41 textural features measured on each nucleus, including the computation of optical density variation, cooccurrence matrices, run length matrices, discrete features and fractal dimension. Samples showing non-neoplastic colorectal mucosa were oriented and thirty nuclei measured within each of three crypt compartments. Analysis of the data revealed that there were significant differences (p<0.05) in nuclear texture between normal rectal mucosa and adenocarcinoma. Mucosa adjacent to, and at a distance from, adenocarcinoma was significantly different (p<0.05) from the values seen in rectal mucosa suggesting that adjacent mucosa was severely altered. Using functions based on multivariate discriminant analysis, examination of crypt compartment data showed an obvious decline in function score from rectal mucosato mucosa adjacent to the lesion. This also illustrates the value of compartmental analysis in colonic mucosa as independent information is offered by the different compartments. Using chromatin texture analysis a unique signature could be derived for each nucleus and an abnormality index computed for each sample. Results suggested that histologically normal appearing mucosa adjacent to colorectal polyps and cancer is a highly diverse set of cells which may directly relate to potential for lesion progression, development and recurrence. This initial study has highlighted the potential of nuclear chromatin texture in determining the biological characteristics of colorectal lesions and their background mucosa.

 


6718

Interaction of Receptor-Tyrosin-Kinases (RTK) of the EGFR-family and related growth kinetics in breast and bladder cancer cells

Gero Brockhoff, University Regensburg, Institute of Pathology, Regensburg, Bavaria, Germany

The Epidermal-Growth-Factor-Receptor (EGFR) family comprises at least four members: EGFR, c-erbB2, c-erbB3, and c-erbB4 receptor. These receptors have different ligand-binding specificity and show a well-defined expression pattern on normal tissue. However, on tumor cells an overexpression of one receptor and/or an abnormal pattern of receptor-coexpression occurs and influences tumor growth in a partially unknown way. The physiological cellular response to ligand-stimulation depends on interaction of several RTK of the EGFR-family, generating homo- and heterodimers respectively, and together activate downstream signal-ing pathways. In this study the specific receptor interactions and crossactivation on several tumor cell lines were investigated, correlated to the coexpression pattern and related to the kinetic growth behavior. Several in vitro cell lines derived from breast and bladder carcinomas were used to investigate receptor interaction via fluorescence-resonance-energy-transfer (FRET). Heregulin (HRG) and Epidermal-Growth-Factor (EGF) were used for cell stimulation. The assembly of receptor aggregates was flow-cytometrically detected with the donor dye phycoerythrin and the acceptor fluorochome cyanin-5. ELISA experiments were done to analyzethe phosphorylation state after receptor crossactivation. In order to ascertain the cellular response and the growth behavior after stimulation the dynamic cell cycle analysis (anti-BrdU technique) was performed. Data demonstrate that the mitogenic, ligand-specific stimulation of tumor cells de-pends on the absolute ratio of receptor coexpression, and the erbB2-receptor play a keyrole for early signal transduction events across the cell membrane. The effect of HRG and EGF depends on specific receptorinteractions, nevertheless similar receptor coexpression patterns can cause varying crossactivation. Depending on the cell types growth factors affect the duration of cell cycles at differ-ent phases and this modulation can be attributed to specific RTKinteractions. Literature: 1) Brockhoff et al. Cytometry 17:73-85, 1994 2) Brockhoff et al. Anal Cell Pathol 11:55-70, 1996 3) Brockhoff et al. Virch Arch 432:77-84, 1998 4) Chow NH et al. Virch Arch 430:461-466, 1997 5) Nagy et al. Cytometry 321:120-131, 1998 6) Nagy et al. J Cell Sci 112:1733-1741, 1999

 


5874

Heterogeneity Of Hla Class I Expression In Cervical Cancer Detected By Multiparameter DNA Flow Cytometry

Willem Corver, Leiden University Medical Centre, The Netherlands

Background. Heterogeneity of HLA class I expression occurs frequently in cervical cancer and other solid tumors. It can be observed as HLA class I positive and negative tumor areas by standard immunohistochemistry. Here we demonstrate the possibility to identify HLA class I positive and negative tumor sub-populations in freshly isolated cervical carcinoma cells by a recently developed four-color multiparameter DNA flow cytometric technique. Methods. Single cell suspensions were prepared from seventeen fresh tumors by a combined mechanical - enzymatical dissociation method. Cells were stained with a four-color multiparameter technique allowing the simultaneous detection of: 1. stromal cells by anti-CD45 or anti-vimentin mAbs (FITC fluorescence), 2. the epithelial cell fraction by anti-keratin mAbs (RPE fluorescence), 3. HLA expression by allele-specific murine or human monoclonal antibodies (mAbs) (APC fluorescence) and 4. DNA content (PI fluorescence). A FACScalibur flow cytometer with dual excitation facility was used for fluorescence analysis. Tissue sections from the corresponding tumors were stained for HLA, keratin, vimentin and CD45. Results. Flow cytometry enabled the simultaneous discrimination of CD45 or vimentin positive cell populations, keratin positive cell populations and measurement of DNA content readily. These normal stromal cells served as intrinsic HLA class I positive as well as DNA-diploid references. DNA histogram quality was good as indicated by low coefficients of variation of the G0G1-peaks (average < 4%). DNA-aneuploidy was observed in 71% of the cases. Intra-tumor keratin positive HLA class I positive as well as keratin positive HLA class I negative sub-populations were identified in 82% of the cases and were found in both DNA-aneuploid as well as DNA-diploid tumors. These tumor sub-populations differed in DNA content or showed a similar DNA content. Losses of allele specific HLA class I expression found by immunohistochemistry were also detected by flow cytometry. Conclusions. We conclude that multiparameter DNA flow cytometry is a powerful tool to study loss of HLA class I expression in cell suspensions from fresh human cervical tumors. The method enables flow-sorting of tumor and normal cell sub-populations based on keratin or vimentin expression, HLA class I expression and DNA content for molecular genetic analysis. Furthermore, our results indicate that loss of HLA class I expression is a relative early event in cervical tumor development.

 


5875

MULTIDRUG-RESISTANT HUMAN OVARIAN CARCINOMA CELLS DISPLAY NUCLEAR TEXTURE CHANGES ASSOCIATED WITH HISTONE H4 ACETYLATION

Sonia Yatouji, Universite de Reims, Laboratoire de Physiologie Cellulaire, Reims, France; Chantal Trentesaux, Universite de Reims, Laboratoire de Biologie Moléculaire, Reims, France; Françoise Liautaud-Roger, Institut Jean-Godinot, Reims, France; Jean Dufer, Universite de Reims, Laboratoire de Physiologie Cellulaire, Reims, France

We showed previously that multidrug-resistant cancer cells displayed nuclear texture changes as assessed by image cytometry. In this work, two tumoral cell lines were studied: the human ovarian carcinoma cell line IGROV1 and its multidrug-resistant variant OV1-VCR selected with vincristine. We confirmed that OV1-VCR cells overexpress P-glycoprotein and its corresponding mRNA without mdr1 gene DNA amplification. Cell smears of these two cell populations were stained with Feulgen method and analyzed by image cytometry. As compared to sensitive cells, OV1-VCR display a chromatin global decondensation as assessed by textural features analysis. Several studies suggest that transcriptionally or potentially active genes exist in an altered conformation in nuclear interphase chromatin. In order to correlate this decondensation with alterations in chromatin structure, DNase I was used as a probe to preferentially digest potentially active genes in chromatin. Genomic DNA was extracted from both cell lines nuclei after treatment with various concentrations of DNase I. OV1-VCR DNA displayed an increased (about 5 fold) DNase I sensitivity, suggesting an increased chromatin accessibility. To investigate further the presence of putative DNase I hypersensitive sites (HS)in the vicinity of the mdr1 gene promoter, genomic DNA from both lines was analyzed by Southern blot after digestion with DNase I and the restriction enzymes Hind III, Bam H1, EcoR1, and Kpn1. According to the mdr1 probe (1.3kb) and enzymes used, no change in DNase HS sites could be visualized. Finally, it has been shown that high levels of chromatin acetylation across complete chromatin domains induces chromatin changes detected as "general DNase I sensitivity". By western blotting using a monoclonal antibody directed against acetylated histone H4, the expression of this acetylated form of histone H4 appears increased in OV1-VCR cells. Furthermore, treatment of IGROV1 drug-sensitive cells with the histone deacetylase inhibitor trichostatin A induces significant mdr1 mRNA levels and leads to nuclear textural changes similar to those observed in OV1-VCR resistant cells.

 


6110

Morphometric measurement of prostate structures as a measure of genomic variability

Bruce Milthorpe, Samir Philips, University of New South Wales, Graduate School of Biomedical Engineering, Sydney, Australia; Allan Jones, Institute of Respiratory Medicine, Allergen Diagnostic Project, Camperdown, USA

The accumulation of clonal diversity in tumorigenesis (Shackney and Shankey, 1997) provides a compelling theoretical basis for the staged developmental behaviour of tumours. Cells slowly accumulate a variety of genomic errors and generate new clones.Those genomic errors that lead to a more rapid accumulation of errors and to errors in cell cycle control generate pre-malignant and malignant states. Even malignant clones, though, will still continue to accumulate errors leading to a variety of malignant clones with varying behaviour. Evidence already exists for poly-clonal tumour development in prostate cancer (Hedley, 1997). We are hypothesising that since the accumulation of genomic errors should be random, then genes affecting intracellular and extracellular morphology should also accumulate errors, leading to a variety of structural morphologies. Thus the accumulation of clonal diversity should be reflected by the accumulation of different structural morphologies in both the cell and the tissue. A measurement of the number or variance of morphological states should, therefore, provide an estimate of the likelihood of a tumour clone(s) developing androgen resistance and/or metastasising. This measure of morphological states could well become an important prognostic indicator. This work aims to determine if such a series of morphological indicators can be extracted from images of prostate cancer biopsy samples. The initial step has been to identify the duct lumens and the stroma in the samples (Fig. 1) using high level analytical packages (WiT and MatLab) with purpose-written subroutines. The shape factor ([perimeter]2/[4 pArea]) and perimeter distributions for duct lumens (Figs. 2 and 3) are distinctly different for normal and cancerous tissue. The next step is to look at the population distributions of properties (eg. internuclear distance, nuclear shape, texture and orientation, cytoplasm size) of the cells surrounding the lumens using the lumen boundary as the datum line. Some significant problems still await resolution. It is difficult for the software to distinguish the small termini of the ducts from blood vessels unless there are erythrocytes present. High magnification, to obtain details of individual cells, obscures organ structure. The presence of both normal and abnormal cells may skew the population distribution results. Notwithstanding these problems, the technique is showing promise and should improve with more development of the analysis software. D.W. Hedley, 1997, Inaugural Sam Latt Conference on Molecular Pathology, Toronto 18-22 June. S.E. Shackney and T.V. Shankey, 1997, Common patterns of genetic evolution in human solid tumors, Cytometry 29(1): 1-27.

 


6810

 

Abstract Withdrawn

 

 

 


Parallel Session I: Informatics

Chair: Johannes Frank


3257

BioScope: Cytometry and the Internet for K-12 Education

J. Paul Robinson, Gerald Krockover, Kelly Kinch, Monica Shively, Gretchen Lawler, Kathy Ragheb, David Whittinghill, Ty Filby, Greg Moore, Steve Barg, Laura Moretti, Purdue University, W. Lafayette, IN, USA; Marshall Overley, West Lafayette J/S High School, USA

Funded by the National Science Foundation, the BioScope Initiative is an exciting program which aims at placing high quality science-biology into every high school biology classroom. The BioScope Initiative develops high quality biology related material, which is scientifically accurate, referenced, and relevant to kids needs. Animations and cartoons are frequently used to interest the kids. Everything operates under standard web browsers, and quality internet sites are linked directly from the CD-ROM. The unique aspect of this program development is that it is intentionally designed to be platform-independent and will therefore operate on all computer systems. BioScope material is based upon the National Science Standards andclosely follows these guidelines. It is designed to identify "handles" that grab students attention and use them to provide a rich learning environment in which the student participates in a journey through science. To the student, BioScope appears to bea WEB site - it operates and fells exactly like "surfing" the net - but the key material is actually on the CD-ROM in the student’s computer - and if the student is connected to the net -they can also be linked to science information anywhere in the world. All sites linked by BioScope are automatically checked on an hourly basis to determine any changes in that site - new sites can be linked at any time, and in this way, the student can experience new pathways to information at any time. The current BioScope Initiative material is aimed at high school student biology (grades 9-12) however, material at all levels is currently under development. Further information can be obtained at http://www.bioscope.org

 


6183

A new dating mining/knowledge discovery method for flow cytometry

Scott Mclaughlin, Lisa Reece, James Leary, Univ. of Texas Medical Branch, Galveston, Texas, USA; Jacob Smith, University of Texas at Austin, USA;

We demonstrate the use of a novel data mining/knowledge discovery program called "subtractive clustering" that searches for the similarities and differences between two or more flow cytometry listmode data files. This is in contrast to most data analysis and clustering algorithms which work on only one datafile at a time. The algorithms can be run in either supervised or unsupervised mode, depending on the needs and knowledge of the researcher. As such it functions as an exploratory data analysis orknowledge discovery program to aid the researcher in seeing the important similarities and differences between two or more data files. While making no assumptions about the data, this program uses variable-sized multidimensional data bins from test and control files, allowing for the "subtractive clustering" of data by a new method that is not limited by data dimensionality or resolution. In contrast, other subtraction methods, such as channel-by-channel subtraction of carefully aligned multidimensional data, are dependent on small variations in cell clusters positions in multidimensional data space. Our method involves algorithms of "similarity" which allow similar, but not exact, datapoints in two or more datafiles to be compared. In supervised mode using a Windows™ interface in the program, the input of the experienced researcher (having information concerning the number of relevant subpopulations) is taken into account. Multidimensional listmode data can be adjusted or normalized over subregions using appropriate control data. Three different types of data sets illustrate the possible uses and range of potential use of this new methodology. Subtractive clustering was able to pull out hidden subpopulations within sample mixtures,while preserving the integrity of all other populations. This new data mining/knowledge discovery method provides a powerful means of finding non-obvious differences between two or more multidimensional data files.

 


6324

Cytometric Data Processing: From Modular Arrangement of Several Software Programs to an Integrative Database-Analysis System

Johannes L.J. Frank, University of Greifswald, Institute for Immunology and Transfusion Medicine, Greifswald, Germany

BACKGROUND: Cytometric data management and analysis gains growing importance with the increasing applications and multidimensionality of cytometry in research and clinical diagnostics. While storage of primary measurement usually is performed in single listmode files (FCS standard format) common cytometry software mainly focus on data analysis (statistical functions, cluster analysis etc.). Exporting of secondary analysis data and / or graphical information (diagrams), further processing, databasing and overviewing of the complete patient clientele of a laboratory ("tertiary data") can become a major problem and is far less standardized. METHODS: Common Windows-software for flow cytometry, databasing, spreadsheet calculation as well as practicable ways and formats for alphanumeric and graphical processing were tested for building up a flexible, routine-fitting management system. In particular, performance, user-friendliness, data storage and hardware requirements were critically analyzed. Finally 3 different solutions for cytometric data processing were developed and discussed. RESULTS: (1) Listmode data analysis with WinList 4.0 (Verity Software House), macro-DDE-driven export of vector graphics and statistical results to Microsoft Excel 5.0 / 7.0, user-friendly labeling, navigating and managing of complete measurement series by VBA-Program. (2) Listmode data analysis and macro-driven export as in (1) and storage and further managing of numeric and graphical data in a Paradox-Desktop-Database by a Windows-application developed with Borland Pascal / Delphi. (3) Integrative, stand-alone cytometry data managing system based on a SQL-server-database storing listmode files and all other information, allowing data analyzing and visualizing on primary (listmode), secondary (statistical results) and tertiary (patient groups) level. DISCUSSION AND OUTLOOK: While exporting and databasing of statistical results alone (ASCII files) can be executed with combinations of common software without greater expense, simultaneous processing and storing of graphical output (histograms, dotplots) requires special soft- and hardware conditions. A practicable stability, performance and flexibility of cytometric data processing from listmode data to patient collective analysis can best be performed in an integrative SQL-server based application system. Beside single sample orientated batch processing such a system allows overview, comparison and analysis of numerical and graphical information on higher levels without missing relation to the primary listmode data. Similar to the standard database retrieval language SQL the development of a simple Cytometry Query Language (CQL) could make possible quick evaluation of even complex cytometric questions.

 


6492

PRINCIPLES FOR CODING DATA ANALYSIS METHODS

Gyorgy Lustyik, Institute of Biophysics, University Medical School of Pecs, Pecs, Hungary; J. Paul Robinson, Purdue University, West Lafayette, USA; Francis Mandy, Bureau of Labs & Research Services, Federal Centre for Aids - NHW, Ottawa, Canada

The expansion of clinical and research applications of analytical cytology has dictated the need for standardization of the sample preparation procedures, reagents, instrument design and data storage formats (Quantitative Fluorescence Cytometry: An Emerging Consensus, Cytometry, 33(2):93-288). The next mandatory step of standardization is the development of standardized tools for the data analysis and processing. In this work, the principles of a new standardization technology for coding computerized data analysis protocols are proposed. We have worked out a system of definitions and formalism for coding the basic steps of data processing. The principle is to define a series of statements and functions that are able to code the data evaluation methodsof any complexity. The application of the principles that we call PCDAM (Principles of Coding Data Analysis Methods) provides the means of performing computerized data processing functions and operations in a unified manner. The proposed definitions and technology are the starting points for discussion. Ideally, the suggested standardization principles will be evaluated and reviewed by software developers, and consensus definitions, functions, and coding principles will be established and ultimately built into the new software applications. This technology will provide a large armamentarium of methods of analyzing data with a standard precision and accuracy. This process may take several years similar to the acceptance of the flow cytometry standard (FCS) data file format introduced in 1982 that is already used by all major manufacturers and software developers. This work was partly supported by grant from National Research Foundation (OTKA T17070), Hungary.

 


6505

Defining a BioInformatics System From FCS Data Files

Charley Bay, Barclay Purcell, George Malachowski, Cytomation, Inc., Fort Collins, USA

Defining a BioInformatics System From FCS Data Files The FCS Data File Standard brought a means of exchanging sample and histogram data across software and hardware platforms. Since its inception, significant technology advances open new possibilities for data tracking and analysis. This poses the issue of (1) expanding the FCS Data File Standard to incorporate these new technologies, or (2) loading existing FCS data files into systems that provide this new functionality. We define guidelines for aBioInformatics system, and how the system may be populated with tabular and spatial data including FCS Data Files. A comparative analysis is made for alternative approaches to data storage, retrieval, and mining; identification and maintenance of relationships and correlations; and data-context driven analytical manipulation. Further exploration is made of how Computer Science approaches analogous problems, the tools available, and outstanding issues. Finally, examples are given for existing tools readily available for Cytometry application in the transition from a single-sample basis to a higher level of data integration.

 


6889

Flow Cytometry Markup Language: A proposal for structured representation of experimental metadata

Adam Treister, Tree Star, Inc., San Carlos, CA, USA

ISAC has adopted the FCS Data File Standard for the common representation of flow cytometric data. This standard is supported by all of the major analytical instruments. Scientists can choose among instruments and software with no major data compatibility issues. However, the standard does not adequately express broader concepts of flow cytometry experiments. This paper proposes an XML-compliant set of interoperable document type definitions, schemas and XSL style sheets to support addition concepts beyond list mode data. This standard would not supplant the FCS data format, but would provide extensible, structured, Internet compliant support of metadata required for advanced experimental design, analysis and presentation.

The goal of FCML isto provide an extensible language for Electronic Data Interchange (EDI) of FCS files and associated data. FCML identifies suitable existing public XML standards for adoption, and adds FCS-specific EDI to support specific concepts required by the field (such as fluorochrome or gate). The concepts of structured markup languages are widely used in the business community, supporting electronic commerce, client-server data administration techniques and search technologies. This architecture is the basis for future expansion of the Worldwide Web. Within the sciences, MathML, the Chemical Markup Language and the Bioinformatic Sequence Markup Language provide precedent examples of the technology.

FCML permits comprehensive description of experiments. It can express external files and data base queries, exploiting distributed computing resources not supported in FCS. Using FCML tools, a scientist can design a set of experiments on the desktop before acquisition and specify all the descriptive meta-data spanning the collection, analysis, presentation and publication phases. With XSL style sheets, journals can provide standard report formats and improve the uniformity and readability of the publications without limiting the authors’ choice of analysis tools. Users can adopt any tool independently. For instance, adoption of FCML’s XSL style sheets for publication need not require perturbation of existing FCS data archives.

FCML supports the structured design of experiments in research cytometry, promoting better scientific practice, easier data administration, and ready exchange of experimental data and analysis. Clinical cytometry benefits from a more rigorous methodology to validate results. All users would gain through interoperable flow-specific EDI tools and external data management services. Open standards permit software developers to expand the capabilities of any tool in any phase, and to adapt the toolkit to support site-specific requirements.


Parallel Session I: Phytoplankton in the Aquatic Environment

Chair: Daniel Vaulot


6432

Analysis of Phytoplankton Communities in Lakes and Artificial Systems by Flow Cytometry and Laser Scanning Cytometry

Wolfgang Beisker, Bernhard Bruckmeier, Ingrid Juettner, GSF-National Research Center on Environment and Health, Neuherberg, Germany; Ute Simon, IGB - Institute of Freshwater Ecology and Inland Fisherie, Berlin, Germany; Christof Schade, Institute of Zoology, Uni-Stuttgart Hohenheim, Germany

A detailed analysis of chlorophyll fluorescence at various excitation and emission wavelengths can reveal the composition of accessory pigments in phytoplankton cells. A rough classification in different pigment groups allows an overview of the phyoplanktic inventory of the water body. Even more, by simultaneous staining with fluoresceinisothiocyanate (FITC), the biomass of each single cell can be measured. However, a three laser flow cytometer is required for such type of analysis. For our analysis we use three laser flow cytometry for chlorophyll pigment group determination and biomass measurement by FITC staining. Because the centrifugation of phytoplanktic cells is very delicate to avoid cell loss of certain populations, we use a filtering technique with a 0.2um polystyrol filter attached to a syringe for staining and cell wash. The cells are fed into the cytometer by an automatic syringe to measure precisely the volume of the sample for concentration determination of the cells. The cytometer, a FACSstar+, is equipped with argon ion lasers of 488 nm and 528nm emission wavelengths and a HeNe laser(632nm). From the fluorescence (emission >665nm) excited by 528nm and 632nm, the pigment analysis is performed, whereas the FITC fluorescence excited by 488nm is measured for biomass spectra. The sorting capability of a cell sorter can be used to precisely identify phytoplankton species by visual microscopy. However, this is a very time consuming method, if samples with many subpopulations have to be inspected. Therefore we use a Laser Scanning Cytometer (LSC) for detailed identification. A pigment analysis after excitation with the appropriate wavelengths is performed on the LSC and galleries of images from the selected cell populations are be stored for detailed visual inspection. This combination of flow and scanning cytometry allows a rapid investigation of the plakton community of lakes and rivers as well as of artificial micro- and mesocosmic systems, which is a prerequisite for ecological studies. As applications for this combination technique we present the seasonal variation of three lakes near Heilbronn, the Ehmetsklinge, the Lake Michelbach and Lake Katzenbach, two of them have been artificially aerated. Depth profiles up to the ground (6m to 8m) have been measured. To show the use in ecotoxicological studies, we demonstrate the influence of nonylphenol on the phytoplankton community in a microcosmic system (a few hundered litres per pond).

 


6606

Oceanic Picoplankton: Large Scale Trends And Small Scale Variability Determined By Flow Cytometry

Daniel Vaulot, Dominique Marie, Jean Blanchot, Stephan Jacquet, Ric Partensky, CNRS, Station Biologique, Roscoff, France

Flow cytometry was first applied about 15 years ago to the study of the smallest size classes of marine unicellular organisms, that constitute what is called the ‘microbial food web’. In particular, flow cytometry has been instrumental for the discovery of key organisms, such as Prochlorococcus, the most abundant photosynthetic organism in the ocean, or Ostreococcus tauri, the smallest eukaryote known to date. Another aspect that is often overlooked, is the capacity of flow cytometry, because of its inherent speed and statistical accuracy, to provide very large sets of high quality data for the abundance of several key components of the microbial food web. To date, at least five microbial groups can be easily separated based on their cellular parameters, in particular pigment fluorescence and DNA content: viruses, heterotrophic bacteria, Prochlorococcus, Synechococcus, and picoeukaryotes. Moreover, some of these groups, such as heterotrophic bacteria, can be further subdivided into separate clusters based on size and DNA content. Over the past ten years, we have accumulated a data base of more than 7,000 oceanic picoplankton measurements that allows to establish global relationships between the different microbial groups, as well as between these groups and environmental parameters such as sea surface temperature or nitrate concentration. The use of conversion factors allows estimates of the contribution of each microbial component to the global oceanic living carbon. By combining flow cytometry with automatic sampling of surface sea water on board ship, we have obtained data on the small temporal and spatial scale variability of oceanic picoplankton. For example, we have conducted basin-wide surface transects throughout the Mediterranean Sea, allowing to precisely delimitate transitions between different trophic regimes. We also followed the circadian variability of photosynthetic picoplankton during several days and assessed some of the factors, such as cloud cover, that affect its population dynamics. These examples demonstrate that flow cytometry has become one of the most important technique to study the ecology of picoplankton in marine waters. Future directions include the development of more sensitive flow cytometers (in particular to detect very small biological particles such as viruses) and of flow cytometers capable of in situ analysis (i.e. very compact and streamlined instruments operating underwater).

 


6661

Monitoring changes in phytoplankton populations in ocean samples using artificial neural network analysis of flow cytometry data

Lynne Boddy, Malcolm Wilkins, Cardiff University, School of Biosciences, Cardiff, UK; Colin Morris, University of Glamorgan, School of Computing, Pontypridd, UK; Richard Jonker, Aquasense, Amsterdam, The Netherlands; Glen Tarran, Peter Burkill, Plymouth Marine Laboratory, Centre for Coastal and Marine Sciences, Plymouth, UK

Phytoplankton are key components of marine ecosystems: collectively they fuel the marine food web; they have been implicated in climate control; and some groups form nuisance blooms. It is, therefore, crucial to be able to identify and quantify phytoplankton populations accurately, rapidly and preferably actually during survey cruises. In the past appropriate data have been obtained by microscopic analysis of samples in the laboratory. This is laborious and time-consuming and since it is performed a long time after sampling, interesting phenomena cannot be resampled or followed up directly. Analytical flow cytometry can provide signatures for micro-algal cells, at rates of about 103 cells sec-1, which allows species to be discriminated. Artificial neural networks (ANNs) are ideal for analysing the vast quantities of non-normal data produced, and can make identifications in near real-time. ANNs are not rule based but learn/train from examples presented. Essentially, thereare two types of training - supervised and unsupervised. With the former, data patterns (in this case flow cytometric signatures) of known identity are repeatedly presented to the ANN and, provided that the parameters measured are sufficiently discriminatory, after many iterations the network will be able successfully to identify classes (in this case phytoplankton species) upon which it has been trained. With unsupervised training, on the other hand, patterns are presented to the network and it forms its own groupings of the data. The supervised approach is appropriate for making identifications, and, of a number of types of supervised ANN, the radial basis function (RBF) ANN or an asymmetric version of it has proved most successful for identification of phytoplankton from flow cytometric signatures. RBF ANNs train rapidly, make identifications rapidly and can reject taxa as unknown to the trained network. The unsupervised approach is useful for reducing dimensionality and revealing clusters in datasets, and can be used to aid obtaining training data from natural populations in the sea. ANN identification of phytoplankton from flow cytometric signatures provides a tool which will allow considerable steps forward to be taken in understanding the primary producers of the world’s oceans. The approach will be illustrated using flow cytometry data on over 60 important marine phytoplankton taxa. Scaling up to cover the many hundreds of species found in the world’s oceans is a non-trivial task; major progress towards this will be discussed.

 


6777

In situ on-line analysis of phytoplankton with CytoBuoy

George B.J. Dubelaar, CytoBuoy, Haarlem, The Netherlands; Peter L. Gerritzen, Datawell b.v., Haarlem, The Netherlands

CytoBuoy is a rugged flow cytometer for autonomous, battery powered in situ operation, intended for the on-line analysis of phytoplankton abundance in surface waters. Reliability and stability were prime considerations of this ‘field’ design, which resulted in both limitations and new exiting possibilities with regard to instrument and data specifications, ways of operation and integration in monitoring systems.

CytoBuoy analyses a large size range of particles, typical for marine coastal zonesand fresh waters (basing on the earlier O.P.A. cytometers). The 5-10 µl/s sample flow rate keeps the analysis time reasonably low at the low cell concentrations in surface water. The sheath fluid contains an anti fouling agent and circulates to eliminate the need for refilling. Its optics and electronics allow for maximal reflection of the particle morphology in the measured signals. Whereas standard cytometers reduce these to single peak or area ‘listmode’ numbers, the signal courses are being preserved fully in CytoBuoy by direct 4 MHz analog to digital conversion. This allows more extended morphological analysis and discrimination. The example shown is a set of pulse courses (one course for each detector) measured from a single particle. This type of raw data is available for all measured particles and can be reduced to standard listmode files.

Besides ambulant and laboratory use, CytoBuoy can be deployed as

1) a buoy based cytometer, mounted in the sensor compartment of a small moored buoy with wireless data transfer to shore, or as

2) part of an instrumentation package in a ‘ship of opportunity’, and also as

3) underwater device with the cytometer lowered on a cable or mounted inside a so-called A.U.V. (autonomous underwater vehicle).

By enabling direct on-line and high frequency estimation of phytoplankton biomass and assessment of population dynamics, flow cytometry may become a key technology in the field of environmental monitoring.

 


6794

A FLOW CYTOMETRY APPROACH TO ASSESS OCEANIC MICROBIAL RESPIRATION

Gérald Grégori, Dominique LefèVre, Michel Denis, Université de la Méditerranée, Laboratoire d’Océanographie et de Biogéochimie, Marseille, France

Microbial respiration in the ocean is considered as the major process representative of the biological oxidation of organic matter (OM). The corresponding metabolic CO2 production rate which characterizes the rate of OM remineralization has been estimated of the order of 22 Gt C y-1 (Packard et al., 1988). However, the in situ respiration rate is generally too low (by several orders of magnitude) to be accessible to the available direct measurement methods.

Some fluorescent probes, such as DIOC6(3) (Molecular Probes), have been shown to be very sensitive to changes in the electrochemical potential difference (*µH+) characterizing membranes (mitochondrial or plasmic) bearing the cell respiratory system (Shapiro, 1997; Novo et al., 1999). In mitochondria, *µH+ is linked to the flux of oxygen uptake by a linear relationship (Rigoulet et al., 1987). However, to our knowledge, no such relationship has been established in the case of whole cells. The objective of this work was therefore to look for a possible relationship between the fluorescence signal of the probe DIOC6(3) and the respiration rate of a phytoplankton culture, with the aim to develop a new method using flow cytometry to determine cell respiration rates. For the direct measurements of respiration rates, we used a high resolution respirometer (OROBOROS). For fluorescence measurements and cell counts, we used the CYTORON flow cytometer (ORTHO Diagnostic Systems). Experiments were run with axenic cultures of Dunaliella tertiolecta. Under these conditions, we have found a linear relationship between oxygen uptake by Dunaliella tertiolecta (dark respiration) and the green fluorescence signal of the probe DIOC6(3). The conversion of fluorescence to respiration rate has been standardized for Dunaliella tertiolecta The establishment of such a relationship for organisms representative of natural populations remains to be developed to open the way to direct in situ measurements.

This new method, easy to use, appears very promising due the high frequency of measurements provided by flow cytometry. It could be easily extended to fresh water environments.

Novo D. et al. (1999) Cytometry 35: 55-63.

Packard T.T. et al. (1988) Deep-Sea Res., 35, 371-382.

Rigoulet et al. (1987) Eur. J. Biochem., 168, 275-279.

Shapiro H. (1997) In: Current protocols in cytometry pp. 9.6.1-9.6.10.

 


6851

Phytoplankton size spectra measured by flow cytometry and imaging in flow in the Gulf of Maine and coastal waters

Michael Sieracki, Terry Cucci, Christian Sieracki, Edward Thier, Bigelow Laboratory for Ocean Sciences, W. Boothbay Harbor, ME, USA

Size spectra is indicative of successional and trophic status in planktonic ecosystems. Until recently, automated measurement of cell size spectra has been limited by inability to discriminate cells from non-living particles, and the lack of adequate dynamic range in one instrument to cover the wide size and abundance ranges observed in natural waters. Standard flow cytometers cannot efficiently sample the larger, less abundant cells in the microplankton (20 - 200 um) that often have the largest effect on size spectra slopes in coastal marine waters. We used a B-D FACScan flow cytometer to count and measure cells in the picoplankton and nanoplankton (0.5 - 20 um) size ranges. Microplankton were measured using a fluorescence-triggered imaging-in-flow system designed and built at Bigelow Laboratory. The two instruments overlap in the 8 - 15 um size range. Continuous monitoring and discrete, semi-weekly sampling was conducted to measure the size distributions and composition of phytoplankton in Boothbay Harbor, Maine, USA over a seasonal cycle. Discrete samples were analyzed from a transect in the Gulf of Maine. In these samples, size calibrations for forward light scatter were made using imaging cytometry. Our seasonal results show the changes in the phytoplankton community size spectrum during the spring bloom and the transition into the temperate summer community. Phytoplankton community structure in the Gulf of Maine is related to the hydrography.


5:15 pm – 7:00 pm

Parallel Session II: Digital Imaging

Chair: Peter Lansdorp


6389

Immunophenotyping of Lymph Node by Laser Scanning Cytometry (LSC)

Attila Tárnok, Peter Jirasek, Peter Schneider, Pediatric Cardiology, Cardiac Center Leipzig, University of Leipzig, Germany; Andreas Gerstner, University of Leipzig, Department of Otorhinolaryngology, Germany; Pal Racz, Klara Racz, Christine Trumpfheller, Bernhardt-Nocht-Institute for Tropical Diseases, University of Hamburg, Hamburg, Germany

Quantitative analysis of cellular subsets such as leukocytes in solid tissues is difficult to perform. Nevertheless it would yield important data in a wide variety of clinical and experimental settings. In lymphatic organs the quantitative analysis of the spatial distribution of the cells would give relevant information about alterations in the course of different diseases and therapy/therapeutic regimen of lymphatic disorders, leukemia or HIV and AIDS. We have set up an automated analysis method for LSC suitable for archived or fresh biopsy material of human lymph nodes and tonsils. Sections are stained with PI after RNAse-treatment followed by simultaneous immuno-fluorescence staining of one or two antigens using indirect methods and fluorescence enhancement (Alexa, Tyramide). Measurement is performed by triggering on DNA-fluorescence (Argon Laser). Due to the heterogeneity in cell density and nuclear condensation in lymphatic tissues the measurements have to be repeatedly performed at different threshold levels. At low threshold levels regions of low cellular density and follicular centers, at high threshold dense regions as the follicular cortex are measured. Data are acquired at single (Ar) or dual-leser excitation (Ar-HeNe) in order to determine single (FITC) or dual staining (FITC, APC), respectively. Aided by imaging, percentage and cellular density of unstained and stained cells can be quantified in different structural regions of the specimen. Using this technique we could quantify CD4 (T-helper), CD8 (T-cytotoxic) and CD1a (dentritic cells) positive cells in lymph nodes from different patients and detect clear differences between different stages of HIV infection. In combination with markers for cell activation (e.g. CD152) or proliferation (Ki-67) we could further subdivide the cells into different subsets. With LSC an semi-automated operator-independent analysis of lymph nodes and tonsils is possible. This technique should yield new insights into processes during diseases involving lymphoid organs and should help to quantify the success of therapeutic interventions.

 


6924

Coherent Multiprobes and Quantitative Spectroscopic Multimode Microscopy for the Study of Simultaneous Intracellular Events

Felix De La Iglesia, Jeffrey Haskins, Parke-Davis Pharmaceutical Research, Ann Arbor, MI USA; Daniel Farkas, University of Pittsburgh and Carnegie Mellon University, Pittsburgh, PA USA; Gregory Bearman, California Institute of Technology, Jet Propulsion Laboratory, Pasadena, CA USA

The cell is a complex set of organelles undergoing dynamic simultaneous changes in response to the extracellular environment. Cellular responses reflect toxicity or adaptation, and up to now there has not been a practical methodology that would allow monitoring of more than a single cellular activity. Analyzing several reactions at the same time in living cells provides a significant advance in the understanding of molecular interactions, enzyme interrelationships or organelle-to-organelle cooperations.Taking advantage of pre-existing technology around the concept of multimode microscopy, our efforts have been directed to add a multiprobe approach that addresses the issue of simultaneous or near-instant analysis of several cell functions.

Our coherent multiprobe methodology involves monitoring at least three compatible fluorophores in metabolically active liver cells. Human and animal hepatocytes are amenable to this purpose because of extensive experience in the isolation and characterization of subcellular functions. Using an adapted and expanded multimode microscope based on the Carnegie Mellon instrument, fluorescent probes were added in combination as a panel. An example of a panel includes Texas Red Phalloidin, MitoTracker Green and Fura-2. This panel helped to characterize membrane integrity, mitochondrial energetics and calcium transport simultaneously and in the presence of a liver cell toxicant.

This technology was successful in the simultaneous evaluation of coherent cellular probes. Capture of closely spaced time reactions and further enhancements could be significant improvements that could result in smaller image capture intervals. With this technique cellular reactions were better aligned and in vitro experiments with hepatocytes showed a definitive correlation between decayed activity and degradation of subcellular function. The instrument enhancement consisted of an acousto-optical filter system that replaced the mercury light source and associated filters anda computed tomography imaging spectrometer (CTIS). This CTIS device provided five to tenfold increased image capture efficiency without including potential benefits from design advances in the excitation-emission cube. This is the first successful application of quantitative spectroscopic techniques to multiprobe fluorescence in living cells. A significant advantage of this technique is that it provides near-instant information on the co-localization of probes in a given organelle, as well as the previous benefits already addressed above.

 


6757

Quantitative in situ hybridization using peptide nucleic acid probes: methodological aspects and applications in telomere biology.

Peter Lansdorp, Prakash Hande, Elizabeth Chavez, Steven Poon, Terry Fox Laboratory, BC Cancer Agency, Vancouver, BC, Canada

Directly labeled peptide nucleic acid (PNA) probes are powerful probes for in situ hybridization because these probes can hybridize to denatured DNA under high stringency conditions that destabilize DNA or DNA/RNA duplex molecules. As a result, PNA probes can be used for quantitative purposes by capturing fluorescence images with a digital camera mounted on a fluorescence microscope. In order to obtain reproducible results that can be used for quantitative measurements of telomere length, a number of variables need to be carefully controlled. These variables include the variation in light output of available light sources for fluorescence microscopes, the bleaching of probe fluorescence, uneven illumination of the microscope field, the focal plane of the fluorescent objects, the optical path (including coverslip, immersion oil, microscope slide and embedding medium containing antifade reagents, etc.). The influence of each of these variables should be tested experimentally in order to control inevitable variation between measurements to an acceptable and documented level. The use of fluorescence calibration beads and biological controls are essential in this process. Such calibration measurements are needed to relate integrated fluorescence intensity values to units of fluorescence or copy number of target sequences. The application of quantitative FISH for the measurement of telomere repeat sequences in human and murine cells and chromosomes has illustrated the power of quantitative FISH in biological studies. These studies have shown that efforts to control variation in fluorescence measurements from FISH slides are very much worthwhile and allow insight into biological processes that cannot be obtained in any other way.

 


6648

Automatic quantification of tumour angiogenesis: comparison of two staining procedures and three colour characterisation methods.

Kim Tran, Nicolas Elie, GRECAN University of Caen, Centre François Baclesse, Caen, France; Jehan-Pierre Signolle, Département Mesures Physiques IUT Caen, Caen, France; Abder Elmoataz, Teddy Coudrey, GREYC-ISMRA, Caen, France; Paule Opolon, Institut Gustave Roussy, Villejuif, France; Paulette Herlin, Centre François Baclesse, Laboratoire d’Anatomie Pathologique, Caen, France; Michel Coster, LERMAT - ISMRA, Caen, France

Angiogenesis is of clinical relevance since microvessel density has been proved to be a prognostic factor in many cancers. Nevertheless, because interactive, currently used methods do not allow a reproducible and rapid quantification of blood vessels(1); few fully automatic procedures are available(2). The aim of the present study is the development of such an automatic method. By use of a slide scanner (SprintScan 35 Plus, Polaroid), image analysis is applied to images of the whole tumour section of three cases of lymphoma and ten cases of ovarian carcinoma. Immunostained blood vessels are counted after a two-step procedure : segmentation and colour characterisation. Comparison with manual counting showed previously that colour characterisation is the most critical step, results depending probably on either to the staining technique used or the chosen method of colour detection (3). In order to improve blood vessel detection, two staining procedures (DAB or AEC) and three colour characterisation methods (colour translation, double thresholding on hue image and fuzzy logic classification) are compared with reference to manual counting. The mean percentage of blood vessels detected by the three image analysis methods (sensitivity : Se) is respectively 80%, 82% and 72% after DAB staining and 72%, 71% and 89% after AEC. The mean percentage of false-positive objects (FP) obtained by the three methods is 38%, 50% and 51% and 43%, 21% and 23% using DAB and AEC respectively. A quality factor Qt defined as Qt = Se x (100 - FP) / 100 shows that the performances obtained by the three image analysis procedures depends on the staining method used and that the most promising combination is given by AEC staining and fuzzy logic classification. 1. Hansen et al. (1998), Lab. Invest., 78 : 1563-1573. 2. Beliën et al. (1999), J. Clin. Pathol., 52 : 184-192. 3. Herlin et al. (1999), Anal. Cell. Pathol., 18 : 27. This work was supported by grants from the departemental comity of the Manche of the ‘Ligue Nationale deLutte contre le Cancer’.

