International Clinical Cytometry Society
Annual ICCS Meeting Plenary Session Reports


Plenary Session I: Novel Approaches to Data Analysis.
Chair: Joseph A. DiGiuseppe, M.D., Ph.D.

Plenary Session II: EuroFlow – Background, Standardization and Panels for Hematologic Neoplasms. 
Chair: Alberto Orfao, MD, PhD.

Plenary Session III: Eclectic Flow Cytometry Applications for the Assessment of Novel Immunological Parameters. 
Chair: Dr. Maurice (Mo) R.G. O’Gorman, MSc, MBA, PhD, D(ABMLI).

Plenary Session IV: Translational Medicine B-cell Biomarkers in Health, Disease and Therapy. 
Chair: Virginia Litwin, PhD.

Plenary Session V: Hematopathology; Insights into intimidating assays. 
Chair: Jeff Jorgensen, MD, PhD.


ICCS, Houston 2010, Plenary Session I: Novel Approaches to Data Analysis.
Chair: Joseph A. DiGiuseppe, M.D., Ph.D.

Presentations:

Beyond Gating:  Analysis of Clinical Flow Cytometric Immunophenotyping Data by Clustering on Statistical Manifolds:  Treating Flow cytometry Data as High-Dimensional Objects.  William Finn MD, University of Michigan, MI, USA.

Identification of Abnormal Hematopoiesis through Competitive Probability Modeling.  Brent Wood MD PhD, University of Washington, Seattle, WA, USA.

Automated Classification of AML by Cytometric Fingerprinting.  Wade Rogers PhD, University of Pennsylvania, Philadelphia, PA, USA.

The 25th Annual Meeting of the International Clinical Cytometry Society (ICCS) officially opened with a Plenary Session focused on novel approaches to data analysis. Dr. Will Finn of the University of Michigan presented his work involving the treatment of flow cytometry data sets as high-dimensional objects.  Of their nature, flow cytometric data are highly dimensional.  For instance, in a simple 4-color experiment in which FSC and SSC are also considered, the list-mode data reflect 6-dimensional information.  In routine analysis, we reduce these data into a series of 2-dimensional representations of hierarchical gates.  Although we may be able to arrive at a diagnosis in most cases, it seems likely that by reducing the data in this way, we are losing potentially valuable information.

In order to pursue this possibility, Dr. Finn and his colleagues have attempted to treat flow cytometric data as intact high-dimensional objects rather than as a series of sequential 2-dimensional projections.  Their approach makes use of “statistical manifolds,” within which the data are embedded using probability density functions.  In a process termed “Fisher information non-parametric embedding” or “FINE,” differences between these high-dimensional probability density functions can be quantified using robust estimates of the so-called “Fisher information distance” to build a neighborhood map on a statistical manifold.  Each point now represents a single patient’s sample analyzed with a single combination of antibodies, and the distance from its nearest neighbor within the statistical manifold is a measure of how similar or dissimilar the two patients’ probability density functions are.  To permit visualization of the information distances between cases, the statistical manifolds are reduced into 2- or 3-dimensional displays.  In a procedure termed “information preserving component analysis,” the axes of these 2-dimensional representations that optimally preserve the information distances measured by FINE are determined.

Dr. Finn illustrated the ability of this approach to distinguish completely cases of CLL from mantle cell lymphoma using only a single tube containing CD23, CD45, and FMC7.  Similarly, most cases of hematogone hyperplasia and B lymphoblastic leukemia were separable on the basis of a combination of CD10, CD19, CD38, and CD45.  Distinction between normal and dysplastic granulopoiesis was somewhat more problematic using 4-color combinations, though.  For example, left-shifted myelopoiesis in a patient with drug-induced agranulocytosis could not be distinguished from dysplastic granulopoiesis.  In principle, these techniques might provide diagnostic support in difficult cases, permit content-based archiving and searching of flow cytometric data, and potentially uncover diagnostic subgroups that cannot be recognized using our current analytical methods.

The next speaker was Dr. Brent Wood from the University of Washington, who discussed his ongoing work involving “Identification of Hematopoietic Neoplasia through Competitive Probability Modeling.”  Dr. Wood introduced this topic by illustrating some limitations of current analytical approaches, including the inferential reasoning required when not all parameters are simultaneously measured.  This latter problem has been ameliorated to some extent by the increasingly common adoption of polychromatic flow cytometry.  Dr. Wood emphasized that the key to recognizing abnormal populations is a thorough understanding of the corresponding normal populations.  For instance, if one has mapped out in phenotypic detail normal B-cell development, recognizing minimal residual B-lymphoblastic leukemia becomes feasible.  However, the disadvantages of this approach include the need to acquire expertise with increasingly complex multidimensional patterns of maturation, and the subjectivity inherent in recognition of abnormal “populations.”

Dr. Wood has developed an approach, which derives from the principle that the probability that a cell belongs to a population is a function of both the distance of the cell from that population in multidimensional space, and the distribution of the population.  In this approach, termed “competitive probability modeling,” points in multidimensional space are assigned to populations on the basis of the distance of those points to modeled normal populations.  However, in normal hematopoiesis, populations are really focal accumulations of events within continuous maturational pathways.  To take this into account, curves representing maturational pathways are fitted to the points and variation about the median determined for each parameter along the maturational pathway.  This permits the identification of abnormal populations as outliers (e.g. >2 S.D.) from the normal pathway in multiparameter space.  The complex multidimensional pathways may also be deconvoluted, that is, converted into a straight line, in which the variation in expression of a given antigen throughout that pathway is represented.  The multiparametric distance and/or probability from the pathway may be visually represented in a similar way.  It is hoped that one might be able to demonstrate and objectively describe the abnormal maturational pathways seen in myeloid neoplasia using this approach.  Dr. Wood is currently undertaking a large-scale characterization of normal hematopoiesis using antibodies against 242 unique cell surface antigens. As a starting point, he is studying normal thymic T-cell maturation and T lymphoblastic leukemia with the aims of generating a complete description of normal T-cell maturation and identifying antigenic differences in T lymphoblastic leukemia, which should facilitate the detection of abnormal T-lymphoblasts after therapy.

