Clinical flow cytometry, a hypothesis‐driven discipline of modern cytomics

Recently, two major books have been published that summarize the historical aspects and recent achievements of practical flow cytometry (1, 2). Both emphasize the role played by this newly developed technical discipline in the development of scientific (1) and diagnostic platforms during late 20th-century medicine (2). Indeed, the gray box called a flow cytometer is the result of a multidisciplinary collaboration between engineers, biophysicists, biochemists, histopathologists, molecular cytologists, hematologists, immunologists, and quality controllers, with a more recent contribution from physicians specializing in the human immunodeficiency virus (HIV), oncologists, and epidemiologists (Table 1). The foresight by the “fathers” has been astonishing (1). In his book, Howard Shapiro reminds us that flow cytometry, used to investigate cells in a flow system instead of on a static microscope, was put into practice by Coulter when he used impedance to count red cells. Rapid cell spectroscopy was introduced by Kamentsky to leukocyte differential counting with computer-assisted displays. Then optical cell counters, fluorescence dies, lasers, and photomultiplier tubes were added by Fulwyler and Jones. Immunologic concepts and reagents were introduced by the Herzenbergs and were fully primed to harvest the gems of the incipient monoclonal antibody revolution. The first commercial cytometer with precision engineering of the flow cells was launched by Goehde. These initial steps were duly followed by the release of a series of cytometers that showed increasing sensitivity, computerization, and practicality in research and routine laboratory work. This is illustrated by science historians Keating and Cambrosio in their book that blends the sociological aspects of medical technological developments with witness accounts (2). Flow cytometry currently is a colorful, practical discipline that has become available to medicine at the appropriate time for various pressing clinical applications, thus rewarding the scientists who had the vision to foresee these needs. These contributions established cytometry, including flow technology, image analysis, and advanced microscopy, among the leading trends of modern biomedicine and health care. After the launch of the Human Genome Project (3), two areas, genomics and proteomics, have been majestically promoted on both sides of the Atlantic and in the East (4). Genomics includes the identification of genes and gene regulatory processes, and proteomics investigates the abundance of proteins simultaneously with the changes associated with alterations of the functional state of the cell. Such a “pseudo-functional” approach aims to extend the study of quantitative changes during differentiation, proliferation, and signaling of different cell types (5). Clearly, there is an enormous, newly generated influx of information here, but it is not certain that a mere analysis of genes and protein structure, even in its extended format that includes the interaction of various biomolecules, will provide all of the necessary information to understand function and regulation at the level of living cells and organisms. Hence, the concept of cytomics has been introduced recently (6, 7), for two reasons. First, cytomics is the cell-oriented analysis of molecules and their functions in cellular systems and/or organs, referred to as cytomes. Second, cytomics is the new, absolutely essential, interface between biosciences and clinical medicine. Importantly, advanced cytometry, including the traits defined in Table 1, represent the driving engine of cytomics. A new paradigm is that cytomics is the combined, interrelated force for cytometry, proteomics, and genomics (6, 7). There are, however, significant differences between these three modern trends. First, it is in the realm of cytometry, with the major contribution of flow systems, where the appropriate functional analysis of cells, the display and regulation of functional molecules, and the differentiation pathways are successfully carried out. Animal models and human diseases provide targets for these studies. Second, it is the cytometric field where the diagnostic tests, for malignancies (cf. 8,9) and the cellular analysis of infectious diseases (10), currently harvest the richest crop of practical results, whereas in the fields of genomics and proteomics only the seeds are sown for the practically useful multiarray systems. Third, these two fields at their current developmental state, are essentially observational sciences that primarily document the observed heterogeneity. As a marked contrast, the research by flow cytometry has been, from the beginning, primarily