 


8047

APPLICATIONS OF MULTISPECTRAL IMAGING IN PATHOLOGY AND CYTOLOGY

Richard Levenson, Clifford Hoyt, Cambridge Research and Instrumentation, Inc., Boston, MA, USA

Background: Color information, either for cell recognition or multiplexed signal detection, is typically appreciated visually or with a conventional RGB-type analogue or digital camera, thereby confining the spectral data to combinations of 3 hues (red, green and blue). However, spectroscopic imaging, with typical spectral resolution of approximately 10 nm per band, can demonstrate the presence of complex and useful spectral information in common pathology, immunohistology and cytology samples.

Design: Recently developed technologies, such as high-throughput liquid crystal tunable filters, along with Fourier-transform interferometry and acousto-optical tunable filters, can be used to generate a complete, high-resolution, optical spectrum at every pixel of a microscopic image. Elegant software tools have been developed to analyze the resultant data sets, including simple spectral similarity mapping, principal component and independent component analysis, and automated clustering algorithms in n-dimensions.

Results: Conventional hematoxylin-and-eosin- or Papanicolaou-stained pathology sections evidence a rich spectral behavior which can be used to delineate cells of different lineage or separate normal from neoplastic cells, using spectral "signatures" and simple segmentation algorithms. Spectral imaging can also be used to quantitatively detect the presence of at least 3 immunohistochemical chromogens even if they overlap spatially. Informative spectral signatures can be derived from a variety of cells stained with a single fluorescent or brightfield dye.

Conclusion: Multispectral microscopy may find applications in histopathology, cytopathology and multi-color immunohistochemistry.

 


6886

Spectral Pathology (SPY) - a novel method for histo- and cytological analysis

Yuval Garini, Applied Spectral Imaging, Migdal HaEmek, Israel; George Mcnamara, Applied Spectral Imaging

Spectral imaging was developed in the last few years and fully combines spectroscopy and imaging. By using a novel interferometer-based spectrometer that was designed for imaging purposes, the full visible light spectrum can be measured for every point of the image.

By applying spectral imaging to pathological samples, it is possible to acquire both structural and chemical information. We call this new method Spectral Pathology or simply SPY.

SPY combines spectral imaging, molecular pathology and bright-field or fluorescence microscopy, to enable the detection of multiple markers. It is well known that cancer is a genetic disease and can be identified by using various molecular probes. By using SPY, multiple chromogens or fluorochromes can be used simultaneously for staining histological or cytological specimens. The spectral images are than analyzed and provide accurate quantitative mapping of each one of the probes on the specimen.

For example, SPY should make feasible the multiplexing of several immunohistochemical stains on a single fine needle aspirate biopsy (FNAB), to help render a definitive diagnosis from limited clinical material. We have also found that SPY provides deeper insight even for common histological staining like Hematoxylin and Eosin (H&E). This may have advantages in facilitating the trend toward minimally invasive biopsies. The figure below demonstrates the ability of SPY to map Hematoxylin and Eosin (the two right images) from a specimen that was stained with both of these stains (left image). The mapping was preformed by using a special algorithm that we have developed called SUN (spectral unmixing).

In this talk we will present the physical and mathematical basics of SPY. Focusing on bright-field applications, we will also present few examples of histological stains (H&E, Papanicolaou), multiple immunohistochemical stains, and mixed histo- and immuno-staining.

 


6771

Data Annotation: The Bridge from Design to Interpretation

David Parrish, Management Science Associates

We will describe a prototype system that demonstrates a framework for the capture and format of information generated in a flow cytometric experiment, including experimental design, reagent selection, sample processing, analysis, and information storage. Our approach is extendable to other technologies and testing methodologies, most notably, molecular assay measurement. By associating the methodological and demographic information with experimental results the assay can be evaluated in the context for which it was designed. The foundation of our approach is based on using Extended Markup Language (XML) documents to serve as an information bridge between the front-end applications, the instrument, and the data archive. By establishing both the document type and standard definition applications, we will be able to retrieve information from or add structured data to the document. The prototype is built around three component applications: a protocol generator, instrument interface, and a structured archive.

PROGEN is a Java application that accesses and maintains a Relational Database Management System (RDBMS). The application will provide support and documentation throughout the experimental design, sample processing, protocol selection, reagent evaluation and staining procedure. When the samples are stained and ready for analysis the program will generate an XML document for each experiment, annotating the information about each protocol in that experiment. Examples of some of the information include institution, investigator, study, sample type, subject species, reagents used (e.g., clone, titer, expiration date) and staining method. The XML file will be passed to an application that parses and adds data about the instrument (e.g., settings and calibration) to the file. When complete, the raw data and the XML file are linked and sent to the archive where the XML form will be used to create a structured catalog of incoming files which are stored in the archive.

 


6902

A high resolution color scanning digital microscope: A new tool for telepathology

J. Paul Robinson, Purdue University, W. Lafayette, IN, USA

This paper describes a new technology in digital imaging. The COSMIC digital microscope is a "cameraless" white light-scanning microscope useful for providing rapid, high resolution images using a traditional bright field microscope base. The instrument based on a Zeiss Axiovert scope using 2.5, 10, 40 and 100x infinity-corrected objectives uses standard photomultiplier tubes (pmt) for detectors. Each detector collects a fixed bandwidth of light (300-490; 500-600; >600 nm via optical filters. There are five unique aspects of the microscope that should be noted.

The light source is a CRT oscilloscope providing 13.8 images per second,

The detectors are PMTs

The instrument contains a 3 x electronic zoom that does not require changing of objectives

There are extensive image processing options available directly on the scope

Images are saved directly over the Internet via 10 Base10 cable as TIF files

The microscope collects images as 24 bit 1280 x 1024 pixel images. Since there are no moving parts on the microscope there are no alignment issues, and no camera or computer compatibility issues to deal with. The microscope has no ocular - however because of the optical path, none is required to find or focus the specimen. All imaging is monitored on a screen. Because of the light path, it is necessary to ensure proper Köhler illumination to obtain sharp well-focused images. There are a number of unique aspects to the system which is ideal for telepathology applications. Presented are a series of images collected with high cost, and low cost color cameras compared with the COSMIC scope images. Also presented are the many variations of image processing available in the system, which allow for electronic versions of dark field, color reversal and gamma and contrast enhancements.


Parallel Session II: Microbiology

Chair: Chris Hewitt


6370

Rapid DNA Fingerprinting of Pathogens by Flow Cytometry

Erica Larson, Janetta Penttila, Hong Cai, James Jett, Stefan Burde, Richard Kellr, Babetta Marrone, Los Alamos National Laboratory, Bioscience Division, Los Alamos, NM, USA

A new method for rapid discrimination among bacterial strains and species based upon DNA fragment sizing by flow cytometry has been developed. This revolutionary approach combines the reproducibility and reliability of restriction length polymorphism (RFLP) analysis with the speed and sensitivity of flow cytometric measurements. For traditional bacterial RFLP analysis, genomic DNA is extracted from bacteria in an agarose plug and cut into defined fragments with a rare cutting restriction enzyme. The resulting population of fragments is analyzed by pulsed field gel electrophoresis (PFGE), which typically requires ~20 hours. The band pattern produced is a restriction digest fingerprint. For flow cytometric analysis, DNA fragment lengths are measured by recording the number of photons emitted by stained fragments as they pass through a focused laser beam. Major differences between this type of measurement and conventional flow cytometric measurements that enable ultrasensitive fluorescence detection are long transit times through the laser beam (~2 msec vs. 10 µsec) and data acquisition by photon counting. Data collected in 10 minutes from small amounts of DNA (~200,000 X lower than PFGE) are processed to generate a DNA fragment size distribution that displays the number of fragments in a size class versus the size of the fragment in absolute terms. Just as the PFGE fingerprints are discriminating, the flow cytometric measurements distinguish among bacterial species and strains. Modifications to the sample preparation protocol have reduced the time required from days to hours. Thus, once a fingerprint library is assembled, it will be possible to identify the species and strain of a bacterial sample in hours if it is in the library or to determine that is a new organism. This work supported by the Department of Energy Office of Nonproliferation and National Security (NN-20) and the NIH National Center for Research Resources Grant RR-01315 for the National Flow Cytometry Resource.

 


6504

Glucose Uptake Dynamics in Single Cells

Friedrich Srienc, Arvind Natarajan, Gortner Laboratory, St. Paul, MN, USA

The fluorescent glucose analog, 2-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)-2-deoxyglucose (2-NBDG) was used to measure with flow cytometry glucose uptake rates in single cells of Escherichia coli. The analog is taken up by the phosphoenolpyruvate phosphotransferase system through an active transport mechanism. It is intracellularly phosphorylated and accumulates, therefore, in the cells to large amounts that are readily detectable even in the small bacterial cells. The measured single-cell fluorescence levels are directly related to the uptake rates mediated by the transporter molecules. Therefore, the carbon flux as it enters the cellular metabolism, representing the first step in metabolism, can be quantitated in single cells. The substrate dependence of the uptake dynamics can be described according to Michaelis-Menten kinetics. Specificity of the glucose transporters to the analog was confirmed by inhibition of uptake of 2-NBDG by D-glucose, 3-o-methyl glucose, and D-glucosamine, and lack ofinhibition by L-glucose. Inhibition of 2-NBDG uptake by D-glucose is competitive in nature. This inhibition effect provides the basis for a novel assay for determining extremely low (submicromolar) concentrations of glucose. This novel single-cell uptake assay is useful for studying the physiology of cell and cell population growth as well as for the isolation of mutants altered in metabolism.

 


6633

Applications of Laser Scanning, Flow and Epifluoresecence Cytometry to the Rapid Detection of Bacteria and to Viability Assessment

Nathalie Parthuisot and Philippe Catala, Observatoire Océanologie; Phillipe Lebaron, Universíté P&M Curie

Flow cytometry is a well established technique for the analysis of many different cellular characteristics of microorganisms. Applications are growing fast concomitantly to the development of new fluorophores to analyse different aspects of cell functions making analytical and research applications similarly growing, offering new approaches for direct cell identification, viability and activity assessment.

During the last 5 years, we used a wide variety of fluorescent stains having different cellular targets and compatibles with compact flow cytometers with a single laser source, to develop rapid and alternative methods to overcome the limitations of cultural methods (time consuming, presence of viable but non-culturable cells,…). Rapid methods have also been developed to allow the real time monitoring of active and/or viable cell counts in aquatic ecosystems. Applications and comparisons of dyes used for industrial and environmental questions will be presented. Examples will include the comparison of different blue nucleic acid dyes for total counts of bacteria and CTC, ChemChrome V6, DiOC, oxonols dyes tested for activity assessment. Other dyes were also used for Gram staining and will be presented. The advantages and limitations of each staining procedure will be discussed based on examples with both pure cultures and natural water samples.

Moreover, flow cytometry does not apply to bacteria fixed on solid supports and neither to the detection of rare events. Recently, a new laser scanning device (LSD) called ChemScan has been developed to fill this gap. We will compare the quantitative and qualitative limits of flow cytometry, epifluorescence microscopy and laser scanning cytometry. This new LSD technique provides an efficient method for the detection of specific and/or viable bacteria present at low densities or diluted in an important background of non specific and/or non viable cells.

 


6664

Application of Flow Cytometry using Fluorescent Probes to Study Substrate and Product Toxicity in the Indene Bioconversion.

Ashraf Amanullah, University College London, Department of Biochemical Engineering, University College London, London, UK; Christopher Hewitt, The University of Birmingham, Centre for Bioprocess Engineering, School of Chemical Engineering, Birmingham, UK; Alvin Nienow, The University of Birmingham, Centre for Bioprocess Engineering, School of Chemical Engineering; Chanyong Lee, Michel Chartrain, Barry Buckland, Stephen Drew, Merck & Co. Inc., Bioprocess R&D, New Jersey, USA; John Woodley, University College London, Department of Biochemical Engineering, London, UK

The chemical synthesis of the key intermediate for Merck’s HIV protease inhibitor, CRIXIVANTM, cis-(1S, 2R)-1-aminoindan-2-ol [CAI] in enantiomeric pure form, is highly technically demanding. Alternatively, CAI can be directly synthesized using either cis-(1S,2R) or trans-(1R,2R) indandiol. The latter compounds can be made using a bioprocess strategy involving the bioconversion of indene. A bioconversion process is attractive because it is possible to control the chirality of the final product. However,the biotransformation of indene is characterised by low product titres and yields with the production of various undesirable mono-oxygenation by-products such as 1-indenol and 1-indanone. The dioxygenation product, cis-indandiol, is also further oxidised to 1-keto-2-hydroxy indan via dehydrogenase activity. The reasons for the poor performance are not well understood, although it is considered that the product and by-products exhibit toxic effects on the culture. In terms of process improvement, one objective may be to construct metabolic pathways which allows the conversion of indene to cis-(1S,2R) indandiol at high yields and prevents its further degradation to 1-keto-2-hydroxy indan or to engineer pathways directly from indene to CAI. Our interest stems from the need firstly, that a better understanding of the toxicity effects of the substrate and products on cell physiology and metabolic activity is essential if a suitable metabolic pathway is to be designed. Secondly, since the bioconversion of indene to indandiol can be conducted using Rhodococcus sp., Pseudomonas sp. and recombinant E.coli sp., the choice of a potential microbial host to carry out this modified bioconversion has to be made on a rational basis. Thirdly, a better understanding of cell physiology in such biotransformations will also help to devise suitable indene feeding strategies to improve process performance. This paper reports on the application of multi-staining flow cytometry as a tool for implementing improvements in the bioprocess. Specifically, cell physiological state (cell membrane integrity and cell membrane depolarisation) and metabolic activity (oxygen uptake rate and optical density) in response to the possible toxic effects of both substrate and products was studied under fermentation conditions. The results provide information on the suitability of the different species as a suitable host to carry out the bioconversion. The toxicity results indicate the sub-toxic levels up to which it is possible to accumulate the desired intermediate compounds by engineering metabolic pathways directly from indene using a model host. The toxicity information has also been used for implementing a suitable indene feeding strategy in the bioconversion.

 


6722

Multi-parameter flow cytometry: Assessment of bacterial physiological state and its application to the study of the scale-up of bacterial fermentations.

Chris Hewitt, University of Birmingham, School of Chemical Engineering, Birmingham, UK; Gerhard Nebe-Von-Caron, Unilever Research, Colworth Laboratory, Bedfordshire, UK; Alvin Nienow, University of Birmingham, School of Chemical Engineering, Birmingham, UK

Fluorescent staining methods developed in our laboratories have lead to a functional classification of the physiological state of individual bacterial cells based on reproductive activity, metabolic activity and membrane integrity. The multi-colour staining approach has improved our understanding of various dyes and has also highlighted the interference of dye efflux systems, a major cause of the erroneous interpretation of the mode of action of many fluorescent stains. Efflux systems relying on non-specific proton anti-port pumps can be harnessed and a deliberate measurement of such, can be used to discriminate for cells generating a proton gradient. Used in combination with membrane potential and membrane integrity stains this allows the simultaneous differentiation of four functional sub-populations in bacterial populations. It is therefore possible to resolve an individual cells physiological state beyond culturability, based on the functionality of these metabolic pumps and the presence or absence of an intact polarised cytoplasmic membrane, enabling assessment of population heterogeneity. Importantly, the reproductive growth capacity of intact cells can be demonstrated by cell sorting. Such multi-staining flow cytometric techniques have been used for the ‘at-line’ study of the scale-up of bacterial fermentations. Scale-up is usually the final step in a research and development programme leading to the large scale synthesis of microbial products by fermentation. It is known that chemical gradients exist in large scale bio-reactors where additions of a concentrated, often viscous, substrate at a single point mean that mixing times are high (>50s). In laboratory scale bio-reactors where much development work is done and mixing times are low (<5s)such gradients do not exist. The composition of a cells micro-environment is a product both of fluid dynamics and a cells physiological response to it. Cells circulating around a large scale bio-reactor will experience rapidly changing micro-environments. Knowledge of how a cell reacts to such changes is essential if mathematical models are to be derived that accurately predict process times and product yields on scale-up. It has been shown using multi-staining flow cytometry that a changing microenvironment with respect to substrate concentration has a profound effect on cell physiology and hence viable biomass yield on scale-up. The relatively poorly mixed conditions in a large scale reactor leads to a low biomass yield with a high viability and small scale simulations of poor mixing gave similar results. However, under the normal well mixed small scale condition the converse was found to be true.


Parallel Session II: Hematopoiesis/Stem Cell Biology

Chair: Ole Didrik Laerum


4393

Fas/CD95 expression level after stem cell transplantation: a simple parameter to monitor immune reconstitution.

Thierry Lavabre-Bertrand, Jean-Paul Bureau, Laboratoire de Biologie Cellulaire, Nîmes, France; Thierry Levayer, Carole Exbrayat, Service d’Hématologie et d’Oncologie médicale, Hôpital Lapeyronie, Montpellier, France; Philippe Poncelet, Isabelle Besson, Biocytex, Marseille, France; Nathalie Fégueux, Jean-François Rossi, Service d’Hématologie et Oncologie médicale, Hôpital Lapeyronie, Montpellier, France; Laboratoire de Biologie cellulaire et Cytogénétique moléculaire, Faculté de Médecine de Montpellier-Nîmes, Nîmes, France

T cell reconstitution after bone marrow transplant (BMT) is characterized, for at least 1 year, by the expansion of T cells with a primed/memory (CD45RAlow/CD45R0high) phenotype. In addition, these cells show a marked propensity to spontaneous and Fas-dependent apoptosis in vitro (1). The distribution of Fas/CD95 staining is also abnormal with a vast majority of T cells displaying a CD95high phenotype. By measuring the expression of Fas/CD95 using quantitative flow cytometry we show that not only the frequency of CD95high T cells but also the mean level of expression of the molecule CD95/Fas, expressed as molecules per cell, is increased in transplanted patients, and that the return to normal levels may require 2 years or more. CD95 expression level was measured using CellQuant Fas kit, (BioCytex, Marseille, F) on CD3-PE+ T cells from patients treated with autologous peripheral blood stem cell transplantation (PBSCT) or allogeneic BMT (aBMT) and compared to age-matched normal donors. The mean level of CD95/Fas on all T cells, a new parameter sensitive to both antigen density per cell and relative proportions of CD95low vs CD95high T cells, was slightly increased vs controls in the few (n=5) tested aBMT patients and highly increased in PBSCT patients (cf. table). The mean level of CD95/Fas on T cells was negatively correlated with the time elapsed from transplantation. When CD3+/CD95low and CD3+/CD95high subpopulations of T cells could be resolved (in 7 / 21 PBSCT patients), CD95 overexpression was noted on both . Higher Fas levels were also noted on both CD45RAlow and CD45RAhigh lymphocytes, suggesting that increased Fas antigen density occurs in addition to the previously described elevated proportions of CD95high T cells.

Thus CD95/Fas expressionon T lymphocytes is abnormally high after BMT until immune reconstitution is completed, not only in terms of frequency of CD95high T cells but also of CD95 antigen density. This suggests that quantitation of CD95/Fas expression should be included in the routine evaluation of immune reconstitution after BMT.

(1) Hébib NC et al. (1999), Blood, 94:1803-13.

 


6235

FLOW CYTOMETRIC HIGH THROUGHPUT SCREENING FOR OPTIMIZING SELF RENEWAL OF PRIMITIVE HUMAN HEMATOPOIETIC CELLS

Clay Smith, Tracy Gentry, Duke University Medical Center, Durham, NC USA

Defining culture conditions which support self-renewal of primitive human hematopoietic stem cells (HSCs) is central to developing successful ex-vivo stem cell expansion and gene therapy strategies for human diseases. In order to systematically and efficiently evaluate the effects of the enormous number of potential combinations of factors which may influence HSC self renewal, we have developed and validated a screening technique, termed flow-high throughput screening (F-HTS), which combines the efficiency and accuracy of automated high throughput techniques with the ability of flow cytometry to measure cellular processes. To evaluate HSC self renewal, small numbers of cells were cultured in 96 well plates under a wide range of different conditions and then each sample was analyzed with FACS for markers of cell proliferation and self renewal. A computer controlled robotic fluid handler was utilized for cell culture, reagent addition and cell staining procedures. A second automated device was used to deliver samples from the 96 well plates to the flow cytometry. FACS data analysis was performed using a software program which facilitates batch analysis of large volumes of flow cytometry data. In an initial test of the F-HTS technique, the effects of Flt-3Ligand (Flt-3L) and IL-3 alone and in combination over a wide range of concentrations were evaluated on HSC self-renewal. The results demonstrated that [Flt-L] >1ng/ml and [IL-3] >5 ng/ml were optimal for proliferation and self renewal of primitive cordblood cells. Since retroviral vector gene transfer is absolutely dependent on cell proliferation, we next tested whether the results of the F-HTS assay would predict the outcome of gene transfer experiments. The optimal concentrations of Flt-3L and IL-3 for retroviral vector gene transfer into cells retaining HSC properties was similar to those defined in the F-HTS assay. Studies are ongoing using the F-HTS technique to evaluate other compounds which may enhance HSC self renewal, expansion and genetic modification. In summary, the F-HTS is more efficient than manual culture and analysis techniques and accurately predicts the results of long term studies. Adaptations of the F-HTS technique may also be useful in other cellular engineering and drug discovery applications.

 


6329

Clinical Interest of a combination of two standardized flow cytometry assays for paroxysmal nocturnal hemoglobinuria diagnosis.

Uta Oelschlaegel, Yvonne Bux, G. Ehninger, Universitatsklinikum Carl Gustav Carus, Dresden, Germany; Isabelle Besson, BioCytex, Marseille, France

Paroxysmal nocturnal hemoglobinuria (PNH) is an anemia caused by complement-mediated hemolysis. It is an acquired, clonal disorder of the hematopoietic stem cells associated to an absence of glycosylphosphatidylinositol (GPI) anchor synthesis. The lack of GPI anchor leads to a lack of the cell surface expression of at least 16 membrane proteins including complement regulatory proteins such as Decay-Accelerating Factor (CD55) and Protectin (CD59). The protein defect is observed on erythrocytes (RBC), granulocytes (PMN), monocytes, lymphocytes and platelets. Various tests are used in routine for the diagnosis of PNH. The Ham’s test and the sugar water test determine the RBC sensitivity to the complement-dependent lysis. Both tests can only screen for the protein defect on RBC, but not on other cells. Several flow cytometry assays have been proposed for the study of GPI anchored molecules. All these tests require the drawing and testing of normal samples as control and interlaboratory interpretation is made difficult by the absence of standardization. A combination of two simple, rapid and standardized FCM assays for the detection of CD55 and CD59 deficient PMN (CELLQUANT CD55/CD59 kit, Beckman Coulter) and RBC (REDQUANT CD55/CD59 kit, Beckman Coulter) is proposed. Thus, 2 different cell types and 4 parameters are analyzed on each sample. Both kits use precalibrated beads to standardize the cut-offs between normal and deficient cell subpopulation. These cut-offs avoid recruiting control samples and provide result reproducibility over-time and across laboratories. The test has been validated on a series of 70 EDTA blood samples from patients without hematological abnormalities (n = 48) and from patients with hematological disorders (n = 22) including confirmed PNH (n = 6). The sensitivity and the specificity of the proposed tests for PNH diagnosis were assessed by defining the negative predictive value (NPV) and the positive predictive value (PPV). The combination of REDQUANT CD55/CD59 and CELLQUANT CD55/CD59 showed a 100% sensitivity (NPV) and a 98% specificity (PPV) whereas the assay performed only on RBC or only on the PMN failed to be effective (67% NPV and 93% PPV, 100% NPV and 88% PPV, respectively). These results confirm the need to evaluate at least two different markers measured on two different cell types as a diagnostic panel for PNH laboratory assessment. In addition, the same test association may be indicated for the monitoring of treated PNH patients. Supportive data are shown on a PNH patient tested three times during one year of therapy. In conclusion the combination of REDQUANT CD55/CD59 and CELLQUANT CD55/CD59 kits is suitable for both the exclusion diagnosis of PNH and as laboratory monitoring tools for PNH therapy.

 


6435

Colony formation by flow sorted hemopoietic stem cells from C57bl mice: Interactions with growth regulatory oligopeptides

Ole Didrik Laerum, The Gade Inst., Dept. of Path. L, Bergen, Norway; Ger Van Den Engh, Department of Molecular Biotechnology, University of Washington, Seattle, WA, USA

The purpose of the present study was to investigate isolated hemopoietic stem cells which have a high proliferative potential (HPP)in vitro and how three different inhibitory oligopeptides interacted with their colony formation. Differentiated cells were removed from bone marrow cell suspensions of female C57Bl mice (femurs and tibiae) by use of magnetic beads and specific antibodies, and a stem cell enriched cell fraction (scal+, lin- cells) was obtained by immunofluorescence and high-speed cell sorting. A HPP-CFC colony assay with 13 days colonies above 0.5 mm was used to test the effects of the inhibitors hemoregulatory peptide (pEEDCK), goralatide (AcSDKP) and MIP 1 alpha. The 3 peptides inhibited colony formation by 15-30 % in concentrations from 10-9 to 10-7 M. No further effect was seen with combination of two or all 3 peptides. The inhibitory effect was released by growth factor secretion of stromal cells (M-CSF and a GRO beta chemokine) upon stimulation of a peptide secretory factor (hemoregulatory peptide dimer), indicating an interplay with stromal cells in the action of stem cell inhibitory peptides.

 


6466

Data Pattern Analysis of Flow Cytometric Blood Leukocyte Immunophenotypes of Donors and Bone Marrow Transplant Recipients

Guenter Valet, Cell Biochemistry Laboratory, Martinsried, Germany; Hanna Kahle, Max-Planck-Institut für Biochemie, Martinsried, Germany; Joost Van Esser, Jan J. Cornelissen, Dept. Hematology Daniel den Hoed Cancer Center Univ. Rotterdam, The Netherlands; C.H.J Lamers, Dept. Clinical & Tumor Immunology Daniel den Hoed Cancer Center Univ. Rotterdam, The Netherlands; R.L.H. Bolhuis, Jan Gratama, Dept. Clinical & Tumor Immunology Daniel den Hoed Cancer Center Univ. Rotterdam, The Netherlands

Aims: The conventional analysis of flow cytometric list mode data from immunophenotyping (IP) assays uses only a fraction of the available information e.g. by selecting lymphocytes and discriminating cells with positive or negative fluorescence (FL). An exhaustive, self-learning strategy was applied in this study to 53 list mode data file sets of healthy donors (HD) and recipients of bone marrow transplants (BMT) with the aim to select antibody (Ab) combinations with maximum discrimination for predictions of clinical complications from IP data in future studies. Methods & Results: List mode files of antibody (Ab) combinations CD3/16+56/45, TCRab/CD8/45, TCRgd/CD4/45, CD57/HLA-DR/TCRab, CD8/95/TCRab, CD45RO/27/4, CD45RO/62L/4 were analyzed with the CLASSIF1 program(Cytometry(CCC) 30:275-288(1997), http://www.biochem.mpg.de/valet/classif1.html). Lympho, mono- and granulocytes(LMG) were autogated on FSC/SSC with >95% of these cells included. FL1/FL2, FL2/FL3, FL1/FL3 histograms of LMG were evaluated as FL positive and negative quadrants with fixed thresholds at 1/3 of the 4 decade logarithmic scales. Cell frequency (%), mean Ab FL on abscissa and ordinate as well as FL ratio and relative FL packing densities were calculatedfor each quadrant. 74 data columns per cell population provided 222 columns for LMG and 7x222=1554 columns for 7 three color assays. The entire databased information was subjected to automated CLASSIF1 analysis. The classification of the total information permitted correct identification of HD and BMT patients between 31-68d, 77-95d and 105-159d post BMT in 100.0, 88.9, 83,3 and 75.0% (n=28/9/12/4). Similar results of 96.4, 77.8, 91,7 and 100.0% were obtained with only the CD8/95/TCRab information on LMG, indicating a significant information redundancy (IR) for the six other Ab panels. Classification of the information from all transplanted patients (31-159d) against HD permitted a 100.0 and 92.9% recognition (n=25/28) for all Ab combinations on LMG with identical results for the single CD45RO/27/4 LMG analysis. HD could be distinguished from matched related donors (MRD) in 84.4 and 100.0% of the cases (n=13/14) using all Ab information from LMG with 100.0 and 86.7% correct identification for the single CD8/95/TCRab LMG analysis. Conclusions: The self learning classification of three color IP blood leukocyte data discriminates between HD and MRD before BMT as well as the post BMT time periods. The detected IR of Ab combinations will permit panel simplification. The results are encouraging for laboratory independent data classifications by telecytometry. This will be essential for the consensus driven international optimization of experimental and clinical multicolor IP panels.

 


6516

Clonal culture of the mouse hematopoietic stem cells for the proof of self-renewal in vitro

Hiromitsu Nakauchi, Yo-Hei Morita, Hideo Ema, Institute of Basic Medical Sciences, University of Tsukuba, Department of Immunology, University of Tsukuba and CREST(JST), Ibaraki, Japan

We have previously shown a long-term hematopoietic reconstitution by single CD34-negative, Kit-positive, Sca-1-positive, and lineage marker-negative (CD34—KSL) mouse bone marrow cells. Using CD34-KSL cells, we directly addressed whether self-renewal occurs when they undergo cell divisions in vitro. Single CD34-KSL cells were cultured under serum-free condition in the presence of Kit ligand and thrombopoietin. After days of culture, a variety of numbers of progeny were generated from single CD34-KSL cells. To examine a correlation between the number of cell divisions and presence of repopulating ability in proliferated cells, groups of graded numbers of cultured cells derived from single cells were transplanted into irradiated congenic mice together with 2x105 syngeneic bone marrow cells. After transplantation, peripheral blood mononuclear cells were obtained and assayed for the presence of donor derived cells. Significant levels of repopulating activity by the donor cells was detected in some recipients of 2 to 16 cultured cells, but none in the recipients of more than 17 cultured cells. Since CD34-KSL cells must have divided at least once, these data provide evidence for in vitro self-renewal of CD34-KSL cells. Under the culture condition examined, however, they lost self-renewal capacity after maximum of 4 cell divisions. Number of cell divisions, rather than the length of culture appeared to be critical in preserving the self-renewal capability in vitro.

 


6681

COMPARISON OF TWO SINGLE PLATFORM METHODOLOGIES WITH EITHER NUCLEIC ACID DYE OR VIABILITY DYE FOR ENUMERATION OF HEMATOPOIETIC PROGENITOR CELL IN BLOOD-DERIVED SPECIMENS

David Bossy, Joëlle Hirn, Immunotech, Marseilles, France; Jamila Tortel, Beckman Coulter Immunotech, Marseilles, France

We compared the use of Stem-Kit CD34+ HPC Enumeration Kit and 7-AAD Viability Dye for HPC enumeration on the Coulter EPICS® XL-MCL flow cytometer to the ProCOUNT Progenitor Cell Enumeration System available on B-D® FACS®. The HPC absolute counts obtained with both single platform methodologies are not significantly different for the fresh mobilized peripheral blood and cord blood tested, but the HPC counts differ significantly for one third of the fresh apheresis tested and for all thawed apheresis and fresh bone marrow tested. ProCOUNT software gives QC messages and no count for a majority of the thawed specimens and fresh bone marrow while, for the same specimens, a HPC count can be obtained with Stem-Kit + 7-AAD on EPICS XL with System II software. The obtained results suggest that viability assessment is a critical parameter for HPC enumeration in frozen/thawed specimens. Moreover, the association of viability assessment to the versatile ISHAGE-based gating strategy allows to obtain a consistent HPCenumeration whatever the specimen to be analyzed.


Parallel Session II: Nano-/Micro-Particle Technology/Rare Cell Detection

Chair: Larry Sklar


6046

Solubilization and display of seven transmembrane receptors on beads for real-time fluorescence and flow cytometric analysis

Larry Sklar, Janeen Vilven, Eric Lynam, Donna Neldon, Teresa Bennett, Eric Prossnitz, University of New Mexico Health Sciences Center and NFCR, Albuquerque, New Mexico, USA

G protein coupled receptors (GPCR) and cellular signaling elements are prime targets for drug discovery. Sensitive real-time methods that expand the analytical capabilities for these elements can play significant roles in basic research as well as drug discovery. We describe in the present work novel approaches for the real-time fluorescence analysis of G-protein coupled receptors. Using the G protein-coupled N-formyl peptide receptor (FPR) as a model system in concert with a fluorescent ligand, we showed the quantitative solubilization of his-tagged FPRs in 1% dodecyl maltoside (DOM). Solublized receptors reconstitute in DOM with a mixture of bovine brain Gi/Go showing an apparent Kd ~ 10 -7 M. Solubilized receptors also bound to Ni2+ -silica particles and were detected in a flow cytometer by the binding of fluorescent ligand. The efficiency of receptor uptake by the particles was in excess of 80% with an apparent affinity for the bead in the nM range. The receptors had largely homogeneous dissociation characteristics, an appropriate Kd for the ligand in the low nM range and a high site number, with several million receptor molecules per particle. These approaches for displaying receptors could prove useful in drug discovery as well as in the analysis of the molecular assemblies in signal transduction and termination.

 


6474

Multiplex analysis of mutations in the RET-gene using suspension arrays

Thomas Langmann, Jochen Kilwinski, Christine Von Landenberg, Karl Lackner, Gregor Rothe, Gerd Schmitz, University of Regensburg, Regensburg, Germany

Germline mutations in exons 10, 11, 13, 14, and 16 of the tyrosine kinase receptor gene RET are responsible for a number of inherited diseases including the cancer sydromes multiple endocrine neoplasia types 2A and 2B (MEN 2A and MEN 2B) and familial medullary thyroid carcinoma (FMTC). Individuals in MEN 2A families carrying the defective gene have a high risk of developing medullary thyroid cancer. DNA analysis has provided a highly specific, reliable and less-intrusive alternative to aid diagnosis. DNA sequencing technology, PCR-RFLP and SSCPs are being widely used but are time consuming. New technology systems with simultaneous amplification and detection of mutation in glas capillaries are very fast but not suitable for multiplex analysis, while DNA-chips are very attractive but not affordable for everybody. We used an alternative multiplex technology that is a combination of flow cytometry, fluorescent bead technology and PCR. Mutation-specific, biotin-labeled oligonucleotides from the Ret-gene have been coupled to fluorescenty dyed, avidin-microspheres and serve as specific detectors for mutations in RET PCR-products. For each specific bead, a complementary oligonucleotide with Cy3 (reporter) is included in the reaction hybridization and the fluorescence is monitored in the Luminex100. Reporter molecules preincubated with appropriate RET PCR-products (competitor) are not accessible to the microspheres and leads to a fluorescence inhibition. We have evaluated this multiplex technology using detector beads for mutations in the Ret-codons 609, 611, 618 and 620 in combination with exon 10-specific PCR-products as competitors and were able to discriminate mutant from wildtype alleles. This powerful technology can be adapted for any application with new and cheap instrumentation.

 


6383

NOVEL METHOD FOR MULTIPLEX HIGH-THROUGHPUT FLOW CYTOMETRIC ANALYSIS

Kodumudi Venkateswaran, Fred Milanovich, Richard Langlois, Lawrence Livermore National Laboratory, Livermore, CA, USA

Flow cytometry is inherently capable of multi-parameter analysis. Flow microsphere immuno assays are comparable in sensitivity with enzyme immuno assays. The sample throughput in enzyme immuno assays in microtiter format is 96 samples in one minute using most of the colorimetric plate readers. Since sample analysis is done serially in flow cytometer, it takes longer time for the analysis of large numbers of samples. Multiplex analysis of up to 64 different analytes using optically-coded microspheres havebeen demonstrated. For increasing the sample throughput to eightfold, recently a flow plug format was reported. We hypothesize a novel method to increase the sample throughput of flow cytometric analysis up to 100 fold by using an array of fluorescent microspheres. Each sample is uniquely coded with a microsphere bead set having the same capture antibody. The immuno assay was performed on individual bead sets, then all the bead sets are pooled and analyzed in a single flow cytometric run. With the possibility of 100 optically encoded bead sets available, it is conceivable that 100 samples can be analyzed within a minute by flow cytometry, comparable to that of enzyme immuno assays in microtiter format. To test this hypothesis, we used an immunoassay for the detection of bacterial spores using 15 different bead sets. Results of the 15-plex bacterial spore detection and the optimal conditions to get the same resolution as the individual sample analysis will be presented. Use of this approach for the high throughput multiplexed PCR amplicon detection will also be discussed. (Work by LLNL performed under the auspices of US DOE contract W-7405-ENG-48)

 


6896

Simultaneous Quantitation of Six Human Cytokines by Microparticle Flow Cytometry

Rudi Varro, Roy Chen, Becton Dickinson Biosciences, San Jose, CA, USA; Lawrence Lowe, Becton Dickinson Biosciences, 2350 Qume Dr, San Jose, CA 95131, USA; Jerry Wilson, PharMingen, San Diego, CA 92121, USA; Edward Morgan, PharMingen, San Diego, CA 92121, USA; Tzeggai Kifle, Becton Dickinson Biosciences, 2350 Qume Drive, San Jose, CA 95131, USA; Frank Vegh, Becton Dickinson Biosciences, 2350 Qume Drive, San Jose, CA 95131, USA

We have developed a microparticle based flow assay system, which is compatible with the Becton Dickinson flow platforms. The system contains a bead set of seven FL3 intensities, each of which can support an individual analyte. It also includes a Mac compatible analysis software, which is analysing imported FCS files, collected by CellQuest.

Here we describe the application of this system, called BD Cytometric Bead Array system, to determine six human cytokines concurrently from a single culture supernatant or serum sample. The six cytokines are human IL2, IL4, IL5, IL10, TNFa and IFNg. Each individual capture antibody is covalently coupled to one discrete intensity FL3 bead. The six detector antibodies are directly labeled with PE. The assay is performed by mixing the six cytokine capture beads in equal amounts, and combining 50 ul of the beads with 50 ul culture supernatant or serum samples. 50 ul of the combined detector antibody PE conjugate is added, and the assay mix is incubated for 2 hours. Dilution series of a freeze-dried, combined cytokine standard are used to generate six simultaneous standard curves, using the same assay format. After data acquisition the FCS files are imported to the analysis software, which constructs a standard curve for all six analytes, and then assigns the cytokine concentration to each sample.

The assay performance (detection limit, accuracy, precision) of this combined assay is comparable to the performance of the individual ELISA assays.

The assay is a useful method to characterize Th1/Th2 responses, and examples will be presented on its use with different body fluids.

 


6609

Flow cytometry with excitation into surface plasmon resonance bands of antibody-linker-gold or silver nanoparticle-aminodextran-polystyrene bead conjugates as white blood cell markers

Olavi Siiman, Kristie Gordon, Alexander Burshteyn, John Maples, Beckman Coulter, Inc., Advanced Technology, Miami, Florida, USA

Gold(Au) nanoparticle-, silver(Ag) nanoparticle-, and aminodextran-polystyrene(PS) latex bead conjugates with antibodies, that have the potential to allow simultaneous enumeration of three mutually exclusive cell populations at one time with two laser excitation lines, were prepared. The linkers between antibody and colloidal metal were an aminotrithiol ligand or aminodextran. A modified flow instrument, including light scatter at other intermediate angle ranges, LMALS (10-20 deg) and UMALS (20-65 deg), together with DC and rf conductivity was used. Two mutually exclusive cell populations, CD4+ and CD8+ lymphocytes, have been simultaneously enumerated in whole blood by using a mixture of CD4-PS, CD8-Au-PS or CD4-Au-PS, CD8-PS beads and one laser line, 633nm excitation. Similar measurements were made with mixtures of CD4-PS, CD8-Ag-PS or CD4-Ag-PS, CD8-PS beads.

 


5881

The role of genetic variation in cervical cancer

Alina Deshpande, Los Alamos National Laboratory, Biosciences Division, Los Alamos, NM, USA; Cosette Wheeler, University of New Mexico, Albuquerque, NM, USA; John Nolan, Life Sciences Division, Los Alamos, USA

Cervical cancer is the second leading cause of cancer related death in women. A persistent infection with Human papillomavirus (HPV) is now known to be the primary risk factor for this disease. Susceptibility of an individual to developing cervical cancer is largely attributed to the type of HPV infecting the cervix. However, there is increasing evidence for a role of host factors in susceptibility/resistance to cervical cancer. Single nucleotide polymorphisms (SNPs) are the most common form of genetic variation in the human genome, and have been shown to be associated with suceptibility to a wide variety of diseases. The genes in the major histocompatibility complex (MHC) which are responsible for individual immune response are highly polymorphic in nature, and these polymorphisms may affect the nature of the immune response to an HPV infection, thereby affecting the progression to cervical cancer. We have developed a sensitive and rapid flow cytometry(FC) assay for scoring SNPs. The assay is a variant of the polymerase mediated primer extension process (also known as minisequencing) adapted to microspheres for analysis by flow cytometry. A tagged primer anneals directly upstream of the site of interest on the sample DNA and is extended by a single fluorescently labeled dideoxynucleotide triphosphate using a thermostable DNA polymerase. The incorporated base reveals the identity of the nucleotide at the site of interest on the template strand. The extended primers are then captured onto microspheres and analyzed by flow cytometry. The assay is multiplexed using commercially available arrays of differently stained fluorescent microspheres that enable the simultaneous interrogation of multiple polymorphic positions in a single sample. We are using this technology to investigate the role of SNPs present in HLA associated genes and the HLA B exon 2 locus, in susceptibility/resistance to cervical cancer. HLA associated genes encode proteins that are involved in antigen processing and presentation. SNPsin these genes, together with SNPs in the HLA class I molecules could affect the nature of peptide presented by the HLA class I molecules thereby affecting the host immune response. We will present genotyping data at the above-mentioned genes from patient and control samples.