The session concluded with a presentation by Dr. Wade Rogers from the University of Pennsylvania.  Dr. Rogers began with a discussion of high-dimensional biology, which brings not only challenges (e.g., need for more data points, increased variance), but also opportunities.  Importantly, the interactions among variables in highly dimensional data may contain more information than the variables themselves.  Dr. Rogers compared the 2-D dot plots with which we are all familiar to shadows cast by a sculpture, and argued that there is a need to use computational methods to overcome the limitations of manual analysis.  To this end, Dr. Rogers and colleagues have created flowFP (FP stands for fingerprinting), which is part of Bioconductor (R), and can be downloaded free of charge.  FlowFP transforms flow cytometric data into a form that can be represented in heat maps, as has become routine in microarray studies.  Building on a technique called “probability binning,” multivariate space is divided into “bins” to create a model of the space.  The number of events in each bin is counted, and the collection of counts is flattened into a 1-D feature vector.  By lining up the feature vectors for a series of samples, a matrix is created in which the rows are samples, and the columns are features (i.e., discrete regions in multivariate space).  Each event is tagged with its bin membership to facilitate downstream analysis.

Fingerprints of samples represent the expected numbers of events in each bin. By comparing fingerprints of samples, one can detect differences among multivariate probability distributions.  Applications of fingerprinting include: QC/QA, distinction between abnormal and normal samples, and detection of changes in an individual over time.  One of several illustrations of the potential applications of fingerprinting presented by Dr. Rogers was a study involving automated classification of AML.  A series of 42 AML patients and 309 normals was divided into a training set and a test set. Using a standard 5-color immunophenotyping panel, fingerprints were computed for each tube in both AML patients and normals.  Then the most statistically significant informative features that distinguished AML from normal were identified.  These fingerprints were subjected to clustering analysis, to yield heat maps that (with one exception) segregated normals from AML cases in the training set.  Using the most informative features, a support vector machine classifier was trained to determine the probability of a case’s being classified as normal.  Then the classifier was challenged with the test data, a mixture of normals and AML cases.  Overall, the sensitivity and specificity of the classifier for identifying the AML cases were 90.5% and 99.5%, respectively.



Joseph A. DiGiuseppe, MD, PhD
Hartford Hospital, Hartford, CT.


ICCS, Houston 2010, Plenary Session II: EuroFlow – Background, Standardization and Panels for Hematologic Neoplasms. 
Chair: Alberto Orfao, MD, PhD.

Presentations:

The EuroFlow Program: Current Achievements and Future Activities.  Jacques J. M. van Dongen MD PhD, Erasmus MC, University Medical Center Rotterdam, Netherlands.

Construction of the EuroFlow Panels for the Diagnosis and Classification of Haematological Malignancies.  Alberto Orfao MD, PhD Department of Medicine, Cancer Research Centre (IBMCC-CSIC/USAL) University of Salamanca, Spain.

EuroFlow Diagnosis and Classification of Acute Lymphoblastic Leukemia.  Ludovic Lhermitte MD, Department of Hematology, Hopital Necker, Paris, France.

How are panels created in your flow cytometry laboratory?
During the EuroFlow session at the 2010 ICCS annual meeting in Houston, Dr. Jacques J. M. van Dongen suggested that panel construction is typically done in individual laboratories based on personal experience.  These panels frequently take a long time to develop and once established, might be modified slightly by replacing individual antibodies, but are rarely replaced.  Thus flow cytometry panels frequently don’t keep up with developments in instrumentation and reagents.  A group of flow cytometry experts in Europe recognized the need for a more objective multidisciplinary translational research approach to panel development utilizing cutting edge technology.  They also noted that despite the discovery of oncoproteins virtually no new markers had been recently introduced into flow cytometric practice.  During 2004 and 2005 this group began to meet to define the requirements and formulate a strategy for a new standardized system.  They recognized the need for a technical standard, guidelines for antibody selection, and software that could effectively analyze data derived from newer flow cytometers capable of assessing more colors.  As a result of these discussions the EuroFlow consortium was established and has been collaborating since 2006.

How does EuroFlow develop flow cytometry panels?
Instead of using a disease oriented approach, EuroFlow panels are designed to address a specific medical indication.  Recommended panels include tubes to distinguish normal, reactive/regenerative, and/malignant populations, and additional tubes to establish a diagnosis, provide prognostic sub-classification, and to evaluate for treatment effectiveness via detection of minimal residual disease.  The current EuroFlow recommendations are based on a standardized 8-color, 3-laser system with defined instrument settings, standardized procedures and analysis performed using InfinicyteTM software.  Antibody combinations are selected using the shared experience of the EuroFlow members and prospective, objective, evaluation.  Multiple antibody clones and antibody-fluorochrome combinations are tested and combinations selected based on brightness, limited spectral overlap, limited need for compensation, and reagent stability.  These standardized panels are tested against existing panels and further evaluated using Infinicyte software to determine which markers provide optimal separation of populations of cells, distinction  between malignant populations and normal or reactive, and discrimination between disease entities.  These efforts have culminated in the development of several panels and a proposed strategy for application, which is outlined in a manuscript scheduled to be published in Leukemia in 2011[1].

Alberto Orfao described the EuroFlow experience developing panels applied to the evaluation of B-cell lymphoproliferative disorders and used this to illustrated analysis using InfinicytTM software [2-4].  This analysis program uses principle component analysis for automated population separation and graphical representation.  This feature can also be used to objectively evaluate the contribution of each parameter in the separation of populations or diseases and hence their utility in the panel.  For example, this tool can determine the most important markers to distinguish chronic lymphocytic leukemia and follicular lymphoma, and surprisingly preliminary information suggests that the answer is not CD5 and CD10.  The software can also be used to compare information from a single patient to a graphical representation of previously analyzed specimens, for example an abnormal B-lymphoblast population can be recognized by comparing with the expected hematogone pattern, or a new malignancy can be compared to a historical archive of disease subtypes to determine which it resembles most closely.  InfinicytTM  software also has the ability to fuse data from multiple tubes that have overlapping information, and calculate missing results.  For example, all EuroFlow tubes used to evaluate B-cell precursor-ALL contain the common, or backbone, parameters CD45, CD19, CD34, forward light scatter, and side light scatter, that are used to identify populations of cells. These populations are then evaluated for staining with other antibodies contained in separate tubes, such as one tube with CD10 and a separate tube with CD22.  Data from these separate tubes is then merged and the expected staining pattern for CD10 versus CD22 calculated.  This tool has the advantage of providing electronic display of information that is often deduced from review of multiple 2D dot plots, but perhaps has the disadvantage of using several colors for repeated backbone markers rather than unique antibodies. 