hypothesis driven and backed up by sophisticated experimental designs and clearly defined clinical needs. In this review, some of the basic hypotheses that have contributed to the introduction of clinical applications are summarized to document the hypothesis-driven nature of flow cytometry research. Six hypotheses are described that have been conceived with the flow cytometric technology in mind to answer biological and/or clinical diagnostic questions. In each case, a short account of the background is followed by the relevant hypothesis. Then the implications of the results are summarized, with additional observations that substantiate the results observed and explain the significance of the findings. A fluorescence-activated cell sorter (FACS) from Becton-Dickinson, the first to arrive to Europe in 1974, was operated by David Capellaro at the ICRF Tumour Immunology Unit, London (Fig. 1). This sorter was used for leukemia and stem cell research by Greaves and colleagues. Around that time it was already known that the membrane markers for murine and human T and B cells were also present on these cells' malignant counterparts (11, 12). As the common form of acute lymphoid leukemia (ALL) remained negative (non-T, non-B ALL), Greaves and Brown made a rabbit anti–non-T, non-B ALL antiserum (13); in this laboratory other antisera to T cells, myeloid cells, and class II antigens also were used for immunodiagnosis (Fig. 2). The fluorescence activated cell sorter (FACS) at University College London, 1973 to 1974, operated by David Capellaro. The reactivity of heterologous anti-ALL, anti-T, and anti–class II antisera with the common form of ALL (cALL), thymic ALL (Thy-ALL), and a “lymphoid” blast crisis of chronic myeloid leukemia (CML-bc). The results of the absorption of the antiserum made against ALL with Thy-ALL and CML-bc show that the antigen recognized by this diagnostic reagent is shared between them (14). This was the common ALL antigen, with CD10 of 100 k (15, 16). Hypothesis 1A: A non-T, non-B ALL antiserum will react with (a small number of) normal precursor cells, the “target cells,” from which this particular form of malignancy derives (14). The anti-ALL antiserum was shown to react with an antigen of 100 kD (CALLA; later proven to be CD10) (2, 16), strongly expressed on most common ALLs (Fig. 2-1), and weakly expressed on some T-ALL (Fig. 2-7), but undetectable on acute myeloid leukemias and chronic lymphoid leukemia. As predicted, this was the first anti-precursor cell reagent; the reactivity is shown in Table 2. Once this cell type, also referred to as the pro-B cell, and its close relative, the cytoplasmic immunoglobulin (Ig) positive pre-B cells were identified, these cell types were characterized extensively by flow cytometry, cell sorting, biochemistry, and gene rearrangements (2, 14) (also see below). Hypothesis 1B: The same early lymphoid precursor also can be involved in malignancies that may develop as a result of different pathologic mechanism(s), e.g., in Philadelphia chromosome positive (Ph′) chronic myeloid leukemia (CML) of pluripotential hemopoietic stem cell origin (13, 17, 18). Anti-ALL serum reacted with the “lymphoid” blasts in CML (Fig. 2-13) (14, 17). When the anti–non-T, non-B ALL antiserum was cross-absorbed with Ph′ lymphoid blasts (Fig. 2-4) or with T-ALL blasts (Fig. 2-5), the reactivity to common ALL was removed and vice versa (Figs. 2-10 and 2-17). Thus, the antiserum detected the same precursor cell-related moieties (the CD10 antigen) (15, 16) in the different diseases. In Ph′-positive disease that the lymphoid blasts were frequently seen admixed with other cell types such as maturing myeloid cells and myeloblasts (17), cell sorting with FACS showed that only the lymphoid blasts had the CALLA-positive phenotype (Fig. 3) (18, 19). Similar cell sorting studies (20) have shown that class II antigens are expressed more widely on lymphoid and myeloid blasts but not on differentiating forms (Fig. 4). These investigations have established the immunophenotyping platform for leukemias (2). Heterogeneity of cell phenotypes in CML blast crisis: analysis using the FACS. Cells are labeled with anti-cALL serum and separated according to fluorescence as indicated in the lower oscilloscope screen picture. Cells of lymphoid morphology sort into the positive fraction and the nonlymphoid cells in to the negative (18, 19). Heterogeneity of cell types in CML blast crisis. A mixed myeloid and lymphoid blast crisis was stained with an anti–class II antiserum and sorted as in Figure 3. a: The negative fraction contains differentiating granulocytic and erythroid cells. b: In the class II-positive fraction, lymphoid and myeloblasts are seen (20). Hypothesis 2: It is suggested that viable monoclonal