 


6812

Rare event detection in human genetic biomonitoring. Enumeration and sorting of micronucleated nascent peripheral blood reticulocytes by immunomagnetic separation and flow cytometry.

Jan Grawé, Stockholm University, Dept. Genetic and Cellular Toxicology, Stockholm, Sweden; Lilianne Abramsson Zetterberg, Gösta Zetterberg, Uppsala University, Department of Conservation Biology and Genetics, Uppsala, Sweden

The rodent bone marrow or peripheral blood erythrocyte micronucleus test is the most used in vivo assay for chromosome damage induced by genotoxic agents. Adaptation of this assay to flow cytometric analysis has rendered it the by far most sensitive test for induced chromosome damage (Grawé et al. Mutat Res 405:199-208, 1998; Abramsson-Zetterberg et al. Int J Radiat Biol 67:29-36, 1995).

The sensitivity and simplicity of this assay would make it a prime candidate for human biomonitoring of induced chromosome damage, but for the fact that in humans, access to bone marrow samples is severely limited, and in peripheral circulation micronucleated erythrocytes are virtually eliminated by the spleen within an hour of their entrance.

We have developed a method to isolate extremely young reticulocytes from peripheral blood samples by immunomagnetic separation based on the residual presence of transferrin receptors on the retics. The mean age in the peripheral circulation of the isolated population (Trf-reticulocytes) is about 10 minutes, and the frequency of this population among all erythrocytes is about 6x10E-5. This population was stained for RNA and DNA content using Thiazole Orange and Hoechst 33342 and analyzed by flow (figure 1).

For a group of healthy donors the baseline frequency of micronucleated cells (MN-Trf-reticulocytes) in the isolated population varied between 0.3 and 3.5 x10E-3, which is similar to baseline values found in rodents. The baseline frequency of the cells of interest among all erythrocytes is thus in the range of 10E-7 - 10E-8. Flow sorting of these revealed that the combination of immunomagnetic separation and flow analysis identifies these cells with 90%+ precision.

Baseline frequencies were found to correlate with donor age and folate status.

Cancer patients treated with cytostatic drugs or radiation and were sampled before and after treatment. For patients where significant exposure of red bone marrow was expected, a strong elevation of MN-Trf-reticulocyte frequency was seen after treatment.

An assay for MN in human erythrocytes could be of great use for the following reasons: uncomplicated sample collection and processing; flow cytometric enumeration gives high sensitivity and possibility to sort large numbers of MN for further characterization eg by FISH; the kinetics of erythrocyte production allows repeated sampling of individuals, e g before and after exposure; and there is a well validated experimental counterpart in the mouse micronucleus test.

The present data indicate that this MN-assay is a promising endpoint for human monitoring.

 


6257

Detection of Circulating Tumor Cells in Peripheral Blood Using Immunomagnetic Enrichment and Automated Cellular Imaging

Blaise Bossy, Karen Yamamoto, William Decker, Michal Peri, Jose Torre-Bueno, Douglas Harrington, Kenneth Bauer, ChromaVision Medical Systems, Inc., San Juan Capistrano, CA, USA

BACKGROUND: The identification of very rare (frequency <= 10e-8) circulating tumor cells in peripheral blood represents an assay with considerable potential clinical utility. This study overviews our experience in devising a system consisting of the combination of tumor cell enrichment (immunomagnetic cell separation), immunocytochemical staining using anti-cytokeratin monoclonal antibodies, and analysis using the Automated Cellular Imaging System (ACIS). The accuracy of the ACIS for circulating tumorcell detection is facilitated by requiring review (and editing, if necessary) of collected cellular images. In addition, the ACIS allows for image comparisons from previous instrument runs (montage compare tool), providing highly consistent between-run classification of cellular objects. RESULTS: Preliminary studies were performed in which 20cc of peripheral blood was spiked with 1-100 carcinoma cells. These studies indicate an assay detection sensitivity of approximately one tumor cell/ 100 million normal hematopoietic cells. Studies comparing ACIS-assisted analysis with manual microscopy revealed that ACIS analysis provides tumor cell detection sensitivity which supersedes that provided by manual microscopy. Replicate analysis of specimens showed that ACIS-assisted tumor cell detection is highly reproducible, whereas comparatively variable results are obtained by manual microscopy. Initial studies of human breast cancer specimens revealed that 4 of 4 cancer patients demonstrated the presence of tumor cells in peripheral blood, whereas 8 of 8 normals demonstrated no tumor cells. A double-blinded study involving 75 breast cancer patients and 25 controls was next performed to relate the presence of circulating tumor cells to stage of disease. Finally, quality control guidelines aimed toward assuring consistent system performance are detailed. CONCLUSION: In conclusion, an analysis system for rare tumor cell detection using the combination of immunomagnetic cell enrichment, immunocytochemical staining, and analysis by automated cellular imaging is described. This system provides highly sensitive and reproducible results. The analysis platform also facilitates cell classification decisions based on established morphologic criteria and provides a convenient means for documenting and archiving data. The system, including quality control procedures, has matured to the point that it can be applied in the clinical laboratory setting. Current clinical feasibility studies focus on assessing the system’s potential for cancer screening and disease monitoring.


7:00 pm – 9:00 pm

Exhibit Hall Open/Grand Opening Reception in the Exhibit Hall


Monday, May 22, 2000

8:00 am – 9:45 am

Technical Plenary I on Microarray and Chip Technology

Chair: Daniel Pinkel


Fundamentals of Oligonucleotide Hybridizations

Edwin Southern

PAM, A Flow Through Approach to Microarrays

Tim Kievits

Microarrays for Analysis of Single Nucleotide Polymorphisms

Ann-Christine Syvänen

The progress of the Human Genome Project has created a need for methods with high throughput for multiplex analysis of DNA sequence variation. We have previously developed the "minisequencing" single nucleotide primer extension method for analysing DNA sequence variants. Recently we have converted this reaction principle into a microarray ("DNA-chip") format. The properties of fluorescence-based genotyping by DNA-polymerase-assisted primer extension on microarrays will be described. Our applications of the system to analyzing human single nucleotide polymorphisms and detecting disease causing mutation in large population studies will also be described. We conclude that primer extension is a robust reaction principle for future high-throughput genotyping on "DNA chips".

 


10:45 am – 12:30 pm

Parallel Session III: Microarrays Technology I

Chair: Paul Meltzer


6264

Use of DNA Microarrays and Functional Proteomics in Drug Discovery.

Alexander Nakeff, Balanehru Subramanian, Frederick Valeriote, Josephine Ford Cancer Center, Henry Ford Health System, Detroit, MI, USA

We have taken a unique approach to the discovery and development of new anti-human solid tumor active agents from natural products and synthetics. Prospective agents are first analyzed with a specialized clonogenic cell disk assay in vitro that involves a carefully selected panel of murine and human leukemia, solid tumor and normal cells that identifies lead compounds selectively cytotoxic to human solid tumors. These compounds are then tested in vivo in SCID mice bearing human solid tumors. Those lead agents with proven in vivo activity become candidates for mechanism of action studies using genomics (DNA microarray) and functional proteomics. H116 human colon adenocarcinoma cells, exponentially-growing in T75 flasks, are treated withLD90 or LD99 cytotoxic doses of the in vivo active lead for various periods of time or left untreated (control) and the cells harvested and lysed. Cell lysates are used to extract RNA for identification of gene targets by cDNA microarrays and to identify and isolate target proteins (proteomics) by 2-D gel electrophoresis, respectively. The RNA is used to generate cDNA probes by reverse transcriptase. The cDNA’s are then labeled with either Cy3 (untreated cells) or Cy5 (drug-treated cells) for competitive hybridization with known genes in a commercially-available DNA microarray. These are scanned for fluorescence intensity as a quantitative measure of gene expression and used to identify specific, drug-regulated genes whose expressions are significantly turned on or off. Lysates from drug-treated and control H116 cells are run on 2-D gels for quantitative analysis of specific, drug-induced changes in the levels of total cellular proteins and identification of proteins that are newly-produced. Genomics/proteomics is a powerful means to identify specific, drug-induced molecular pathways and define the mechanism of action of new anti-human solid tumor active drugs.

 


6379

Employing DNA microarrays for the analysis of plant gene expression

David Galbraith, Elizabeth A. Pierson, Michael Deyholos, Wenying Xu, Rene Feyereisen, Hans Bohnert, University of Arizona, Dept of Plant Sciences, Tucson, AZ, USA; Virginia Walbot, Herrin Laboratory, Stanford University, Stanford, USA

Microarrays provide the means for the large scale analysis of the regulation of gene expression patterns within living organisms. My laboratory is involved in three projects which employ microarrays for characterization of plant gene expression. The first project, funded by the NSF program in Plant Genomics, involves analysis of the changes in plant gene expression that accompany imposition of drought and salinity stress. We are employing microarrays to unravel hierarchical patterns of gene regulation in response to imposed stress. We are also using novel methods for identification and characterization of genes that are coordinately-regulated within specific tissue types, based on flow cytometric analysis and sorting of nuclei. The second project, funded by USDA and the Human Frontier Science Program, involves characterization of the functions and expression patterns of the approximately 340 members of the cytochrome P450 gene superfamily in Arabidopsis thaliana. We are employing microarrays to systematically determine the extent of cross-hybridization between different superfamily members, in order to define sequences that might provide member-specific expression data. We are using this information to examine the tissue specificity of expression of individual P450 genes, as well as their responses to biochemical stimuli. The third project, also funded by the NSF Plant Genome program, aims to sequence at least 50,000 ESTs and cDNAs from maize (Zea mays), and provide this information to the public sector. Our responsibility is the production of microarrays comprising unigene sets from these sequenced ESTs, of which ~25,000 have been produced. We are working to resolve technical problems as well as derive biological information specific to the different projects. This includes the identification of suitable substrates for printing, the reduction of costs associated with microarray production and analysis, debugging programs for commercial arrayers, establishing sensitivity and reproducibility criteria, implementing high throughput methods for probe production, and methods for sample tracking and for data analysis, reduction and archiving. Progress in these technical areas, as well as results from the individual projects, will be discussed.

 


6868

Array-based CGH for the differential diagnosis of renal cell carcinoma.

Monica Wilhelm, Joris Veltman, Frederic Waldman, UCSF Cancer Center, San Francisco, CA, USA; Joseph Presti, VA Medical Center, San Francisco, CA, USA; Gyula Kovacs, University of Heidelberg, Urology - Molecular Oncology, Heidelberg, Germany

Renal cell carcinoma (RCC) can usually be separated histologically into one of four major types: conventional (clear cell), papillary, chromophobe and oncocytoma. These tumor types show widely varying biological and clinical behaviors. However, RCC can show mixed morphology rendering conventional diagnosis by histology unreliable. Genetic alterations within a tumor appear to be specific for histologic type, and a classification system has been proposed which uses genetic analysis to define each type. We have applied array-based comparative genomic hybridization (CGH) to a set of tumors to test the utility of this approach for diagnosis of renal tumors. DNA arrays were prepared by robotic spotting in quadruplicate of 96 target clones with a focus on chromosomal regions altered in RCC (chromosomes 1,2,3,5,7,8,9,13,14,16,17 and 21). 40 well characterized RCC DNAs including all four RCC types were analyzed in a blinded study. Gains and losses were detected which identified the four histologic patterns, and which agreed with microsatellite based LOH and cytogenetic analyses. This DNA chip can be applied to the genetic characterization of RCC cases, whose diagnosis based solely on histology is questionable. As more information about candidate genes and prognostic markers involved in RCC tumorigenesis become available, it will be possible to add those targets to the analyses creating a new and powerful diagnostic/prognostic tool.

 


6870

Technical approaches for producing and analyzing DNA microarrays.

Greg Hamilton, Joe Gray, Donna Albertson, Dan Pinkel, UCSF Cancer Center, San Francisco, CA, USA; Arthur Jones, Don Uber, Donn Davy, Tony Hansen, Robert Nordmeyer, Dave Wilson, Joe Jaklevic, LBNL, Engineering, LBNL, Berkeley, CA USA

We have developed an integrated microarray production and analysis facility for quantitative comparisons of DNA copy number and RNA expression. The overall process is managed by a custom Oracle database that tracks clone acquisition and preparation of printing solutions, manages array printing and image acquisition, and stores results of the fluorescence analysis.

The overall goal of the array printing process is to keep the array size small in order to minimize the amount of specimen required for analysis, while employing a large number of printing pins to increase printing rate. The pins made from quartz capillary tubing (outside diameter of ~ 0.4 mm) which has been heated and pulled have an opening of ~ 20-40 µm. Maximum load of a pin is ~ 0.2 µl, enough to print > 1,000 spots. Solutions can be printed from any standard microtiter plate, up to and including those with 1536 wells. Routine printing is performed from 864 well plates with center to center spacing of the spots ~ 100 µm. A 12 mm square array can be printed using 16 pins.

The array imaging system uses a 16 bit CCD camera, arc lamp illumination, computer controlled excitation and emission filters, and custom designed lenses. Image data from the entire array is obtained with an exposure of several seconds or less for each fluorochrome. The lenses are chromatically corrected from 450 to 750 nm, and they provide flat images over the entire field. Emission from at least 4 fluorochromes can be independently measured in a single specimen. These typically include DAPI, which is used as a counter stain to permit spot localization, and three labels, for example fluorescein, Cy3 and Cy5, in the specimen DNAs or RNAs. Fluorescence ratios are calculated using custom software that automatically segments the spots based on the DAPI image, determines local background, and integrates signal intensities within each segmented spot.

Work supported by NCI grant CA83040 and Vysis Inc.

 


7234

An expression microarray for investigation of genes involved in loss of heterozygosity on chromosome 16 in breast cancer

Anne-Marie Cleton-Jansen, Hetty Van Beerendonk, Cees Cornelisse, Leiden University Medical Center, Departments of Pathology, The Netherlands; Marlies Van De Berg, Johan Den Dunnen, Gert-Jan Van Ommen, Leiden University Medical Center, Department of Human and Clinical Genetics, Leiden, The Netherlands

Molecular-genetic as well cytogenetic studies have revealed that loss of heterozygosity (LOH) at the long arm of chromosome 16 (16q) belongs to the most frequent genetic alterations occurring in breast cancer. In a previous study we identified two separate smallest regions of overlap (SRO) involved in LOH at 16q22.1 and 16q24.3. A detailed investigation of genes in the SRO at 16q24.3 has so far not led to the identification of the gene targeted by LOH. We now want to use a different approach that will lead to the identification of the tumour suppressor gene(s) thereby making use of the wealth of data generated by the human genome project and the novel technique of cDNA expression arrays. Microarray construction and scanning is performed at the Leiden Genome Technology Centre.

We are constructing a microarray containing partial cDNA clones of genes and ESTs (expressed sequence tags) that are mapping to the region which is most frequently involved in LOH, i.e. 16q22.1-qter. Information on the presence of transcribed sequences at this locus is obtained from the public databases such as Genemap ’99, and from the large scale sequencing efforts of Los Alamos National Laboratories. Special effort is put into obtaining the genes from the two SROs at 16q22.1 and 16q24.3. For the latter one a detailed map has already been constructed previously. Genes and ESTs are PCR-amplified from a 40,000 clone library constructed by the I.M.A.G.E. consortium and PCR products are spotted onto microscope glass slides in an arrayed fashion.

The microarrays are hybridized with labeled RNA from normal mammary epithelium versus breast cancer cells obtained from cultured cells or primary tumours. The result shows which genes are expressed in mammary epithelial cells, both normal and tumour cells, which is a prerequisite for a gene to be candidate tumour suppressor. Furthermore it is anticipated that in a subset of the tumours a gene shows repeatedly decreased transcription, either through transcriptional inactivation or due to an inactivating mutation. This increases the probability that this gene is the targeted tumour suppressor. This microarray analysis will facilitate the selection of potential candidate genes, which will be further investigated in tumours showing LOH atchromosome 16.

 


6829

Analyzing gene expression data

Lodewyk Wessels, Marcel Reinders, Eugene Someren, Delft University of Technology, Faculty of Information Technology and Systems, Delft, The Netherlands

A thorough understanding of cell processes is central to the development of therapeutic drugs for the treatment of, for example, various kinds of cancer. Such tailor-made treatments are desirable since their specificity maximizes effectiveness while minimizing undesirable side-effects. High-throughput screening technology can be employed to monitor the time dependent behavior of thousands of genes simultaneously, providing valuable insight into cell processes. A wide range of data processing techniques can be employed to analyze expression data of this kind, and clustering of time signals is currently one of the most widely used techniques. In this paper, we outline preprocessing techniques, similarity measures and various clustering techniques relevant to the extraction of interesting information from gene expression data. We reflect on their relative merits and illustrate how a gene expression analysis can be optimized in an objective fashion. The empirical results are based on several publicly available data sets, but are primarily focused on expression data provided by the UCSF Cancer Center. This data set captures differential gene expression patterns of 214 human apoptosis related genes after treatment with the PI3-kinase inhibitor LY294002. The clustering process identifies sets of genes (possible functional groups) that react similarly when the conditions in the cell are perturbed. In addition to the cluster analysis, parametric modeling approaches, aimed at extracting (causal) relationships between genes (functional groups) have also been implemented on the data sets. A particularly interesting configuration involves the application of the clustering step to perform dimensionality reduction (many genes to few functional groups) followed by a parametric genetic modeling step to identify relationships between the functional groups. This approach is also illustrated on various data sets.

Acknowledgment We would like to thank Joe Gray and Russ Baldocchi from the UCSF Cancer Center for providing us with the data set and the opportunity to work with them. This work was funded by the Intelligent Molecular Diagnostic Systems program of the Delft Inter-Faculty Research Center at the TU Delft.


Parallel Session III: Clinical Immunology

Chair: Guenter Valet


6344

SIMPLIFIED QUANTITATION OF MYELOID DENDRITIC CELLS IN PERIPHERAL BLOOD USING FLOW CYTOMETRY

Denis Snider, Dept. of Pathol. Mol. Med., Hamilton, Ontario, Canada; Hong Liang, McMaster University, Dept. of Pathol. Mol. Med., Hamilton, Ontario, Canada; John Upham, Univ. of Western Australia, Dept. of Medicine, Perth, Australia

Background: Recognition of the importance of dendritic cells (DC) in the initiation of T-cell dependent immune responses has led to increasing interest in methods for the identification of DC within the circulation. We sought to develop a flow cytometric method that would allow the reliable enumeration of absolute myeloid DC counts in minimally manipulated blood samples. Methods: Myeloid DC were identified by three color staining of whole blood leukocytes as a discrete population of mononuclear cells expressing high levels of HLA-DR and CD33, yet having little or no expression of CD14 and CD16. This method was analyzed for reproducibility, variation in blood DC number during typical clinical day hours and after exercise. The new method was compared to an established commercial kit method. Results: FACS sorting of the CD33+ DC showed that they morphologically resembled immature DC, and developed cytoplasmic projections typical of mature DC following overnight culture in GM-CSF. Within peripheral blood, these DC were found at a mean concentration of 17.4 ± 5.4 x 106 per litre, corresponding to 0.93 ± 0.27 % of mononuclear cells. Comparison of duplicate samples stained and analyzed in parallel showed that the intra-sample variability was very low, with an intra class correlation coefficient of 0.95. The frequency of CD33+ myeloid DC and their light scatter characteristics were similar to that of CD11c+ myeloid cells, identified using a commercially available method of DC staining. Numbers of blood myeloid DC, identified by CD33 staining, showed no significant variation during standard laboratory hours, however their numbers rose significantly during vigorous exercise, in parallel to other blood cells. Conclusions: The method described herein is rapid, reproducible, requires only small volumes of blood, can be readily used by a clinical immunology laboratory, and requires fewer antibodies than a currently available commercial method.

 


6248

Quantitative assessment of T cell function using an integrated calibration and software package and a 1:1 CD69:PE reagent: implications for clinical use to document immune recovery following bone marrow transplantation

Lawrence Lamb Jr., W. Bradford Lindsey, P. Jean Henslee-Downey, South Carolina Cancer Center; Gerald Marti, NIH, Bethesda, MD, USA

CD69 is one of the earliest activation antigens expressed on the surface of activated peripheral blood T lymphocytes. Measurement of T-lymphocyte activation in the clinical setting has several advantages over the widely used method of radiolabeled thymidine incorporation following mitogen stimulation. These include reduced specimen volume, elimination or radioactive isotopes, ability to subtype activated T cells using additional surface and/or cytoplasmic markers, and the ability to assess constitutive(non-stimulated) activation state. However, the use of a quantitative method for measurement of CD69 expression is essential to eliminate subjectivity in determining the percent of positive cells when CD69 expression shows a "smeared" or incremental pattern of expression. Our laboratory determined antibody binding sites per cell (ABC) values for nonstimulated, phytohemagglutinin (PHA)-stimulated expression in healthy volunteers and recovering bone marrow transplantation (BMT) patients using a quantitative flow cytometry calibration and software package developed by BD Biosciences (San Jose, CA) and Verity Software House (Topsham, ME) and a 1:1 antibody:PE preparation of CD69. Resting CD69 expression on CD3+ T cells from healthy volunteers totaled 464+/-150 ABC. PHA-stimulated specimens (n=14)increased to 12,000 +/- 3000 ABC. Variation in ABC from specimens collected from a single donor at multiple time points and processed at the time of collection were negligible (+/- 500 ABC). Interestingly, specimens that were held for four hours at room temperature displayed a 28% +/- 7% decline in ABC values following PHA stimulation compared to samples processed at collection. Specimens from recovering BMT patients showed significantly decreased response to PHAstimulation (1200 +/- 400 ABC) at 28 days (n=5) post-BMT but had increased to near-normal levels at 100-180 days. Resting levels of CD69 expression was also increased in some patients possibly due to allogeneic activation of engrafted donor T cells. In conclusion, quantitative determination of CD69 expression using a combined calibration and software package and a 1:1 antibody:PE reagent may be a useful clinical test for determination of T lymphocyte activation if strict sample collection procedures are standardized.

 


6348

Preoperative Prediction of Post-Cardiotomy Syndrome (PCS) in Children Cardio-Pulmonary Bypass Surgery: Use of Cytometric or Humoral Parameters ?

Gunter Valet, Cell Biochemistry Laboratory, Martinsried; Michal Pipek, Jörg Hambsch, Peter Schneider, Attila Tárnok, Pediatric Cardiology, Cardiac Center, Univ. Leipzig, Pediatric Cardiology, Cardiac Center, Leipzig, Germany; Jozsef Bocsi, 1st Institute Pathology, Semmelweis Univ. Budapest, Hungary

Aims: Children cardiopulmonary bypass surgery for the correction of congenital heart defects may lead to inflammatory PCS (morbidity >30%) with pericardial, pleural or peritoneal effusions, liver swelling and edema or to the more severe capillary leak syndrome (CLS) with capillary hyperpermeability causing generalized edema, protein extravasation into the interstitial space and acute respiratory distress syndrome (ARDS). There is a substantial clinical interest to identify risk patients prior to surgery. Methods & Results: Our earlier work has indicated that preoperative (possibly previous infections or allgery) rather than surgical conditions are responsible for PCS formation. The CLASSIF1 classification (Cytometry(CCC) 30:275-288(1997), http://www.biochem.mpg.de/valet/classif1.html) of 50 clinical chemistry parameters from blood or urine identified a data pattern of complement components (C1Inh, C3, C5), cytokines (IL10) soluble adhesion molecules (L-selectin, ICAM1) as well as hemoglobin, potassium and total serum protein as predictive parameters. Further refinement with >90% correct predictions of postoperative complications was achieved by classification of the selected parameters with the SPSS-program system. The need for significant amounts of blood, the time requirements (24h) for the determination of some of the parameters and the high total costs prompted for the investigation of more suitable cellular parameters as an alternative. In any case, flow cytometric parameters are probably closer to the biochemical mechanisms involved in PCS formation than humoral parameters. The CLASSIF1 classification of three colour CD45/14/HLA-DR and CD18/11a/3 manually quantitated flow cytometric measurements provided >95.5% correct preoperative predictions from the CD45, CD14 and CD18 expression on monocytes as well as from CD45, CD18 and CD11a/CD18 expression on neutro-, eosino- and basophils. Similarly, the CLASSIF1 classificationof four colour assays: CD45/14/HLA-DR/3, CD25/54/3/19, CD45RA/45RO/4/3, CD19/69/3/55 and CD3/16+56/19 on mono- and granulocytes provided >90.0% correct predictions. It is likely that the automated and exhaustive list mode analysis of all peripheral leukocyte populations e.g. including lymphocytes will further improve the results. Conclusion: Flow cytometric measurements of CD-antigen expression on peripheral blood leukocytes may well replace humoral parameters in the preoperative risk assessment for PCS development in children cardiac surgery. Low amounts of required blood, decreased costs and a higher analysis speed, at similar or better predictive capacity, are likely to favor the flow cytometric approach in the clinical environment.

 


6890

Neutrophil CD64 Expression in Acute Inflammatory Arthritides and in Systemic Infections.

Antony Bakke, Oregon Health Sciences University, Pathology and Medicine; Eduardo Paiva, Everette Allen, Atul Deodhar, Oregon Health Sciences University, Medicine/Arthritis and Rheumatic Disease Division, Portland, OR, USA

Fever, malaise, myalgia and arthralgia, common symptoms of systemic infections are non-specific and are also seen in acute flares of autoimmune inflammatory conditions, such as rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE). Differentiation between these two groups of diseases is vitally important as the treatment differs dramatically. The usual laboratory tests employed to diagnose systemic infections, e.g. leukocyte number, presence of immature forms, and erythrocyte sedimentation rate can also be misleadingly high in patients being treated with corticosteroids or spuriously low in patients on cytotoxic drugs. Microbiologic cultures require an extended culture period and may be falsely negative. Often patients are empirically treated with both, broad spectrum antibiotics and high dose corticosteroids.

Fcg1 (CD64) enhances phagocytosis and is up regulated on neutrophils within hours of a systemic infection. Previous studies have demonstrated up regulation in 97%of culture positive patients. Although the specificity of CD64 up regulation has not been studied closely, expression on neutrophils from patients with RA and SLE is normal. The purpose of this study was to determine the sensitivity, specificity and positive predictive value of CD64 up regulation in patients with acute flares of their inflammatory disease, patients with gout, patients with culture proven infection and patients with non-inflammatory rheumatic conditions. A quantitative assay of CD64 expression using CD45 gating and QuantiBrite beads was employed (kindly provided by Becton-Dickinson, San Jose, CA, USA).

At this point in the study we can define three levels of CD64 expression: <1000, 1000-2000, and >2000 molecules/cell. Our results are presented in the table below for patients with infection, inflammatory arthritides and non-inflammatory conditions.

Although the number of patients is still small, there is a clear trend for high levels of CD64 (>2000 molecules/cell) predicting infection with a 95% specificity for inflammatory disease. Only one active RA patient expressed >2000 molecules of CD64 on their neutrophils. These results indicate that monitoring expression of high levels of CD64 on neutrophils may be clinically useful to diagnose infection in patients with chronic inflammatory diseases and institute appropriate therapy.

 


6415

Measurement error of absolute CD4+ T lymphocyte count double platform evaluations for the assessment of laboratory performance in lymphocyte immunophenotyping quality control (QC)

Annalisa Kunkl Phd, M.P. Terranova, Lorenzo Mortara, Laura Bottone, Dept. Immunology, San Martino Hospital, Genoa, Italy; Domenico Risso, Pasquale Bruno Lantieri, Dept. Health Science, Biostatistics Unit, Univ. of Genoa, Genoa, Italy; Mauro Girotto, Cascina Bose, Magnano, Italy

We have studied critical errors of WBC count (AbsWBC), % lymphocyte (%Lymph), lymphocyte count (AbsLymph), % CD4+ T cell (%CD4+) and CD4+ count (AbsCD4+) determinations on comparative analyses of 24 HIV+ and HIV- blood samples by 18 labs participating to the Ligurian Region QC program. For each sample we calculated the 99.9% confidence interval of the mean of the distribution of each parameter, excluding outliers and data without internal consistency (some 300 valid data for each parameter). Regression of the upper and lower confidence limits over the means of the 24 samples were used to interpolate confidence intervals for any theoretical "true" mean in the tested range. The regression equations of upper and lower limits were: y=1.053 x+149.02 and y=0.947 x-149.02 for AbsWBC, y=1.0576 x+2.1036 and y=0.942 x-2.1036 for %Lymph, y=1.0372 x+156.42 and y=0.9628 x-156.42 for AbsLymph, y=1.0436 x+0.9721 and y=0.956 x-0.9721 for %CD4+, y=1.1262 x+7.6444 and y=0.8738 x-7.644 for AbsCD4+. We validated these boundaries of acceptable error by studying lab performance on comparative analyses of 33 blood samples by 42 labs of the Piemonte Region QC program. All samples were evaluated by the double platform method. Since lab performance in QC is more readily evaluatedin terms of error over the median of the distribution of all data than from the consensus mean, we extrapolated acceptable errors over mean (A) and over median values (B) and we compared B vs A vs the Paxton’s criteria (C). These identify outliers as data outside lower and upper fences defined by the 25th percentile - 1.5*interquartile range (IQR) and the 75th percentile + 1.5*IQR respectively. The counts (n) of unacceptable data are reported in table. Moreover, the degrees of concordance of (A) vs (B ), (A) vs (C) , (B) vs (C) by the Cohen’s k statistics were: 0.90, 0.57, 0.55 for AbsWBC; 0.87, 0.51, 0.54 for % Lymph; 0.88, 0.48, 0.50 for AbsLymph; 0.89, 0.39, 0.38 for % CD4+; 0.90, 0.31, 0.30 for AbsCD4+ respectively. Criteria A and B were more severe than criteria (C). Moreover, criteria (A) and (B) showed a similar impact on lab grading, as indicated by their high degree of agreement. We propose definition of Lab performance by criteria (B) to reduce inter-laboratory variation.

 


6479

Ceramide Binds to and Signals Through the LPS Receptor CD14 that is Upregulated in Patients with Myocardial Infarction

Gerd Schmitz, Alexandra Götz, Evelyn Orsó, Gregor Rothe, University of Regensburg, Regensburg, Germany

Ceramide (Cer), a degradation product of sphingomyelin, mimicks LPS with regard to cellular activation, suggesting that LPS and Cer dependent cellular activation is induced via a common receptor molecule, such as the LPS-receptor (CD14). As CD14 is a GPI-anchored protein, signal transduction is assumed to occur via associated transmembrane proteins such as the ß2-integrin (CD11b/CD18). The purpose of this study was to characterize the involvement of CD14 in cellular responses to LPS and Cer using fluorescence resonance energy transfer (FRET) with regard to the activation dependent assembly of CD14 with CD11b/CD18. In resting monocytes no significant FRET occured indicating no close association between CD14 and CD11b. LPS as well as Cer, but not PMA or FMLP induced a significant FRET between the CD14 and CD11b, suggesting a conformational change of the receptor complex due to ligand binding. Moreover, CD14 also clustered the Integrin Associated Proteins (IAP, CD47 and CD81) and the FcgRIII (CD16) into a multimeric functional receptor complex. LPS and Cer were also comparable regarding time and concentration kinetics, and stability of the induced conformational changes. These data demonstrate Cer as an endogenous ligand that induces sustained co-assembly of the CD14 associated receptor complex similar to LPS. In the acute phase response LPS and Cer have been shown to be associated to HDL lipoproteins, and this association seems to be mediated via a family of lipid transfer proteins including the LPS-binding protein (LBP) and the phospholipid transfer protein (PLTP). Recently, an apo-AI/LBP containing particle has been demonstrated, facilitating cellular responses to LPS. Based on our data regarding the interaction of CD14 with LBP/LPS or LBP/Cer complexes we propose CD14 to be a receptor for LPS/Cer enriched apo-AI/LBP-containing HDL-particles in the acute phase response. The increase in LBP and PLTP activities in inflammation may further enhance the transfer of LPS to HDL. Together with the recent report of an increased expression of CD14 on monocytes in myocardial infarction, this suggests the targeting of proinflammatory HDL particles to the Cer/LPS-receptor complex on monocytes, which may also include further proteins such as CD16 or other IAPsto form an innate immunity receptor complex, in order to achieve a profound enhancement of cellular activation in the acute phase response.

 


6797

Flow cytometry evaluation of peripheral blood leucocytes functions in silicotic patients

Jean François Subra, Toulouse, France; Gilles Renier, Stephane Mattmann, François Carrere, Alain Chevailler, CHU Angers, Immunology, laboratoire d’immunologie, Angers, France

An increased prevalence of connective tissue diseases such as rheumatoid arthritis or systemic sclerosis has been demonstrated in silica exposed patients.

Few data are available about systemic immunological abnormalities in silicotic patients, therefore we conducted a cross-sectional study to compare peripheral blood leucocytes (PBL) phenotype and functions among 58 former slate workers and 41 healthy volunteers. Silicosis was assessed in all 58 patients according to the International Labor Office’s criteria. In the same time, clinical and biological data were collected. A flow cytometric evaluation of the lymphocytes subsets was performed using isotypic controls monoclonal conjugated antibodies. The granulocyte oxidative burst was estimated using the fluorescent probe 2’7'-dichlorofluorescin diacetate and phorbol myristate acetate induced activation, phagocytosis was evaluated using 2µm microspheres.

No differences were observed in the number of total leucocytes and platelets or CD4+/CD8+ T cells ratio between the two groups. When the absolute numbers were considered, a very significant decrease of CD3+, CD3- CD19+, CD4+ CD8-, CD4- CD8+ lymphocytes as well as natural killers was discovered in silicotic patients compared with healthy volunteers. Conversely the number of activated T cells (CD3+ HLA DR+) was greater yet not significantly, in silicotic patients : 106/µL versus 87/µL related to an increased activated T cells / T cells ratio (12,3% versus 6,5% : p<0,0001) in the silica exposed group.

Furthermore the silicotic patients group had a greater prevalence of connective tissue diseases (7/58 versus 0/41 antinuclear antibodies : p = 0,02 Fisher exact test) and of antinuclear autoantibodies. In the two groups, phagocytosis and oxydative burst assessment were normal in all individuals.

In conclusion our results show that in our silicotic patients, absolute number and proportions of PBL are altered. Whether the reduction of circulating T lymphocytes could be related to increased apoptosis and could predispose to autoimmune or malignant disorders needs further investigations.

This work was supported by the Programme Hospitalier de Recherche Clinique 1994.

 


6823

Quantitation of antigen-specific T-cells in patients with cancer using VB-specific antibodies

Markus Maeurer, Kirsten Freitag, Henryk Pilch, Johannes Gutenberg University, Department of Medical Microbiology, Mainz, Germany; Elke Jaeger, Krankenhaus Nordwest, Medical Clinic II, Frankfurt, Germany; Antje Necker, Beckman Coulter, Immunotech, Marseille, France

A number of clinical trials have started to implement antigens in the form of antigenic peptides in order to activate or to augment T-cell responses directed against the autologous tumor in patients suffering from cancer. Some of these patients experience objective measurable responses to therapy, some of them progress despite repetitive immunization. One of the surrogate markers for successful immunization represents evaluation of T-lymphocytes reacting to the immunizing peptide(s). In general, the T-cell response directed to a given antigen may be described by combining three categories I. Specificity and effector functions of a T-cell population, II. The quantity of T-cell responses (i.e. number of T-cells reacting to the antigen) and III. The quality of T-cells defined by the T-cell receptor (TCR) structure. Several methods to measure anti-peptide/tumor T-cell responses are now available including evaluation of T-cell precursors using limiting dilution, the ELISPOT-assay, ex vivo T-cell receptor (TCR) variable (v) segment analysis determined by flow-cytometry and TCR-CDR3 length analysis (spectratyping) as well as identification of peptide-specific T-cells using MHC-class I tetramers containing appropriate peptides. Up to now, only a limited set of MHC/peptide complexes are available as tetramer-complexes. We demonstrate ex vivo detection of antigen-specific CD8+ or CD4+ T-cells in patients with melanoma or with cervical cancer by a combination of spectratyping, functional analysis and the implementation of an antibody panel directed against 21 individual VB TCR chains, which cover 65% of the T-cell repertoire. Detection of these T-cell populations by flow cytometry in longitudinal studies is instrumental in order to define the magnitude of a CD4 or CD8 T-cell response directed against tumor-associated antigens.


Parallel Session III: Cell Growth & Differentiation

Chair: Piet VanErp


6848

Multiparametric Flow Cytometry of the Cell Cycle

R. Michael Sramkoski, James Jacobberger, Case Western Reserve University, Cancer Research Center, Cleveland, Ohio, USA; Susan Wormsley, Pharmingen, San Diego, USA,

Cycling HeLa cells were pulsed with BrdUr, fixed with MeOH, washed, then subjected to the SBIP technique. UV-treated, Tdt-reacted cells were washed, then labeled indirectly monoclonal antibody MPM2 (mitotic marker) and Cy5-conjugated antibody. The cells were then labeled with PE-anti-BrdUr (Phoenix Flow, San Diego) and FITC-anti cyclin B1 (GNS-1, Pharmingen, San Diego) direct conjugate monoclonal antibodies, and 7-Actinomycin D to label DNA. The cells were analyzed by flow cytometry with two lasers. Analysis demonstrated that the combination of marking BrdUr, cyclin B1, and mitotic cells provides a complete cell cycle analysis without inclusion of a DNA dye. This leaves UV-excited dyes (e.g., Alexa 350) and far red dyes for additional markers. Further, preliminary data indicates a possibility that by pulse-chase and/or continuous labeling, the kinetics of S to G2 and M may be accurately quantified, thereby providing a means to associate cyclin B1 expression with time. If this works, then it is our intent to test the hypothesis that cyclin B1 levels could be used as a surrogate marker for population G2 transit time.

 


6820

Combined analysis of Ki-67, MCM3 and p27 protein expression: A snapshot of cell proliferation and differentiation

Elmar Endl, Thomas Scholzen, Johannes Gerdes, Research Center Borstel, Immunology and Cell Biology, Division of Molecular Immunology, Borstel, Germany

Proliferation and differentiation are accompanied by a characteristic expression of proliferation associated proteins. The Ki-67 protein is one of the most commonly used markers for proliferation assessment in cell biology and cancer research. Other proteins like the minichromosome maintenance proteins (MCM), which are involved in the control of DNA replication, and the cyclin dependent kinase inhibitors such as p27/KIP1 are currently under investigation for their use as diagnostic markers. The question we address was how the expression of these three proteins is related to cell differentiation and proliferation on routinely processed human tonsil sections. For this purpose we used established antibodies against the Ki-67 protein (MIB-1) and the p27 protein and generated a new monoclonal antibody directed against the MCM3 protein. We applied multicolor immunofluorescence, to get a detailed insight in colocalization and reciprocal expression patterns of all three proteins. According to the spatiotemporal organization of germinal centers and oral mucosa the expression of the p27 protein was closely related to differentiated cells, whereas MCM3 and Ki-67 predominantly localized to the regions of proliferating cells. However, it is important to note that a considerable amount of cells that were growth arrested as confirmed by the absence of the Ki-67 protein, stained positive for the MCM3 protein. These results were further confirmed using 3T3 mouse cells that were growth arrested in vitro. The MCM3 protein is therefore also expressed in cells that have ceased to proliferate i.e. do not express the Ki-67 protein, but are not terminally differentiated according to the absence of p27 protein expression. In conclusion a combined analysis of Ki-67, MCM3 and p27 protein expression may provide a tool for a more detailed analysis of cell proliferation and differentiation processes in cancer research and new insights in the complexity of individual tumor growth.