Dr. Ludovic Lhermitte elaborated on the EuroFlow recommendations for the diagnosis and classification of acute lymphoblastic leukemia [1].  Initial evaluation of acute leukemia is performed with the following acute leukemia orientation tube: cytoplasmic (cy) CD3 (pacific Blue), CD45 (Pacific Orange), cyMPO (FITC), cyCD79a (PE), CD34 (PerCP-Cy5.5), CD19 (PE-Cy7), CD7 (APC), surface membrane (sm) CD3 (APC-H7).  In this 8-color combination CD45, CD19, and CD34 serve as the backbone for comparison with additional tubes when evaluating B-cell precursor ALL and cyCD3, CD45 and smCD3 serve as the backbone for comparison with additional tubes when evaluating T-cell precursor ALL  The antibody panel for B-cell precursor ALL includes four additional 8-color tubes with the goal of further characterization for diagnostic and prognostic purposes, including determination of the maturation stage of the leukemic cells (CD10, smIg-kappa, smIg-lambda, cyIgmu), detection of an ambiguous lineage (CD33. CD13, CD15, CD65, CD123, CD117), detection of phenotypes associated with molecular aberrations (CD66c), and markers that are useful in the assessment of minimal residual disease following therapy (CD58, CD22, CD123, CD 81, CD9, CD21). 

In addition to the development of standardized panels and optimized analysis, EuroFlow has been working on the development of multiplex immunobead assays for the detection of fusion proteins and oncoproteins for diagnosis and classification of leukemia.  Jacques van Dongen described the development of an immunobead assay for the detection of the BCR-ABL fusion transcript [5].  This initiative required the development of new antibodies that can recognized the fusion protein and a procedure that incorporates protease inhibitors and a check of household proteins to monitor protease activity.  In comparison to the molecular assays for fusion genes, immunobead assays have the advantage of the speed and availability of flow cytometry and, unlike PCR and fluorescence in situ hybridization studies, are not dependent on the position of the breakpoint.  However, the flow cytometric assay is currently not designed to assess minimal residual disease.  The EuroFlow consortium is currently developing multiplex assays for fusion genes associated with AML and ALL that could be incorporated into the flow cytometric evaluation of these diseases.

In conclusion, there is clearly a need for standardization of flow cytometric immunophenotyping for the evaluation of hematologic neoplasms.  EuroFlow has made significant progress in reaching consistency across several laboratories in Europe.  Although there will undoubtedly be further discussion about which, and how many, reagents to utilize, the accumulated data from specimens handled with the standardized EuroFlow protocol provides a valuable resource to assist in answering these questions.



Fiona E. Craig, MD
University of Pittsburgh Medical Center, Pittsburgh, PA.


References:

1.     EuroFlow antibody panels for standardized n-dimensional flow cytometric immunophenotyping of normal, reactive and malignant leukocytes. JJM van Dongen, L Lhermitte, S Bottcher et al. Leukemia (in press).

2.     Automated pattern-guided principle component analysis vs expert-based immunophenotypic classification of B-cell chronic lymphoproliferative disorders: a step forward in the standardization of clinical immunophenotyping.  ES Costa, CE Pedreira, S Barrena et al., Leukemia 2010, advanced online publication.

3.     Generation of flow cytometry data files with a potentially infinite number of dimensions.  CE Pedreira, ES Costa, S Barrena et al., Cytometry Part A; 73A:834-846, 2008.

4.     Probabilistic approach for the evaluation of minimal residual disease by multiparametric flow cytometry in leukemic B-cell chronic lymphoproliferative disorders.  CE Pedreira, ES Costa, J Almeida, C Fernandez et al., Cytometry Part A, 73A; 1141-1150:2008.

5.     Flow cytometric immunobead assay for the detection of BCR-ABL fusion proteins in leukemia patients.  Leukemia 2009;23:1106-1117.  F Weerkamp, E Dekking, TT Ng et al. 


ICCS, Houston 2010, Plenary Session III: Eclectic Flow Cytometry Applications for the Assessment of Novel Immunological Parameters. 
Chair: Dr. Maurice (Mo) R.G. O’Gorman, MSc, MBA, PhD, D(ABMLI).

Presentations:

A Quality Assurance Program for Monitoring Intracellular Cytokine Secretion Assays in Clinical Trials.  Patricia D’Souza, PhD, NIAID, NIH, Bethesda, MD, USA.

The Human Immunome: Where are we and where are we going? J. Philip McCoy Jr., PhD, NIH, Bethesda, MD, USA.

Measuring Telomere Length by Flow Cytometry in Bone Marrow Failure Syndromes.  Peter Lansdorp, MD, PhD, University of British Columbia, BC, Canada.

The Role of Flow Cytometry in the Development of New Therapies to treat Metastatic Melanoma.  Luis Vence, PhD, MD Anderson Cancer Center, Houston, TX, USA. 