 


6319

Flow cytometric characterization of hyperproliferation-associated keratinization in psoriatic skin

Piet Van Erp, Roland Mommers, Hans Goossen, Peter Van De Kerkhof, University Hospital Nijmegen, Department of Dermatology, The Netherlands

Keratins are a group of cytoskeletal proteins that are found in human epidermis and other stratified squamous epithelia. Several different types of keratins have been described. Keratin 10 (K10) is a keratin that is expressed in well differentiated, suprabasal keratinocytes, and keratin 6 (K6) is a keratin that is associated with hyperproliferation. Psoriasis is a chronic inflammatory skin disease, and besides inflammation, disturbed differentiation and hyperproliferation are its hallmarks. In order to study the hyperproliferation-associated keratinization in both well-differentiated and disturbed differentiated keratinocytes, and in order to assess the proliferative activity of all K10 and K6 subpopulations, simultaneous assessment of K6, K10, and DNA content is required. In the present study, we established a novel protocol for simultaneous measurement of K6, K10, and DNA content, which enables the characterization of the proliferative activity of several cellular subpopulations in epidermis. From 16patients with psoriasis and from 15 healthy volunteers, punch biopsies were obtained. After preparation of single cell suspensions, cells were stained with the anti-keratin 10 IgG1-isotype monoclonal antibody RKSE60, with the anti-keratin 6 IgG2a-isotypemonoclonal antibody LHK6B, and with the DNA fluorochrome TO-PRO-3 iodide. Isotype specific secondary antibodies conjugated with PE and FITC were used as the second step in the staining procedure. Owing to the IgG specificity of RKSE60 and LHK6B, no cross-reactivity was observed between these antibodies. The triple staining with RKSE60, LHK6B, and TO-PRO-3 iodide showed subpopulations of K10 expressing cells, K10/K6 co-expressing cells, and K6 only expressing cells. There was a significant difference in the proportion of K6 expression and K10/K6 co-expression between psoriatic and normal skin. Moreover, the proliferative activity of these subpopulations could be quantified by this protocol. We concluded that a triple staining protocol for the assessment of K6, K10, and DNA content, using the monoclonal antibodies LHK6B, RKSE60, and TO-PRO-3 iodide, supplies reliable and reproducible data for cellular studies on these keratins and for studying the proliferative activity of the subpopulations of these keratins in epidermis. Moreover, the present study showed that with respect to the proportion of K6, significant differences are present between psoriatic and healthy human skin.

 


6241

Phosphorylated Histone H3 as a marker of mitosis and the monocytic phenotype

Frank Traganos, Zbigniew Darzynkiewicz, New York Medical College, Brander Cancer Research Institute, Hawthorne, NY, USA; Gloria Juan, Sloan-Kettering Cancer Institute, NY, NY, USA

Phosphorylation of the nucleosome core histone H3 (H3) on serine 10 is thought to be a prerequisite for chromatin condensation during mitosis. Changes in chromatin structure involving local areas of condensation also occur during interphase and especially during differentiation and mitogenic activation. H3 phosphorylation status allowed identification of mitotic human leukemic cells by flow cytometry when a monoclonal antibody (anti-H3-P) was used in combination with the DNA dye propidium iodide (PI). Simultaneous analysis of H3 phosphorylation status and expression of cyclins A and B1 determined that all H3-P positive (H3-P+) cells were cyclin A negative (with the exception of prophase and prometaphase cells) while cyclin B1 expression was threefold higher in these compared to G2 cells. The status of H3 phosphorylation was also assayed in individual human lymphocytes during mitogenic stimulation and in human myeloid leukemic HL-60 cells induced to differentiate by all-trans-retinoic acid (RA), 1,25-dihydroxyvitamin D3 (vit D3), dimethyl sulfoxide (DMSO), and phorbol myristate acetate (PMA). Multiparameter flow cytometry was used to correlate H3 phosphorylation with cell cycle position. H3-P+ cells were identified during interphase in HL-60 cells and in normal lymphocytes but the intensity of H3-P+ fluorescence was several fold lower than during mitosis. While no changes in H3 phosphorylation were noted during lymphocyte stimulation, unexpectedly, H3 phosphorylation levels were fourfold higher in normal monocytes in these cultures. Differentiation of HL-60 cells was accompanied by a rise in H3 phosphorylation which was higher after induction with RA, vit D3 and PMA (~ threefold) compared to DMSO (~20% higher). This data indicates that in addition to being a critical event during chromatin condensation at mitosis, H3 phosphorylation plays a role during chromatin changes accompanying monocyte differentiation. The high level of H3 phosphorylation in monocytes may serve as a marker of these cells and preliminary data suggests may be a useful tool in identifying cell lineage in leukemias.

 


6824

DIFFERENT EXPRESSION PROFILES OF HUMAN CYCLIN B1 IN NORMAL PHA-STIMULATED T LYMPHOCYTES AND LEUKEMIC T CELLS

Jean François Viallard, France; Francis Lacombe, Laboratoire de greffe de moelle, Bordeaux, France; Maryse Dupouy, France; Francis Belloc, Laboratoire D’Hematologie, Pessac, France; Josy Reiffers, France

In a previous work, we demonstrated using flow cytometry (FCM) methods that accumulation of human cyclin B1 in leukemic cell lines begins during the G1 phase of the cell cycle (Exp Cell Res 1999;247:208-219). In this work, FCM was used to compare the localization and the kinetic patterns of cyclin B1 expression in normal and leukemic T-cells.

Methods: Cell synchronization was performed in G1 with sodium n-butyrate (1 mM final concentration), at the G1/S transition with thymidine (2 mM) and at mitosis with colchicine (0.2 mg/mL). Cells (leukemic cell line Jurkat or PHA-stimulated human T-lymphocytes) were stained for DNA (10 mg/mL of propidium iodide) and cyclin B1 (clone mouse monoclonal GNS-1; Pharmingen; 1:200 dilution) and then analyzed by FCM. Western blotting was used to confirm results.

Results: Under asynchronous growing conditions and for both cell populations, cyclin B1 expression was essentially restricted to the G2/M transition, reaching its maximal level at mitosis. When the cells were synchronized at the G1/S boundary by thymidine or inside the G1 phase by sodium n-butyrate, Jurkat cells accumulated cyclin B1 in both situations, whereas T lymphocytes expressed cyclin B1 only during the thymidine block. The cyclin B1 fluorescence kinetics of PHA-stimulated T lymphocytes was strictly similar when considering T lymphocytes blocked at the G1/S phase transition by thymidine and in exponentially growing conditions. These FCM results were confirmed by Western blotting. In cells sorted in G1phase, the detection of cyclin B1 by Western blot showed that cyclin B1 was present in leukemic T cells but not in normal T lymphocytes. Cyclin B1 degradation was effective at mitosis, thus ruling out a defective cyclin B1 proteolysis.

Conclusions: Our data support the notion that human cyclin B1 is expressed in G1 phase of leukemic T cells but not normal T lymphocytes. We hypothesize that after passage through a restriction point differing in T lymphocytes and in leukemic cells, the rate of cyclin B1 synthesis becomes constant in the S and G2/M phases and independent from the DNA replication cycle.


Parallel Session III: Molecular & Clinical Genetics

Chair: Dorothy E. Lewis


6495

Isolation of Fetal Nucleated Red Blood Cells for Non-Invasive Genetic Analysis

Dorothy E. Lewis, Jeffrey Scott, Amie Bryson, Farideh Bischoff, Joe Simpson, Baylor College of Medicine, Immunology Section, Houston, TX, USA

For genetic diagnosis of fetal abnormalities using a non-invasive technique, we have focused on isolation of fetal nucleated red blood cells (NRBC’s) obtained from maternal blood. Anti-coagulated blood is first ficoll separated and cells stained with CD45, (to mark maternal white blood cells) permeabilized for staining with gamma globin and then sorted. To optimize isolation procedures, we obtained fetal livers at 16-24 weeks gestation and analyzed them for gamma and beta globin expression. We found that 62 ± 10% of the nucleated cells in the fetal liver (n=3) expressed gamma globin, which is the fetal form, whereas only 6.3 ± 2.8% expressed beta globin, which is the predominant adult form. Because maternal NRBC’s are beta globin+ gamma globin-, whereas fetal NRBC’s are predominantly beta globin- gamma globin+, we performed spiking experiments to optimize enrichment conditions. We found that more fetal liver nucleated RBC’s were isolated using a 1.119 g/ml ficoll density gradient at 700 x g than with a conventional 1.077 g/ml density gradient at 400 x g. Comparison of the 1.119 g/ml with the 1.077 g/ml ficoll density gradient on maternal specimens (n=5) showed that an average of 3.46 ± 1.8% of the isolated cells from the 1.119 g/ml gradient were gamma globin+ compared to 0.6 ± 0.6% using the 1.077 g/ml ficoll gradient (p> .02). These results suggest that recovery of fetal NRBC’s is improved by the use of fetal liver cells as representative of NRBC’sof fetal origin likely to be found in maternal blood.

 


7948

Chromosomes participating in translocations typical of malignant haemoblastoses are also in frequent interchanges induced by densely ionising radiation

Emilie Lukasova, Stanislav Kozubek, Andrea Mareckova, Eva Bartova, Institute of Biophysics, Academy of Sciences, Brno, Czech Republic; Michal Kozubek, Magdalena Skalnikova, Faculty of Informatics, Masaryk University, Brno, Czech Republic; Vaclav Kroha, Institute of Nuclear Physics, Academy of Sciences, Prague, Czech Republic

Repeated triple-colour fluorescence in situ hybridisation was used for the detection of exchange aberrations among 10 selected chromosomes of human lymphocytes irradiated with fast neutrons and He ions of mean energy of 7 MeV and 30 MeV respectively. In each hybridisation two different pairs of chromosomes were stained. Defined stage positions of metaphases on a slide were stored on a hard disk and an automatic scan of images according to these positions was performed after six successive hybridisations. In this way, six different images of the same metaphase with differently stained pairs of chromosomes and all centromeres were obtained.

The comparison of these images enabled the identification of mutual exchanges between selected chromosomes 1, 2, 3, 4, 8, 9, 12, 14, 18 and 22. The frequencies of exchanges were not linearly proportional to the DNA content of interacting chromosomes. The most significant were exchanges between chromosomes 14/18, 14/8, 18/8, 8/3, 1/14, 1/8, 3/18, 3/14 and 9/22 in the first mitoses after irradiation of lymphocytes with fast neutrons.

The results indicate significant interactions between chromosomes involved in translocations in B-cell non-Hodgkin’s lymphoma and chronic myeloid leukaemia. We propose that the reason for the high frequency of exchanges between some chromosomes is the proximity of their territories in the majority of cell nuclei. It may also be one of the reasons for the induction of translocations between the specific sequences of nucleotides leading to malignant transformation of the cell. The results show also the stability of some frequent interchanges several days after irradiation.

 


6364

FEATURES OF CHROMOSOMES IN HUMAN TERMINALLY DIFFERENTIATED BRONCHIAL CELLS PROVIDE A BLUEPRINT FOR NUCLEAR ORGANIZATION.

Leopold G. Koss, Montefiore Medical Center, Dept. of Cytology, Bronx, NY, USA

With FISH painting technique of terminally differentiated ciliated bronchial and goblet cells, it was shown that the two homologues of each chromosome occupy a separate domain on or adjacent to the nuclear membrane. One of the homologues was usually more "compact" than the other which was more "open," displaying fiber-shaped extensions. The differences between the territories of homologues 1 and 7 were statistically valid (p<0.0001). In some arrays of bronchial cells, the position of the homologues was either identical or formed a mirror image, suggesting that the position of the chromosomes may be constant. To confirm this observation, the angles formed by the homologues X, 1 and 7 were measured in oval nuclei. In about 2/3 of the nuclei, the two homologues formed angles of 150, 157 and 148 degrees, identical to those formed by the same chromosomes in prometaphase rosettes of diploid human fibroblasts (Nagele et al., 1995). In about one third of the nuclei, the same homologues formed angles of 89,72, and 94 degrees, and occasionally an angle of 180 degrees. A three-dimensional computer reconstruction of the nuclei was performed using the data for X chromosomes. By cinematographic technique, it was shown that the angles separating the 2 homologues depended on the rotation of the nucleus. The cause of the rotation is speculative at this time. It is suggested that the position of chromosomes in interphase human cells is constant and that the 2 homologues of each chromosome are not identical. These observations may have considerable bearing on chromosomal translocations, the formation of the nuclear membrane and of nuclear pores.

 


6911

DYNAMICS OF DNA REPAIR IN LIVING CELLS.

Adriaan Houtsmuller, Deborah Hoogstraten, Jan Jh Hoeijmakers, Wim Vermeulen, Erasmus University Rotterdam, The Netherlands

UV-induced DNA damage is removed from the genome by nucleotide excision repair (NER). Questions concerning in vivo nuclear organization and dynamics of NER-constituents have hardly been addressed. It is largely unknown in which way repair complexes identify and get access to DNA injury.

To address these questions, we developed and applied fluorescence recovery after photobleaching (ref. 1) (FRAP) assays for the confocal microscope to study in living cells: 1) nuclear diffusion rate (ref. 2), 2) immobilisation and 3) duration of immobilisation (ref. 3) of DNA-repair enzymes tagged with green fluorescent protein (GFP). Before and after induction of DNA damage by UV-irradiation, fully functional GFP-coupled repair proteins (XPA, XPB and ERCC1) showed a homogeneous distribution in nuclei of living cells. Confocal FRAP revealed high mobility of all investigated NER proteins, at diffusion rates expected for molecules of their respective molecular weights. In addition, induction of DNA-damage did not alter diffusion rates. These findings argue against the presence of preassembled DNA-scanning NER-holocomplexes either before or after DNA damage induction. A confocal FRAP immobility assay especially designed for studying fluorescent molecules in the cell nucleus (ref. 3), showed that a fraction of the studied repair-proteins became immobilized after DNA-damage induction. Since non-repair-proteins did not respond to UV-irradiation in this way, we conclude that this immobilization was due to actual engagement in DNA repair. When fluorescence recovery (of ERCC1-GFP) was monitored for longer periods of time after photobleaching, it appeared that individual immobilised proteins regained full mobility within five minutes. These results support a model of ‘on-the-spot-assembly’ of repair holocomplexes by repair proteins that arrive at DNA-lesions by random diffusion, and transiently bind during a single DNA repair event. The combined application of GFP-tagging and confocal FRAP methodology can provide insight into the dynamics of any inducible nuclear process in the living cell.

References:

1. D. Axelrod, D.E. Koppel, J. Schlessinger, E. Elson and W.W. Webb. (1976) Mobility measurement by analysis of fluorescence photobleaching recovery kinetics. Biophys. J. 16, 1055-1069.

2. N.B. Cole, C.L. Smith, N. Sciaky, M. Terasaki, M. Edidin and J. Lippincott-Schwartz. (1996) Diffusional mobility of Golgi proteins in membranes of living cells. Science 273, 797-801.

3. A.B. Houtsmuller, S. Rademakers, A.L. Nigg, D. Hoogstraten, J.H.J. Hoeijmakers, W. Vermeulen. (1999) Action of DNA repair endonuclease ERCC1/XPF in living cells. Science 284, 958-961.

 


6793

A new application for the Laser-Scanning Cytometer: Fast detection of aneuploidy induction in sperm of mice and humans

Adolf Baumgartner, Thomas Schmid, GSF - National Research Center, Oberschleissheim, Germany; Michael Nuesse, GSF - National research Center, Flow-Cytometry Group, Neuherberg, Germany; Ilse-Dore Adler, GSF National Research Centre of Environment and Health (GSF-Forschungszentrum fuer Umwelt und Gesundheit GmbH), Institute of Mammalian Genetics (Institut fuer Saeugetiergenetik), Neuherberg, Germany

Laser-Scanning Cytometry (LSC) as a diagnostic research tool automatically measures laser-exited fluorescence at multiple wavelengths (blue light up to infrared with 4 photomultipliers) on slides featuring relocation of every single cell. The cells were spread and fluorescence labeled directly on the slide. Particular chromosomes were tagged with fluorochromes being attached on centromeric or locus specific DNA-probes whereas the chromatin was counterstained with DAPI enabling multicolor visualization and cytometric assessment of those chromosomes. Missegregation of chromosomes during mitosis or meiosis, respectively, causes aneuploidy in daughter cells leading possibly to severe syndromes in humans. For comparing frequencies of spontaneous aneuploidy rates, FISH signals in epididymal sperm of young adult mice and human sperm were manually evaluated by fluorescence microscopy and automatically detected by LSC. Preliminary LSC data showed no significant differences from the expected sex ratio of 1:1 between normal sperm carrying chromosomes X or Y. The frequencies of hyperhaploid sperm in mouse obtained by LSC were 0.030% for chromosome 8, 0.010% for the X chromosome and 0.020% for the Y chromosome. The corresponding frequencies obtained by manual microscopy were 0.018%, 0.016% and 0.006%, respectively. The rates of disomic sperm evaluated by LSC in human sperm were 0.080% for chromosome 13, 0.048% for the X chromosome and 0.026% for the Y chromosome. The equivalent frequencies obtained manually by fluorescence microscopy were 0.050%, 0.044% and 0.010%, respectively. It can be concluded that LSC analysis has the powerful potential to replace manual microscopic FISH analysis of aneuploidy in sperm. In fact, rapid LSC was able to reduce the time-consuming manual scoring of fluorescent spots in a minimum of 10.000 sperm per slide from approximately 6 hours to marginal 30 minutes. Furthermore the Laser-Scanning Cytometer facilitates because of its capability of relocating suspected aneugenic cells detailed studies of aneuploidy induction in male meiosis caused by aneugenic chemicals or confounding factors like smoking or age.

Research funded by EU-contract ENV4-CT97-0471

 


3948

Identification of genetic markers for prostate cancer progression.

Herman Van Dekken, Pieter Jaap Krijtenburg, Wim C.J. Hop, Ries Kranse, Fritz H. Schroder, Janneke C. Alers, Erasmus University Rotterdam, The Netherlands; Hans J. Tanke, Carla Rosenberg, Leiden University Medical Center, Department of Molecular Cell Biology, The Netherlands;

Despite the high incidence of prostate cancer, at present only limited data are available on genes or chromosomes specifically involved in the initiation and progression of prostate cancer. We have applied comparative genomic hybridization (CGH) to routinely processed, paraffin-embedded, tissues at different timepoints in prostatic tumor progression to screen the tumor genome for gains and losses. Our panel included specimens derived from 56 patients (23 patients with primary, prostate-confined, carcinomas, 18 patients with regional lymph node metastases, 15 patients with distant metastases). Chromosome arms that most frequently showed losses, included 13q (55%), 8p (48%), 6q (43%), 5q (32%), 16q (25%), 18q (20%), 2q, 4q, 10q (18%), and Y (16%). Gains were often seen of chromosome arms 8q (36%), 17q, Xq (23%), 7q (21%), 3q, 9q (18%), 1q, Xp (16%). We were further able to delineate the ‘minimal overlapping region’ of a number of frequently altered chromosomal sites. A significant accumulation of genetic changes in distant metastases was observed, e.g. loss of 10q (P=0.03) and gain of 7q (P=0.03) sequences. In addition, investigation of a potential biomarker identified in previous studies by our group, i.e., extra copies of # 7 and/or #8, revealed a high prevalence of 7pq and/or 8q gain in the distant metastases (P=0.02). Importantly, gains were observed more frequently in tumors derived from progressors after radical prostatectomy, than in non-progressors (mean time of follow-up 74 months). Specifically, gain of chromosome 7pq and/or 8q sequences appeared an accurate discriminator between the progressors and non-progressors. Multivariate analysis showed that the number of chromosomes with gains, as well as gain of chromosome 7pq and/or 8q, were significantly related to progressive disease, also when grade and stage were taken into account (P= 0.003 and P=0.008, respectively). Additionally, an increase in the number of chromosomes with gains per case was related to a decrease in biochemical progression-free survival (Ptrend <0.001). More specific, the presence of gain of 7pq and/or 8q sequences markedly reduced the biochemical progression-free survival (P<0.001). In conclusion, this study has, firstly, documented the spectrum of chromosomal alterations insubsequent stages of prostate cancer, a number of which had not been described previously. It allowed us to identify chromosomal regions related to advanced tumor stage, i.e. loss of 10q24 and gain of 7q11.2 and/or 7q31 sequences. Secondly, gain of 7pq and/or 8q was identified as a genetic discriminator between progressors and non-progressors after radical surgery.


2:00 pm – 4:00 pm

Workshops A

Microarray Technology

Facilitators: Daniel Pinkel, Jeffrey M. Trent, Ian T. Young

Analysis of HIV Disease

Facilitators: Francis Mandy, Janet K.A. Nicholson

Technological Advances in Image and Flow Cytometry

Facilitator: James H. Jett

Analysis of Cell Cycle Regulation

Facilitators: James Jacobberger, Frank N. Traganos

Science Education: Linking With Students at all Levels

Facilitator: J. Paul Robinson


4:00 pm – 6:00 pm

Parallel Session IV: Microarrays Technology II

Chair: Ann-Christine Syvänen


5860

Oligonucleotide-array-based Comparative Genomic Hybridization

Russ Baldocchi, Dave Kowbel, Colin Collins, University of California at San Francisco, Cancer Center, San Francisco, CA, USA; Richard Glynne, Ed Tom, David Mack, Eos Biotechnology, South San Francisco, CA, USA; Joe Gray, UCSF - Mount Zion Cancer Research Bldg., San Francisco, CA, USA

Array-based Comparative Genomic Hybridization (Array CGH) provides a higher-resolution and more quantitative alternative to chromosome CGH for the assessment of genomic copy number abnormalities. In array CGH analyses published to date, copy number abnormalities are mapped onto arrays of cloned DNA sequences such as P1s, BACs or cDNAs. These methods, while much superior to chromosome CGH, are limited by the need to manage and manufacture the cloned DNAs and to suppress hybridization of interspersed repeated sequences. In addition, results may be distorted by elements in the array targets that are present at multiple copies in the normal genome. We have developed a PCR-oligonucleotide array approach (oligo-array CGH) that overcomes some of these limitations (see figure below). In oligo-array CGH, fluorescently labeled probes are prepared from test and reference DNA samples by multiplex PCR amplification. Primers are taken from dbSTS (currently 59K entries) or dbSNP (18K entries). These entries include mapping data and intervening sequence information, the latter enabling the design of 50-mer oligos for use as array elements. In our studies, four array elements are synthesized for each PCR primer pair: 2 sense and 2 anti-sense sequences that do not overlap each other or the primer sequences. In a "proof-of-principle" approach, multiplex PCR was performed for 10 or 24 loci. This method requires as little as 200 pg of starting genomic DNA template, and is thus amenable to analysis of micro-dissected tissue samples. Shown below are copy-number measurements (relative to a normal genome) that are similar to those obtained using the well-established FISH methodology. Amplifications of ZNF217 and AIBC1 in MCF7 and CCND1 in600MPE were clearly detected demonstrating the feasibility of this approach. The large-scale application of oligo-array CGH to breast and ovarian cancer samples will be demonstrated. Having 2 sense/anti-sense pairs of arrayed oligonucleotides per locus also enabled analysis of oligonucleotide specific hybridization. Interestingly, the intensity of hybridization varied as much as 100-fold between sense and anti-sense targets. We are exploring the hypothesis that this variation is caused by secondary structure of the probe molecules rather than by the DNA sequence of the array element oligonucleotides.

 


6056

FLUOROCYTOMETRIC ARRAY SYSTEM FOR BLOOD GROUP SEROLOGY, NUCLEIC ACID TESTING, AND IMMUNOASSAYS PERFORMED BY BLOOD BANKS

Girish Vyas, Manish Gandhi, University of California, San Francisco, CA, USA; Kodumudi Venkateswaran, Richard Langlois, Lawrence Livermore National Laboratory, Biology and Biotechnology Research Program, Livermore, CA, USA

Each blood donation requires screening for (1) blood group serology by hemagglutination assay since 1935, (2) viral infection markers by immunoassays since 1970, and (3) viremia by gene amplification technology since mid-1999. The purpose of our current research is to create an integrated diagnostic test system designed for multiplexed assays needed for screening the 14 million blood donations made annually in the U.S.A. Our research design is based on the accomplishments of(a) flow cytometric technology for the detection of HIV antigens, antibodies, and PCR-amplified nucleic acids of the blood-borne viruses, a UCSF-NHLBI joint project, (b) multiplexed blood grouping, a UCSF-LLNL joint project, and (c) Bioengineering, Advanced Microtechnology and Medical Technology Programs at LLNL. Our system is designed to employ an automated but modular system for laser-based fluorocytometric analyses, as a common technology platform applicable to the three disparate blood screening technologies, so as to meet the testing requirements of different sized blood centers. Currently, blood samples require screening with at least (i) 7 individual tests for determining the ABO and Rh red cell antigens, expected iso-antibodies, and any unexpected allo-antibodies, (ii) 7 individual immunoassays for HBsAg, HIV-p24 antigen, anti-HCV, anti-HBc, anti-HIV (1&2), anti-HTLV (I&II), and syphilis, and (iii) 3 nucleic acid tests for detection of amplified HIV, HCV, and HBV sequences. Our design embodies one test for blood grouping, one test for serum iso-/allo-antibodies, one test for microbial antibodies, one test for viral antigens, and finally, one test for detecting HCV-RNA, HIV-RNA, and HBV-DNA. The proposed fluorocytometric array technology permits analyses of blood groups and microbial antigens, antibodies and nucleic acids using a new generation of multiplex reagents that incorporate monoclonal antibodies, recombinant proteins, and oligonucleotides bound to micro-spheres. Thus, we can reduce the total number of routine blood screening tests by 70%, resulting in significant decrease in human errors, labor costs, and bio-hazardous medical waste. The modular design allows for the addition or deletion of tests, and flexibly that permits both manual and semi-automated screening in small blood centers and fully automated high-throughput systems in large blood centers. With demonstrated feasibility of this system in our laboratories, we ultimately anticipate a fundamental transformation of technology in blood screening to enhance worldwide safety of blood transfusion.

 


6119

Multiplexed Single Nucleotide Polymorphism Analysis by Flow Cytometry

Marie A. Iannone, J. David Taylor, Jingwen Chen, May-Sung Li, Philip Rivers, Kimberly Slentz-Kesler, Michael P. Weiner, Glaxo Wellcome, Research Triangle Park, NC USA

Single nucleotide polymorphisms (SNPs) are single base changes in genomic DNA. In humans, SNPs are the most prevalent form of genetic variation. As the sequencing of the human genome is completed, the need for a rapid SNP genotyping method will increase. We describe a system for SNP analysis that employs flow cytometric analysis of fluorescent microspheres. This system of SNP genotyping uses PCR-amplified genomic DNA that encompasses the SNP (‘target’ DNA), a synthetic ‘capture’ oligonucleotide probeand a fluorescent ‘reporter’. The capture probe contains both a sequence that is complementary to the target sequence and a 25-base ZipCode sequence. A green fluorescent reporter molecule is covalently attached to the capture probe using one of two methods, oligonucleotide ligation assay (OLA) or single-base chain extension assay (SBCE). In either method, after thermocycling, multiplexing of SNP analysis is made possible through the use of fluorescent microspheres where each set of microspheres is identified by its unique profile of red and orange fluorescence. Microsphere sets, each set coupled to a specific complementary ZipCode oligonucleotide sequence, are hybridized to the capture probes. Flow cytometric analysis of the microspheres simultaneously identifies both the microsphere type and the green fluorescent signal associated with the SNP genotype. Application of this system will be demonstrated with multiplexed genotyping of DNA samples for SNP markers located near the ApoE locus on chromosome 19. Multiplexed SNP genotyping by OLA or SBCE with flow cytometric analysis of fluorescent microspheres is an accurate and rapid method for the analysis of SNPs.

 


6325

Compact Disk player based fluorescent reader of magnetically aligned objects in an array

Arjan G.J. Tibbe, Bart G. De Grooth, J. Greve, University of Twente, Enschede, The Netherlands; Leon W.W.M. Terstappen, G.J. Dolan, Immunicon Corp., Huntingdon Valley, USA

A schematic representation of the reader is shown in Figure 1. 6 micron polystyrene beads labeled with ferromagnetic nanoparticles as well as the fluorescent dye Cy5 are placed in a specially designed chamber that is placed between two wedge-shaped magnets. The optically transparent uppersurface of the chamber contains ferromagnetic lines of nickel (Ni) deposited by lithographic techniques. The magnets create a gradient moving magnetically labeled objects upwards to the top of the chamber. When they reach the top of the chamber, they become subject to a high local gradient induced by the Ni lines and align in between them. Non magnetic ones move down under the influence of gravity. In this way, the beads labeled with magnetic particles are well aligned at the upper surface of the chamber. A 635nm laser-diode is focussed on the Ni lines through an objective of a common compact disk player. A feedback system using the laser light reflected from the Ni lines was developed to keep the laser focussed on a line of beads while the chamber is moved in the Y-direction. After one line is scanned, the laser jumps to the next line and the chamber is moved in the opposite Y-direction. By moving the chamber the beads pass under the laser focus one after each other. For each bead the fluorescence signals generated are analyzed and the position recorded. The sensitivity of the system was compared to that of a flow cytometer by analyzing a mixture of beads that contained beads with four different Cy5 intensities. Figure2 shows clearly that the new method can resolve all four populations whereas the flow cytometer could only resolve the brightest labeled beads.

 


6405

An Applciation of Genetic Algorithm to Oligonucleotide Hybridization

Hiroshi Dozono, Shigeomi Hara, Yoshio Noguchi Dr., Faculty of Engineering Saga University, Saga, JAPAN

Recently, many kind of DNA chips are developed and it will be possible to analyze DNA sequences raplidly from the results of hybridization between them. We propose an implimatation of Genetic Algorithm(GA) for oligonucleotide hybridaization using the DNA chip which is constructed as an array of N-mer DNA probes. GA uses the reconstructed sequences as individuals and the fitness function is given as follows fitness=num+CS*num2/len-CM*miss where num is the number of reconstructed probes, len is length of the sequence and miss is the number of miss hybridization. CM and CS is the constant which evaluate the efficency of the probe usages and penalty of the miss hybridization respectively. We introuced special setups to operates the relatively long individuals. At first grouping GA to keep the diversity of the individuals by evolving each group almost independently. Secondly, constraint based crossover, which searches better sequences keeping the constraint of probe connections by combining the parental sequences. We made some experiments of sequence reconstructions by computer simulations changing the length of the sequenes and length of the probes. The changes of the number of the reconstructed probes and sequence length is shown in Fig.1. We used a sequence of 11166 DNAs from Genbank data base and probes of 9 mers. As the results, the avarage of the length of the obtained sequences becomes 10790.4. The Oligonucleotide hybridization method can not reconstruct completely same sequence by nature. So we evaluate the accuracy of the result by the indexes for sub-sequences of motif length. Table 1 shows the results. Preservation index is the rates of the completely matched subsequence accuracy index is the avarage rates of matched DNAs in specified length subsequences between the original sequence and reconstrcuted sequences. Using sufficiently long probes, the sub-sequences are preserved to check the existences of motifs.

 


6459

Optical transduction of structural changes of proteins using surface enhanced nano cluster resonance

Thomas Schalkhammer G.M., Norbert Stich, Georg Bauer, TU-Delft, Delft, The Netherlands; Roland Palkovits, Fritz Pittner, University of Vienna, Vienna, Austria

A surface enhanced metal nano cluster based chip device is established as a platform for gaining, analyzing and managing complex biorecognitive and structural information in order to improve structural studies on protein and DNA molecules as well as screening, diagnosis, monitoring and treatment of diseases. The system consists of disposable probe arrays containing proteins or gene sequences as well as small ligands on a chip, an instrument to process the biochip arrays, a scanner, and software to analyzeand manage biorecognitive and structural information from the chip. Device production integrates thin-film and semiconductor fabrication techniques, solid phase surface chemistry, biorecognition techniques, molecular biology, software and robotics. Conformational and structural changes of ultra thin protein layers deposited on the chip (caused by changes in microenvironment) were transduced into an optical signal and observed directly as a colour change of a biochip surface. We have successfully coated proteins as thin films of 10 to 300 nm (about 2-100 molecules thick) onto optically reflecting chip-surfaces via various techniques including microdotting, spin-coating and photocrosslinking. A novel optical resonance effect, surface enhanced absorption, was triggered by deposition of metal nano-clusters on top of the protein molecules. The response of this protein biosensor array was measured spectroscopically in the visible and IR range of the spectrum via an optical scanner at 5-9 micron resolution and a data-rate of 60 MBytes/minute. This set-up enabled us to quantitatively and reversibly transduce changes of biomolecule conformation into a signal directly visible to the human eye.

 


6511

High Throughput Molecular and Genomic Analyses Using Microspheres

John Nolan, Life Sciences Division, Hong Cai, P. Scott White, Alina Deshpande, La Verne Gallegos, Rhiannon Nolan, Yolanda Valdez, Sabine Lauer, Bioscience Division, Los Alamos National Laboratory, NM, USA

The imminent completion of the human genome project promises great opportunities, as well as great challenges to biological researchers. Primary genomic sequence contains a wealth of information, but faster and more efficient experimental methods are required to extract meaningful biological insights. Recently, we have developed improved approaches for high throughput genetic and molecular analysis using microspheres and flow cytometry. Among the insights emerging from the human genome project is that the DNA sequence from any two individuals is 99.9% identical, and that the phenotypic differences between individuals are conferred largely by the 0.1% of the sequence that is different. The vast majority of this sequence variation is in the form of single nucleotide polymorphisms (or SNPs), sites in the genome where a single base varies between chromosomes in the same individual or between different individuals. As genetic markers, SNPs have great potential for use in disease diagnostics and drug discovery, but traditional methods of genotyping are too slow and expensive. We have developed a high throughput method for large scale SNP scoring based on single base extension (SBE) of oligonucleotide primers using microsphere arrays. This system provides accurate genotyping in a flexible format with ten fold higher throughput and ten fold lower costs than conventional genotyping methods. The human genome project will also identify thousands of potential genes, which code for proteins of unknown function. Efficient and quantitative large-scale approaches to measuring protein function are essential elements in studying both known and predicted proteins. Flow cytometry has several advantages for the analysis of molecular assembly and function, including sensitivity, speed, and flexibility. We are applying these advantages to the large scale screening of molecular interactions as well as the detailed mechanistic analysis of protein function. The combination of quantitative fluorescence measurements with sensitive screening capabilities in a flexible platform with multiplexed analysis capabilities make microsphere-based flow cytometry a powerful complement to cell-based assays for addressing large scale biological questions in the post-genomic era. Supported by NIH, DOE, and LANL.


Parallel Session IV: Cytometry for the Industrial Laboratory


4060

Flow Cytometric Cell Sorting - An Indispensable Tool for Drug Discovery

Lisa Green, Fred Chadwell, Mary Beth Anania, Larry Mann, Jonathan Lee, Philip Marder, Eli Lilly & Co., Indianapolis, Indiana, USA

The discovery of new pharmaceuticals is being fueled by two new disciplines: genomics and combinatorial chemistry. Genomics has identified an abundance of novel molecular targets while combinatorial chemistry has produced an explosion of new molecules to screen. To meet these challenges, Pharma has developed high throughput screening procedures. The increasing complexity of target systems requires more sophisticated methods for HTS, such as cell-based assays. FACS sorting can enhance the specificity and sensitivity of cell-based assays. In this presentation we report how our laboratory has used cell sorting strategies based on mobilization of intracellular calcium and transcription of a reporter gene to generate cell lines amenable to HTS screening. The calcium flux assay is a popular HTS method because it is a rapid and direct signal for biological activity of many receptors. Using fluorescent calcium indicators (Indo-1, Fluo-3) and a cell sorter fitted with Cytek Time Window module, we have developed cell lines with superior receptor-coupled calcium responses. A direct comparison of the sorted and unsorted cultures was done using a fluorometric imaging plate reader. Our results indicated that the sorted cultures had a10- fold increase in signal window over unsorted cells. Assays of transcription reporter genes are another important tool for HTS. Novel cell lines that express the recombinant receptor, transcriptional response elements, and a reporter gene can be used to screen compounds for effects on various signal transduction pathways. The beta lactamase reporter system (Aurora Biosciences) utilizes a FRET-based, fluorescence ratio readout of reporter enzyme activity. We used FACS to isolate cell lines with low constitutive activity and transcriptional responses specific to targeted receptors. These sorted cell lines had 50 fold response windows when tested in a HTS assay format. Flow cytometry increases productivity of cell line development efforts through the efficiencies of automated cloning and screening of many thousands of cells for a desired phenotype. Furthermore, we have shown that flow cytometric cell sorting provides sensitive, reliable cell lines with superior signal to noise windows in cell-based assays.

 


4102

Comparison of transient transfection efficiency using commercially available transfection reagents.

O. Joseph Trask, Jr. and Thomas H. Large, Sphinx Pharmaceuticals, A Division of Eli Lilly and Company, Research Triangle Park, NC

The importance of determining transfection efficiency and cell survival of transiently transfected living cells is critical for evaluating exogenous protein expression. In this study, five different commercially available transfection reagents were tested in several cell lines using green fluorescent protein (GFP) reporter to measure transfection efficiency. GFP expression was analyzed by flow cytometry and morphological characteristics of GFP-transfected cells were observed by fluorescent microscopy. Cell viability was determined by using the nucleic acid dye propidium iodide (PI). Using the two-color flow cytometry analysis of GFP and PI provides a fast and quantitative measurement of the transfection efficiency and cell survival, respectively. This study documents differences in transfection efficiency and cell survival of various cell lines using different commercially available transfection reagents.

 


5849

FLOW CYTOMETRY APPROACHES TO ASSESS STRESS AND FOOD PRESERVATION EFFECTS ON BACTERIAL PHYSIOLOGY AND SURVIVAL

Joerg Ueckert, Unilever Research Vlaardingen, Microbiology & Preservation, Vlaardingen, The Netherlands

To evaluate effects of moderate single or combined preservation steps, methods are required which provide information about viability and damage distributions within bacterial populations as a function of treatment, culture condition and media composition. Flow cytometry in conjunction with selected fluorescent probes have been chosen to assess physiological and structural effects of low medium pH, physical stresses and antimicrobial activity in food spoilage and pathogenic bacteria. Several approaches are examplarily presented and critically discussed: The multi-purpose probe carboxyfluorescein (diacetate) succimidylester CF(DA)SE allowed proliferation tracking of Lactobacillus plantarum populations for up to 8 generations and direst determination of lag times induced by mild heat treatments. A combination of CFSE and propidium iodide (PI) disclosed resistant/susceptible subpopulations upon (combined) nisin and lysozyme treatments. Recovery of acid stressed Listeria and E. coli spp. could be linked todifferences in internal pH, determined by CFSE fluorescence (using the ratio of a pH-dependent and a pH-independent emission wavelength) and to membrane permeability variation, evidenced by ethidium bromide (EB) uptake. Acid stress also decreased membrane potential levels significantly, which was shown by DiOC6(3) fluorescence. Action of antimicrobial compounds like carvacrol or sorbic acid on the outer and cytoplasmic membrane of E. coli could separately be assessed by combined application of PI and NPN (1-N-phenylnaphtylamine). The latter enhances its fluorescence in hydrophobic environments, e.g. phospholipids. This dye combination not only revealed induced damage, but also allowed identification of differences in cellular targets, depending on the agent. References Helander, I.M. et al (1998). Characterization of the action of selected essential oil components in Gram-negative bacteria. J. Agric. Food Chem. 46:3590-3595. Davey, H.M. and Kell, D.B. (1996) Flow cytometry and cell sorting of heterogeneous microbial populations - the importance of single-cell analysis. Microbiol. Rev. 60: 641-696. Ueckert, J. et al. (1997). Flow cytometric analysis of Lactobacillus plantarum to monitor lag times, cell division and injury. Lett. Appl. Microbiol. 25:295-299.

 


5957

Cell Encapsulation and Flow Cytometry - Incorporation into Cell Line Development Programmes

Allison Vernon, Hilary Metcalfe, John Birch, Lonza Biologics, Berkshire, UK; Adrian Knight, USA

The productivity of a cell line selected for the manufacture of a biopharmaceutical has a major influence on the economics of the production of that molecule. Expression systems, including those based on the glutamine synthetase system, can yield highly productive stable cell lines. It is nevertheless important when developing such cell lines to monitor cell line stability. Cell encapsulation technology, such as that offered by the One Cell gel microdrop system, can be used to rapidly monitor changes within a cell population by flow cytometry. The cell encapsulation technology was first developed using relatively low producing hybridoma cell lines. When this system was first tested with a highly productive recombinant GS-NSO cell line, leakage of excess product out of the microdrop masked the presence of non-producing sub-populations. Following optimisation, the technique could quantify the presence of non-producing cells when they were present at 5% of the population. The limit of detection for the presence of non-producing cells within a population was approximately 0.1%. This methodology has been developed to complement more traditional techniques for examining cell line stability. A study determining its value in cell line development programmes will be presented.

 


6133

Application of Flow Cytometry to Bioprocess Development, Validation and Control

Paul Holmes, David Lloyd, Mohamed Al-Rubeai, University of Birmingham, School of Chemical Engineering, UK

Analysis of cellular components using intercalating fluorescent agents to stain nucleic acid or analysis of intracellular proteins using fluorochrome labelled antibodies are common place. However, it is now possible to easily stain cells according to quantity of secreted product as opposed to intracellular product. Our method, affinity capture surface display (ACSD) uses a biotin/avidin affinity system to localise capture antibodies at the surface of the cell to bind with secreted product. Cross feeding of secreted product between cells is prevented by artificially elevating the viscosity of the medium. Following a period of secretion, secreted product is detected using fluorescent labelled antibodies and analysed/ sorted using conventional fluorescent activated cell sorting procedures. This method enables rapid and objective isolation and cloning of the highest producing cells from a heterogeneous population. Indeed, the process can be completed within 6 weeks compared with at least 8 months for traditional limiting dilution and colony expansion screening techniques. Also, this technique can be adapted to use a fluorescent labelled antigen as the detection agent enabling selection of antigen specific antibody secreting cells which represents a considerable time and cost saving compared to conventional methods. Industrial cell line stability is frequently validated over a 100-day period with the cells passaged at normal intervals. As an alternative to measuring supernatant product concentrationusing assays such as ELISA, cell line stability can be monitored in terms of single cell secretion rate using ACSD. The ability of FC to monitor large numbers of individual cells, compared with ELISA methods which determine the population mean values, allows instability to be rapidly detected as a lower productivity sub-population. Such timely information can aid the decision to re-clone the cells if it is believed that the instability is a result of overgrowth of the producer cells by a lower productivity variant. For cost effective production, fed batch or continuous cultures are favoured over simple batch processes. Optimising the feeding strategy is critical to ensure the best growth and productivity. There are several ways to monitor the physiological status and nutrient requirements of the cells. Most of these parameters have significant lag between sampling and obtaining data on which to base process decisions. The percentage of S phase cells within a population correlates with specific growth rate and therefore, FC measurement of cell cycle enables reliable estimates of specific growth rate to be rapidly obtained. Thus enabling tighter control of bioprocess operation and significantly increasing productivity.

 


6307

Flow Cytometry as a Tool for Pharmaceutical Assay Development

Thomas Knapp, Aurora Biosciences Corporation, San Diego, CA, USA

We have used flow cytometry as a tool for efficient pharmaceutical assay development using functional readouts including calcium flux, GFP expression, membrane potential and reporter gene expression. This presentation will focus on the use of ß-lactamase, a highly sensitive reporter gene with no known endogenous mammalian background, using the cell-permeant, ratiometric, fluorogenic substrate CCF2/AM. In addition, a comparison is made of the ß-lactamase and ß-galactosidase reporter gene systems. The robust and reliable detection of single positive cells and the correspondingly low incidence of false positive events have facilitated the use of ß-lactamase as a reporter gene for flow cytometry, including applications such as functional genomics and the reliable selection of engineered cell lines with functional clones occurring at frequencies <10-4. Overall, the ß-lactamase reporter system has been combined with flow cytometry to produce a large number of clonal cell-based screens for a wide range of therapeutic targets for industrial scale drug screening.

 


6484

Murine Bone Marrow Toxicity Assessment in Pre-Clinical Drug Development by Flow Cytometry

Padma Kumar Narayanan, Donna Williams, Melanie Quinlan, Cindy Zhang, Terry Sellers, Timothy Hart, Kent Gossett, SmithKline Beecham Pharmaceuticals, King of Prussia, PA USA

Microscopic evaluation of bone marrow for the delineation of erythroid and myeloid composition is time consuming, labor intensive, and subject to human error or bias. Limited numbers of cells are evaluated and only semiquantitative morphologic information is obtained. In contrast, flow cytometry can assess phenotype and function at the single cell level in a rapid, reproducible, and objective manner. We exploited this potential of flow cytometry to evaluate murine bone marrow toxicity in pre-clinical drug development. A panel of monoclonal antibodies (CD45, CD3, CD11b, TER119, CD117, Ly6E-A, CD34) was used to phenotype mouse bone marrow progenitors of the lymphoid, myeloid, and erythroid series and stem cells. Consistent with flow cytometric evaluations of human bone marrow, murine bone marrow cells could be uniquely identified on a cytogram of CD45 fluorescence plotted against forward angle scatter. The maturity and lineage characteristics of these discrete areas were then determined by electronic gating and evaluation for the presence or absence of CD34 with and without the lineage markers. Microscopic evaluation of Wright’s stained cytospin preparations of sorted cells was used to confirm flow cytometric classification of proliferating and non-proliferating lymphoid, myeloid, and erythroid populations. This methodology facilitated identification of selective modulation of specific bone marrow populations by novel xenobiotics in drug development in a rapid and cost-effective fashion.

 


6585

Sorting, In Vitro Differentiation and Phenotyping of Mouse and Human Osteoclast Precursors

Gary Elliott, Laura Chiu, Raffi Manoukian, W. J. Boyle, D. L. Lacey, Victoria Shalhoub, AMGEN, Inc., Thousand Oaks, California, USA

Osteoclasts (OC’s) are bone resorbing cells derived from hematopoietic precursors that, along with osteoblasts, are important in maintaining bone morphogenesis and remodeling. Because increased osteoclast activity has been implicated as a principle feature of osteoporosis and other osteopenic diseases, it is important to understand OC development and differentiation. However, studies using flow cytometry to characterize OC’s and their precursors (OCP’s) have been limited due to the difficulty in obtaining and establishing cultures of OC’s. In addition, although some studies have been carried out to characterize murine OC precursors, very little is known about human osteoclastogenesis. In several recent reports, investigators from AMGEN have reported the cloning and characterization of two key positive and negative regulators of osteoclastogenesis. Osteoprotogerin (OPG) and its ligand (OPGL), are two new members of the tumor necrosis family of proteins that inhibit and promote OC growth differentiation, respectively. Here, we report on the application of flow cytometry and confocal microscopy to aid in the determination of OPG-Fc binding, expression-cloning of OPGL and subsequent use of a FITC-labeled OPGL to specifically select and phenotype OCP’s directly from normal mouse bone marrow and human peripheral blood. Additional studies utilizing FITC-OPGL have demonstrated that the OCP is derived from a pure monocyte fraction (as determined by CD14 dual staining) and also expresses the receptor activator of NFkB (RANK). All of the cells positive by FITC-OPGL staining were also positive for CD11b, CD38, CD45 and CD54, but were negative for CD34, CD3, CD20, CD19, and CD56 consistent with the surface staining characteristics of a myelomonocytic precursor cell. These studies will be beneficial in further characterizing OC differentiation and have implications for the clinic as it will be important to follow effects of such potential therapeutics as OPG on OCP’s and other peripheral blood cell populations.


Parallel Session IV: Cell Death I

Chair: Zbigniew Darzynkiewicz


5549

Square wave pulse ectropermeabilization of cytochrome c induce apoptosis

Etienne Morel, Franck Sureau, Patrice X. Petit, INSERM U129, Paris, France; Bruno Gabriel, Justin Teissié, IPBS, CNRS, Toulouse, France

It has been shown that cytochrome c is released from mitochondria during apoptosis [1] following a drop in mitochondrial membrane potential and generation of superoxide anions [2,3]. Cytochrome c release from the mitochondrial intermembrane space is associated with the opening of the so-called transition pore and/or disruption of the outer mitochondrial membrane [4]. Cytoplasmic cytochrome c interacts with dATP and Apaf-1 in a complex call ‘apoptosome’ to activate procaspase 9 which in turn activate the procaspase 3 in active caspase-3. To establish wether cytochrome c can primarily induce apoptosis we used direct electropermeabilization of cytochrome c into CHO cells, jurkat cells and hepatocytes in suspension culture. Cells were permeabilized by square wave pulse under hyperosmotic conditions which allow high post-pulse viability (over 85%) either at short time (1h) or at 24h. Apoptosis was detected by annexin-V-FITC (exposure of phosphatidyl serine), viability (Propidium iodide), DNA fragmentation (Tunnel assay), hypoploidy (subploidy) and caspase-3 measurements. The mitochondrial membrane potential was also measured with DiOC6(3). At 4h post-pulses aberrant phosphatidyl serine exposure is apparent with a lower viability of the cells. These avents are concomittent to caspase-3 activation. Surprisingly, the mitochondrial membrane potential is also altered indicating of a feed-back action of the activated caspases [5] onto the mitochondrial membranes. Whereas, apoptosis induced by fas ligandis blocked by Bcl-2, exogenous cytochrome c induced apoptosis is only partially controlled by bcl-2. The transition pore inhibitors, i.e. cyclosporin A and bongkrekic acid, also fails to protect mitochondria from caspase activation. The system described offers a novel and simple approach for investigating regulation of apoptosis by apoptogenic proteins. [1] Liu, X.et al. (1996) Cell 86, 147-157 [2] Petit, P.X. et al. (1995) J. Cell. Biol. 130, 157-167 [3] Petit, P.X. et al. (1996) FEBS Lett.396, 7-14 [4] Petit, P.X. et al. (1998) FEBS Letters 426, 11-116. [5] Marzo, I. et al. (1998) FEBS Letters 427, 198-202.

 


6121

New approaches in analysis of cell death by flow and laser scanning cytometry

Zbigniew Darzynkiewicz, Xun Li, New York Medical College, Hawthorne, NY USA; Elzbieta Bedner, Zaklad Patomorfologii, Ul. Unii Lupelskiej #1, Poland

Thus far the existing cytometric methodologies did not allow one to directly correlate, within the same cells, functional cell attributes that are revealed supravitally, with features that to be detected or measured require cell fixation and permeabilization. Laser scanning cytometer (LSC) provides the unique capability to analyze the same set of cells repeatedly, record the data in a list mode fashion including the XY position of each measured cell on the slide, and integrate ("merge") results of sequential mesurements into a single file. Taking an advantage of this feature of LSC we have been able to correlate the supravital changes that occur during apoptosis, namely drop in mitochondrial transmembrane potential (DFm), decrease in intracellular pH, and generation of reactive oxygen intermediates (ROIs)with the features revealed by analysis of fixed cells: cell cycle position, PARP cleavage and DNA fragmentation. Furthermore, by virtue of the LSC capabilities that allow to probe cell morphometry the events that precede apoptosis such as Bax accumulation in mitochondria and activation of NF-kB (its translocation from cytosol to nucleus) also could be measured. Several experimental models of induction of apoptosis were studied, including apoptosis triggered by DNA damage by antitumor drugs and by the cell surface ligands such as TNF- a,CD95 or TRAIL in different cell lines (HL-60,Jurkat,MCF-7,LNCaP). Using this approach it was possible to map the sequence of events associated with the triggering and execution stages of apoptosis and to reveal whether the preceding events are essential for the subsequent ones to occur. Interestingly, for example, in some cells PARP cleavage and DNA fragmentation was observed with no evidence ofdissipation of DYm or drop in ROIs. This suggests that the permeability transition or oxidative stress are not always a prerequisite for initiation of the execution stages of apoptosis. When possible, the observations by LSC have been correlated with cells’ analysis by multiparameter, multilaser flow cytometry. Our approach opens a possibility to study causal and temporal relationships, within the same cells, between cellular changes detected by functional assays of live cells and the changes that cannot be analyzed supravitally. This allows one to map the sequence of molecular and biochemical events that occur not only during apoptosis but also in the course of mitogenic stimulation, differentiation, carcinogenesis or cell response to drugs.

 


6192

Continuous monitoring and analysis of the apoptotic process in individual cells

Naomi Zurgil, Yana Shafran, Menachem Kaufman, Mordechai Deutsch, Jerome Schottenstein Cellscan Center, Bar-Ilan University, Ramat-Gan, ISRAEL

Apoptosis is a dynamic process of variable duration. The time window during which the death process is identifiable varies depending on the measured parameter, cell type, or the nature of it’s inducing agent. Since high variability exists within cell populations with regard to their kinetics and the terms of their death course, the ability to continuously detect the death process occurring in single or sub-groups of cells is crucial in identifying apoptotic cells within a complex population using several parameters. The Cellscan Mark-S (CS-S) apparatus was used in the present report for the continuous monitoring of the apoptosis process. The CS-S is a multiparametric multilaser scanning cytometer system that performs sequential optical measurements of individual living cells. A unique cell carrier allows up to 10,000 viable cells to be maintained, manipulated and monitored repeatedly, over extended periods in addressable locations. Using the CS-S, fluorescence intensity, polarization, kinetic measurements, and cluster analysis of sub-population were carried out utilizing various fluorescence probes (FDA, Annexine V, Rhodamine 123, fluorescent caspase substrate, and PI). Intact living lymphocytes or fresh thymocytes were loaded onto the cell carrier and apoptosis was induced by introduction of H2O2, dexamethasone or anti Fas antibodies. The death process was monitored on-line by making continuos sequential measurements of the same individual cells, of their functional apoptotic parameters such as the changes in the cellular viscosity, the cellular plasma membrane asymmetry and integrity as well as in the mitochondrial membrane potential. Additional structural data was obtained following chemical fixation of the cells onto the cell carrier and staining with fluorescent nucleic acid probe. The results show a distribution in kinetics of the cellular changes in the sub-populations or individual cells undergoing apoptosis.

 


6330

Individual radiosensitivity: search for genetic markers

Annette Schmitz, Commissariat a l’Energie Atomique, Fontenay aux roses, France; Jan Bayer, Benedicte Thuille, Fabienne Dufour, Lucien Cazes, Claudia De Toma, Gilles Thomas, Fondation Jean Dausset - CEPH, Paris, France

The stability of genomes can be compromised by environmental factors such as ionizing radiation. Lacking efficiency of repair mechanisms can lead to cell death, or transmission of corrupt genetic information. Many of the transmitted genetic alterations will have little effect, but in general will at least contribute to aging of cells and organisms. Different observations suggest that inherited factors contribute to radiation sensitivity variability within the population. Those observations justify the development of a strategy aiming at the identification of genes involved in this trait. Modern techniques allow to perform genetic linkage studies , without any hypothesis on the nature of the genes involved. The only prerequisite is the development of a reliable test of individual radiosensitivity, allowing a quantitative evaluation of in vitro responses to irradiation. Environmental factors influencing the test should be kept under control, to be able to measure inter-individual differences related to the genetic constitution of the individual. We focused on the development of such a quantitative test of individual radiosensitivity, in the aim of applying it to normal individuals’ blood samples. Our test assesses dose-dependent induction of apoptosis inspecific lymphocyte subsets, after in vitro irradiation. Dose-effect curves are established for each of the subsets of each of the individuals tested, and the slopes of these curves are calculated and compared. We used a combination of subset markers and annexin-V in a four-color flow cytometric assay, allowing us to simultaneously determine radiosensitivity in the major T-cell subsets, as well as in the B-cell compartment. We also applied our test to B-lymphoblastoid cell lines (BLCL) of CEPH-family members, and mono- and dizygotic twin-pairs. Using these different cell models, we assessed reproducibility and discriminating power of the test. We obtained highly reproducible measures of individual radiosensitivity in serial blood samples of five individuals, taken at 2-week intervals. We observed a wide range of reactivities between different individuals. The studies on the BLCL showed that within a given family, the levels of radiosensitivity observed for the children can largely exceed those observed for the parents. Our twin studies indicate the heritability of the trait, although the number of twin-pairs included in this study does not yet allow us to reach statistical significance. The presented approach will be applied to a large cohort of normal volunteers (approx. 700 persons) of an insular population with a high degree of consanguinity, allowing the study of genetic linkage in relation to individual radiosensitivity.

 


6403

Monitoring Cell Death by Laser Scanning Cytometry

Ed Luther, Louis Kamentsky, CompuCyte Corp., Cambridge, Massachusetts, USA

Non-confocal laser scanning cytometry allows automated multiparameter analysis of cells attached to a microscope slide. As the location of cells on the slide is recorded as a feature, cells can be relocated and visualized, and slides can be rescanned and the two resultant data files merged to obtain population data changes for different experimental conditions, including scanning the same slide at different time points. In necrobiology, a recently proposed term for the disciplines related to cell death, it is important to distinguish between active cell death or suicide (apoptosis) and death caused by extrinsic factors (necrosis). Morphological examination of cells can be used to aid in distinguishing between the two pathways. In these experiments, HeLa cells or human skin fibroblasts (HSF) were cultured in microscope chamber slides. The slides were stained with a live/dead staining cocktail of propidium iodide and Syto16™, the chambers were removed, and a coverslip was placed over the gasket on the slides, with staining media remaining between the coverslip and the slide. The slides were analyzed on the LSC, incubated at 37 degrees C, and reanalyzed at various time points ranging from one hour to three days. In some experiments, camptothecin was added to the cell cultures to induce apoptosis. The methodologies possible in these types of experiments include comparing the population data from the various time points, obtaining galleries of cells selected on their population features in the initial analysis and obtaining images of the same cells at later time points, and viewing cells at a later time point based on their features as they were measured at an earlier time point. Figure 1 shows galleries of images of cells relocated from regions in a scattergram showing different cell morphologies. Figure 2 shows a time course analysis of cells relocated based on their staining characteristics in the initial analysis.

 


6690

Dynamics of growth and death in antigen specific swine T-cell subset proliferation in vitro.

Kristi Harkins, Ray Waters, Josep Bassaganya-Riera, Michael Wannemuehler, Iowa State University, Ames, Iowa, USA

Understanding the means by which protection is induced with vaccination is important to defining the nature of the vaccine design. We have developed a 5-color flow cytometric assay for measuring proliferation (PKH2) and cell death (Annexin V and 7AAD) in conjunction with two antibody markers to T-cell subsets in an effort to evaluate the effect of bacterial and viral antigens on the dynamics of T-cell activation and death in vitro. Peripheral blood lymphocytes collected from pigs vaccinated with Brachyspira hyodysenteriae or pseudorabies virus were PKH2 labeled and cultured in the presence of antigen, concanavalin A or media only. Samples were taken over a seven day culture period and the lymphocyte subsets measured included CD8a vs CD4, CD8a vs CD8b, CD8a vs CD8gs and CD3 vs IgM using phycoerythrin and ECD tandem antibody combinations. Dual laser excitation with spatially separated 488 nm and 633 nm laser lines allowed the measurement of PKH2, phycoerythrin, ECD tandem, Annexin V-APC and 7AAD fluorescence. Annexin V positive cells were found to be smaller and more granular in nature with few apoptotic cells observed in the normal lymphocyte gate. The CD markers were readily discernible on the Annexin V positive cells which facilitated detection of lymphocyte subsets within the proliferative or nonproliferative populations which were dying. This in turn allowed us to characterize the dynamics of antigen specific T-cell growth and death with time in vitro. Experimental parameters such as fluorescence resonance energy transfer, fixation prior to analysis and the observed population dynamics of antigen response will be discussed.

 


7018

Prooxidant and Antioxidant effect of Melatonin on Human Liver Cell Line: Assessment using Microplate Cold Light Cytofluorimetry (MCCM) Directly on Live Adherent Cells

Raouf A. Osseni, Patrice Rat, Andre Bogdan, Jean-Michel Warnet, Yvan Touitou, Laboratoire de Biochimie Médicale - CHU Pitié Salpêtrière, Paris, France

Flow cytometry (FCM) needs monodispersed or suspension cells for oxidative stress evaluation. This study presents the role of cold light cytofluorimetry (MCCM) to assess different oxidative stress targets (glutathione, free radicals), directly on live adherent cells in microplates.

The aim of this study was to evaluate melatonin cytotoxicity in adherent liver cells, by measuring its effects on various cellular targets. Cell viability, intracellular reduced glutathione (GSH) level, and reactive oxygenspecies (ROS) production were assessed in the human adherent liver cell line (1) (HepG2), after incubation, with increasing melatonin concentrations (0.1 - 10,000 µM). The incubation times tested were 24, 72, and 96 h for cell viability and intracellular GSH level, and 15 and 45 minutes for ROS production. Cellular target evaluations were possible in living cells by means of a new microplate cold light cytofluorimeter (Fluorolite -DYNEX°) which permits to scan and detect fluorescent signals directly in live adherent cells on 96-well microplate. This technology (2) was suitable for the assessment of cell viability, GSH level, and ROS overproduction with, respectively, neutral red, monochlorobimane (mBCl), and 2˘,7˘-dichlorofluorescin diacetate (DCFH-DA) fluorescent probes (1). At the lowest melatonin concentrations (0.1 - 10 µM) and for a relatively short incubation time (24 h), the antioxidant effect of melatonin was revealed by an increased intracellular GSH level. After longer incubation (72 and 96 h), cell viability decreased and GSH was depleted with a very significant cytotoxic effect. Moreover, high melatonin concentrations (1,000 - 10,000 µM) induced oxidative stress with increased ROS production and GSH depletion. This dual effect is strong evidence that, in vitro, the effect of melatonin on the human liver cell line can be both antioxidant and prooxidant, depending on the concentration and incubation time. Moreover Microplate Cold Light CytofluoriMetry (MCCM) technology and MiFALC tests (3) directly on live adherent cells are adapted for oxidative stress assessment and pharmacotoxicology screening study.

References:

(1)- R.A. OSSENI et al. (1999) Toxicol in Vitro, 13 : 683-688

(2)- P. RAT et al. (1995) Methods Enzymol., 252 : 331-340

(3)- P.RAT et al. (1997) Animal Alternatives Welfare and ethics (Van Zutphen and M. Balls Eds) elsevier : 813-823


Parallel Session IV: Basic Immunology

Chair: Alexander Scheffold


6315

IDENTIFICATION OF A HUMAN RECENT THYMIC EMIGRANT PHENOTYPE

Richard Mcfarland, UT Southwestern Medical Center at Dallas, Dallas, Texas, USA; Daniel Douek, Richard Koup, Louis Picker, UT Southwestern Medical Center at Dallas, Department of Internal Medicine, Dallas, Texas, USA

The ability to measure human thymic output would be an invaluable tool for the study of the development of the naïve T cell repertoire, as well as naïve T cell regeneration after intensive cytotoxic chemotherapy or effective anti-retroviral therapy of progressive HIV infection. We have previously demonstrated that quantification of T cell receptor rearrangement excision circles (TREC) within peripheral T cell populations provides insight into the frequency of recent thymic emigrants, (RTE) and therefore into thymic function (Douek et al, 1999, Nature, 396:690-694). However, measurement of RTE by this approach is complicated by the fact that TREC levels are also determined by turnover within the naïve T cell compartment. Here, we report a flow cytometric approach to RTE phenotype measurement. We demonstrate that CD103 expression is upregulated very late in thymic development on CD8+/CD4- thymocytes, and also defines a distinct subset of naïve CD8+ T cells in the periphery. The latter subset is differentiated from circulating CD103+ mucosa-associated memory T cells by its naïve T cell phenotype (CD45RO-, CD62Lbright, CD27bright, CD11adim, CD95dim), and its high amount of TREC. Indeed, sorted CD103+ naïve CD8+ cells display higher levels of TREC than their CD103- naïve counterparts, and, as would be expected of RTE, these cells demonstrate an age-related decline in frequency which is significantly enhanced by thymectomy. Taken together, these data suggest CD8+ T cells bearing the CD103+, naïve phenotype represent a population of RTE, and that quantification of their frequency allows direct assessment of human thymic output.

 


6360

Identification of live antigen-reactive T cells by flow-cytometry

Alexander Scheffold, Till Muzzulini, Desiree Kunkel, Andreas Radbruch, Deutsches Rheumaforschungszentrum, Berlin, Germany

The flow-cytometric analysis of antigen-specific T cells has always been hampered by their low frequency and the low affinity of the T cell receptor to its ligand the MHC/peptide complex. We have established a simple assay to directly analyse antigen-reactive cells ex vivo by measuring the specific proliferative activity on the single cell level following antigen-restimulation in vitro. All cells are uniformly labelled with the stable fluorescent marker carboxyfluoresceine-diacetate-succinimidylester (CFDA-SE) and cultured with antigen for 72 hours. Proliferating T cells can then easily be identified by flow-cytometry according to their loss of fluorescence intensity which is halfed with each single division step. In a model system of OVA-TCR transgenicT cells mixed with normal Balb/c spleen cells, we could show that it is possible to detect specific cells at frequencies as low as 0.1-0.01% and that all specific cells react in this assay. Specific T helper cells from normal Balb/c mice immunized with Ovalbumin as a model antigen can be identified. Importantly the specific cells can be separated alive and they can directly be used for further functional studies. This method can be used for the fast isolation, expansion and functional analysis of antigen-specific T cells reacting to specific peptides or proteins or even less defined antigenic mixtures.

 


6369

Sensitive detection of APC subpopulations according to the presentation of a specific antigen

Desiree Kunkel, Andreas Radbruch, Alexander Scheffold, Deutsches Rheumaforschungszentrum, Berlin, Germany

For control of immune reaction, T helper lymphocytes have to recognize antigenic peptides, processed and presented on MHC class II molecules by antigen-presenting cells (APCs). Cytometric identification of peptide presenting cells is still a challenge due to the low number of specific MHC/peptide complexes displayed on antigen-presenting cells and the high degeneracy of MHC molecules loaded with different peptides making the generation of MHC/peptide specific antibodies difficult. We have developed a method to identify APCs according to the presentation of specific peptides by flow cytometry. The antigenic peptide is labelled with hapten and surface-bound peptide is detected by staining with magnetofluorescent liposomes for fluorescence signal amplification. In addition, labelled APCs can be isolated by magnetic cell sorting. The sensitivity of detection correlates with the T cell stimulatory capacity of these APCs and quantification of the fluorescent signal reveals a detection limit of less than 300 peptides per cell. After in vivo administration of antigen we have identified and phenotypically analyzed the specific peptide-presenting cells. Following different forms of immunization we compared peptide presentation by various APC subtypes as well as the kinetics of antigen presentation. Using this approach APCs which present peptides at very low but physiological relevant levels can be identified ex vivo on the single cell level for the first time.

 


6475

IDENTIFICATION OF A NOVEL CD56+ NON-CYTOTOXIC NK CELL SUBSET

Loris Zamai, University of Urbino, Italy; Giorgio Zauli, Institute of Normal Morphology, University of Chieti, Italy; Rossana Trotta, Jefferson Medical College, Kimmel Cancer Center, Philadelphia, PA USA; Barbara Cononico, Francesca Luchetti, Stefano Papa, Cytometry and Cytomorphology Center, University of Urbino, Italy, Urbino, Italy; Bice Perussia, Thomas Jefferson University, Philadelphia, PA USA

Natural killer cells were induced to differentiate from cord blood hematopoietic progenitor cells cultured on stromal cell line expressing the membrane bound form of SCF, in the presence of IL-2 or IL-15. Both culture conditions generated a phenotypically and functionally mature (CD161+/CD56+/IFN-gamma+), cytotoxic NK cell subset, while an additional immature (CD161+/CD56-/IFN-gamma-), non-cytotoxic NK cell population was generated only in cultures with IL-2. When immature CD161+/CD56-/IFN-gamma- cells were purified and seeded in secondary cultures supplemented with combinations of IL-2, and IL-15, they generated a novel (CD56+/CD2+/CD11b+/CD8-/CD16-/KIR-/CD3-) NK cell subset that neither mediate granule release cytotoxicity nor produce IFN-gamma, indicating that, during NK cell differentiation, the surface expression of CD56 and CD2 molecules is not concomitant with the acquisition of the two most important NK cell functions, and likely precedes them. This novel NK cell subset acquired the capacity to secrete IFN-gamma, but not granule release- and Fas ligand-mediated cytotoxic activity, when IL-12 was added to the secondary cultures in combination with IL-2 or IL-15, suggesting that the NK cell ability to produce IFN-gamma depends on the presence of IL-12 and possibly precedes the acquisition of spontaneous cytotoxicity and FasL expression. Because most of specific cytotoxic granule-associated proteins and adhesion molecules are already expressed by immature CD56- NK cells, the inability to display spontaneous cytotoxicity even after secondary cultures with combination of IL-2, IL-15 and IL-12 likely reflects defective expression of killer activatory receptors in this subset of NK cells.

 


6926

Normal Human B-Cell Differentiation: Quantitative Analysis of the Levels of Bcl-2 Expression

Pablo Menendez, Sergio Roa, Gema Mateo, Adelaida Vargas, M. Luz Sanchez, Juana Ciudad, Marta Martín, Universidad de Salamanca, Servicio General de Citometría y Departamento de Medicina, Salamanca, Spain; Luis Escribano, Servicio de Hematología, Hospital Ramón y Cajal, Madrid, Spain; Jesús F. San Miguel, Servicio de Hematología, Hospital Universitario de Salamanca, Salamanca, Spain; Alberto Orfao De Matos, Hospital Universitario de Salamanca, Servicio General de Citometría, Salamanca, Spain

Overexpression of bcl-2 protein has been found in different hemopoietic malignances such as precursor B-ALL, monoclonal gammapathies, B-CLL and other B-cell chronic lymphoproliferative disorders. However, the information on the levels of bcl-2 expressionin normal B-cell differentiation is scanty.

Our aim was to explore the quantitative expression of the bcl-2 protein in normal human B-cell differentiation using a four-color stainings analyzed at flow cytometry. A total of 34 samples from healthy donors were analyzed: peripheral blood (PB) (n=10), bone marrow (BM) (n=11), spleen( n=8) and lymph nodes (n=5). Expression of cytoplasmatic bcl-2 was assessed in CD19+ subsets and/or CD38+++, established according to the CD34 and CD38 antigens or the CD23 and CD38 markers in BM and PB, spleen and lymph nodes, respectively. The following B-cell subsets were identified: 1) BM samples: CD34+/CD38++/CD19+ (CD34+ B-cell precursors), CD34-/CD38++/CD19+ (CD34- B-cell progenitors), CD34-/CD38-/+/CD19+ (B-lymphocytes), and CD34-/CD38+++/CD19-/+ (plasma cells) and 2) PB, spleen and lymph node samples: CD23+ and CD23- B-cells; additionally CD38+++ plasma cells were detected both in normal spleen and lymph nodes but not in PB. An important degree of variability in the levels of bcl-2 expression in the different B-cell subsets was detected. The more immature BM B-cell precursors (CD34+) showed a low bcl-2 expression (10.8±6 -mean fluorescence intensity±standard deviation)); as these cells differentiated into CD34/CD38++ B-cell progenitors they slightly decreased bcl-2 expression (6.1±4 MFI) which increased in more mature BM B-lymphocytes (44.3±18 MFI). These values were similar to those detected in the B-lymphocytes from PB ( CD23+ subset:40.2±12 MFI; CD23- B-lymphocytes: 57.1±19 MFI), spleen (CD23+ subset: 39±12 MFI; CD23- subset:48.5±35 MFI) and lymph node (CD23+ subset: 35.5±32 MFI; CD23- subset: 57.5±51 MFI). Regarding plasma cells two different subsets were identified in the spleen and lymph node samples analyzed, which showed different bcl-2 expression (spleen: 115±54 and 7.5±6 MFI; lymph node: 163±66 and 36±8); BM plasma cells displayed the highest levels of bcl-2 expression: 161.2±48 MFI.

In summary, our results show that expression of bcl-2 largely varies in normal individuals along B-cell differentiation. Thus, the assessment of the presence of aberrant levels of bcl-2 expression in hemopoietic malignances involving B-cells should take into account the levels of bcl-2 protein expressed by their normal counterpart.

 


6547

Confocal analysis of the nuclear translocation of NFkB family members during the maturation of monocyte derived dendritic cells

Edmond Kahn, Frederique Frouin, Andrew Todd-Pokropek, INSERM U494, CHU-Pitie-Salpétrière, Paris, France; Youssef Bacri, Bruno Canque, Jean-Claude Gluckman, Laboratoire d’Immunologie cellulaire et immunopathologie de l’EPHE, CHU Pitié-Salpétrière, Paris, France

At difference with immature dendritic cells (DC), mature DC display post-entry block in human immunodeficiency virus (HIV) replication cycle. Because NFkB family members play a major role in the control of HIV transcription, we analyzed the influence of different DC maturation conditions on their basal expression level and subcellular localization. Expression of NFkB members p50, p52, p65, c-Rel, RelB and IkB-a was analyzed by laser scanning confocal microscopy (CLSM). Fluorescence spectral emission characterization and factor analysis of medical images (FAMIS) of confocal series of images were performed. Tetramethylrhodamine isothiocyanate (TRITC) was chosen to detect the expression of NFkB members by immature and mature cells. Phycoerytrin (PE) was used to stain cell membranes. Cell nuclei were counterstained with propidium iodide (PI) at low concentration. Excitation of each fluorochrome was obtained by selection of specific bands (488nm, 514nm) of the argon laser of a confocal microscope (SARASTRO CLSM 2000, Molecular Dynamics). Spectral characterization was performed on the tandem TRITC-PI. FAMIS was applied on spectral and temporal series of images to estimate images corresponding to specific stains in the case of the tandem PE-PI. Fluorochromes were clearly distinguished in each case and images showed localization of the proteins in the cytoplasm or the nucleus. On mature and immature cells, it was possible to analyze differences in both expression levels and subcellular localization of proteins. FAMIS and 3D reconstruction methods were also applied to study 3D-localization of cellular proteins from Z-series of images obtained on the CLSM and it resulted in interpretable images to analyze the expression of proteins via reconstructed focal images and projections. References : B. Canque et al, Blood, vol 93, n°11, 1999, 3866-3875; E. Kahn et al, J. Microscopy, vol 193, n°3, 1999, 227-243.

 


Tuesday, May 23, 2000


8:00 am – 9:45 am

Technical Plenary II on Microarray and Chip Technology

Chair: Daniel Pinkel

Analysis of DNA Copy Number Variations Using Array CGH

Donna Albertson

Interpretation of Expression Array Data

Presenter not yet confirmed


10:45 am – 12:30 am

Symposium on In Vivo Molecular and Cellular Imaging

Chair: Christopher H. Contag


Revealing Gene Expression Patterns in Living Animals Using Photoprotein Imaging

Christopher H. Contag, Stanford University School of Medicine

The future of molecular and genetic medicine lies in the ability to effect and then evaluate biological changes at the level of cells and molecules in vivo. Until recently a fundamental difficulty in biomedical research has been obtaining spatiotemporal information about in vivo processes as they occur, e.g. monitoring changes in gene expression patterns in response to various stimuli in intact living animal models. An imaging strategy based on the use of light producing enzymes, photoproteins such as luciferase, as metabolic reporters that are detected using low light imaging systems provides access to biological information in living rodent models of human biology and disease. Following expression of photoprotein reporters is a rapid and accessible method that has the capability of making connections between the ex vivo derived data and in vivo biological events. As such this approach has broad applications in the fields of oncology, infectious disease and in evaluating DNA-based therapeutics and gene expression patterns. Refining animal models using noninvasive imaging tools will reduce the number of animals needed for biological assessment of therapeutics, improve data quality and increase the predictive power in estimating the human response.

 


Towards In Vivo Molecular Imaging

Ralph Weissleder

In vivo imaging technologies are at a stage at which in vivo imaging can occur at near µm resolutions with image specificity at the physiological, cellular and molecular level. The current presentation will focus on new developments in magnetic resonance, nuclear and optical imaging. Several examples will be used to illustrate the power of in vivo imaging techniques. These examples will cover areas such as in vivo imaging of gene expression, cell tracking and angiogenesis.

 


Optical Bioimaging

Daniel L. Farkas


2:00 pm – 4:00 pm

Workshops B

Biosafety

Facilitators: Ingrid Schmid,

Annalisa Kunkl

CLINICAL ACCEPTANCE OF DNA ANALYSIS

Facilitator: John Lawry

Cytometry and Signaling Pathways

Facilitator: Paul J. Smith

Optimization and Performance in Confocal Microscopy

Facilitators: Robert M. Zucker, Nicholas H.A. Terry

Data Analysis Advances

Facilitator: Mario Roederer

 


Wednesday, May 24, 2000


7:45 am – 9:45 am

BiologICAL Plenary Session on Cell Cycle Regulation and Analysis

Co Chairs: James Jacobberger, Andrew Koff

4D Control of Mitosis

Jonathan Pines

The Activity and Expression of Cyclin D1 in Individual Cycling Cells

Dennis W. Stacey

Cell Cycle Kinetics and Transient Transfection

Thomas Delhory

Measurement of Kinetic Variability Using Three Color Flow Cytometry

R. Allen White

Ultraviolet-induced Detection of Halogenated Pyrimidines in DNA Replicating Cells

Hans Hammers


1045 – 1230

New Investigator/Student Symposium

Chair: Alexander Nakeff

Exceptional Student Award Presentations

President’s Award for Excellence Presentations


6340

ULTRAVIOLET-INDUCED DETECTION OF HALOGENATED PYRIMIDINES IN DNA REPLICATING CELLS (UVID)

Hans Hammers, Holger Kirchner, Peter Schlenke, University of Luebeck, Institute for Immunology and Transfusionmedicine, Luebeck, Germany

A new method is described which allows the detection of halogenated thymidine analogues in DNA replicating cells without the need of extensive denaturation with acid/base or additional enzymatic treatments. Brief irradiation periods with UV-A in the presence of Hoechst 33258 or UV-B alone and a subsequent treatment with a hypotonic buffer makes BrdUrd accessible to sensitive immunological detection with monoclonal antibodies. The photolysis of incorporated BrdUrd in DNA with UV-light is sequence dependent and results into radical mediated nuclear damage allowing the detection of BrdUrd under hypotonic conditions. However, treatment with other inducers of single or double strand breaks such as g radiation or hydrogen peroxide did not allow for BrdUrd detection. The new methodology is compatible with mild crosslinking fixation, i.e. aldehydes, or coagulative fixation, i.e. alcohols, giving the opportunity to combine BrdUrd detection with a wide range of additional markers such as immunophenotyping or apoptosis.As shown in pulse-chase experiments, this approach enables BrdUrd detection in all compartments of the cells cycle and suggests its use in kinetic analyses as well. The method has been applied successfully in identification of CD34+, CD138+ or CD19+ cells out of heterogeneous cell suspensions and allowed their cell-cycle analysis. Results correlated very well with acid denaturation (r = 0.938). The average CV of G1 in the DNA histogram was smaller than 5% resulting in good preservation of DNA distribution and the signal-to-noise ratio was almost twice as high for as for 2N acid denaturation at RT allowing convenient discrimination of BrdUrd positive cells. Besides allowing for the simultaneous detection of cellular markers this method lacks any enzymatic treatment making the method fast, convenient, reliable and inexpensive and should further its use in basic and clinical research of cancer, immunology and pharmacology.

 


6216

Visualisation Of Aortic Elastin Structure By Selective Tissue Digestion And Confocal Microscopy

Deanna Jones, Alberto Avolio, University of New South Wales, Graduate School of Biomedical Engineering, Sydney, NSW Australia

Arterial elastic properties are determined by the lamellar and inter-lamellar structures of the elastic fibre network. Conventional light microscopy of stained arterial sections is adequate for visualisation of the lamellae, but is limited for high-resolution visualisation of the inter-lamellar fibres because of interference by other tissue components such as collagen and smooth muscle. Although no stain is imparted to these non-elastin components, they are not transparent and diffract the light passing through the specimen, causing resolution to be lowered.

Confocal microscopy can offset this limitation by removing out-of-focus information. However, it cannot entirely compensate for the effect of diffraction by unstained elements. Furthermore, the fluorochromes available for elastin staining are not as selective as conventional light microscopy stains such as resorcin-fuchsin and orcein. This study aims to obtain high resolution images of both lamellar and inter-lamellar elastic fibres, without interference from unwanted tissue components, by the use of selective tissue digestion in combination with confocal microscopy.

Hot formic acid digestion is an established technique for removal of non-elastin components. However, once its supporting framework has been removed, the elastic tissue becomes difficult to handle and further processing for histological observation has been restricted to freeze fracture in combination with SEM. We adapted the formic acid digestion treatment to develop an on-slide digestion method, which allowed in situ removal of non-elastin tissue. Subsequent fluorescence staining with eosin and confocal imaging revealed crisp detail of both the lamellar and inter-lamellar fibres. Furthermore, the signals for the two types of fibres were directed into separate channels: the lamellar fibres were primarily imaged in the red channel (560-640 nm) and the inter-lamellar fibres in the green channel (506-538 nm). This may indicate a difference in relative content of the elastin protein and microfibril constituents between the two types of fibre.

This technique is useful in imaging potentially degenerative structural changes, which are seen with increased age and total number of cardiac cycles throughout life or withearly development of aortic aneurysms. Such changes can be seen at the inter-lamellar level of the elastic network of the aorta before they are observed in the lamellae.

 


6451

Characteristics and Dynamics of Bacterial Populations within Poloxamer-Hydrogel Biofilm-Constructs

Stephanie Ann Sincock, J.Paul Robinson, Purdue University Cytometry Laboratory, West Lafayette, IN, USA; Bartek Rajwa, Jagiellonian University, Institute of Molecular Biology, Krakow, Poland

Bacterial infections associated with indwelling medical devices often demonstrate intrinsic antimicrobial resistance. Possible explanations for this resistance include relatively slow growth rate of cells deep within the biofilm, the production of thick exopolysaccharide glycocalyx, poor penetration of antimicrobial agents due to limited diffusion within the biofilm, and the development of a specific biofilm-phenotype. A simple, reproducible, and rapid biofilm model system was investigated to examine biofilm-phenotype cells in a 3-dimensional environment. A thermoreversible hydrogel [liquid at temperatures <15°C, gel at temperatures >15°C] was formed using a di-block copolymer of polyoxyethylene and polyoxypropylene. Polymer flakes were hydrated overnight at 4°C using trypticase soy broth (30% w/v) and autoclaved. Sterile, chilled poloxamer was inoculated with stationary phase bacteria (104 cells/ml), placed in chambered coverslips, and incubated at 37°C for 3 hours. Within the hydrogel, bacteria grew to high cell densities, formed microcolonies, and exhibited a biofilm phenotype. 2D, 3D (2D over time or real 3D) and 4D (3D over time) confocal backscattered (reflected) light microscopy (TLR) was used to monitor bacterial growth within hydrogel biofilm-construct. Several series of x, y sections and series of x, y, z arrays were collected at frequent intervals. TLR imaging allowed analysis of 3-D structural details observed during microcolony formation. Confocal fluorescence microscopycombined with vital dyes was also used to investigate the viability of the cells within the microcolonies. The simple addition of cold buffer to the hydrogel collapsed the biofilm-construct and allowed easy recovery of cells without inflicting additional damage; biofilm-phenotype cells were then enumerated and characterized using flow cytometry. A model system that reproduces the physical and biological characteristics of biofilms and allows easy recovery of cells for further analysis should be a useful tool for the study of antimicrobial resistance of biofilm-phenotype cells.


President’s Award for Excellence Presentations


4398

Investigation of Cardiac Myocyte Cytoplasm Structural Changes During a Contraction Cycle by Means of Intracellular Fluorescein Fluorescence Polarization.