Dr. Patricia D’Souza received her PhD in Biochemistry from the University of Glasgow in Scotland and is currently Team Leader in the Vaccine Clinical Research Branch of the Vaccine Research Program in the Division of AIDS, NIAID.  Dr. D’Souza serves as an NIH liaison with scientists around the world through her directions of domestic and international programs.  To determine whether or not a laboratory can generate reliable data for the assessment of magnitude and frequency of antigen-specific T cell subsets to candidate vaccines tested in humans, tools must be developed.  Dr. D’Souza presented the results of an international proficiency testing program to monitor the intra- and inter-assay variability of flow cytometry based intra-cellular cytokine measurements in 16 different laboratories.  The assay included both a 4 color and a 7 color antibody cocktail, with standardized reagents including pre-configured lyophilized stimulation and staining plates.  In each round of proficiency assessment, frozen PBMC were sent to each of the laboratories and they had 4 weeks to perform the flow cytometry assays and upload their calculated cytokine positive lymphocyte subset results and FCS files via a website.  Analysis of all of the data following seven rounds of this program allowed for the identification of key factors contributing to inter-laboratory variation.  From the beginning to the end of the assay, factors included cell processing (assessed as cell viability and cell recovery), instrument set up (compensation) and acquisition (total number of events collected) and finally data analysis (gating strategies).  Review of each of these key factors with the individual laboratories following each proficiency round, combined with centralized analysis resulted in significantly improved inter-laboratory variability.  A scoring system was developed to accurately grade each laboratory in each of the key factors such that a laboratory’s performance could be assessed.  Overall, as was expected, inter-assay variability was higher than intra-assay variability however the level of inter-assay variability was lower than had been reported for other types of immune function assays.  The improved precision was attributed to availability of standard operating procedures along with the use of common reagents.  It is believed that the adoption of a similar strategy along with criteria defining pass/fail could be co-opted for other flow cytometry based assays used to measure antigen specific immune responses.

The second lecture was delivered by Dr. Philip McCoy.  Phil received his PhD in Microbiology from the University of Miami in Florida and then went on to complete 2 post doctoral fellowships, one at the University of Connecticut and the second at the University of Michigan.  He is currently a Senior Scientist and the Director of the Flow Cytometry Platform Center for Human Immunology, Autoimmunity and Inflammation at the NIH.  Phil  was also a past president of the ICCS.  Phil’s lecture began with a video of traveling the entire universe designed to give the audience a perspective on the magnitude of what researching the “Immunome” will entail.  A simple definition as “all of the genes and proteins that are associated with an immune system” also gave the audience a sense of the breadth and scope of the project.   Dr. McCoy presented his particular angle to this project which was the initial development of several 15 color immunophenotyping panels each designed to define one small component of the “immunome” For example one 15 color tube is devoted to examination of T-regulatory cells, a second to Th17 cells, a third to naïve and memory B cells, a fourth to dendritic cells, and so forth.  Given that 15 colors can identify thousands of subsets, the project will require major software analysis.  Phil presented several newly developed software applications designed to crunch such complex list mode files and he has obtained some exciting preliminary results.  Dr. McCoy’s presentation illustrated a novel approach attempting to characterize the complexity of the immune system, the problems inherent in attempting to characterize such a complex system and the potential for the attainment of new information by such an approach.  We look forward to new information that will inevitably be generated as Phil and colleagues continue their quest to understand the “immunome”.

Dr. Lansdorp obtained his MD from the Erasmus University in Rotterdam and his PhD from the University of Amsterdam.  He is currently a professor at the University of British Columbia in Canada. Since 1985 he has been working in the area of stem cell biology at the Terry Fox Laboratory in the B.C. Cancer Agency in Vancouver. He is well known for his work in establishing stem cell assays and reagents, most notably HPCA-2 which most of us use as the CD34 Ab clone to measure hematopoietic stem and progenitor cells. Dr. Lansdorp discussed the biology of telomeres and telomerase in human health and disease. Haploinsufficiency for genes encoding the minimal components of the telomerase enzyme complex including the reverse transcriptase protein (TERT) and the telomerase RNA component (TERC) are now known to give rise to various disorders including dyskeratosis congenita, aplastic anemia and pulmonary fibrosis. The association between mutations in TERT and TERC and these clinical syndromes was discussed.  He also discussed the use of multicolor flow cytometry to measure the average telomere length in a variety of nucleated cells and specific leukocyte subsets. Such assays have proven to be useful to detect carriers of mutations in genes encoding telomerase components. The "flow FISH" assay uses fluorescence in situ hybridization (FISH) with directly labeled peptide nucleic acid probes (18-mers specific for telomeric DNA) and a limited number of antibodies (recognizing epitopes that can withstand the harsh conditions required to denature DNA, i.e. 15 minutes at 87 degrees Celsius in 70% formamide). He explained that inclusion of internal bovine control cells with long telomeres in every test tube has greatly reduced variation in telomere length estimates within and between experiments. Flow FISH has become the method of choice to screen for heritable telomerase disorders and for the study telomere length in individual patients.

Dr. Vence earned his PhD. in Immunology at the Louis Pasteur University in Strausbourg, France and the Joslin Diabetes Center at the Howard Medical School in Boston, MA.  He went on to complete a post-doctoral fellowship in the Laboratory of Jacques Banchereau at the Baylor Institute for Immunology Research in Dallas, TX.  He is currently a Senior Research Scientist and the Co-Director of the Immune Monitoring Laboratory at the M.D. Anderson Cancer Center right here in Houston, TX.  Dr. Vence reviewed the results of 3 clinical trials involving multicolor flow cytometry applications for the evaluation of immunomodulatory effects of targeted and immune based therapies for the treatment of one of the most common and deadliest cancers - malignant melanoma.  Luis discussed the role of immune monitoring of the effects of these therapies in both the understanding of how these novel treatments work as well as in terms of improving their efficacy.  Measurement of early thymic emigrants in a clinical trial involving the addition of a proposed thymic involution inhibitor (Leuprolide) to a melanoma peptide vaccine trial identified an increase in CD8+ TREC enriched emigrants (CD8+/CD45RA+RO-/CD27+/CD103+) and not CD4+ early thymic emigrants (CD4+/CD45RA+RO-/CD31+), without an increase in peptide-specific CD8 T cell response.  The second trial involved the measurement of phosphorylated STAT-1 and intracellular cytokine measurements in patients being treated with high dose interleukein-2.  They observed an interesting relationship between a deficiency in the phosphorylation of STAT-1 and impaired interferon-gamma secretion by NK cells. Interestingly, phosphorylation of STAT-5 was not defective in these melanoma patients. The last clinical trial involved the development of a comprehensive 8-color immunophenotyping panel designed to investigate potential immunological effects of a new tyrosine kinase inhibitor.  Their results indicated that the new targeted BRAF V600E inhibitor had no apparent ill effects on the immune system.

The plenary session was followed by several excellent questions and a lively discussion.



Maurice RG O'Gorman, MSc, MBA, PhD, D(ABMLI)
Feinberg School of Medicine, Northwestern University
and  Children's Memorial Hospital, Chicago, IL



ICCS, Houston 2010, Plenary Session IV: Translational Medicine B-cell Biomarkers in Health, Disease and Therapy. 
Chair: Virginia Litwin, PhD.