Dror Fixler, Reuven Tirosh, Mordechai Deutsch, Jerome Schottenstein Cellscan Center, Physics Department, Bar-Ilan University, Ramat-Gan, ISRAEL; Asher Shainberg, The Otto Meyerhoff Drug Receptor Center, Department of Life Sciences, Bar-Ilan University, Ramat-Gan, ISRAEL

Intracellular structural changes, occurring in a cardiac myocyte during a contraction cycle is investigated by means of intracellular fluorescein fluorescence polarization (IFFP), in comparison to cytoplasmic concentration of ions measured by Indo-1. A simple physical model is presented. It assumes a bi-phase intracellular matrix, differing in its potency to restrict hosting fluorescent probe mobility. The first is a mobile non-restricting phase, mostly made of aqua (aqua zone), while the second is a mobile-restricting phase, allocated mainly at the proximity of the filament sites. Their physico-chemical properties such as [Ca++], viscosity and pH, may differ and thus influence differently the hosting probe fluorescence characteristics. These possible influences are experimentally examined. Based on the experimental data, the model enables the evaluation, to first order of approximation, of the relative number of fluorescent probes populating the two phases and the time variation viscosity, of the mobile-restricting filament zones, taking place along a contraction cycle. To the best of our knowledge, this is the first time that cardiac myocyte contraction is monitored by IFFP measurements on an individual cell basis within a population.

 


5853

High-resolution assessment of DNA loss in RIP-Tag mouse pancreatic tumors using comparative genomic hybridization to DNA microarrays.

Graeme Hodgson, Jeff Hager, Colin Collins, Meredith Wernick, Heather Werhane, Daniel Pinkel, Douglas Hanahan, Joe Gray, University of California Cancer Center, San Francisco, CA, USA; Donna Albertson, Lawrence Berkeley National Laboratory Life Science Division, Berkeley, CA, USA

Comparative genomic hybridization to DNA microarrays (CGHa) allows for rapid and quantitative assessment of DNA copy number changes in tumor cells. Additionally, the resolution to which DNA copy number changes can be accurately measured is significantly improved over chromosome-based comparative genomic hybridization (CGH). We are using CGHa to fine structure map 2 recurrent regions of DNA loss in RIP-Tag mouse insulinomas. In this model, transgenic expression of the SV40 large T-antigen in pancreatic islet bcells results in a distinct pathway leading to tumor formation. The usually quiescent islet cells begin to hyperproliferate, induce angiogenesis, and finally down regulate apoptosis resulting in solid tumor formation. Recurrent regions of deletion on chromosomes 9 (LOH9) and 16 (LOH16) have been identified by loss of heterozygosity (LOH) and CGH analyses. These losses are respectively associated with induction of angiogenesis and down regulation of apoptosis in tumor progression. This suggests that tumor suppressor genes (loh9 and loh16) reside within LOH9/16 whereby loss of loh9 protects cells against apoptosis and loss of loh16 allows for induction of angiogenesis. We estimate that LOH9 and LOH16 span 9Mb and 30Mb respectively. To refine these regions using CGHa we have generated DNA microarrays containing 72 different BAC/P1 clones spotted in quadruplicate onto GAPS coated glass slides. All clones have been FISH mapped and the set constitutes 37 control clones that map close to the centromeres and telomeres of each mouse chromosome, 16 LOH9 clones, and 19 LOH16 clones. This provides a resolution of 1BAC/500kb for LOH9 and 1BAC/1.5Mb for LOH16. We are able to detect single copy deletions at LOH9/16 in RIP-Tag tumor derived cell lines by performing CGH on these arrays. Normalized fluorescence ratios for control clones are 1.0+/-0.13, whereas ratios for deleted clones are 0.55+/-0.05 (mean+/-s.d.). We will report on the screening of a large number of RIP-Tag tumors to distinguish those containing partial deletions at LOH9/16 and therefore identify the minimal region of DNA loss. BAC clones belonging to this minimal region will serve as seeds for contig assembly and will be sequenced to facilitate gene discovery.

 


6072

Activation- and expression level-dependent association of erbB proteins in breast cancer cells

Peter Nagy, Attila Jenei, Laszlo Balogh, Barok Mark, Sandor Damjanovich, Janos Szollosi, University Medical School of Debrecen, Department of Biophysics and Cell Biology, Debrecen, Hungary; Achim Kirsch, Thomas Jovin, Max Planck Institute for Biophysical Chemistry, Department of Molecular Biology, Göttingen, Germany

Four erbB proteins constitute the erbB family of transmembrane tyrosine kinase receptors. ErbB2 seems to fulfill a central role in signal transduction through this system, and is often overexpressed in breast cancer with poor prognosis. Previously we have established that erbB2 shows a high level of homoassociation in unstimulated breast tumor cells, and the cell surface distribution of the homoassociation is uneven: some membrane areas with a diameter of about 1 mm show anomalously high homoassociation. Activation-dependent changes in the homoassociation of erbB2 were detected, which were influenced by the relative expression level of erbB proteins. In our present work we wanted to find out if the cell surface distribution of erbB2 is uneven in the sub-micrometer range (large-scale association), and to correlate the extent of small-scale association (e.g. dimer formation) of different erbB proteins with their expression level. We have used the shared-aperture mode of a near-field optical microscope and confocal microscopy to characterize the large-scale association pattern of erbB2 proteins labeled by fluorescent anti-erbB2 antibodies. Flow cytometric energy transfer was used to detect the small-scale association of different erbB proteins labeled by donor- and acceptor-tagged antibodies. ErbB2 was found to be concentrated in membrane patches (clusters) with a mean diameter of 0.5 mm in nonactivated SKBR3 breast tumor cells. The average number of erbB2 in a single cluster on unstimulated SKBR3 cells was about 1000. Activation of erbB2 on SKBR3 cells with epidermal growth factor (EGF), heregulin and an anti-erbB2 antibody led to an increase in the mean cluster diameter to about 0.6-0.9 mm. The EGF-induced increase in the cluster size could be blocked by an EGF receptor specific tyrosine kinase inhibitor. In order to correlate the large-scale association of erbB2 with the expression level of the protein we chose three different breast tumor cell lines (SKBR3, MDA453, MCF7), and found that the mean cluster diameter was independent of the level of erbB2 expression. On the other hand, the small-scale association of erbB2, erbB1 and erbB4 are very much dependent on their expression level. The degree of erbB2 homoassociation and erbB2-erbB4 heteroassociation were smaller in a cell line in which the relative expression level of erbB1 is high. We hypothesize that erbB1 competes for erbB2 as an association partner. The possible biological significance of the hierarchical association of transmembrane receptors is the increased efficiency of signaling through receptors which are concentrated in localized membrane domains. This work was supported by the following grants: OTKA F022725, FIRCA 1R03-TW-00871-01A2.

 


2:00 pm – 4:00 pm

Workshops C

Data File Standards

Facilitator: Phillip Dean

HIGH THROUGHPUT Sample Handling AND KINETICS

Facilitators: Bruce Edwards, Larry A. Sklar

Bead-based Analyses

Facilitators: Francis Mandy, John P. Nolan

Apoptosis

Facilitators: Carolyn Dive, David Hedley

Differential Cell Counts in Flow Cytometry/Hematology

Facilitators: Francis Lacombe, Bruce H. Davis


1600 – 1800

Parallel Session V: Image Cytometry: Technical Advances

Chair: Bruce Milthorpe


4705

Segmentation of cell nuclei using immunolabeled nuclear membrane proteins

Carlos Ortiz De Solorzano, Life Sciences Division, Berkeley, CA, USA; David W. Knowles, Ravi Malladi, Stephen J. Lockett, Lawrence Berkeley National Laboratory, Berkeley, CA, USA

The nuclear lamina forms a protein mesh that underlies the nuclear membrane. In most mammalian cells it contains the intermediate filament proteins, lamins A, B and C. Using fluorochrome-tagged antibodies against one or a combination of these lamins we can obtain 2D or 3D images of cell cultures or tissue where the domain occupied by each nucleus is nicely delimited, no matter how clustered the cells might be. This can be use to overcome the traditional inefficiency of nuclear segmentation methods, when applied to highly clustered cells in culture or tissue. Our approach to nuclear membrane segmentation consists of defining unique nuclei seeds (curves in 2D, surfaces in 3D) and expand them using curvature and gradient driven flows. These flows move each surface until it finds the nuclear membrane of the nucleus that contains it. Seeds can be defined interactively by manually drawing the seeds, or automatically, using a Hough Transform-based algorithm that detects the centers by projecting the image gradients in the direction normal to the membrane. Doing this, we obtain a reduced version of the nucleus that can be used as a seed. Seed expansion is done by embedding its surface as the level set zero of a higher dimensional function and solving the Partial Differential Equation (PDE) that describes the evolution of this new function under a force term that depends on the curvature of the front and the gradient of the image at any given point. The force term that we use is one that produces an outward movement that attracts the surface to high image gradients and that does not allow high curvature fronts, as those which are produced by shot or specular noise. We have tested our algorithm using 2D and 3D images of cell cultured cells, with promising results. Once the parameters of the PDE are defined for one image of a type, the algorithm can automatically segment any image of the same type. Future work will be devoted to automating parameter selection and improving the implementation to reduce algorithmtime.

 


6786

Direct Neural Network Application for Automated Cell Recognition

Bruce Milthorpe, Zheng Qing, Graduate School of Biomedical Engineering, Sydney, NSW Australia

Automated cell recognition from histological images is a very complex task. Traditionally, the image is segmented by some method chosen to suit the image type, the objects are measured and then a classifier is used to determine cell type from the object measurements. A variety of classifiers have been used with reasonable success, including neural networks, working with the morphological information.

This project is examining the possibility of directly classifying cell types by using a neural network on the pixel intensity information in an area of the image. The most appropriate type of neural network, and its ability to work with cells in a complex patterned background require to be assessed, as well as the accuracy of classification. Classification by direct application of neural networks has the potential for much faster cell image processing.

Materials and Methods:

Sample images have been made from a variety of histological sources of increasing complexity. These include manually segmented cells against a black background, leukocytes in blood smears, and, from rabbit models, inflamed cornea and biomaterials implanted in paravertebral muscle.

Neural networks under investigation include 3-layer back-propagation, a 4-layer backpropagation and a two-layer shared-weights architecture.

A training set and an independent test set of cells from each tissue sample have been prepared and assessed manually for cell type eg monocytes from blood smear (figure 1). Image subsamples of 32x32 pixels are used to provide RGB inputs to the neural network. Test set cells are never used for training.

Results and Discussion

The 3-layer back propagation neural network has been fully trained on all 4 types of image sources (see figure 2). After training on 68 cells of four different cell types from the blood smear, the neural net accurately classified all cells from the training set. Whilst the single eosinophil in the test set was accurately identified, the overall classification accuracy from the test set was 76% (Table 1). This is lower than desirable, but comparable to the accuracy of this type of neural network in handling handwritten numeric characters..

The 4-layer and shared weights neural networks are expected to perform better in terms of classification accuracy once training has been completed on them.

Current problems with the systems that are being addressed are increasing the number of cells in each test and training set and improving the quality of the images from the rabbit cornea samples.

The number of cells in the training set determines the training time to a large degree and the minimum number of representative cells required for training remains to be properly determined.

 


6800

Fuzzy classification of color images: application to quantitative Immunohistochemistry

Sophie Schüpp, Abderrahim Elmoataz, Daniel Bloyet, University of Caen, Sciences, Caen, France; Paulette Herlin, Centre de Lutte Contre le Cancer François Baclesse, anatomie pathologique, Caen, France

Quantitative Immunohistochemistry of proliferation and differentiation markers using image analysis is a promising tool for the evaluation of cancer prognosis and potential responses of tumors to therapy. The value of image analysis for immunoquantitation was pointed out as early as 1987 (1). Nevertheless, nine years later, Cohen (2) underlined that "The technique of quantitation of immunostaining is still in its infancy". The reasons for this slow evolution can be found, firstly, in the complexity of problems to be solved, secondly, in the use of insufficiently sophisticated image analysis operators. The last reason can be found in the low performance of computers in the recent past, which has limited the use of more advanced segmentation procedures. In this paper, we present a method based on fuzzy classification for color segmentation. The fuzzy classification principle is the following: from a chosen number of color classes and an initializing image of class centers, an iterative process assignsa degree of membership of each pixel to each class, up to minimizing the distance between the mobile center and the members of a class. The resulting image of segmented objects is obtained by computing the maximum of each class. We illustrate this classification procedure on two different examples: quantification of vascular and nuclear immunostained markers. In the first example, fuzzy classification is used to split the objects of interest and the rest of the image, in the second example the method isused at two levels: delineation of cancer cell clusters and characterization of immunostained and counterstained nuclei.

1. TUCKER J, SCHENK, EIERMANN W, BURGER G, Quantitative Image Analysis on Immunocytochemically labelled cells. Clinical cytometry and Histometry, 1987, Burger G, Ploem JS, and Goertler K, Academic Press, London.

2. COHEN C, Image Cytometry Analysis in Pathology. Human Pathology, 1996, 27(5), 482-493.

This work was supported by grants from the "comités départementaux de la ligue de Lutte Contre le Cancer du Calvados, de la Manche et de l’Orne".

 


6822

Glass slide scanning: a fast and inexpensive answer to the problem of tumor tissue heterogeneity for quantimetry.

Paulette Herlin, Kim Tran, Nicolas Elie, Jacques Chasle, Université de Caen, Centre de Lutte Contre le Cancer François Baclesse, anatomie pathologique, Caen, France; Benoît Plancoulaine, Institut Universitaire de Technologie, Mesures Physiques, Caen, France; Sophie Schüpp, Université de Caen, Sciences, Caen, France; Daniel Bloyet, ISMRA, GREYC, Caen, France

The introduction of 2D morphometry and quantimetry in clinical practice is frequently limited by the well known problem of tumor tissue heterogeneity. Quantification of structures in visually selected "representative areas" and "hot spots" is subject to poor reproducibility and inter-observer variability. Moreover, the information collected in these conditions cannot reflect the whole tumor behavior for heterogeneous samples and does not allow an appraisal of the topographical distribution of a marker. Two strategies are frequently proposed to overcome this problem: systematic sampling of microscopical fields and systematic acquisition of neighboring fields for building a mosaic image of the whole tissue section. The precision of the measure obtained by the first strategy is highly dependant on the number of fields sampled and topographical information is limited. The second strategy is very time consuming and tedious and requires high image storage capacities.

A fast and inexpensive solution is offered by scanning the whole tumor tissue section at an adequate resolution, using a flat scanner, or a slide scanner provided with a glass slide holder. Scanning images can be used either for direct quantification of large objects or for delineation of fields of interest, automatic driving of a fully automatic microscope and high magnification mosaic image building. Two examples are given: the direct estimation of the volume fraction of stroma and blood vessels on scanned images of histological sections of ovarian carcinomas and the appraisal of the proliferative compartment of these tumors on selected large field microscopical images.

This work was supported by grants from the "comités départementaux de la ligue de Lutte Contre le Cancer du Calvados, de la Manche et de l’Orne". Iconofast (France) Olympus (France) and Polaroid (France) companies are gratefully acknowledged.

 


7971

Advances in high-resolution cytometry of FISH dots in interphase cell nuclei

Michal Kozubek, Magdalena Skalnikova, Pavel Matula, Petr Matula, Irena Koutna, Faculty of Informatics, Masaryk University, Brno, Czech Republic; Stanislav Kozubek, Emilie Lukasova, Eva Bartova, Pavla Jirsova, Alena Cafourkova, Institute of Biophysics, Academy of Sciences, Brno, Czech Republic

Recently, we have described a high-resolution cytometry (HRCM) technique (Cytometry 36:279-293) as a reasonable compromise between flow cytometry (FCM) and laser scanning cytometry (LSCM) on one hand and confocal laser scanning microscopy (CLSM) on the other hand. Whereas FCM/LSCM and CLSM enable measuring of either large number of cells with low resolution or small number of cells with high resolution, respectively, HRCM can analyze quantities comparable to FCM/LSCM measurements (using overnight acquisition) with an accuracy comparable to CLSM.

The HRCM system consists of a fully motorized fluorescence microscope and a cooled digital CCD camera fully controlled by a high-performance computer which performs both acquisition and related on-line image analysis. It is capable of analyzing microscope slides with FISH-stained interphase nuclei in 2-D as well as in 3-D. Using HRCM, it is possible to analyze multi-color preparations including UV-excited dyes as well as repeatedly hybridized preparations re-acquiring individual nuclei. The speed of the acquisition and analysis is about 50 nuclei per minute in 2-D and 1 nucleus per minute in 3-D but depends on the density of nuclei on the slide; the precision of the lateral and axial measurements is approximately 100 nm.

Very recently we have built a second HRCM instrument (see picture at http://www.fi.muni.cz/lom) which enables switching between conventional and confocal modes. The confocal mode is realized using a 5% Nipkow disk optimized for fluorescence imaging. It offers a better resolution but at a lower speed. The user can choose which fluorochromes are imaged in confocal mode. Thus, the system can be configured for a variety of imaging modes ranging from quick 2-D analysis up to high-resolution confocal 3-D studies.


Parallel Session V: Solid Tumors

Chair: Herman Van Dekken


3947

Towards a model for neoplastic progression in Barrett’s esophagus.

Herman Van Dekken, Kees J. Vissers, Janneke C. Alers, Peter H.J. Riegman, Erasmus University Rotterdam, Department of Pathology, Wim C.J. Hop, Erasmus University Rotterdam, Department of Epidemiology and Biostatistics, Rotterdam, The Netherlands; Eric Geelen, Carla Rosenberg, Leiden University Medical Center, Department of Molecular Cell Biology, Leiden, The Netherlands; Hugo W. Tilanus, Erasmus University Rotterdam, Department of Surgery, Rotterdam, The Netherlands

The incidence of adenocarcinoma of the esophagus has increased rapidly over the past decades. At present, it equals the number of squamous cell carcinomas of the esophagus in the Western world. These adenocarcinomas have an extremely poor prognosis, since metastases are frequently present at diagnosis. They almost exclusively develop in columnar lined (Barrett’s) esophagus, a precursor condition termed metaplasia. Barrett’s esophagus has a 30-40 fold increased risk for the development of adenocarcinoma. However, only a small proportion of patients with Barrett’s esophagus will eventually develop cancer. This transformation is preceded by dysplasia, an important tool in the clinical surveillance of Barrett’s esophagus. Histopathology alone is, however, notable to predict the malignant transformation to adenocarcinoma. Malignant transformation to adenocarcinoma in Barrett’s esophagus is characterized by three well-defined premalignant stages: metaplasia, low-grade and high-grade dysplasia. A unique genomewide overview, based on comparative genomic hybridization, is presented here, enabling comparison of preneoplastic lesions and their malignant counterpart. The frequency of losses and gains significantly increased in the subsequent stages of malignant transformation. More specific, loss of 5q21-q23, 9p21, 17p12-13.1 and 18q21 could be revealed in low grade dysplasias, subsequently followed by loss of 7q33-q35, as well as gain of 7p12-p15, 7q21-q22, and 17q21 in high-grade dysplasia. In the invasive cancer numerous additional alterations were detected. In addition, loss of 7q33-q35 was found to represent a distinction between low-grade and high-grade dysplasia. Loss of 16q21-q22 and gain of 20q11.2-q12 were disclosed as discriminators between high-grade dysplasia and adenocarcinoma. In conclusion, genomic events during transformation in Barrett’s esophagus were revealed and a model was generated ordering the genetic changes and candidate genes involved. Further, chromosomal alterations were distinguished that might serve as clinical biomarkers for neoplastic progression in Barrett’s esophagus.

 


6296

Flow Cytometry Aids Histology in Identifying Low and High Risk Subsets of Patients in Barrett’s Esophagus

Peter Rabinovitch, Corinna Palanca-Wessels, Patricia Blount, University of Washington, Seattle, WA, USA; Douglas Levine, AstraZeneca, Wayne, PA, USA; Gary Longton, Brian Reid, Fred Hutchinson Cancer Research Center, Biostatistics, Seattle, WA, USA

Baseline histologic and flow cytometric findings were examined as predictors of cancer as an outcome in a prospective study of 322 patients with Barrett’s esophagus from 1983 to 1998. The mean follow-up time was 3.9 years among patients without cancer. Baseline increased 4N fractions, aneuploidy, and high grade dysplasia (HGD) were all strongly associated with cancer risk, with five-year cumulative cancer incidences of 56% (95% CI: 37,77), 43% (28,62), and 59% (44,74), respectively. While the presence of high grade dysplasia was the strongest predictor of cancer (relative risk [RR] of 12.6 [95% CI: 5.3, 30] in a multivariable model which accounts for the effect of cytometric abnormalities), the presence of either elevated G2/tetraploid (4N) fractions oraneuploidy was also a significant predictor of cancer (RR of 3.0 [1.4, 6.7]) when adjusting for the contribution of high grade histology. Most importantly, in the subset of patients without HGD, histologic subcategories were not significantly predictive, while the presence of a DNA content abnormality conferred a substantially increased cancer risk (RR of 19 [4.7, 78]). Patients who had neither HGD nor flow cytometric DNA content abnormalities at the baseline evaluation had a 5-year cumulative incidence of cancer of 0 (0,4.7). An elevated 4N fraction was defined as 6% of cells or greater. 4N fractions of 6%-15%, 4N fractions of 15% or greater and the presence of aneuploidy were equally strong predictors of cancer outcome. S phase fraction did not, however, have independent predictive value. Study of Barrett’s epithelial cells in culture suggests that elevated 4N fractions are associated with an abnormality in the centrosome cycle. Since elevated 4N fractions precede aneuploidy in Barrett’s esophagus by 17 months, on average, centrosome abnormalities may produce the genomic instability that ultimately leads to aneuploidy and cancer in Barrett’s esophagus. In conclusion, a baseline endoscopic biopsy protocol using both histology and flow cytometry more accurately identifies subsets of patients with Barrett’s esophagus that are at low and high risk for progression to cancer. Patients whose baseline biopsies are negative, indefinite or LGD without increased 4N or aneuploid populations may have surveillance deferred for up to five years. Patients with cytometric abnormalities or HGD merit more frequent surveillance.

 


6185

GENETIC ANALYSIS OF DYSPLASIAS AND CANCERS OF THE UTERINE CERVIX USING CGH, FISH AND RT-PCR.

Masaru Sakamoto, Hiroko Sakamoto, Keiko Kawasaki, Tsukasa Akiya, Hiroshi Iwabuchi, Tetsuya Muroya, Tadashi Sugishita, Yoshio Tenjin, Sasaki Institute Kyoundo Hospital, Tokyo, Japan; Tetsuo Noda, Cancer Institute, Tokyo, JAPAN; Tadao Tanaka, Jikei Medical University, Tokyo, JAPAN

In order to investigate genetic changes associated with carcinogenesis and tumor progression, gains and losses of DNA sequences were measured in 66 normal, dysplasia and cancer of the uterine cervix using Comparative Genomic Hybridization (CGH). HPV DNAs were also detected and typed using PCR. DNA sequence copy number abnormalities (CNAs) were not detected in the normal uterine cervix. Total number of CNAs increased as tumor progress from dysplasia to cancer. Losses of DNA sequences were observed more frequently in the carcinogenesis step, whereas gains were observed more frequently in the invasion step. Total number of CNAs in Severe Dysplasia was greater than that in Mild and Moderate Dysplasia, and almost the same level as CIS. Twenty five common CNAs were found in uterine cervical cancers. Among those CNAs, decreased copy number in four regions, i.e., 3p24-pter, 6q25-qter, 10p13-pter, and Xq26-qter seemed to be associated with carcinogenesis, whereas increased copy number in 8 regions, i.e., 1q41-qter, 5p14-pter, 5q31-qter, 12q24, 17q23-qter, 19q13.2-qter, 20q13.1-qter, and 22q13, and decreased copy number of 16q12-q22 were thought to be associated with invasion of the uterine cervix. No significant correlations were found between HPV status and common CNAs in the cancers. Statistical analysis among those genetic events suggested the possibility that there might be two pathways for carcinogenesis of the uterine cervix; one type of CIS with 3p loss would finally progress to invasion, another type with 6q loss or Xq loss would remain in situ without invasion. Among genes located at 20q13.2, ZNF217 gene was selected for further examination. Quantitative RT-PCR for ZNF217 mRNA revealed frequent overexpression of ZNF217 mRNA in 70% of uterine scervical cancer. FISH analysis with ZNF217 gene specific DNA probe revealed frequent amplification in 50% of uterine cervical cancer. Therefore, ZNF217 amplification might be involved in the progression of uterine cervical cancer. This work was supported in part by Grants-in-Aid from the Ministry of Health and Welfare of Japan.

 


7220

Proliferative Potential of Neuroblastoma Cells Disseminated in the Hematopoietic System

Gabor Mehes, Department of Pathology, Pecs, Austria; Claudia M. Hattinger, Peter F. Ambros, Childrens Cancer Research Institute, St. Anna Kinderspital, Vienna, Austria; Thomas Lörch, MetaSystems GmbH, Altlussheim, Germany

Introduction: Early tumor cell dissemination and minimal residual disease are of increasing clinical interest, however, little is known about the biological impact of rare tumour cells.

Aims: In this study, automatic microscopical analyses were performed to elucidate the proliferative status of minimally disseminated tumour cells in the bone marrow of neuroblastoma patients.

Methods: Automated fluorescence microscopical analyses (MRDetect, MetaSystems) of bone marrow samples from multiple punction sites of 18 neuroblastoma patients were performed. A total of 76 cytospin preparations was automatically scanned and analysed for GD2/TRITC and Ki-67/FITC immunolabelling to demonstrate the Ki-67+ fraction of GD2+ tumour cells present in the bone marrow. Demonstration of tumor cell specific cytogenetic aberrations and exclusion of false GD2 immunostaining could be performed by sequential FISH analysis of the selected cells after automatic relocation.

Results: Through the automatic scanning approach, as few as 1 GD2 positive tumour cell out of 106 bone marrow mononuclear cells (MNCs) could be identified and quantified. The overall Ki-67 labelling index in these cells was found to be between 0.00 and 0.78. However, marked differences were found between samples obtained before and after CT. Tumor cell populations in the initial samples showed a high Ki-67 labelling index (mean 0.35±SD 0.15), independent of the tumor cell load in the bone marrow. In samples taken after CT, a significant decrease of the Ki-67 labelling index (0.18±0.16) was observed. BM samples with less than 100 GD2+ tumor cells/106 MNCs showed a significantly lower GD2+/Ki-67+ fraction, than samples with a tumour load over 100 tumour cells/106 MNCs (0.12±0.13 vs.0.37±0.09, respectively). Despite of a general decrease, in 8 out of 31 samples with rare tumor cells, a retained Ki-67 labelling index (>0.2) was observed.

Conclusions: Rare tumour cells in the bone marrow may represent cell clones with high proliferative capacity. The determination of the Ki-67 labelling index, with special emphasis on residual tumour cells after CT, might be of therapeutical importance in the follow up of stage 4 tumours.

 


6702

DEVELOPING A UNIFIED PROGNOSTIC MODEL FOR NODE NEGATIVE BREAST CANCER PATIENTS FROM FLOW CYTOMETRIC DNA HISTOGRAMS

Charles Bagwell, Verity Software House, Inc., Topsham, ME, USA; Par-Ola Bendahl, Stal Olle, Richard Kelander, Bo Baldetorp, University Hospital, Department of Oncology, Lund, Sweden; Ursula Falkmer, Karolinska Hospital, Department of Oncology, Stockholm Sweden

Is it possible to obtain a small random cell sample from a primary tumor and, with the appropriate methodologies and mathematics, make some predictions on a tumor’s virulence? Is it also possible to create a breast cancer prognostic model that works well and similarly for any laboratory in the world? The major purpose of this study is to begin to answer these questions.

The specific goals of the study are to 1) identify and correct for factors that may have mitigated the usefulness of S-Phasefraction (SPF) and DNA ploidy in other studies, 2) develop a prognostic model for node negative breast cancer patients, 3) identify possible improvements to current flow cytometry methodology and 4) demonstrate transference of the technique to other laboratories.

A total of 350 univariate DNA histograms from three separate centers (center 1, Lund; 2, Linkoping; and 3, Stockholm) were included in the analysis. Frozen samples of primary tumor were stained with propidium iodide using each center’s own staining protocol. The clinical correlates were primary size, number of positive nodes, estrogen receptor status, treatment, time to relapse and relapse status with a median follow-up period of 67 months. Cell cycle analysis was performed by ModFitLT 3.0 with debris and aggregate compensation turned on.

Prognostic analysis was initially restricted to centers 1 and 2 data, totaling 109 node negative specimens. Two compensations were applied to the S-Phase fraction (SPF) prior to prognostic modeling to eliminate confounding correlations between ploidy and SPF. Nodal status was determined to violate the Cox proportionality assumption and was only used as a stratification variable. The three independent prognostic factors for node negative patients were found to be SPF (p<0.0002), ploidy (p<0.02) and aneuploid fraction (p<0.04) in decreasing order of importance. The relative risk (+1SD for each factor compared to population average) was 5.23 (see Figure 1A).

When the Cox prognostic model derived from centers 1 and 2 is applied to center 3 specimens, similar prognostic stratification is observed (see Figure 1B). Other notable conclusions of the study are 1) total SPF is not recommended as a prognostic variable and 2) estimated tumor fraction for the DNA diploid samples may significantly improve the performance of the prognostic model.

In conclusion, information derived from DNA histograms can provide strong prognostic information for node negative breast cancer patients and has the potential for standardization.

 


7991

ADENOCARCINOMA CELL PHENOTYPING BY FLOWCYTOMETRY

Claude Lambert, K. Passabosq, G. Li, P. Blanc, C. Genin, Immunology Lab, University Hosp, St Etienne, France

Adenocarcinoma cell immuno-analysis is of great interest in tumour diagnosis and experiment. For that, Flowcytometry (FCM) should be a tool of choice but is rarely used.

The aim of that study was to validate multiple parameter immuno-analysis of human adenocarcinoma cells by flow-cytometry.

Methods: 15 tumour cell lines, 30 Human solid tumours (kidney n = 15, colon n = 15) and 20 tumoral effusions were tested with more than 20 Antibodies commercialised for immuno-histology (Dako AS, Cymbus ltd, Zymed ltd and a kind gift from do Oostervich Nederland)). A four colour Coulter-Beckmann XL FCM was used. Cytoplasmic analysis were performed through permeabilisation with saponin (0.1%) following cell fixation with paraformaldéhyde.

Results have shown that a few antibodies (ESA, EMA, HEA, MNF116 and CD45) were useful to distinguish accurately cells from epithelial and hematopoïetic origins. Two clones (ESA-1VuD9-, MNF116) were most sensitive to detect a large panel of epithelial cell types (Colon, pancreas, liver, avary, breast, lung, kidney). None of the tested antibodies (even CA 15.3, CA19.9, CA 125) were specific for the tissue origin of the epithelial tumor. But, the probability of immuno-labelling was very different for each antibody according the type of tumour cells. Thus a combination of 3 to 4 antibodies did enhanced the diagnosis accuracy. A statistical program have been established do define the best combinations of 3 antibodies and several combinations are proposed. (ie: kidney cancer cells were G250high, ESA+, CD45- while Colon cancer cells were ESA+, CA15.3-, CA19.9+ and CD45-).

That technique will allow to analyse prognosis parameters on the tumour cells such as adhesion molecule and metastasis risk, or the degree of differentiation. Oncoproteins could also be analysed.

In conclusion, FCM has a fabulous potential application for analysis of cancer cell other than the cell cycle and the ploidy.

 


6801

Has Tpot Gone To Pot?

George Wilson, Gray Laboratory Cancer Research Trust, Mount Vernon Hospital, Middlesex, UK; Nicolas Paschoud, Onco-Hématologie FMH, 2, Lausanne, Switzerland; Philippe Coucke, Centre Hospitalier Universitaire Vaudois, Department of Radiotherapy, Lausanne, Switzerland

The flow cytometric measurement of potential doubling time (Tpot) was hailed as a major advance in human tumour cell kinetics with the first measurements being performed in vivo in 1985. Since that time, several thousand patients have been assessed with a variety of neoplasms and a wide range of treatments including surgery, chemotherapy and radiotherapy. Many of these studies were purely to increase biological knowledge and assess the prognostic significance of this different measure of proliferation. Perhaps, the most interesting and widely studied application of Tpot was its potential as a predictive test to identify patients, particularly head and neck cancer sufferers, who were likely to fail with conventional fractionation. There was a clear rational to believe that fast proliferating tumours would be at a disadvantage. The reality of Tpot has not lived up to its promise. The results from several published studies are at best ambiguous with some centres reporting highly significant correlations with outcome and others failing to find any association with local control or survival. In addition, a recent re-evaluation of all the currently available data in head and neck cancer also failed to reveal any impact of Tpot although the simple labelling index (LI) did show a weak correlation with local control. Why should this be? Tpot is theoretically a better parameter of proliferation than LI or markers such as Ki-67. The answer is multifactorial and includes our inability to measure the parameter correctly, the problem of proliferative misclassification of diploid tumours, the problem of heterogeneity and the design of the clinical study. Each of these problems could be overcome or at least minimised and this can be achieved by understanding the areas of conflict in data analysis, combining flow cytometry and immunohistochemistry, taking multiple samples and undertaking larger scale studies. There is already a trend to dismiss Tpot but this is premature, there is still one major studied to be completed, NCI T92-0045 trial in which 1000 patients are being studied. The criteria for this trial were that the tumour should be head and neck, cervix or rectum and be treated by what is considered to be conventional treatment in each centre and, most important, that a Tpot measurement should be made. The accrual to this trial has finished. There are several reasons why this trial will be the definitive Tpot study, these include the large number of patients and that that each Tpot measurement is critically reviewed and analysed by expert observers. In addition, where possible, more than one sample has been studied from the patients. This presentation will focus on the data being generated by this trial.

 


6966

Uterine Sarcomas: A Clinico-Pathologic, DNA Flow Cytometric and Microsatellite Evaluation of 47 Cases

Jeanette Drew, University Dept. of Obstetrics & Gyn., Subiaco Perth, Western Australia

Uterine Sarcomas comprise less than 1% of gynaecologic malignancies and 2-5% of all uterine malignancies. These tumours arise primarily from two distinct tissues: 1) leiomyosarcoma from myometrial muscle, and 2) mesodermal (mullerian) and stromal sarcomas from endometrial epithelium. The only documented etiologic factor in 10-25% of these malignancies is prior pelvic radiation, often administered for benign uterine bleeding 5 to 25 years earlier.

The prognosis for uterine sarcoma is poor and primarily dependent on the extent of disease at the time of diagnosis. The five-year survival for patients with stage I disease confined to the uterine corpus is approximately 50% versus 0-20% for the remaining stages.

The relative infrequency of these tumours has precluded significant studies on clinical and genetic evaluations of these tumours. At King Edward Memorial Hospital for Women, in Perth, Western Australia (the state referral center for gynaecological neoplasm’s), 47 women with a diagnosis of uterine sarcoma were identified in the pathology archives to be examined as part of a pilot study of clinical, pathological and genetic factors in the pathogenesis of this group of neoplasm’s.

The majority of neoplasm’s were Carcinosarcomas / Malignant Mixed Mesodermal Tumours (MMMT) [31], with 9 patients having Leiomyosarcoma and 3 patients with Endometrial Stromal Sarcoma.

Clinico-pathological parameters of patient age at presentation, tumour stage, grade, myometrial invasion and survival outcome were documented and correlated in the univariate and multivariate analyses of the variables under examination.

DNA ploidy and tumour proliferation was measured by flow cytometric techniques (modification of Hedley technique) on paraffin-embedded biopsies. Microsatellite Analysis for Instability or Loss of Heterozygosity was undertaken using 32P-PCR amplification of 10 selected microsatellite loci putatively involved in oncogenesis (hMLH1, hMSH2, APC, MTS1, Rb, AT, BRCA1, BRCA2, and p53).

Flow cytometric sorting of neoplasm’s with heterogenic ploidy for subsequent molecular evaluation of microsatellite regions was employed to study questions of genetic divergence at the DNA ploidy level and to provide an association for these gross chromosome changes with concomitant changes at the molecular level.

The results of these studies are presented.


Parallel Session V: HIV/AIDS

Chair: FRANCIS MANDY


6753

Single Platform T-cell Enumeration Involving the Multiple Sites of the Canadian Quality Assessment Programme for Immunophenotyping

Tracy Minkus, Michèle Bergeron, Frank Mandy, Health Canada, National Laboratory for Analytical Cytology, Bureau of HIV/AIDS/STD and TB, Ottawa, Ontario, Canada; The Participating Laboratories Of The Canadian QAP, Canada

Introduction:

The National Laboratory for Analytical Cytology (NLAC) provides a Quality Assessment Programme (QAP) for HIV Immunophenotyping for the laboratories of the Canadian HIV Trials Network. In order to address the problem of high interlaboratory variability in absolute T-cell counts using the traditional dual platform methodology, a single platform methodology was adapted for introduction into the Canadian QAP. After several workshops, conducted for the purposes of technology transfer,and an 8 stage initial trial programme, participating labs of the Canadian QAP were asked to report absolute T-cell counts using the new single platform methodology.

Objective: Introduction of a single platform absolute count methodology into the multi-site Canadian QAP for HIV immunophenotyping, in which there is variability in instruments, sample preparation methods, and reagents used.

Method: Absolute T-cell counts, using single platform technology, were reported for a total of five multi-site QAP sendouts. Three fresh whole blood samples (1 HIV- and 2 HIV+) were included in each sendout, making a total of 15 samples for all 5 sendouts. Reporting labs used either TRUCOUNT™ or FLOWCOUNT™ calibrator beads. A universal template, combining fluorescence triggering and anchor gating, was used to determine CD3+4+, CD3+8+, and CD3+ absolute counts. To measure interlaboratory variability, the % coefficient of variation (%CV) for each cell population was determined for each sample (N=15). The reporting labs used a variety of flow cytometers, lysing methods, and antibody combinations (including 3-colour and 4-colour).

Results: On average, approximately 25% of the participating labs in each sendout reported single platform T-cell counts. All %CVs were <12%, with the majority being <10%. The %CVs ranged from 3.62 to11.63 for CD3+, 4.85 to 11.89 for CD3+4+, and 4.76 to 11.84 for CD3+8+.

Conclusion: The single platform method, using a universal analysis template to accommodate all instruments and protocols, is effective for T-cell enumeration in the Canadian multi-site QAP.

 


6395

Simultaneous immunofluorescence and in-situ PCR for determination of HIV viral loads in immune cell subsets

Lisa Reece, James Leary, University of Texas Medical Branch, Division of Infectious Diseases, Galveston, Texas, USA; Kiran Chennupati, University of Texas, Austin, Texas USA; Michelle Nichols, University of Texas at Southwestern, Dallas, Texas USA

While plasma viral load and CD4 counts continue to be important parameters for monitoring the status of AIDS patients, there is a need to develop new ways to monitor the cellular viral load in different cell subsets in different organs (e.g. peripheral blood, bone marrow and lymph nodes) of the patient. Using new in-situ PCR amplification of HIV gag sequences to incorporate fluorescent nucleotides within cells to indicate viral infection and subsequent surface antigen immunofluorescence labeling to indicate cell subset identity, it may be possible to identify the fraction of latently infected or uninfected subsets of different cell subpopulations by flow cytometric analysis. This study utilized CytoChex(TM) fixed 8E5 cells that contained one fully integrated proviral copy of HIV-1 and a single primer pair that generates a 300 base pair sequence of the gag region. Following in-situ PCR and simultaneous incorporation of fluorescein-conjugated dUTP, and staining with phycoerythrin-conjugated anti-CD7 monoclonal antibody, the cells were analyzed by flow cytometry. An important part of the overall process included simultaneous use of Trypan Blue (which also fluoresces in the far red part of the spectrum) to assay cell permeability on a cell-by-cell basis. Differential permability from cell-to-cell is one of the most vexing problems of in-situ PCR as a methodology. An immediate, practical application of this methodology is the development of new ways to monitor cell population fractions latently infected with HIV. Current analyses of viral load from plasma of HIV-infected individuals reflect only on the extent of productive infection and not the number of infected cells. Knowledge of the fraction of latently infected cells, could providea basis for more effective monitoring of HIV-positive patients undergoing treatments. Long-term implications of this research are the identification and isolation by flow cytometry of non-infected hematopoietic stem/progenitor cells for gene therapyand subsequent autologous transplantation back into HIV-infected individuals. For example, live early stem/progenitor cells expressing coincident markers characteristic of cells likely to be uninfected by HIV could be used for immune system re-constitution, with gene therapy making a population of healthy, mature lymphocytes resistant to HIV infection.

 


6501

The Use of In Cell Viral Load (ViroTect TM) as a Tool to Study Infected Subpopulations Seen in HIV-Infected Individuals

Keith Shults B.S., Sue B. Martin, Cytometry Associates, Inc., Brentwood, Tennessee, USA; Bruce K. Patterson, MolPath Inc., Chicago, IL USA

The use of gag-pol specific antisense olgionucleotides to the mRNA of HIV has been described in the literature as a tool to study active infection in intact lymphocytes and monocytes. In an attempt to use the multiparametric power of flow and laser scanning cytometry as a tool to study the multiple dynamic changes seen in HIV infection, we have recently adapted a series of modifications to the assay designed to decrease background, increase the S/N of the probe, optimize fixation conditions to maintain surface antigen quantification and adapt all of the above to an anti-coagulant that will preserve message for extended periods of time( 48-72 hours). We recently confirmed our ability to identify and quantify HIV-infected cell populations by analyzing i) PBMCS infected with T- tropic HIV isolates, ii) HIV positive individuals off therapy and iii) HIV negative normal volunteers ( n=3 of all groups). Studies of the infected PBMCS revealed infected CD4+, CD45RO+ cells with no reactivity present in the CD14+population as would be expected using a T- tropic isolate. The CD4+, CD45RO + population was the predominant productively infected cell with 47% of this population expressing multiple copies of gag-pol mRNA. The expression of the CD4 antigen in the infected cells was 45% less than the uninfected cells with no change seen in the CD14 + monocytes. The HIV positive patients revealed the same pattern of infection with the CD4+,CD45RO+ cell being the predominant infected cell (mean= 18.5%). In 1/3 patients, the monocytes exhibited high levels of gag-pol reactivity and in all cases the CD4 antigen expression was decreased in the infected population. Using the laser scanner, the positive cells revealed a globular "speckled" pattern within the cell as seen with RNA detection. In the normal population, the gag-pol probe revealed less than 0.6% reactivity in the CD4+, CD45RO+, CD4+, CD45RO- and CD14+ populations. In conclusion, we have developed a tool capable of detecting mRNA of HIV in intact cells with the added feature of quantitating copy number/cell in individual cell populations. The technique is robust enough to reveal changes in surface antigen density and should be "plug and play" using other marker combinations. The assay is currently being used to study HIV vaccine efficacy and HIV infection kinetics.

 


6401

 

Abstract Withdrawn

 

 

 


7186

Redox, activation and apoptotic status of monocytes from HIV-infected patients

Carole Elbim, Sylvie Pillet, Marie-Anne Gougerot-Pocidalo, Laboratoire d’Immunologie-Hématologie, Paris, France

Monocytes are precursors of tissue macrophages, which are major targets of HIV-1 infection. Although few blood monocytes are infected, their resulting activation could play a key role in the pathogenesis of HIV disease by modulating their transendothelial migration and by inducing the production of reactive oxygen species (ROS). We used flow cytometry to study the expression of adhesion molecules, actin polymerization, redox status and apoptosis of whole-blood monocytes from HIV-infected patients at different stages of the disease. We demonstrated that circulating monocytes from HIV-infected patients spontaneously produce increased amounts of H2O2, to a degree that correlates with viral load. This high basal H2O2 production was associated with an increased ex vivo oxidative response to bacterial N-formyl peptides and endotoxin, which could participate in the oxidative injury occuring in HIV disease. This increased ROS production was associated with changes in the expression of the antiapoptotic/antioxidant compounds Bcl2 and thioredoxin along the course of the disease. This modulation could result from a dual regulation by oxidative stress and viral load, and could explain, at least in part, the constant rate of DNA fragmentation whatever the stage of the disease. In addition, we found alterations of adhesion molecules expression and increased actin polymerization which could play a role in transendothelial migration of these activated monocytes that already committed to apoptosis.

 


6480

Evaluation of the flow cytometric quantitation of CD95 (APO-1/FAS) as an alternative to CD38 quantitation in HIV+ young children.

Frederic Monsonis, Isabelle Besson, Philippe Poncelet, Michel Canton, Biocytex, Marseille, France

The expression level of CD38 on CD8+ T cells has been proposed as a powerful marker for both prognosis and response to therapy in AIDS. A kit for CD38 quantitation (CELLQUANT CD38/CD8-PE, Beckman Coulter, Miami, FL) was designed in order to standardize the measurement of CD38 expression not only on virally-infected but also on normal blood samples. The rationale was that the test would be sensitive enough to monitor HIV+ patients who initially respond to treatment but may escape it despite keeping a low blood viremia. The low CD38 level expressed on normal sample(300 to 1,600 sites/cell) required the development of an optimal calibrator offering a low detection limit of 300 CD38 sites/cell. Studies of HIV+ patients under treatment are confirming the interest of the sensitive measurement of CD38 in adults. On the other hand, CD38 expression level may not be the most appropriate parameter to be used in HIV+ young children, since CD38 expression is well-known to be elevated in the low age. Indeed, we have observed high values of CD38 mean density (3,500 sites / CD8+ cell) in healthy babies, a value that is typically out of the normal range in adults. On the contrary, the expression of CD95 (APO-1/Fas) is low in babies, increases with aging but also in case of pathologic chronic T cell activation as encountered in viral infections or bone marrow transplantation (BMT). Among T cells Fas is normally distributed in two subpopulations (Faslow/naive and Fashigh /activated-memory, Fig1) that can be distinguished on the basis of their Fas expression. In normal donors, these two subpopulations express <1,500 and 10,500 +/- 2,000 Fas sites per cell, as measured using CELLQUANT Fas kit (Biocytex, Marseille, F). In addition, the mean level of Fas on the whole T cell population (a simple parameter combining the percentage and level of expression of each subpopulation) ranges from 1,700 to 8,600 Fas per cell in healthy donors, a large variation mainly due to the differences in percentages. Invirally infected adults, values over 8,000 are frequent because T cells are often made of > 90% Fashigh cells with a normal level (Fig.1). During immune reconstitution following BMT, this parameter can exceed 25,000 Fas per cell because of the absence of Faslow T cells and the abnormally high Fas expression. Since the level of Fas is normally low in young children and elevated in all HIV+ subjets, further extensive studies could be initiated to validate this parameter as an alternative to CD38 expression for monitoring therapeutic trials in young children.

 


6458

EEPRESSION OF CHEMOKINE RECEPTORS CCR5 AND CXCR4 AS NEW MARKERS OF HIV/AIDS DISEASE IN HEMOPHILIA PATIENTS

Koji Osada, Tokyo’s Women’s Medical University, Japan

Background: The chemokine receptors CCR5 and CXCR4 have been identified recently as major HIV entry sites into CD4+T cells. Our goal was to assess the relationship between HIV/AIDS disease status in patients with hemophilia and expression of these receptors, using lymphocytes from HIV negative blood donors and HIV/AIDS patients using by flow cytometry. Methods: We studied 13 HIV/AIDS patients from the UC Davis Hemophilia Center and 10 asymptomatic HIV infected patients with hemophilia from Shizuoka Children Hospital in Japan. After informed consent, the following data were determined: CD4+ T cells absolute count, HIV RNA viral load, and the presence of CCR5 and CXCR4 on CD4+T cells using anti-human monoclonal antibodies for CCR5 and CXCR4(Pharmingen)and flow cytometry by FACSCalibur (Becton Dickinson). These results were compared with 14 HIV negative blood donors as controls. Results: The receptor CXCR4 on CD4+T cells of American patients was slightly reduced compared with American donors( 87 }7.3% ; 94 }3.3% positive) Japanese patients versus Japanese donors were the same tendency (61 }3.2% ; 50 }2.7 positive). CCR5 expression on CD4+T cells was slightly increased in HIV/AIDS patients compared with HIV negative blood donors (American patients 29 }14% positive; American donors 25 }14% positive, Japanese patients 23 }12% positive; Japanese donors 21 }10% positive). They were not significantly different between patients and donors. But CCR5 expression in cases of Japanese asymptomatic patients were significant reduced (mean 23% positive) when compared with American AIDS patients of the late stage(mean 48% positive). However, CCR5 and CXCR4 on CD4+T cells was not significant different among CD4+ T cells absolute count and HIV RNA viral load. Conclusions: It is interesting that Japanese asymtomatic hemophilia patients infected with HIV had significantly reduced CCR5 numbers. Further studies are needed to determine whether these individuals had slower progression of their HIV as a result of the reduced CCR5 receptors. These new markers using flow cytometry may help in evaluating the status of HIV/AIDS disease.

 


6392

HHV-6 induction of HIV receptors and co-receptors on human hematopoietic stem/progenitor cells

Suimin Qiu, James Leary, University of Texas Medical Branch, Division of Infectious Diseases, Galveston, Texas, USA

Pancytopenia is one of the most common features in overt AIDS patients. The involvement of the hematopoietic stem/progenitor cells during HIV-1 infection has been proposed to lead to the pancytopenia. Whether the CD34+ stem/progenitor cells in vivo are infected with HIV still remains controversial. The project was designed to answer the following three questions: (1) do the CD34+ hematopoietic stem/progenitor cells express HIV receptor (CD4) and co-receptors (CCR5 or CXCR4)? (2) if not, can they be induced? (3) does the induction of the CD4 and CCR5 confer the susceptibility to HIV-1 infection in-vitro and in-vivo? The CD34+ stem/progenitor cells were purified using a combination of magnetic activated cell sorting (MACS) and multi-color flow cytometry and cell sorting. The expressions of CD34/CD4/CCR5/CXCR4 of primary stem/progenitor cells and two stem cell lines (KG-1 and TF-1) were determined by RT-PCR. The infectious HHV-6 viruses used to infect stem cells were derived from an HHV-6 transfected stable cell line (HSB-2/HHV6). The results showed that CD34+ stem cells did not normally express CD4 nor CCR5, but did express CXCR4 (fusin). In the presence of HHV-6 infection, however, CD4 and CCR5 were induced. Viral challenge results indicate thatCD34+ stem/progenitor cells are not normally susceptible to HIV-1 infection because of lack of CD4 and CCR5, but that the induction of CD4 and CCR5 confers susceptibility to HIV infection. To test the in-vivo relevance of these results, we measured the molecular phenotype of developmental antigens and HIV receptors and co-receptors on hematopoietic stem/progenitor cells from peripheral blood samples of HIV-infected patients. 5 of 5 pancytopenic (and CD4<200/mm3) HIV-infected patients expressed HIV-1 co-receptors (CD4 and CXCR4) on their CD34+ hematopoietic stem/progenitor cells in contrast to none of 20 of the non-pancytopenic (and CD4>200/mm3) HIV-infected patients or non-infected controls. While this is obviously a limited study that will need to be confirmed in a larger study, the results may help to explain the rapid decline of the immune system in some AIDS patients. It may also indicate that active infection with HHV-6 may actually make the stem cell compartment vulnerable to direct HIV infection.


Parallel Session V: Molecular Pharmacology & Cellular Physiology

Chair: Rachel Errington


6181

Detecting altered metabolic states in living cells by multiparameter flow cytometry

Martin Poot, Department of Pathology, Seattle, WA, USA; Robert H Pierce, Wright-Patterson Medical Center, Wright-Patterson AFB, Dayton, OH USA

Conventional biochemistry is limited to determining the rate of substrate flux through one metabolic pathway per experiment. Thus, one cannot determine whether metabolic changes occur independently or in concert. Multiparameter flow cytometry, in contrast, allows, after staining cells with a combination of fluorescent dyes of different emission wavelengths, monitoring of several metabolic parameters simultaneously. Thus, by monitoring several parameters in the same cell, several biochemical "experiments" can be performed simultaneously. In addition, one can distinguish metabolic states of cells according to their coordinated changes in several metabolic parameters. For instance, changes in mitochondrial NADH level (UV-excited blue fluorescence) and mitochondrial membrane potential (MMP) can be detected in the same cell by staining with MitoTracker Green (green fluorescent independent of MMP) and with CMXRosamine (red fluorescent sensitive to MMP). The non-fluorescent, reduced form of CMXRosamine (H2-CMXRos) emits red fluorescence as a function of mitchondrial turnover. Thus, mitochondrial turnover can be determined together with NADH level. Concurrent changes in the levels of reduced thiol level (by monobromobimane staining) and mitochondrial membrane lipid (cardiolipin; by nonyl acridine orange (NAO) staining) in individual cells can also be detected in the same cell. A concurrent decrease in cardiolipin and reduced thiol levels and a red-shift of NAO fluorescence are hallmarks of apoptosis. We found this in breast cancer cells exposed to all-trans-retinoic acid (atRA) and to the synthetic vitamin A derivative 4HPR. In cells exposed to atRA, we found the "signature pattern" of a transient increase in mitochondrial oxidative turnover and NADH level. Exposure to 4HPR, in contrast, lead to a concurrent decrease in cardiolipin level and mitochondrial oxidative turnover. In contrast to atRA, 4HPR only weakly affected the NADH level and the MMP. After prolonged incubation, both drugs induced cell death by apoptosis. Thus, two pathways of apoptosis were detected. Cells exposed to drugs that induce apoptosis separated in two distinct metabolic states: an apparently "normal", but functionally altered, and a "compromised" state. Thus, by multiparameter flow cytometry "quantum leaps" between metabolic states can be shown. With this approach, we characterized the biochemical mechanisms of cytotoxicity of estrogen derivatives in normal mammary epithelial cells, of clinically used cytostatic drugs and of oxidative stress. The views expressed in this abstract are those of the author(s) and do not reflect the official policy or position of the United States Air Force, Department of Defense, or the US Government.

 


6472

Subnuclear Localisation And Single Cell Pharmacokinetics Of Topotecan In Human Breast Tumour By Time-Lapse Multiphoton Laser Scanning Microscopy

Rachel Errington, Miriam Jones, Terence Hoy, Marie Wiltshire, Sharon Davies, Anthony Campbell, Paul Smith, University of Wales College of Med, Cardiff, Wales, UK; Michael Shelley, Velindre Hospital NHS Trust, Cardiff, UK; Michael Chappell, School of Engineering, University of Warwick, Coventry, UK

DNA topoisomerase I is a ubiquitous enzyme that forms reversible DNA single-strand breaks (cleavable complexes) as part of its enzymatic manipulation of DNA topology. The enzyme plays a role in transcription, DNA replication, and repair and is the specific target for the anticancer camptothecins. The camptothecin derivative topotecan (TPT) acts by stabilising a covalent bond between a tyrosine residue on the protein and the 3˘-phosphoryl end of the single-strand of DNA it breaks. The ternary complexes formed can trigger cell cycle arrest and cell death through replicon collision or the disruption of transcription centres leading to a selectivity of the drug for S phase cells. The factors that permit and modulate the targeting of topoisomerase I by the camptothecins are important issues given the activity of new derivatives against human solid tumours. Based on free solution studies, the low affinity of camptothecins for DNA has favoured drug-protein rather than drug-DNA interaction models for the genomic localisation and the stabilisation of ternary complexes. Using single-photon UV flow cytometry and 2-photon (800 nm) laser scanning microscopy to detect intracellular drug molecules, we have identified the ability of bioactive TPT to locate at chromatin sites in intact human breast tumour cells (MCF-7). Significant heterogeneity exists in nuclear accumulation and cytoplasmic equilibration of TPT through the operation of a dynamic process dependent upon verapamil-sensitive cellular efflux pathways. Protocols have been established to track the equilibration rates in cell compartments and with DNA to provide pharmacokinetic parameters for individual cells. This represents the first report of direct a camptothecin-DNA interaction being detectable in intact tumour cells. The non-invasive time-lapse multiphoton imaging approach enables the linkage of drug targeting in individual cells with the subsequent biological response. (Supported by the UK Medical Research Council and the Association for International Cancer Research).

 


6222

Determination of the Michaelis-Menten constant (Km) of intracellular enzymatic reaction for individual live lymphocytes

Merav Sunray, Yana Shafran, Naomi Zurgil, Mordechai Deutsch, The Schottenstein Cellscan Center, Physics Department, Bar-Ilan University, Ramat-Gan, ISRAEL

Cell based assays of enzyme reaction kinetics which determine the enzyme activity in individual intact cells have become of great interest in recent years. The Cellscan Mark-S (CS-S) apparatus is a laser scanning cytometer incorporating a unique cell carrier which allows repeated, high-precision measurements to be made on the same individual living cells under physiological conditions. On-line reagent addition or changes in the experimental conditions (changing buffers, ions, substrates and inhibitors)can be accomplished, and the dynamic changes in individual given cells is monitored in real time. In the present study, the CS-S was used to determine the individual Michaelis-Menten constant (Km) of FDA hydrolysis by intracellular esterase in lymphocyte population. The Km constant is extracted from the quantitative relationship between the substrate concentration and the rate of an enzymatic reaction. Any given enzyme reacting with a specific substrate has a characteristic Km constant that may change in the presence of inhibitors. The captured cells on the cell carrier of the CS-S were exposed to the first substrate solution with an FDA concentration, and repetitive measurements were immediately carried out in order to trace the staining dynamics of each individual cell. The first solution was then replaced with an FDA solution with higher concentration solution, that resulted in a different staining rate for each one of the cells. This procedure was repeated four times yielding four hydrolysis rates for each cell. Using this data the apparent individual cell Km values were calculated. The results indicated heterogeneity in reaction rates within the lymphocyte population, a fact which may suggest that the Km of each individual cell can be used asa new tool for kinetic parameter classification. Although a large number of scans per cell were performed, no discernable fading was observed in the control measurements. Furthermore, the accuracy of the Quantitative Fluorescence Cytometry (QFCM) was high despite the repetitive manipulation. The described technique and analysis were applied specifically for tracing differences in the Km distribution within individual lymphocytes in a population following mitogenic activation with PHA. The ratio Vmax to Km of stimulated cells was greater than that of control cells. A measurable difference between the two populations in their intracellular fluorescence polarization (IFP) vs. the reciprocal of Km function was found.

 


6267

HISTIDYLATED OLIGOLYSINES INCREASE THE TRANSMEMBRANE PASSAGE AND THE BIOLOGICAL ACTIVITY OF ANTISENSE OLIGONUCLEOTIDES

Pichon Chantal, Mahajoub Bello-Roufai Jr, Michel Monsigny Sr, Patrick Midoux Sr, Centre de Biophysique Moléculaire-CNRS, Orléans, France

One of the limiting factor of the use of antisense oligonucleotides (ODN) activity is that once internalized by cells, they are confined inside vesicular compartments. Only a small fraction is able to reach their targets located in the cytosol and/or the nucleus. Knowing that poly-L-histidine mediates an acid-dependent fusion and leakage of negative charged liposomes after a protonation of imidazole group (1 ; 2) and H5WYG peptide permeabilizes cell membrane at a pH 6.4 (3); we designed histidylated oligolysines to increase the uptake and the cytosolic delivery of ODN (4). Flow cytometry analysis showed a 10-fold enhancement of the ODN uptake in the presence of histidylated oligolysines. The intracellular localization of ODN and histidylated oligolysines were investigated by confocal microscopy. In the presence of histidylated oligolysines, ODN were localized in vesicles as well as in the cytosol and in the nucleus. While, histidylated oligolysines were only restricted into vesicles. In contrast, in the absence of oligomers, ODN were neither observed into the cytosol nor in the nucleus but were found only in vesicles scattered throughout the cytoplasm. These data indicate clearly that histidylated oligolysines favore the cytosolic delivery of ODN and thereby their nuclear accumulation. The consequence of the cytosolic and nuclear delivery of ODN is an increase of the biological activity of antisense ODN more than 20 fold. That was shown with a transient as well as a constitutive gene expression at mRNA and proteins levels. Substantial activity of histidylated oligolysines was still observed in the presence of serum. Prevention of the lumen acidification of endosome by Bafilomycin A1 abolished the benefit effects of histidylated oligolysines. This data suggets an involvement of the protonation of the imidazole group of the oligomer in the ODN cytosolic delivery. Taken altogether, our results show that histidylated oligolysines and futures derivatives are promissing vectors for enhancing the biological activity of antisense ODN. (1) Wang CY and Huang L. (1984). Biochemistry 23, 4409-44016; (2) Uster PS and Deamer WD (1985). Biochemistry 24, 1-8; (3) Midoux P, et al., (1998). Bioconjugate Chem. 9, 260-267; (4) Pichon C et al., (1999). Nucleic AcidsRes. (revised).

 


6337

Flow cytometric quantitation of integrin expression on cultured fibroblasts or keratinocytes for the study of cosmetic active ingredients.

Christine Garcia, Christophe Chesne, Biopredic, Rennes, France; Philippe Poncelet, Frederic Monsonis, Biocytex, Marseille, France

Adhesion molecules play an important role in regulating various metabolic activities of skin cells These include cell adhesion, efficacy of the dermo-epidermis junction, mechanic properties of dermis and relationship between cells and the extracellular matrix. In order to improve the efficacy of cosmetic active ingredients to modify the expression of integrin molecules on skin cells, we developed an in vitro test using cultures of human skin fibroblasts or keratinocytes. The expression of the alpha integrins (VLA family) by quantitative flow cytometry following immuno-staining with monoclonal antibodies. TGF-beta1 served as a positive control. This growth factor plays an essential role in controlling adhesion of fibroblasts to the extracellular matrix and is known to induce over-expression of VLA-2 ( Hertle et al., J. Invest. Dermatol., 104, 1995, 260-265 ) Primary cultures of human dermis fibroblasts or keratinocytes originating from abdominal surgery (Biopredic International, Rennes, F ) were grown up to ~80% confluency, incubated for 24 hrs with various concentrations of the tested compound (X), TGF-betal ( 10 ng/ml) or medium alone. They were then detached under mild conditions and stained with optimally titered monoclonal antibodies using indirect immunofluorescence. Detachment conditions (trypsin 0.025%-EDTA 0.01%, 10mn, 37°C) were initially verified as having no deleterious effect on the quantitative expression of integrins and several other cell-surface antigens. Calibration beads (Qifikit, DakoFrance) were treated in parallel to standardize immuno-staining and permit absolute measurement of the mean number of integrin molecules per cell. Cells and calibrators were fixed after staining and sent to the flow laboratory by carrier. Analysis were done during the week following staining. Several active ingredients were thus tested towards VLA-1, VLA-2 and VLA-5 (resp. CD49a, b and e) plus their common beta chain CD29, often resulting in significantly increased (e.g. > 120%) expression of some or all of these markers. Typical results of one test on fibroblast are provided in the following table. Whereas internal reproducibility was good, absolute levels of expression varied between experiments performed on primary cultures coming from different donors. Quantitative immunofluorescence flow cytometry is an appropriate adjunct to primary skin cell culture to provide information on the capacity of various cosmetic active ingredients to modulate integrin expression.

 


6184

Mathematical analysis of the [Ca++]i time response curve in Cardiac Myocyte and its diagnostic aspects.

Dror Fixler, Reuven Tirosh, Mordechai Deutsch, Jerome Schottenstein Cellscan Center, Physics Department, Bar-Ilan University, Ramat-Gan, Israel; Asher Shainberg Ph.D, The Otto Meyerhoff Drug Receptor Center, Department of Life Sciences, Bar-Ilan University, Ramat-Gan, ISRAEL

In this study the [Ca++]i time response curve, monitored by Indo-1 during a cardiac myocyte relaxation phase following contraction, is described by an infinite sum of exponential functions of the form: A+SBi*exp(-t/Ci) The coefficients B and A stand for [Ca++]i at the end of contraction (t=0) and at the resting phase (t->infinity) time points correspondingly, and C is the relaxation (or Ca++ release) time constant. The analysis of hundreds of relaxation curves of normal individual myocytes indicates a single exponential component decay behavior having a remarkable repeatability and similarity for many experiments of the coefficients A, B and C, intra and inter cellular. The above coefficients were found to distinguish between normal, abnormal, and diseased myocytes. The effects of adriamycin (ADR) upon calcium accumulation by myocyte sarcoplasmic reticulum (SR), as well as that of IB-MECA and Cl-IB-MECA are examined, utilizing the following proposed analysis method. It is proposed to utilize this approach for the analysis of any response curve obtained by other markers used for the measurement of the contraction-relaxation cycle.

 


Thursday, May 25, 2000


7:45 am – 9:45 am

Clinical Plenary on Hematology

Co-Chairs: Martine Raphael, Hélèné Merle-Béral, Francis Lacombe

Cytogenetically Aberrant Cells in Stem Populations in Hematologic Malignancies

Maria Pallavicini

Immunophenotyping of Heamatological Malignancies: Conventional and New Clinical Applications

Alberto Orfao de Matos

Antitumor Strategies Based on Modulation of NFkB Pathway: Enhancement of Cell Propensity to Apoptosis by Onconase

Zbigniew Darzynkiewicz

Dynamic Fluorescence Imaging: T Cell Activation

Alain Trautman


1045 – 1230

Parallel Session VI: Flow Cytometry: Technical Advances

Chair: David Kaplan


6440

A method for differentiating between diverse fluorophores of similar spectra based on the measurement of their fluorescence polarization

Menachem Kaufman, Naomi Zurgil, Mordechai Deutsch, The Jerome Schottenstein Cellscan Center, Physics Department, Bar-Ilan University, Ramat-Gan, ISRAEL

The utilization of multiparametric assays for biological research and application in single cells is gaining momentum in recent years. Simultaneous labeling of cells with several different fluorescent markers is a common method for performing multiparametric measurements. The standard techniques in use for discriminating between these fluorophore labels are by measuring either fluorescence intensity (FI) at a characteristic wavelength bandwidth (such as the case in most flow cytometers) or fluorescence lifetime. The drawbacks of these methods in this regard are that it is impossible to discriminate between fluorophores with extensively overlapping spectra by FI measurements, and due to sampling time limitations lifetime measurements capability is not provided by high rate individual cell measurement apparatuses. One solution to this problem may be using measurements of fluorescence polarization (FP). The fact that fluorophores with extensive overlapping spectra may yet differ in one or more parameters such as fluorescence lifetime, molecule size, shape, charge, or the sites to which it is constrained and so on, may in many cases affect their FP. This effect was successfully exploited by the CellScan Mark S (CS-S) to perform dual labeling measurement with markers with extensive overlapping spectra. The CS-S is a multiparametric laser based scanning cytometer incorporating a unique cell carrier holding up to 10000 cells in addressable locations that is capable of performing a fast repeated and high precision measurements of FI and FP on intact living cells under tightly controlled but viable physiological conditions. Applying this method we were able to make cell classification on the basis of multi cell staining using standard probes pairs with overlapping spectra regions such as the red wavelength fluorescing propidium iodide (PI) and picoerithrin (PE), and the green fluorescing pair fluorescein and rhodamine 123. Since each on of these probes may be used to indicate a specific cell function, the utilization of this method can greatly enhance and simplify the process of multiparametric analysis both in single cells utilizing static and flow through cytometers, as well as in bulk measurements in cuvette.

 


6509

Digital Data Acquisition System for Flow Cytometric Analysis

Jeffrey Rodriguez, David Galbraith, Bo Xia, Kusnadi Kusnadi, Shiva Murthi, Sundararajan Sankaranarayanan, University of Arizona, Tucson, AZ, USA; Tim Yopp, Palo Verde Software, Inc., Tucson, AZ, USA

Conventional flow cytometers perform cell analysis and sorting through the use of analog circuitry. We have built a prototype of a new digital data acquisition system that interfaces to a commercial flow cytometer. Our system intercepts the analog signal from the photomultiplier, performs analog-to-digital conversion, does digital signal analysis to extract various features, and then feeds these extracted features into one of several pattern classification algorithms. Digital computation offers immense flexibility in the selection of features to compute and allows a variety of pattern classification algorithms to be employed. Based on results with a previous prototype, we are optimistic that the use of a broader set of features will lead to improved classification rates. The new system includes the following key features: * user-selectable amplification factor * user-controlled linear/logarithmic amplification * 40 MHz 12-bit analog-to-digital conversion * digital comparator for pulse detection with user-controlled threshold * dual 1K-word FIFO memories for double buffering of pulses * TMS320C31-60 DSP microprocessor with 32K words of local RAM * 8K words of dual-port RAM for communication with PC. We have completed a preliminary analysis of the processing speed of the new system. We first tested the board using computer- generated, 5-us pulses arriving at constant speed. When we program the microprocessor to merely count the number of arriving pulses (i.e., no feature extraction or classification), the achieved throughput exceeds 500k pulses/s. When we program the microprocessor to read the digitized pulses from the FIFO and calculate three features (peak, width, integral), the throughput is 3800 pulses/s. We performed a second experiment to assess the performance for the more realistic case where the pulses have variable inter- arrival times. Here, we generated pulses as a Poisson random process, where the inter-arrival times are exponentially distributed. The system can be mathematically modeled as a queuing system using a Markov chain. If a new pulse arrives at a time when the FIFOs (the queue) are full, then that pulse is ignored, or lost. Again, we calculating three features per pulse. The results show that the capture rate of the dual-FIFO system is about 20 percentage points greater than that of the single-FIFO system over a broad range of average pulse speeds. A more detailed analysis using real data is currently in progress. Also, we are evaluating the saliency of a variety of features (Fourier transform, wavelet transform, etc.) for use in discriminating cell populations. In addition, we are investigating several approaches (neural networks, isodata, etc.) for pulse classification based on digitally derived features.

 


6627

Enzymatic Amplification Staining for Flow Cytometric Analysis of Cell Surface Molecules

David Kaplan, Dawn Smith, Case Western Reserve University, Department of Pathology, Cleveland, Ohio, USA

Flow cytometric analysis is a powerful technique for the single cell assessment of cell surface expression of selected molecules. A major deficiency of flow cytometry has been its relative insensitivity. Only molecules expressed in abundance have been readily observed. We have developed an enzymatic amplification procedure for flow cytometry that increases the fluorescence signal between 10 and 100 fold thereby producing a significant enhancement in the resolving power of the technique. Using this technique we have been able to detect the presence of molecules that could not be observed by the standard procedure. For instance, we stimulated human peripheral blood mononuclear cells to express the apoptosis-inducing molecule, Fas ligand. Fas ligand expression was confirmed by a sensitive functional assay. The standard flow cytometric analysis of these cells did not demonstrate a definitive population of positive cells; however, enzymatic amplification staining showed that 10-15% of the cells were expressing enough Fas ligand to define a separate postive subpopulation. It should be noted that the amplification method differed from the standard procedure only by the inclusion of the enzyme-catalyzed reporter deposition steps. To ascertain that the staining we observed by the amplification procedure was significant in terms of the function of the cells, we sorted the positive population and the negative population and assessed their cytotoxic potential against Fas expressing Jurkat cells. The results demonstrate that the amplification procedure successfully identified functionally significant levels of Fas ligand on the cell surface.

 


6772

HOW TO DO FOUR COLOR COMPENSATION

Carleton Stewart, Sigrid Stewart, Roswell Park Cancer Institute, Laboratory of Flow Cytometry, Hartwick College, Buffalo, NY USA

Currently, there are two procedures for performing Four-Color Immunophenotyping. In the first method, introduced by Beckman Coulter, a single laser is used to excite FITC, PE, PE-TR and PE-CY5 or PerCP. In the second method, introduced by Becton Dickinson Biosciences, two lasers are used. The first excites APC or CY5 and the second excites FITC, PE and PE-CY5 or PerCP. In both approaches, tandem complexes of either PE-TR or PE-CY5 are used. While Beckman Coulter is currently the only supplier of antibodies directly conjugated with PE-TR (called ECD), there are several companies that supply PE-CY5 antibody conjugates. These reagents exhibit a high degree of inter- and intra-batch variation in the amount of PE leakage from the construct. This requires that each fluorochrome-antibody conjugate must be individually compensated to prevent acquisition of erroneous data.

In both systems, it is also very easy to either over- or under-compensate the instrument because single labeled particles are used and this is inadequate to obtain the correct compensation. By using a bright PE conjugated antibody like PECD8, a bright PE-TR antibody like PE-TR (ECD) CD4 and a dim PE-CY5 antibody, such as PE-CY5-CD2, for labeling lymphocytes, correct compensation can be verified between the separate CD8+ and CD4+ populations. Since both are CD2+, and there is an additional population that is CD2+ CD4- CD8-, correct compensation between PE, PE-TR and PE-CY5 can be verified. For the dual laser system, compensation between FL3 and FL4 is also very sensitive to over- or under-compensation. A very bright APC reagent, such as CD45, can be combined with a PE-CY5 conjugated Ab, such as CD8 and with a PE conjugated antibody such as CD4, to verify the correct compensation.

Because of the variation in CY5 excitation in the PE-CY5 construct (or PerCP) by the diode laser, it is also necessary to compensate each PE-CY5 conjugated antibody uniquely. To perform this task conveniently, we recommend using software compensation for FL2-%FL3 and FL4-%FL3. Careful attention to the details of compensation can provide reliable data for any combination of fluorochromes using any instrument to measure 4-colors of fluorescence.

Supported by # 5P30CA16056-22, 5RO1CA602006 from NCI. and the New York State Department of Health

 


6867

A Compact Solid-State 488 nm Laser For Cell Analysis Applications

Mark Gitin, Paul Ginouves, Matthias Schulze, Wolf Seelert, Jurgen Rosperich, Jurgen Pfaff, Luis Spinelli, Coherent, Inc., Santa Clara, CA, USA

The biological suitability of the argon ion laser is well established. The widespread acceptance of 488 nm excited fluorescent molecules such as FITC and PE sustains the use of the argon ion laser as a light source in cell analysis.

There are obvious drawbacks associated with the comparatively large size and poor electrical efficiency of the ion laser when compared to available solid-state alternatives with outputs at other wavelengths. Many attempts have been made to develop solid-state laser technologies that duplicate the 488 nm output characteristics of the ion laser, including techniques such as upconversion fiber lasers and direct frequency-doubling of diode laser sources. These attempts have been commercially unsuccessful due primarily to unresolved reliability and performance issues.

A new diode laser based technology has been developed that produces 488 nm output with performance characteristics that mimic or exceed those of air-cooled argon ion lasers. This laser is based on a new semiconductor technology derivative that for the first time allows the high quantity production of a robust high power solid-state source at 488 nm. The advent of this new laser technology will allow the production of compact clinical and research instruments with high stability, greater reliability, low power consumption and minimal installation requirements.

The details of this new 488 nm solid-state technology will be discussed for the first time.

 


6871

A Simulation of Chromosome Sorter Using Self-Organizing Map

Shigeomi Hara, Yoshio Noguchi, Hiroshi Douzono, Saga University, Saga, Japan

The high speed sorting of chromosomes is desired in medicine, genome analysis and so on. Many groups of universities and research institutes have been trying to develop the sorting methodology of chromosomes, but at present none has achieved the completely successful sorting.

We call the fluorescence intensity patterns measured by the slit-scan flow cytometer chromosome fluorescence profiles. We use the data of chromosome fluorescence profiles measured by the chromosome sorter developed by Y. Noguchi, one of the authors. Examples of chromosome profiles are shown in Fig. 1. The sorter uses reference profiles representing each 23 classes of chromosomes in order to classify chromosomes. (Since the sample chromosomes are extracted from incubated cancer cells of a woman, there are 23 types of chromosomes.) The selection of the reference profiles strongly affects the correctness of the sorting result of the sorter.

We introduce 23 chromosome models whose form is straight bar, on which there are some fluorescence substances (see Fig. 2). The characteristic parameters of each chromosome model are its length and the intensities of the fluorescence substances. We adjust these characteristic parameters of chromosome models to the data of chromosome profiles measured by the sorter, using T. Kohonen fs self-organizing map (SOM) algorithm. Since SOM algorithm can be regarded as a steepest descent method minimizing an error function, we can improve SOM algorithm so that it may adjust the parameters of chromosome models. We take into account the knowledge about relative lengths of chromosomes, and the intensity distribution of laser beam by Gaussian beam optics. We can construct 23 reference profiles using adjusted chromosome models.

When chromosomes are flowing down in the sorter, the tilt of chromosomes is seems to be inevitable. The tilt of chromosomes make changes of the chromosome profiles, and these changes make the classification of chromosomes more difficult. So we consider the rotation of chromosome models, which also generate reference profiles of corresponding class. Thus we classify chromosome profiles taking into account the rotation of chromosomes.

Adjusted chromosome models also make it possible to simulate the action of the chromosome sorter. The acquisition of sample chromosomes is difficult, and it is also difficult to check how correctly the sorter sorts chromosomes. By computer simulation, we can try many sets of reference profiles, and can estimate the correctness of sorting. Moreover we can determine the performance of the sorter, such as the sampling interval and the precision of measuring fluorescence intensity, which is necessary to sort chromosomes correctly.

 


6888

Routine six color flow cytometry using a 9 parameter, 3 laser benchtop analyzer.

David Coder, University of Washington, Seattle, WA, USA

As the number of parameters that can be measured by flow cytometry increases, the complexity and cost of the required instrumentation increases proportionately. Such difficulties have put complicated analyses beyond the reach of many small- to medium-sized laboratories. The LSR (Life Science Research) is a new benchtop analyzer from Becton Dickinson Bioscience (San Jose, California, USA). Initially released as a dual laser (15mW Ar (488nm) and 8mW HeCd (325nm), 8 detector instrument, there is sufficient space to accommodate a third laser; optical filters can be substituted as needed. Simple modifications in my laboratory included the addition of a HeNe laser (633nm), filters optimized for APC and APC-Cy7, and a simple division of detector outputs to obtain signals from both lasers. Combined with CellQuest software, stored instrument configurations and settings as well as screen displays can be retrieved for a variety of user applications. With the configuration described above, a wide variety of analyses can be performed. Among them are: DNA analyses (PI, Hoechst or DAPI), six color immunofluorescence, immunofluorescence combined with DNA, and some calcium flux studies including four surface markers. Example data to be presented include: six color immunofluorescence (using FITC, PE, PECy5, PE-Cy7, APC, and APC-Cy7), DNA (DAPI) combined with four color immunofluorescence (using FITC, PE, PE-Cy5, and PE-Texas Red), and calcium flux (indo-1) combined with four surface markers. Although the standard instrumentation provides for pulse processing (pulse height, area, and width, and parameter ratios) from signals excited by both lasers, as well as fluorescence compensation for fluorescence from all signals derived from excitation by a single laser, inter-laser compensation is performed off-line on the uncompensated list mode data. The latter is the preferred method for complex multi-color analyses. The LSR accommodates multi-color fluorescence analyses with an instrument that requires no routine alignment. With the selection of proper filters, six color analyses can be performed after the lasers have reached stabile operating power levels. Such ease of operation simplifies instrument operation for the casual user. Moreover, the instrument can act as a complement to multi-laser, multi-parameter sorters in that most fluorochromes employed on the sorter can be used on the benchtop analyzer permitting development of cell labeling combinations on user-operated instrument while leaving the complex sorter for doing what it does best.

 


6929

Development of a Platform Independent, Public Domain FCS File Reading and Writing Package

Robert Murphy, Carnegie Mellon University, Pittsburgh, PA, USA; C. Bruce Bagwell, Verity Software House, Topsham, USA; Mario Roederer, University of California, San Francisco, CA, USA; Gary Salzman, Los Alamos National Laboratory, Life Sciences Division, Los Alamos, NM, USA; Larry Seamer, Bio-Rad, Inc., USA; James Wood, Coral Springs, USA

The Flow Cytometry File Standard (FCS) was first proposed in 1984 and since 1986 has been uniformly used by all flow cytometry system and software vendors. FCS 3.0 was adopted by the ISAC Council as the official flow cytometry data file standard of ISACon March 4, 1998. Version 3.0 was developed primarily to allow files larger than 99 MB and to provide support for alternative (UNICODE) character sets. Since its release, manufacturers have been slow to fully adopt FCS 3.0. The reasons for this appear to include the absence of a pressing need for the new features in FCS 3.0 and concern about the tighter compliance requirements in the new standard. A proposal to abandon FCS may have also contributed to the delay. In order to speed utilization of FCS3.0 features, improve compatibility between vendors, and provide an easier path to adoption of future file standards, the Data File Standards committee of ISAC has been working to develop a platform independent, public domain FCS file reading and writing package. The status of those efforts, including the draft application-programmer interface (API) and a specification for FCS 3.0 in Backus Naur Form, will be presented.


Parallel Session VI: Cell Death II

Chair: Francis Belloc


6787

Apoptotic Cell Death: Correlated cellular changes in multiple apoptotic signals in individual cells.

Gary Bright, Hemang Barbhaiya, Greg Larocca, Valerie Puig-Antich, David Premkumar, Cellomics, Inc., Pittsburgh, PA, USA

Apoptosis pathways involve many biochemical, morphological, and cell physiological events, with the exact pattern of events, pathways, and targets dependent upon the particular cell type and stimulus. Biochemical targets include caspase activation and DNAfragmentation, cell physiological indices include changes in mitochondrial function, and morphological events include nuclear fragmentation and condensation. We have assembled an array or profile of assays including changes in nuclear morphology, such as fragmentation and condensation, cytoskeleton organization, such as f-actin content, mitochondrial physiology, such as membrane potential and mass, and enzyme activation, such as caspase activation and endonuclease activity. High Content Screening (HCS) is a multiplexed functional screening approach based on high resolution fluorescence imaging of multiple targets in living cells. We have applied HCS in the systematic assessment of signals that occur during apoptosis in several cell types in response to a library of apoptosis-inducing compounds. Using HCS we can automatically screen multiple target signals in large numbers of individual cells, enabling a better understanding of how drug candidates affect apoptotic processes. The HCS platform provides an integrated system of specific multiparameter target assays, automated image analysis, data management, and knowledge management.

 


6188

Analysis of laser scattering pattern as an early measure of apoptosis

Zeev Schiffer, Naomi Zurgil, Yana Shafran, Mordechai Deutsch, Jerome Schottenstein Cellscan Center, Physics Department, Bar-Ilan University, Ramat-Gan, Israel

Cell dehydration followed by a change in cell shape and size is one of the early events in apoptosis. Cell shrinkage was found to be a fundamental and universal characteristic of apoptosis. Moreover, recent studies point to the role of the loss of cell volume in activation of the cell death process. Light scattering pattern analysis (LSPA) was applied in the current study, for accurate and sensitive detection of subtle changes in cell size which occur in mouse thymocytes undergoing apoptosis. The method, based on analyzing the scattering pattern as a whole, within a wide range of angles, was found to be more sensitive than flow-through systems or microscope image analysis. This was proved by performing measurements of cell size alteration caused by osmotic changes. The decrease in cell diameter as measured by LSPA, was found to be an early signal of apoptosis preceding the externalization of phosphatydilserine on the outer membrane. When apoptosis was induced by dexamethasone, the change in cell size was dose and time dependent, and could be blocked by pre-treatment of the thymocytes with N acetyl cysteine. (NAC). This implies that the scattering pattern, when combined with fluorescent markers such as annexine-V, may be a powerful tool for early detection of apoptosis. Another advantage gained by the use of this method is the ability to trace repeatedly the same cells and to monitor the kinetics of their size changes.