Presentations:

Characteristics of Steady State Versus Antigen-Specific Plasma Cells in Humans.  Martin Perez-Andres MD, University of Salamanca, Spain.

B Cell Abnormalities in SLE Patients.  Peter Lipsky PhD, formerly NIAMS, NIH, Bethesda, MD, USA.

B Cell Abnormalities in HCV Patients.  Lynn Dustin PhD, Rockefeller University, NY, NY, USA.

Generation of Stable Monoclonal Antibody-Producing B Cell Receptor-Positive Human Memory B Cells by Genetic Programming.  Tim Beaumont PhD, AIMM Therapeutics and AMC/UVA, Amsterdam, Netherlands.

Translational Medicine
The NIH Roadmap for Clinical and Translational Science describes translation medicine as a process to shepherd biomedical discoveries into clinical applications in order to advance both basic and clinical research.1, 2  This transition of information from “bench-to-bedside” can be reduced to three fundamental stages: basic research; clinical research; and drug development.3 

The study of B cell biology provides an illustration of the translational medicine approach. The highly coordinated and precisely regulated processes of B cell development were first defined in basic research laboratories.  Elegant murine models consisting of inbred strains and genetically altered (transgenics, knock-outs, knock-ins) mice have been essential tools in elucidating the mechanisms of B cell development.4  Flow cytometry used in both phenotypic characterization and isolation of B cells at defined differentiation stages was a critical analytical tool in this research. The details of murine B cell development served as a blueprint for basic research on human B cells.  Similarities and differences between the murine and human systems were identified by the analysis of developing human B cells accomplished by immunophenotyping of mixed cell populations using polychromatic flow cytometry and genetic analysis of homogenous cell subsets obtained by fluorescence activated cell sorting (FACS).5  In addition, the use of retroviral vectors encoding green fluorescent protein-tagged proteins and FACS has been used to construct in vitro models of human B cell development. 

With an understanding of the developmental pathways of B cells in healthy populations, clinical researchers have investigated the implications of B cell abnormalities which manifest as B cell malignancies, immunodeficiency and autoimmunity. Pharmaceutical researchers then utilized the many decades worth of information to identify targets for therapeutic intervention and pharmacodynamic biomarkers to monitor response to treatment.  B cell-directed therapies fall into two main categories, B cell depleting compounds and compounds which inhibit B cell developmental and functional pathways.  CD20, CD22, CD19 CD40, BAFF, APRIL are among the targets for B cell-directed biological therapies.6, 7

B-cell Biomarkers in Health, Disease and Therapy
The Plenary Session “Translational Medicine: B-cell Biomarkers in Health, Disease and Therapy” aimed to highlight the important role of flow cytometry in translational medicine.  Each of the four speakers represented one of the key components of the translational medicine continuum.  Dr. Martin Perez-Andres’ presentation, "Immunophenotypic Characterization of B-Cell Subsets in Peripheral Blood in Adults: Relationship with B-Cells Derived from the Bone Marrow and Lymphoid Tissues", represented the Basic Science arena.  The role of B cells in disease pathogenesis was addressed in the presentations by Dr. Peter Lipsky (B cell Abnormalities in SLE Patients) and Dr. Lynn Dustin (B cell Abnormalities in HCV Patients).  Dr. Tim Beaumont’s presentation “Generation of Stable Monoclonal Antibody-Producing B cell Receptor-Positive Human Memory B cells by Genetic Programming” described a novel method of producing human monoclonal antibodies with potential applications as biotherapeutics.

B cells in Healthy Populations
B cells differentiate in the bone marrow, enter the peripheral blood in an immature (or transitional) form and, travel to the spleen where maturation occurs.  Activation and differentiation of mature B cells occurs in the secondary lymphoid organs. Insofar as the trafficking between the lymphoid tissues occurs in the peripheral blood, the composition of B cell subsets in the periphery is thought to be representative of the immune status of an individual. Martin Perez-Andres characterized four subsets of peripheral blood B cells: immature (CD10+,CD19+,CD20+,CD27–,CD38+); naïve (CD10–,CD19+,CD20+,CD27–,CD38–); memory (CD10–,CD19+,CD20+,CD27+,CD38–); and plasma cells (CD10–,CD19+,CD20–,CD27++,CD38++).8  Additional immunophenotyping of these B cells subsets revealed that CD5 expression was restricted to the immature compartment and the expression of markers associated with B cell receptor signaling were decreased in the plasma cell compartment.  Results from B cell subset immunophenotyping in the peripheral blood of 106 healthy volunteers were used to generate the following reference ranges: immature (3-11 cells/µL); naïve (60-130 cells/µL); and memory (26-65 cells/µL); and plasma cells (1-3 cells/µL).  A comparison of B cell subsets and age revealed that both the relative percentage and absolute counts of memory B cells and plasma cells declined with age.

B cells in Pathogenesis
Systemic lupus erythematosus (SLE) is associated with multiple polyclonal B cell abnormalities and the production of auto-antibodies. The checkpoints which control the tightly regulated B cell differentiation process are altered in SLE, however; it is not clear how the perturbations contribute to pathogenesis.  Peter Lipsky reviewed the role of polychromatic flow cytometry in evaluating abnormalities in B cell maturation associated with SLE.9  Patients with SLE display a decrease in the number of circulating naïve B cells and a concomitant increase in immature, memory and plasma cells.  Subsets of the memory B compartment which are expanded in SLE patients but rare in healthy populations have been described.  Moreover, a correlation with an increase in circulating levels of a CD27-,IgD-,CD95+ memory B cell subset and SLE disease activity was noted. 

Challenges in identifying disease signature biomarkers for SLE were discussed. One is that although B abnormalities have been indentified, how these abnormalities actually contribute to autoimmune disease immunopathology remains unclear.  Additional challenges are the incomplete understanding of the phenotypic expression patterns of the various human peripheral B cell subsets and a lack of established, standardized reporting of cellular phenotypes.