 


6008

TNF-alpha and Interferon-gamma pathways to apoptosis in human tumor cell line Me-180

Heinz Baisch, Institute of Biophysics, University of Hamburg

We have previously shown that TNFa induces apoptosis in Me-180 cells. About 50% of the cells were detached from the culture flask and showed apoptitic characteristics as measured by Acridin Orange staining, Mitrotracker staining, Annexin-V, PARP-cleavage, and morphology. Interferon g (IFNg) induced nearly no apoptosis, however, it stoped the cells in the cycle. The combination of TNFa and IFNg resulted in a dramatic increase of the percentage of apoptotic cells. This points to a crosstalk between TNFa- and IFNg-pathways. Two TNF pathways seem to be involved in apoptosis. Upon binding of TNFa-trimer to TNF-receptor I the intracellular Death Domain recruits TRADD and FADD. This complex cleaves pro-caspase 8 to form an active dimer, which triggers a caspase cascade. Downstream caspase 3 cleaves PARP, which is regarded as an important step in apoptosis. Another pathway starts by binding of FAN to a domain of the intracellular part of the TNF-R1 close to the cell membrane. Sphingomyelin is transformed into ceramide, which activates MEKK1, JNK, and eventually caspases. An anti-apoptotic signal is also mediated by the TNF via NFkB, which acts as a transcription factor inducing the expression of a variety of genes. One or several of the expressed proteins are inhibitors of apoptosis. The IFN g receptor binding activates STAT-1, which are translocated into the nucleus. There they bind to a variety of promoters as transcription factors. One of the IFN inducible proteins p202 is a putative inhibitor of AP-1 andNFkB transcription factors and thus may interfere with the TNF pathway. In Me-180 cells we have studied NFkB content in whole cells and in the nucleus using antibodies and flow cytometry. We found increased NFkB content in the nuclei after TNFa treatment and decreased NFkB content after IFNg treatment. In the combined treatment the NFkB level was found to be considerably lower than for TNFa single treatment. Further experiments were performed with NFkB inhibitors to reveal if the reduction in the NFkBlevel is correlated with a similar increase in apoptosis induction as found in the TNF+IFN experiments. Despite the downregulation of NFkB by inhibitor SN50 (Biomol), one of the most specific inhibitors of NFkB, the TNF-induced apoptosis was not increased. We conclude that interference of IFNg pathway with TNFa pathway at NFkB level is not responsible for the strong increase of apoptosis.

 


6785

CASPASE 3 ACTIVATION AS A MARKER OF PRE-APOPTOTIC LEUKEMIC CELLS

Francis Belloc, Laboratoire d’Hématologie, Françoise Durrieu, Elizabeth Bascans, Gerald Marit, Francis Lacombe, Hôpital du Haut Leveque, Pessac, France; Marc-Antoine Belaud-Rotureau, Villenave D’Ornon, France

Caspase 3 is a constitutive enzyme which is activated during apoptosis. Because the active caspase 3 is a common effector in several apoptotic pathways, it could be a good marker of (pre)-apoptotic cells. Apoptosis was induced in U937 cells by either Daunorubicin (DNR) or Camptothecin (CPT). Viable and apoptotic cells were sorted by flow cytometry (FCM) on the basis of either FITC-annexin V binding or DiOC6(3) accumulation. DEVDase activity was measured by spectrofluorometry using a fluorogenic substratein sorted populations. Significant DEVDase activity was detected in viable CPT and DNR treated cells demonstrating that partial caspases 3 and/or 7 activation occurred earlier than phosphatidylserine exposure and mitochondrial membrane potential alteration during chemo-induced apoptosis. Kinetic analysis of the effect of DNR treatment showed that the cleavage of caspases 3 and 7 as detected by western blot was contemporary with the increase in DEVDase activity and the detection of positive cells by FCM after in situ labeling with a polyclonal anti-active caspase 3 antibody. Sorted viable and apoptotic cells were also labeled with this antibody and analyzed by FCM. Low amounts of active-caspase 3 were detected in the treated viable U937 cells thus confirming that this antibody is a good alternative to fluorogenic substrates to detect caspase 3 activation by FCM. The same antibody was then used on blood samples from acute leukemia patients during their treatment, in combination with CD45 gating, to detect in vivo caspase activation by FCM in blast cells. In most cases, but not all, this method evidenced a fraction of apoptotic blast cells accompanying the decrease in circulating leukocytes. The percentage of blast cells exhibiting activated caspase 3 greatly varied pending of treatment and patient. These results on U937 confirm that activated caspase 3 is an early marker of chemo-induced apoptosis in vitro. However, the results on patients suggest that in vivo, in some cases, either cell killing occur by a caspase independent mechanism, or the clearance of apoptotic cells precedes caspase activation, or apoptotic blast cells are not released from the bone marrow into the circulation. Nevertheless, this FCM method can help to understand the in vivo effect of chemotherapy and to monitor treatment in leukemia.

 


6424

A multi-parameter flowcytometric analysis of the action of melittin, a bee venom toxin, on nucleated mammalian cells

Gopal Pande, P.S. Manogaran, Centre for Cellular and Molecular Biology, Pradesh, India

We have used multi parameter flow cytometric analysis to study the membrane active properties of melittin -the peptide-toxin from bee venom. Our study reveals that melittin actions on the plasma membrane of nucleated mammalian cells are concentration and time dependent and comprise a sequence of events, which have so far not been reported. We found that at low concentrations (>1 mM) melittin causes the loss of asymmetrical distribution of phosphatidyl serine (PS) residues in the plasma membranes, as shown by Annexin V staining, without causing any significant change in the size or permeability of the membranes. At slightly higher concentrations (>1.6 mM) changes in the PS distribution are followed first, by changes in the size of the cell, as shown by forward angle light scatter, and then increase in the permeability of the membranes, as shown by staining with propidium iodide. At even higher concentrations (>4 mM) we observed no changes in the distribution of PS residues in the membrane and the change in the permeability and cell size occurred much faster, and almost simultaneously, after addition of melittin. At this concentration we also observed changes in the orthogonal light scatter of the cells, which was indicative of the changes in the cytoplasm of the cells. Interestingly, at low concentrations of melittin (<1 mM) a percentage of cells remained unaffected by melittin action but at >1mM concentration all the cells were affected by melittin. These differences between melittin activity at various steps can be exploited to understand the mechanism of action of membrane active peptides on nucleated cells

 


6441

Programmed cell death specific of the unicellular life of Dictyostelium discoideum relies on apoptosis

Damien Arnoult, INSERM U82, Paris, France; Irène Tatischeff, LPBC, Université Paris VI, Paris, France; Etienne Morel, Mathilde Girard, Franck Sureau, Patrice X. Petit, INSERM U129, Paris, France; Marc Dellinger, MNHN, Paris, France; Jean-Pierre Tissier, INRA, LGPTA, Villeneuve d’Ascqu, France

Considerable efforts are underway to improve our understanding of apoptosis and its phylogenic origin and to identify the underlaying machinery a part of which may have been conserved through the evolution, from the procaryotes to the eukaryotes and especially through mitochondria. Dictyostelium discoideum (Dd) is one of the unicellular eukaryotes, for which Developmental Programmed Cell Death (D-PCD) has been described [1]. Recent data strongly suggests an evolutionarily conserved PCD pathway from unicellular to multicellular organisms [2]. PCD can be viewed either as a specific cell proportion regulator during multicellular development or as an abnormal cell suppressor during aging or after various stresses during growth. Due to its pivotal position of primitive unicellular eukaryote during growth trying a first attempt towards multicellularity during development, Dd might give the opportunity to question the dual role of PCD. We questionned whether this reported PCD, was the only PCD depicted by this unicellular eukaryotic protist. The other alternative was that beside ordinary death by necrosis, one might expect an other more general and stress-relevant PCD to be operating during growth. We report conditioned media experiments inhibiting aggregation of Ax-2 cells, which were originally aimed to evidence formation of Dd microcysts. At the light of the present knowledge upon mammalian apoptosis, these experiments strongly suggest that Dictyostelium discoideum, when blocked in development, might be endowed with another PCD specific of its unicellular phase of life. In these conditions, electron microscopy enabled us to evidence both early and late features characteristic of PCD, as recently observed in multicellular eukaryotic organisms. Flow cytometry analysis reveal also an early drop in mitochondrial membrane potential as frequently observed in mammalian apoptosis [3] followed by exposure of phosphatydyl serine and cell shrinkage. But this apoptosis seems to be caspase-independent despite of cytochrome c release from the mitochondrial intermembrane space. We also confirm the presence of other apoptogenic protein in Dd like the Apoptosis Inducing Factor (AIF) which role in the present sequence of events is questionned. Our working assumption is discussed in the light of new experiments for trying to define the death of Dd cells during physiological senescence under axenic growth conditions. 1 Cornillon, S., Foa, C., Davoust, J. et al. (1994) J. Cell Science 107, 2691-27042 Ameisen, J.C. (1996) Science 272, 1278-9 3 Petit, P.X., Lecoeur, H., Zorn, E. et al. (1995) J. Cell. Biol. 130, 157-167

 


Session VI: Hematological Malignancy

Chair: Trond Stokke


6447

Automated Classification of Peripheral Blood (PBL) and Bone Marrow Aspirates (BMA) from Chronic Lymphatic Leukemia (B-CLL) by Three Colour Flow Cytometric Immunophenotyping

Guenter Valet, Cell Biochemistry Laboratory, Martinsried; Guenter Brittinger, University of Essen (Universitaet Gesamthochschule Essen), Division of Hematology, Essen, Germany; Astrid Franke, Div.Hematology University Magdeburg, Klinik für Hämatologie/Onkologie, Magdeburg, Germany; Heinz-Gert Hoffkes, Med.Klinik III Klinikum Fulda, Fulda, Germany

Aims: The correct analysis of flow cytometrically determined light chain restriction in B-CLL requires thorough investigation because the interpretation of results in case of weakly expressed immunoglobulines may be ambiguous. It was therefore investigated whether CD45/14/20, CD8/4/3, kappa/CD19/5, lambda/CD19/5 flow cytometric list mode files were suitable for automated classification by the CLASSIF1 triple matrix algorithm (Cytometry (CCC) 30:275-288(1997), http://www.biochem.mpg.de/valet/classif1.html). Methods & Results: BMA and PBL of normal and APAAP/histologically proven B-CLL patients served as learning set. After completion of the self learning process, BMA were correctly classified as either normal, kappa or lambda B-CLL in 100.0% of the cases (n=22/26/27) with predictive values or 100.0, 100.0 and 96.3%. Similarly, BMA against PBL, as an easier control, classified correctly with: 100.0, 96.2 and 96.3% (n=58/26/27) at predictive values of 100.0, 94.6 and 94.6%. PBLs were correctly classified in 98.3, 91.2 and 100.0% of the cases (n=58/34/38) with predictive values of: 91.8, 91.2 and 94.1%. As a first test set, manually well classifiable clinical cases of kappa and lambda B-CLL in BMA (n=12) and PBL(n=10) samples, unknown to the previously learned classifiers, were correctly classified in 91.2% of the BMA (n=11/12) and 100.0% of the PBL (n=10/10) samples. A second test set with 4 BMA and 6 PBL samples, not well classifiable by routine flow cytometric/histological analysis and equally unknown to the classifiers were unambiguously classifed by automated CLASSIF1 analysis except for one sample which was assigned a double classification. The correctness of the classifications could be internally controlled because the samples consisted of pairs of BMA and PBL samples from 4 patients taken on different days and of two repetitive blood samples on different days from a 5th patient. The classification of the two samples of this patient as well as of the BMA and PBL samples of 3 other patients patients were logically consistent, while one patient showed a lambda B-CLL classification for BMA and a kappa/lambda B-CLL double classification for the PBL sample. Conclusion: CLASSIF1 classifications are well coincident with manual classifications in clinically clear cases. The algorithm classifies also weakly positive and otherwise unclassifiable samples in majority correctly. The computer classification is significantly faster than either manual or automated flow cytometric measurements. Thus, the criteria for on-line result interpretation during flow cytometric measurements are fulfilled.

 


6431

A Novel Cell Permeant and Far Red-Fluorescing DNA Probe, DRAQ5, for Blood Cell Discrimination by Flow Cytometry in Immunophenotyping Procedures

Terence Hoy, Marie Wiltshire, Sharon Davies, Paul Smith, University of Wales College of Medicine, Department of Haematology, Cardiff, UK; Laurence Patterson, University of London

The discrimination of intact nucleated cells in complex populations is an important aspect of many multiparameter analytical procedures including: the enhancement of assay specificity and simplicity, exclusion of cell doublets and debris, the maintenanceof cellular integrity for epitope presentation or the monitoring of cellular biochemistry. Conversely, the exclusion of nucleated cells is essential in assays involving mature red blood cells and platelets. We describe the application of a novel anthraquinone (DRAQ5), designed for use in a range of fluorescence detection technologies, in the discrimination of nucleated cells in typical immunophenotyping procedures. The deep red fluorescing agent (>670 nm emission) is a synthetic compound with a high affinity for DNA and a capacity to rapidly enter living cells or stain fixed cells. DRAQ5 is opimally excited by red-light emitting sources and yields a deep red emission spectrum which extends into the low infra-red. DRAQ5 shows excitation at sub-optimal wavelengths including the 488 nm line and the multi-line UV wavelengths emitted by argon-ion lasers. We have demonstrated utility of DRAQ5 in the analysis of nucleated cells in whole blood and density gradient-separated blood cell preparations using a conventional bench-top flow cytometer equipped with a standard low power argon ion laser. Single beam (488 nm) flow cytometry has been used to demonstrate the utility of DRAQ5-nuclear DNA fluorescence as a discriminating parameter for human leucocytes and lymphoma cells, in combination with fluorochrome-labelled antibodies for the detection of surface antigens and subpopulation recognition. DRAQ5 fluorescence was found to reflect cellular DNA content as evidenced by cell cycle distribution profiles for asynchronous and cell cycle-perturbed populations. We have compared the ability of DRAQ5 and LDS-751 to stain viable cells, provide resolution of DNA content, and to demonstrate a far-red fluorescence signature. The intensity distributions for cells stained with DRAQ5 or the far red stain LDS-751 were also compared with that achieved for propidium iodide stained cells. DRAQ5 achieves resolution of G1, S phase and G2 DNA contents of intact cells comparable with that obtained using propidium iodide in fixed cells. No DNA content resolution was achieved using LDS-751. Importantly DRAQ5 can be used in combination with FITC and RPE-labelled antibodies without the need for fluorescence compensation due to the large Stoke’s shift of the former. (Supported by the UK Medical Research Council and the Association for International Cancer Research).

 


6932

Minimal Residual Disease Investigation in Multiple Mieloma by Immunophenotypical Analysis

Jesús F. San Miguel, Julia Almeida, Gema Mateo, M. Dolores Caballero, Lourdes Vazquez, Alberto Orfao De Matos, Hospital Universitario de Salamanca, Salamanca, Spain; M. Jesús Moro, Servicio de Hematología, Hospital Virgen Blanca, León, Spain; Joan Blade, Servicio de Hematología, Hospital Clínico, Barcelona, Spain; José Hernández, Servicio de Hematología, Hospital General, Segovia, Spain;

Immunophenotyping has proved to be an attractive technique for minimal residual disease (MRD) investigation in acute leukemias (AL) based on the presence of phenotypic aberrancies that allow the discrimination between leukemic and normal cells. In multiple myeloma (MM) the morphological distinction between myelomatous plasma cells (PC) and their normal counterpart is even more difficult than in AL, and thus the availability of other more specific criteria for such distinction would be of great value in order to improve the sensitivity and specificity of complete remission (CR) status assessment. We have used a high sensitive multiparametric flow cytometry method based on triple antigen combinations and a two-step acquisition procedure, through a SSC/CD38+++ CD138+ "live gate", in which a total of 105-106 are screened. In order to assess the incidence of phenotypic aberrancies in MM we have studied 61 cases with a large panel of 21 monoclonal antibodies (CD38, CD56, CD19, CD40, CD28, CD80, CD138, CD10, CD13, CD33, CD20, CD45, CD22, CD23, HLA-DR, CD117, CD34, sIg, FMC7 and CD9. Overall, 87% of MM patients displayed an aberrant phenotype with respect to normal PC: these mainly including antigen overexpression of CD56 (62%), CD28 (16%) and CD33 (6%) and asynchronous expression of CD117 (28%), sIg (21%) and CD20 (10%) on PC. To evaluate the in vitro sensitivity of immunophenotyping serial dilutional experiments were performed in 14 cases and allowed the identification of one aberrant PC among 104-105 normal bone marrow cells (sensitivity limit: 10-4). In 31 MM who underwent autologous stem cell transplantation (Trx) we have studied the number of myelomatous as well as normal PC at four different time points: diagnosis, before Trx (after 4 cycles of induction chemotherapy); and 3 and 12 months after Trx. The mean number of myelomatous PC significantly decreased along the four time points (13%, 3%, 0.3% and 0.1%) In paralell, the mean number of normal PC increased and after Trx the distribution of myelomatous/normal PC evolved to a similar pattern to that observed in MGUS (>97% normal PC) in 63% of the patients. Follow up of these patients is not long enough in order to analyze whether or not these changes could contribute to predict relapse. In summary, immunophenotyping is a useful technique for discrimination between myelomatous and normal PC and could contribute to a better assessment of remission status and monitoring MRD following stem cell Trx.

 


6434

Genotypes and phenotypes in B-cell Non-Hodgkin’s lymphoma.

Trond Stokke, Harald Holte, Harald B. Steen, The Norwegian Radium Hospital, Oslo, Norway; Paula Deangelis, Dept. of Pathology, The Norwegian National Hospital, Norway

We have studied different genetic aberrations and phenotypes in a panel of 96 B-cell Non-Hodgkin’s lymphomas (NHL). Deletions and amplifications were studied by comparative genomic hybridization (CGH) and fluorescent in situ hybridization (FISH), balanced translocations involving the BCL-2, c-MYC, BCL-6, CCND1 (BCL-1) genes by PCR and long-range PCR, point-mutations in TP53 and ATM by constant denaturant gel electrophoresis (CDGE) and sequencing, and methylation of INK4A by restriction enzyme/PCR analysis. Molecular phenotypes studied included expression of bcl-2, c-myc, bcl-6, cyclin D1-3, p16, p53, pRB, p21, gp170, cyclin A, mcl-1, and PCNA by immunoblotting and immunohistochemistry. We have tried to identify the genes and gene products involved in establishing malignancy-associated cellular phenotypes like increased proliferation and decreased apoptosis during lymphomagenesis. The impact of the different genetic aberrations on patient survival and treatment response (clinical phenotypes) was studied by multivariate Cox regression analysis, and compared to classical prognostic parameters like histological grade, IPI, etc. Translocations increased the expression of the respective gene involved, but amplification of CCND1 was found more frequently as a cause of cyclin D1 overexpression than t(11;14). Loss of pRB expression was only found in 1 case, which had lost both RB1 alleles. The cases with 1 RB1 copy had normal pRB expression when the proliferative state was taken into consideration. p53 was typically expressed in cases with deletion of and/or mutation in TP53. Three cases had clonal expression of wt p53. Aberrations in TP53/p53 were associated with a high S phase fraction and short survival (relative risk: 7.9). The other aberrations that had independent prognostic value were deletions at 11q21-23.1 and amplifications of 6p. Several positive as well as negative associations between the genetic aberrations were found, giving some simple clues to the organization of the molecular pathways employed in suppressing malignant growth in B-lymphocytes. Employing genomic as well as expression arrays, the challenge will be to map these pathways in detail. The ultimate goal with respect to clinical management will be to establish prognosis and treatment strategy from (well-defined) genetic aberrations, rather than from poorly defined phenotypes like histological grade, proliferation, IPI, etc.

 


6721

Flow cytometric detection of clonality in mature T-cell-malignancies by use of a novel T-Cell Receptor Vbeta antibody kit

René VanDen Beemd, Henk Wind, Anton Langerak, Jacques Van Dongen, Erasmus University, Department of Immunology, Rotterdam, The Netherlands; Antje Necker, Beckman Coulter, Immunotech, Marseille, France

In contrast to the easy diagnosis of clonality in mature B-cell malignancies via Ig light chain expression, the diagnosis of mature T-cell malignancies is frequently hampered by the lack of markers for clonality at the protein level. Therefore we evaluated the use of Vbeta specific antibodies for detection of single Vbeta domain expression in mature T-cell malignancies by use of a new test kit.

The kit contains 21 fluorochrome-conjugated Vb specific antibodies covering 65% of the alpha/beta repertoire. A conjugation trick allows the simultaneous analysis of 3 Vb regions in a same tube, so that the whole analysis could be performed in 7 tubes.

Frozen PBL samples of 20 TCRa/b+ leukemias were analysed with the antibody kit by flow cytometry as well as by Southern blotting and RT-PCR of the TCRB genes. The molecular results and the antibody results were fully concordant. All large T-cell expansions detected by V beta antibodies were confirmed for being clonal by the use of the other techniques. Vbeta domains could be identified with the antibodies in 15 of 20 cases tested, whereas in the remaining 5 cases RT-PCR analysis detected a Vb gene segment which was not covered by the antibody kit. In those cases, the substantial distortion of the repertoire (large fraction of negative cells)was still suspicious of a clonal T-cell population.


6273

Clinical significance of the apoptotic proteins in acute myeloid leukemia (AML)

Giovanni Del Poeta, Luca Maurillo, Francesco Buccisano, Dept of Hematology, Ospedale S.Eugenio, Roma, Italy; Adriano Venditti, Ro, Italy; Anna Tamburini, USA; Beatrice Del Moro, USA; Anna Maria Epiceno, USA; Sergio Amadori, USA

The involvement of mitochondria in apoptosis is demonstrated by the crucial interactions between bcl-2 and pro-apoptotic related oncoproteins (bax, bad and bak). Moreover, the monoclonal antibody (MoAb) APO2.7 which reacts with a 38 kDa mitochondrial membrane protein (7A6 antigen) highlights an early event of apoptosis. In order to evaluate the clinical significance of spontaneous apoptosis in AML, we investigated 7A6 and bcl-2 expression in 60 patients, 26 females and 34 males, median age 56 years. All patients were treated with intensive chemotherapy regimens (EORTC/GIMEMA AML10 and AML13 protocols), except for the FAB M3 cases, treated according to the GIMEMA AIDA protocol. Bcl-2 and 7A6 expressions were determined on flow cytometer (Epics XL, Coulter) using an anti-bcl-2 124 FITC MoAb (Dako) and APO2.7 PE Moab (Immunotech), respectively. Bcl-2 and APO2.7 were evaluated as mean fluorescence intensities (MFI), calculated as the ratio of bcl-2 or APO2.7 MoAbs mean/negative controls mean. The results were expressed as an index (APO) obtained by dividing MFI APO2.7/ MFI bcl-2. The threshold for considering AML cases as "apoptotic" was set at the APO median value > 1.5. APO < 1.5 patients were associated both with immature FAB M0-M1 classes (P=0.039) and with CD34 positivity (20/29; P=0.018). There was a significant correlation between low or high cytogenetic risk class and APO index > or < 1.5 (P=0.009). With regard to clinical outcome, a significant difference in complete remission (CR) rate was found between APO < 1.5 and APO > 1.5 patients (51.7% v. 88.5%; P=0.003). Overall survival and CR duration were significantly longer (P=0.017 and = 0.007, respectively) in APO >1.5 patients. In multivariate analysis, only APO (P=0.003), age (0.009) and WBC (P=0.043) were confirmed independent prognostic factors for CR achievement. In conclusion, low 7A6/bcl-2 ratio might explain the poor outcome of patients treated with very intensive regimens and might induce to treat the APO <1.5 cases with apoptosis enhancer drugs.


6746

Immunophenotypic characterisation of myelodysplastic syndromes: results of a multicentric analysis by flow cytometry.

Jean Feuillard, Université Paris, Hôpital Avicenne, Bobigny, France; Marc Maynadie, C.H.R.U. LE BOCAGE, Dijon, France; Françoise Picard, Hopital Cochin, Paris, France; Alex Dromelet, Clinique Universitaire UCL Mont Godinne, Belgique; Bernard Chatelain, BELGIUM; Bernard Husson, Hôpital de Jolimont, Belgique; Lydia Campos, CHU St Etienne, Hôpital Nord, St Etienne, France; Vincent Levy, Hôpital Saint Louis, Paris, France

Myelodysplastic syndromes (MDS) constitute a group of heterogeneous diseases characterised by persistent cytopenia and clonal abnormalities of a bone marrow myeloid precursor. Within the GEIL group, we are currently conducted an immunophenotypic study ofMDS. The first aim of this study is to clustered MDS according their immunophenotype, in conjunction with the FAB classification and cytogenetic data. The second aim of this study is to established a data base in order to recognise prospectively the prognostic significance of each of the marker tested. Six French and Belgian centres are currently participating to this study (Bobigny, Dijon, Paris (Cochin hospital), Saint Etienne in France; Yvoir, Haine St Paul in Belgium). This multicentric study has started in February 1999, and the first results are presented.

Immunophenotyping of whole bone marrow cells of MDS was done according to a pre-established consensus method. This method was based on a primary gating for CD45 antigen expression and side scatter (CD45/SSC). Triple immunolabellings with the CD45 marker in combination with 2 other lineage markers allowed to analyse blast cells, neutrophils and monocytes in the same step. Markers tested were CD2, CD7, CD11b, CD11c, CD13, CD14, CD15, CD16, CD19, CD33, CD34, CD36, CD38, CD56, CD71, CD117, HLA DR and PgP. Internal controls were inserted in the panel by testing twice CD36 and three times CD34 expression on the same sample. Contamination of the "blast gate" by erythroblasts was estimated by the percentage of CD36+/CD11c- events. Immunophenotype of blast cells was declared to be undetermined when CD36+/CD11c- events within the "blast gate" exceeded 20%. For each marker, results were expressed as the ratio between the mean of the fluorescence signal for the marker and the mean of auto-fluorescence. Positivity for a given marker was declared for a ratio>5.

Seventy five primary MDS (17 RA, 14 RARS, 27 RAEB, 8 RAEB-T and 9 CMML) were analysed. The CD34 marker was positive on blast cells for 73% of the RAEB plus RAEBT group and for 54% of the RA plus RARS group (p=0,001). CD117 and CD34 expression on blast cells were correlated except in CMMLs (p=0.02). CD36 and CD71 markers were positive on neutrophils for 75% of RARS and for 29% of RA (p=0.01). CD11c was expressed on monocytes of 100% of RA and 64% of RARS (p=0.01)

These first results show that immunophenotyping of MDS can be correlated to the FAB classification, suggesting that, like for acute leukaemia, flow cytometry may useful in the diagnosis of MDS. Therefore, our series of patients will be extended in order to be able to cluster MDS according their immunophenotype, as well as the FAB classification and cytogenetic data.

 


6394

ABSCENCE OF DETECTABLE Ph+ CELLS IN CD34- LIN- "SP" BONE MARROW PROGENITOR CELLS FROM PATIENTS WITH CHRONIC MYELOID LEUKEMIA (CML).

Michael Andreeff M.D., Douglas Weidner, Rui-Yu Wang, Shourong Zhao, Hagop Kantarjian, University of Texas M.D. Anderson Cancer Center, Houston, Texas USA; I. Kuenle, Malcolm K. Brenner, Margaret Goodell, Baylor College of Medicine, Houston, Texas, USA

Recent reports have identified CD34- normal hematopoietic stem cells from several species based on their Hoechst 33342 dual emission staining profile. "SP"=side population cells exclude Hoechst 33344 (Hoechstlow) (Nature Med. 3:1337, 1997; J. Exp. Med. 183:1797, 1996). CD34- stem cells were shown to engraft in NOD/scid mice and in fetal sheep. CML is thought to originate in hematopoietic stem cells. Much work has been done to characterize these cells and to identify a leukemia-free stem cell compartment. To determine if the "SP" compartment contained normal or leukemic cells in patients with CML, we sorted "SP" cells that were lineage negative using a cocktail of 14 different monoclonal antibodies as markers of lineage differentiation (lin- "SP") and performed fluorescence in situ hybridization (FISH) for the detection of Ph+ cells and Real-time (RT)-polymerase chain reaction (PCR) analysis (Taqman) to determine levels of cells containing the b3a2 transcript found in Ph+ CML. Primary bone marrow (BM) samples were obtained from 18 patients (pts.) with CML. In every case, a reduction in the percentage of Ph+ cells was found in lin- "SP" cells as compared to non-"SP" cells. In 9/18 samples, the level of Ph+ cells in the "SP" population was at or below background positivity for the FISH assay (11%). When the lin- "SP" population was further analyzed, decreasing numbers of Ph+ cells were found in regions with decreasing levels of Hoechst 33342 fluorescence. The non-"SP" population of the same samples contained 20-55% Ph+ cells. Taqman quantitative RT-PCR analysis of bcr-abl mRNA was performed on 10/18 BM samples: 9/10 samples showed reduced levels of bcr-abl transcripts in sorted lin- "SP" cells as compared to non-"SP" cells. The difference in b3a2 transcript levels was 1 log (9.35%+8.3%, range: 2.5-25 fold) between lin- "SP" and non-"SP" cells in the same sample. Given the potential contamination of the lin- "SP" compartment with bcr-abl mRNA expressing non-"SP" cells (assumption of 90% purity), 3/9 samples expressed levels of bcr-abl mRNA within the expected background range (4%, 4%, 9%). Conclusion: Results suggest that cells containing t(9;22) in CML do not originate from lin- Hoechstlow "SP" cells. This compartment was found to be free of detectable t(9;22) positive cells in a major subset of patients with CML.

 


Parallel Session VI: Aquatic Bacteria in the Microenvironment

Chair: Philippe Lebaron


6426

Detection of Functional Subgroups in Heterotrophic Bacterioplankton by Flow Cytometry

Stefan Andreatta, Manfred M. Wallinger, Roland Psenner, University of Innsbruck, Institute of Zoology and Limnology, Innsbruck, Austria

Flow cytometry offers the opportunity to analyze large numbers of heterotrophic bacteria from environmental samples quantitatively, quickly and reliably for physiological properties such as DNA, RNA and protein content. However, when connecting these measurements to ecological data they will most often be reduced to simple means or abundances. The appearance of distinct subgroups of bacteria in terms of fluorescence have been described extensively for DNA measurements (DNA-high and DNA-low group) but have also been found for other parameters. We show that a combination of stains such as DAPI (4˘, 6˘-diamidino-2-phenylindole) and SYTO 13 (Molecular Probes) specific for DNA and nucleic acids respectively, can reveal several distinct subgroups of heterotrophic bacteria in environmental samples. These functional subgroups can then objectively and reproducibly be discerned by means of automated data analysis, thus providing a more detailed image of the natural bacterial community and a strong basis for ecological investigations.

 


6482

Flow and image cytometric analysis of protozoan grazing pressure on a bacterial community in a two-stage flow-through system (chemostat)

Manfred M. Wallinger, Stefan Andreatta, Thomas Posch, Roland Psenner, University of Innsbruck - Institute of Zoology and Limnology, Innsbruck, Austria

The grazing impact of three different protozoan species on a mixed bacterial community was studied by means of a simplified and functionally reproducible experimental microbial food web in a two-stage flow-through system. In the first stage the algae Rhodomonas sp. was grown on an inorganic medium together with its accompanying bacterial community growing on algal exudates. This mixture of the algae and bacteria was led into four second stage vessels: (I) a control and three vessels inoculated with(II) a scuticociliate, Cyclidium glaucoma, (III) a mixotrophic flagellate, Ochromonas sp. and (IV) both predators. Flow and image cytometry have been used to count and measure daily samples of DAPI (4', 6'-diamidino-2-phenylindole) stained bacteria for 17 days. Both protozoans caused a clear but distinctly different change to the bacterial community. While Cyclidium glaucoma appears to be grazing predominantly on large, DNA-rich bacteria, Ochromonas sp. grazing is less specific. Flow and image cytometry both have specific benefits. Image analysis readily gives precise information about bacterial cell dimensions, whereas flow cytometry enables the fast quantification of physiological properties such as DNA-content for a large number of cells.

 


6716

Mediterranean coastal water eutrophication and patterns of microorganism abundances as estimated from flow cytometric measurements.

Marc Troussellier, Centre national de la Recherche Scientifique, Laboratoire d’Hydrobiologie, Université Montpellier II, Montpellier, France; Claude Courties, Lab. Modeles En Biologie Cellulaire Et Evolution, Observatoire Oceanologique, Banyuls-sur-Mer, France; Jean-Marc Deslous-Paoli, Nabila Mazouni, IFREMER, Sète, France

Coastal lagoons of the Mediterranean Sea have both ecological and economic importance. As most of the world’s coastal ecosystem they are and will be submitted to the growing anthropic pressure, inducing an increase in their level of eutrophication which may compromise their functions and derived goods (such as fisheries or aquaculture products) and services (such as waste degradation). Different planktonic microorganisms (bacteria, pico- and nanophytoplanktonic cells) are involved in the eutrophication process, especially because they can contribute to the increase of organic matter standing stock and production in the water column. The nature and the size of the dominant microorganism can greatly affect the ecosystem trophic structure and dynamics. We have used flow cytometry (FCM) to simultaneously estimate abundance of microorganisms with different sizes in different locations (Thau and Prevost lagoons, open sea reference) of the French Mediterranean coast with different eutrophication and economic status (exploited versus unexploited lagoon) during a one year survey. Our results clearly showed that bacteria and especially phytoplankton abundances were positively correlated with inorganic nutrient concentration. The main limiting factor for bacteria and phytoplankton appeared to be inorganic phosphorus (P) rather than nitrogen. Among phytoplanktonic cells, FCM allowed to discriminate between pico- (< 2 µm) and nanophytoplanktonic (> 2 µm) cells. The general trend was that only the largest cell abundance significantly increased with the increase in P concentration. As a consequence, while in oligotrophic Mediterranean seawater, picoplanktonic cells dominated over larger cells, we have observed a progressive inversion of this pattern with nutrient richness increase. However, departure from this general trend was observed for areas with large-scale shellfish farming. In such exploited areas, we provide experimental evidences of a preferential retention of nanophytoplanktonic cells by oysters. On the other hand, the prey size-spectra of filter-feeder community appeared greatly dependent of the composition of this community (oyster / epibiota ratio). Conceptual consequences and management implications for shallow coastal areas will be discussed.

 


6733

Cytometry and environmental microbiology

Philippe Lebaron, Laetitia Bernard, Fabien Joux, Université P & M Curie (Paris VI), Laboratoire Arago, Banyuls-sur-Mer, France; Nathalie Parthuisot, Philippe Catala, Claude Courties, Observatoire Océanologique, Laboratoire Arago, Banyuls-sur-Mer, France

The long response time and failure of culturing methods to detect total, viable and active bacteria in the aquatic environment has led to the development of rapid and alternative methods, based on the use of fluorescent dyes and cytometry techniques. Flow cytometry has revealed a powerful tool that can be used for cell enumeration, activity and heterogeneity assessment and cell cycle analysis of bacterial populations and communities. However, the standardization of these rapid methods remains still difficult because (i) each staining procedure cannot be clearly related to the functional role of a cell nor to viability, (ii) different quantitative detection limits between culture and cytometry techniques do not allow many applications and, (iii) most of the staining protocols are not universal and do not apply to natural communities and therefore, have limited applications in the field. Double staining procedures are of great interest (i) to better characterize the physiological state of individual bacterial cells by combining two different functional probes and, (ii) to detect specific cells depending on their physiological state by combining taxonomic and physiological probes. Population ecology is an emerging theme in the aquatic environment to analyze temporal and spatial variations of specific populations of bacterioplankton. Recently, flow cytometry has proven very useful to investigate the relationships between taxonomic, genetic and functional components of the diversity of bacterial communitiesin both natural fresh- and marine-waters. We will emphasize these recent investigations and those in the field of sanitary microbiology. The combination of molecular techniques with cell sorting of labelled cells by flow cytometry allows the detection and identification of active species in natural waters which are not culturable. This approach allows new investigations of the relationships between genetic, taxonomic and functional diversity.

 


6796

Applications of Laser Scanning, Flow and Epifluorescence Cytometry to the Rapid Detection of Bacteria and to Viability Assessment

Nathalie Parthuisot, Philippe Catala, Laboratoire Arago, Banyuls sur mer, France; Philippe Lebaron, Universite P & M Curie (Paris VI),Banyuls sur mer, France

Flow cytometry is a well established technique for the analysis of many different cellular characteristics of microorganisms. Applications are growing fast concomitantly to the development of new fluorophores to analyse different aspects of cell functions making analytical and research applications similarly growing, offering new approaches for direct cell identification, viability and activity assessment.

During the last 5 years, we used a wide variety of fluorescent stains having different cellular targets and compatibles with compact flow cytometers with a single laser source, to develop rapid and alternative methods to overcome the limitations of cultural methods (time consuming, presence of viable but non-culturable cells,). Rapid methods have also been developed to allow the real time monitoring of active and/or viable cell counts in aquatic ecosystems. Applications and comparisons of dyes used for industrial and environmental questions will be presented. Examples will include the comparison of different blue nucleic acid dyes for total counts of bacteria and CTC, ChemChrome V6, DiOC, oxonols dyes tested for activity assessment. Other dyes were also used for Gram staining and will be presented. The advantages and limitations of each staining procedure will be discussed based on examples with both pure cultures and natural water samples.

Moreover, flow cytometry does not apply to bacteria fixed on solid supports and neither to the detection of rare events. Recently, a new laser scanning device (LSD) called ChemScan has been developed to fill this gap. We will compare the quantitative and qualitative limits of flow cytometry, epifluorescence microscopy and laser scanning cytometry. This new LSD technique provides an efficient method for the detection of specific and/or viable bacteria present at low densities or diluted in an important background of non specific and/or non viable cells.

 


6831

Picophytoplankton ingestion by the appendicularia Oïkopleura dioica in natural and artificial environment

Jean Blanchot, Antenne IRD Station biologique de Roscoff, Roscoff, France; Dr. Gabriel Gorsky, Observatoire Oceanologique, Villefranche/mer, France

The content of houses, animal and faecal pellets of the filter feeding Oïkopleura dioica (Tunicata appendicularia) were analysed by Flow cytometry. In laboratory experiments with monoalgal populations Prochlorococcus, Synechococcus and the picoeucaryote Isochrisis galbana were retained by houses, transited through the digestive tract and were aggregated in faecal pellets.

In experiment with mixed food populations all populations were retained by houses but only Synechococcus and picoeucaryotes were always observed in the digestive tract and in the faecal pellets. Prochlorococcus were only observed in the digestive track of 14% animals and in any faecal pellet. The percentage of Synechococcus increased from houses to the faecal pellet. The picoeukaryotes were highly concentrated in the houses but were actively destroyed in the gut.

In laboratory experiment with natural algal concentration of bay de Villefranche/mer all populations are retained by houses. The cell percentage of the three populations were the same in the field and in houses. Prochlorococcus was not detected in faecal pellets and the proportion of Synechococcus when compared to the picoeukaryotes increased.

Picophytoplankton is actively retained and ingested by O. dioica. The Prochlorococcus and the picoeukaryotes are efficiently digested whereas Synechococcus transit through the gut without any apparent degradation and are aggregated in the faecal pellets.

 


2:00 pm – 4:00 pm

Workshops D

Antibodies, Acquisition, Analysis and Reporting for Leukemias and Lymphomas

FacilitatorS: Carleton C. Stewart, SIGRID J. STEWART

Core Manager’s Workshop

Facilitators: Sue DeMaggio, Julie Ager

World Wide Web Utilization: Have We Made Progress?

Facilitators: David M. Coder, Guenter K. Valet

Intratumor Heterogeneity — Is It Important?

Facilitators: Stanley Shackney, T. Vincent Shankey

Functional Analysis

Facilitator: George Babcock

Laser Scanning Cytometry

Facilitator: Zbigniew Darzynkiewicz


5:30 pm – 6:30 pm

Robert Hooke Distinguished Lecture


Where Science Is Heading

Sir John Maddox

As at the end of the nineteenth century, science is at the point when conceptual advances are needed further to deepen the understanding of the world. Here are some of the issues to be dealt with:

• The problem of cosmology and the constitution of matter, to which the key is a quantum theory of gravitation.

• How did life begin on the surface of the Earth? Little progress has been made since the question was first defined by Ernst Haeckel in 1863.

• The Human Genome Project will be complete in 2003, but the normal functioning of all the genes will not be revealed. Then some body will have to construct a numerical model of the cell.

• How did human beings evolve from their common ancestor with the Great Apes? More will be learned from the construction of the human genome than from the fossils being recovered from East Africa.

• How does the brain work? The further detailed study of neurons will not account for thinking.


6:30 pm – 7:30 pm

AWARDS


 


7:30 pm – 9:30 pm

FINAL NIGHT EVENT