B cell abnormalities also contribute to the pathogenesis of certain infectious diseases. Most patients infected with hepatitis C virus (HCV) become chronically infected and experience liver fibrosis, cirrhosis, and hepatocellular carcinoma as well as a host of extrahepatic diseases such as non-Hodgkin B cell lymphoma and mixed cryoglobulinemia (MC).10  Lynn Dustin described a subset of abnormally activated B cells found in the peripheral blood of HCV patients. For the most part, these cells were clonally expanded, IgM+k+, CD27+, and expressed low levels of CD21. Moreover, they expressed an IgM rheumatoid factor typically utilizing VH1-69, JH4, and Vk3-20 immunoglobulin gene segments. The immunoglobulin genes were somatically hypermutated, suggesting antigen-driven selection.

The clonally expanded B cells were identified with a monoclonal antibody, G6, which recognized B cells expressing VH1-69 derived immunoglobulins.  B cells from HCV+MC+ patients isolated by immunomagnetic bead cell separation and FACS were used in functional assays and for gene expression analysis. The CD21low subset of clonally expanded B cells appeared to be anergic, whereas; the CD21high subset of clonally expanded B cells was capable of activation and differentiation in vitro. The presence of clonally expanded B cells expressing similar immunoglobulin genes in different HCV+MC+ patients strongly supported a model in which B cell expansion occurs in response to a specific antigen rather than a polyclonal stimulus.

Manipulating B cells for Therapy
Tim Beaumont described a novel process for generating cloned lines of antigen-specific B cells which can serve as a source of human monoclonal antibodies with potential applications as biotherapeutics compounds.11  This methodology can also provide a tool for the study of B cell biology and signal transduction through antigen-specific B cell receptors (BCR). 

Briefly, Bcl-6 and Bcl-xL genes were introduced into CD27+ memory B cells which were then cultured in the presence of two factors produced by follicular helper T cells, CD40 ligand and interleukin-21.  This process resulted in memory B cells with high proliferative capacity and the ability of secrete immunoglobulin.  The transduced cells expressed BCR and displayed characteristics of germinal center B cells, such as the expression of activation-induced cytidine deaminase (AID).  Additional characterization by polychromatic flow cytometric revealed that the cells expressed CD19, CD20, CD21 and CD22, the activation markers CD25, CD30, CD70, CD80, CD86, CD95, and CD132 (γc) and IL-21R. 

BCR expression on the immortalized human B cells allows for the selection of antigen-specific cells on the basis of binding of antigen to the BCR, whereas; production of antibodies enables selection of B cell clones on the basis of the functional activities of the secreted antibodies.  Proof of concept experiments included generating cloned lines of respiratory syncytial virus antigen-specific B cells which produced neutralizing antibodies.

Summary
The translational medicine approach is now routinely applied during drug development in order to increase the success rate, and decrease the timelines and costs associated with delivering novel therapeutics.12  The flow cytometry platform which enables the analysis of heterogeneous cellular systems, and provides multiparametric information at the single cell level is a critical analytical tool in drug development and translation medicine.13  Recent advances in flow cytometry including: innovations in sample collection and throughput, increased multiparameter capacity, enhanced event resolution, and novel data analysis tools have greatly increased potential applications of the technology.



Virginia Litwin, Ph.D.
Covance Central Laboratory Services, Inc.
Indianapolis, IN


References

1. http://nihroadmap.nih.gov/clinicalresearch/overview-translational.asp

2. Lean MEJ, Mann JI, Hoek JA, Elliot RM and Schofield G (2008). Translational Research: from Evidence-based Medicine to Sustainable Solutions for Public Health Problems. British Medical Journal; 337: a863.

3. Litwin, V., Andahazy, J.  Monitoring the Cellular Components of the Immune System during Clinical Trials-- A Translational Medicine Approach, in: Litwin, V., Marder, P. (Eds.), Flow Cytometry in Drug Discovery and Development. Wiley-Blackwell, John Wiley & Sons, Inc., New Jersey, pp. 189-204.) 2010.

4. Hardy, RR., Hayakawa, K.  B cell development pathways.  Annual Review of Immunology, 19: 595, 2001.

5. LeBien, TW.  Fates of human B-cell precursors.  Blood. 96:9, 2000.

6. Levesque, MC.  Recent advances in B cell-directed biological therapies for autoimmune disorders .  Clin Exp Immunol 157: 198, 2009.

7. Leandro MJ, de la Torre I. The pathogenic role of B cells in autoantibody-associated autoimmune diseases – lessons from B cell-depletion therapy. Clin Exp Immunol 157:191, 2009.

8. Caraux A, et al. Circulating human B and plasma cells. Age-associated changes in counts and detailed characterization of circulating normal CD138– and CD138+ plasma cells. Haematologica 95:1016, 2010.

9. Dorner, T., et al. Abnormalities of B cell subsets in patients with systemic lupus erythematosus J Immunol Methods. Jun 17. [Epub ahead of print], 2010

10.  Charles ED, Dustin LB.  Hepatitis C virus-induced cryoglobulinemia. Kidney Int. 76:818,  2009

11.  Kwakkenbos, MJ, et al. Generation of stable monoclonal antibody-producing B cell receptor-positive human memory B cells by genetic programming. Nat Med. 16:123, 2010.

12.  Feuerstein, GZ., et al.  The Role of Translational Medicine and Biomarker Research in Drug Discovery and Development. Am. Drug Discovery; 2_1 23, 2007.

13.  O’Hara, D. et al.  Recommendations for the validation of flow cytometric testing during drug development: II assays. J Immunol Methods. Oct 11. [Epub ahead of print], 2010.



ICCS, Houston 2010, Plenary Session V: Hematopathology; Insights into intimidating assays. 
Chair: Jeff Jorgensen, MD, PhD.

Presentations:

ZAP-70, Jeff Jorgensen MD PhD, MD Anderson Cancer Center, Houston, TX, USA.

Myelodysplasia, Sa Wang MD, MD Anderson Cancer Center, Houston, TX, USA.

PNH, Andrea Illingworth MS, Dahl-Chase Diagnostics, Bangor, ME, USA.

T- and NK-Cell Large Granular Lymphocyte Proliferations, Dragan Jevremovic, MD, PhD, Mayo Clinic, Rochester, MN, USAS.

Plenary session V, Hematopathology: Insights into intimidating assays focused on 4 challenging areas at the intersection of flow cytometry (FC) and hematopathology.  

In the first segment, Dr. Jeff Jorgensen discussed approaches to ZAP-70 testing in chronic lymphocytic leukemia (CLL).  CLL, the most common mature lymphoid leukemia, has a heterogeneous clinical course.  While some patients do well for years without definitive therapy, a subset of patients do relatively poorly and progress despite therapy.  Many groups have undertaken the task of identifying variables with prognostic significance in CLL.  One of the features demonstrated to have clear prognostic significance in CLL is mutational status of the variable region of the immunoglobulin heavy chain.  However, assay of mutational status is both time consuming and labor intensive rendering it difficult to perform on a standard basis in most clinical laboratories.  Gene expression profiling studies have demonstrated that expression of ZAP-70 (zeta associated protein of 70 KD) has prognostic significance, with expression of ZAP-70 by CLL being associated with an adverse prognosis. ZAP-70 expression can be assessed by FC and when present (using a cut of expression on 20% of CLL cells or greater1), is associated with an adverse prognosis in CLL with some studies suggesting that ZAP-70 may provide prognostic data beyond that provided by assessment of mutational status.  Since its description as a prognostic marker, several methods have been utilized in assessing ZAP-70 expression by FC.  However, development of a FC assay for ZAP-70 has been difficult given a combination of antigen related (ZAP-70 is a cytoplasmic, non-abundant antigen that shows a continuum of expression), assay related (appropriate sample and processing, appropriate antibody clone to use, controversy in defining positive versus negative), and patient related (lack of normal B cells and/or T cells in CLL patients hindering determination of a threshold) factors.  Upon this background, Dr. Jorgensen described and provided preliminary results using an assay for ZAP-70 testing created to minimize some of he problems that have hindered other described assays.  The 5 color assay he described draws from the work of others1-3 to optimize staining conditions, maximize the signal to noise ratio with the choice of the SBZAP clone, and to determine objective and reproducible cutoffs distinguishing positive and negative cases.  As many patients lack adequate numbers of normal B cells, or, in some cases, T cells to serve as adequate negative and positive controls, samples were spiked with normal blood.   Rather than determine a % positive cut off, this assay investigated the ratio between the MFI of ZAP-70 on normal B cells, NK and T cells, and CLL cells, and, determined a Z index (which normalized the level of ZAP-70 expressed on the abnormal cells to that expressed on the normal background B cells and T cells).   The Z index correlated well with mutational status and appears to be a reliable method for ZAP-70 evaluation in CLL.  Dr. Jorgensen wrapped up by noting that this assay is promising but that there is more that needs to be done in optimizing and standardizing this “intimidating” assay, a sentiment echoed during the question and answer session. 

In the next segment, Dr. Sa Wang, also of MD Anderson Cancer Center, presented insights in the use of FC in the evaluation of myelodysplastic syndromes (MDS).  MDS is a clonal myeloid stem cell disorder characterized by ineffective hematopoiesis in which the peripheral blood shows cytopenias in the setting of a marrow that is typically hypercellular with evidence of morphologic dysplasia.  Blasts may be variably increased in MDS (although by definition are less than 20%) and a subset of cases show cytogenetic abnormalities which, when present, may aid in diagnosis.  However, neither dysplasia nor cytopenia is specific for MDS and a substantial subset of cases may have normal cytogenetics making diagnosis difficult.  FC has long been recognized as a potential tool in the evaluation of MDS with many citations in the literature over the past 15 years demonstrating the efficacy of this modality in MDS.  FC was even recognized in the 2008 World Health Organization Classification of hematopoietic neoplasms, which notes the that the presence of 3 or more FC aberrancies is suggestive of MDS.  However, methods used to evaluate MDS by FC are heterogeneous and there is a lack of standardization in the field.  Dr Wang described two studies conducted in her laboratory demonstrating the utility of FC in evaluating MDS.  The first study4 evaluated  FC in detecting myeloid stem cell disorders v.  In this study, the following patient groups were studied: 180 MDS, 31 patients with (MDS/MPD), 37 non-MDS cytopenia and 20 MPN.  Patients were categorized by FC as having positive, intermediate, or negative FC based on evaluation of myeloid blast, maturing myeloid , or monocytic populations.    When considering cases that are FC positive only, this assay showed a sensitivity and specificity of 84% and 97% respectively for detecting myeloid stem cell neoplasms.  If both FC positive and intermediate cases are considered, the sensitivity of the assay increased, 98%, while the specificity decreased, 78%.  Dr. Wang also discussed a study conducted by her group5 indicating that with use of the same FC assay in patients with cytopenias, the absence of significant FC abnormalities had a very high negative predictive value of 95%.   As a caveat to these encouraging data, Dr. Wang noted that an understanding of normal myeloid maturation is critical to this type of analysis and provided several interesting examples (including a case of PNH) that could be misinterpreted as abnormal maturation by an inexperienced observer.  One approach that Dr. Wang described that appears to simplify analysis and standardization of FC for MDS involves focusing on CD34+ blasts.  Dr. Wang described how this method was applied at MD Anderson Cancer center with favorable results. This talk highlighted many exciting insights into FC for MDS with one of the most exciting points emphasized during the Q and A session being the observations that both a negative and positive FC result are helpful in evaluating patients for MDS and, that FC correlates with morphology and cytogenetics but appears to also provide additional information.  Therefore, patients with normal morphology who have abnormal FC should be followed up because a subset of these patients will develop MDS.  However, in order for FC in MDS to become routine practice, there is a need for standardization in this field and currently, experience with maturational patterns of blasts, myeloid and monocytic cells is required to accurately perform this assay.

The third segment of the session was presented by Ms. Andrea Illingsworth, a newly elected ICCS council member from Dahl-Chase Diagnostic Services, who spoke about PNH assays.  Paroxysmal Nocturnal Hemoglobinuria (PNH) is an acquired clonal disorder caused by mutations of the PIG A gene leading to lack of production of the GPI anchor and resulting in lack of expression of GPI linked proteins.  FCI is the method of choice for diagnosis and testing was evaluated in detail at a workshop following the 2008 ICCS meeting in Portland.  As a result of a discussions at that workshop, a consensus guideline on which Ms. Illingsworth is a co-author was created and published this year in Clinical Cytometry.6   This guideline is designed to assist labs in bringing on PNH testing and to enable labs already performing this testing to improve their methods. Of note, an accurate diagnosis of PNH is now especially critical as a specific therapy, SolirisÒ, is available.  Ms. Illingswrth notes that much progress has been made in increasing the sensitivity of PNH assays.  While assays clinically available in 2004 were able to detect PNH populations present at >1% of the white blood cells, assays available beginning in 2008 utilizing FLAER (a bacterial toxin, aerolysin, that binds to GPI) in combination with absence of GPI linked proteins, are able to detect PNH like populations present at levels as low as 0.01%.  PNH like populations of less than 1% are noted to be PNH like clones rather than being indicative of PNH.  Such clones may be detected in patients with aplastic anemia or MDS and may indicate patient groups more likely to respond to immunosuppressive therapy.  Despite these advances, data from proficiency testing surveys suggest that labs show a range of success in performing this assay with common problems relating to the fact that PNH is a rarely encountered disease, there is variability in testing across labs (including differences in sample processing and antibodies and clones used), there is a relative lack of controls and QC material, and, there is some difficulty in the terminology of reporting.  Ms. Illingsworth reviewed the ICCS guidelines for PNH testing, described the assay as it is performed in her laboratory7, noted several potential pitfalls in testing, and reviewed important elements to include in reporting of PNH results.   She noted that future directions in this rapidly progressing field should focus of standardization and improvement of commercially available controls and survey material to reduce interlab variability. 

The final segment of plenary session V was given by Dr. Dragan Jevremovic from Mayo Clinic who discussed indolent T and NK cell large granular lymphocyte (LGL) proliferations.  These disorders are characterized by a persistent (greater that 6 months) increase in circulating cytotoxic T cells or NK cells in the setting of unexplained cytopenias.  The cell of interest is typically cytologically bland and resembles normal reactive LGLs by morphology and immunophenotype8. T cell LGLs typically express CD3 and CD8 with variable loss of CD5 and CD7.  These proliferations typically express the alpha-beta T cell receptor (TCR) although approximately 10% of cases may express the gamma-delta TCR.  CD57 is usually positive and CD56 may be expressed as well.   NK LGLs express CD2 and CD7 but lack CD3 and CD5 similar to reactive NK cells.   These cells express CD16 and/or CD56.   Clonality can be established in T cell LGL proliferations using T cell receptor gene rearrangement studies or TCR vbeta analysis by FC.  In NK cells, clonality may be suggested by analysis of KIR receptor expression.  KIR receptor skewing may also be seen in a subset of T cell LGL proliferations.  In addition to the peripheral blood, these proliferations can be seen in the marrow where they form sinusoidal or intravascular infiltrates, with immunohistochemistry required to establish the pattern of involvement.  After providing a clinical overview to the spectrum of these disorders Dr. Jevremovic presented 2 cases, a rheumatoid arthritis associated T cell LGL proliferation and an NK cell LGL proliferation that illustrated several features associated with LGL proliferations including assessment of clonality by flow cytometry using KIR receptors and TCR vbeta chain analysis, and, differences in expanded CD16dim/CD56bright and CD16bright/CD56dim NK cell populations with the later group more frequently associated with cytopenias and restricted KIR expression.  In addition, he emphasized that these are typically indolent proliferations; therefore, in patients presenting with a more aggressive disease, alternative diagnoses such as hepatosplenic T cell lymphoma or aggressive NK cell leukemia should be considered.  Finally, Dr. Jevremovic closed with the controversial question of whether this class of diseases should be considered true neoplasms of if they would be better classified as unusual reactive processes.

Plenary session V provided a stimulating look at a variety of topics in the overlapping spheres of FC and hematopathology.  Each talk, although it focused on a different area of hematopathology, shared a common theme relating to the need for increased standardization in FC, both in the mechanics of testing, and, in test interpretation.  PNH has provided one example of how the efforts of ICCS and the society’s members can have a positive impact on standardization. 



Sindhu Cherian, MD
University of Washington, Seattle, WA.


References:
1. Crespo M, Bosch F, Villamor N, et al. ZAP-70 expression as a surrogate for immunoglobulin-variable-region mutations in chronic lymphocytic leukemia. N Engl J Med. May 1 2003;348(18):1764-1775.
2. Shankey TV, Forman M, Scibelli P, et al. An optimized whole blood method for flow cytometric measurement of ZAP-70 protein expression in chronic lymphocytic leukemia. Cytometry B Clin Cytom. Jul 15 2006;70(4):259-269.
3. Gachard N, Salviat A, Boutet C, et al. Multicenter study of ZAP-70 expression in patients with B-cell chronic lymphocytic leukemia using an optimized flow cytometry method. Haematologica. Feb 2008;93(2):215-223.
4. Stachurski D, Smith BR, Pozdnyakova O, et al. Flow cytometric analysis of myelomonocytic cells by a pattern recognition approach is sensitive and specific in diagnosing myelodysplastic syndrome and related marrow diseases: emphasis on a global evaluation and recognition of diagnostic pitfalls. Leuk Res. Feb 2008;32(2):215-224.
5. Truong F, Smith BR, Stachurski D, et al. The utility of flow cytometric immunophenotyping in cytopenic patients with a non-diagnostic bone marrow: a prospective study. Leuk Res. Aug 2009;33(8):1039-1046.
6. Borowitz MJ, Craig FE, Digiuseppe JA, et al. Guidelines for the diagnosis and monitoring of paroxysmal nocturnal hemoglobinuria and related disorders by flow cytometry. Cytometry B Clin Cytom. Jul 2010;78(4):211-230.
7. Sutherland DR, Kuek N, Azcona-Olivera J, et al. Use of a FLAER-based WBC assay in the primary screening of PNH clones. Am J Clin Pathol. Oct 2009;132(4):564-572.
8. Morice WG, Kurtin PJ, Leibson PJ, Tefferi A, Hanson CA. Demonstration of aberrant T-cell and natural killer-cell antigen expression in all cases of granular lymphocytic leukaemia. Br J Haematol. Mar 2003;120(6):1026-1036.