hybrids carry surface Ig with antibody specificity and that, after having been labeled with antigen-coated fluorescing microspheres, the progenitors can be sorted into specific antibody secreting fractions to facilitate the selection of “wanted” clones (22). In this study, viable and dead cells could be distinguished by flow analysis. Rare viable hybrids with antigen binding were recognized and enriched by sorting. The antigen-coated fluorospheres adhered to the hybrid cells that expressed antigen-specific receptors, providing a signal bright enough to do the sorting. At the time, only fluorescent microspheres provided a signal bright enough for this purpose. Subsequently, fluorochromes of higher intensity proteins also could be labeled with fluorochromes and used for positively selecting hybridomas of known specificity (23). In culture, the enriched progenitors developed into clones that secreted specific antibodies against the antigens (mouse Ig-1a; Table 3). The beauty of this experiment is that it combined basic science with a practical purpose: to prepare better reagents and to use these in flow cytometry. The successive Leucocyte Typing Workshops marshaled monoclonal antibodies into CD groups (16, 24), and multiparameter analysis became feasible by the tandem use of monoclonal antibodies labeled with different fluorochromes (25). Mike Loken and Leon Terstappen, in collaboration with Curt Civin, had an organized approach to characterizing the expression of these CD groups. The description of erythroid (26) and the B-lymphoid lineage (Fig. 5) (27, 28) was soon followed by the characterization of the neutrophil lineage (29), and monocytic differentiation from CD34+ cells in the bone marrow and cord blood was described by Knapp et al. (30, 31). Stages of B-lymphoid differentiation in the normal bone marrow, investigated with multiparameter analysis using combined staining with monoclonal antibodies labeled with different fluorochromes. The reagents used were CD34, CD10, anti–terminal deoxynucleotidyl transferase, anti–class II (HLA-DR, -DP, and -DQ), CD19, CD20, CD21, and IgM (for membrane labeling; SmIgM). B-lymphoid precursors at phase II contain cytoplasmic IgM (not shown); see Greaves and Janossy (14). From Loken et al. (27, 28). Hypothesis 3: Flow cytometry, with its multiparameter analytical approach, is ideally suited to indicate the steps of differentiation along a given cell lineage in situ (e.g., in the bone marrow) (27, 28) and to prove this scheme by following the development of lineage precursors in vitro (30-32). The example of B-lymphoid development from early precursors, through CALLA (CD10)–positive pro-B cells to Sigma+ B lymphocytes (Fig. 5) is a model for the highly controlled sequential acquisition of cell surface antigens during B-lymphocyte development. The most immature cells identifiable in the bone marrow express CD34 and HLA-DR. The earliest recognizable B-lineage cells (CD19+, bright CD10+) also express CD34 and terminal deoxynucleotidyl transferase in the nucleus and are proliferating. The progression of cells from stage II to stage III is marked by the acquisition of CD20, HLA-DQ, and SmIgM. A caveat is that SmIgD+ B cells (last stage) may not be an integral or sequential part of this linear progression but perhaps represent cells that seed back to bone marrow from the periphery (33). The cultivation of CD34+ precursor cells in vitro has confirmed these B-cell and myeloid differentiation schemes (30, 31). Hypothesis 4: Differentiating normal cells express their markers in an orderly fashion when studied with three-color (35) or four-color (36, 37) antibody combinations, whereas leukemia cells frequently show aberrant dysregulation that is their characteristic feature. In patients retested at the end of remission induction in full morphologic remission with sensitive techniques to identify MRD, the presence of aberrant cells may have a profound prognostic significance (38). The regularity of marker expression in normal bone marrow for CD10/CD20/CD19 (Fig. 6a) and CD34/CD38/CD19 combinations (Fig. 6b) is depicted. The three populations of normal precursors cell types such as the early (CD34+, CD10+) precursors, intermediate forms and the newly emerging (CD10−, CD20+) B cells are shown as dots (with colours in ref. 35). In the white unallocated areas, aberrant leukemic phenotypes reside, representing 59.7% and 53.9% of ALL cases; the frequency of ALL cases that exhibit the individual aberrant combinations is also shown (35). Identification of leukemic blast cells of the B-lymphoid lineage by irregular aberrant marker expression. Normal B-cell precursors provide a regular differentiation pathway that can be investigated by different combinations such as CD10/CD20/CD19 (a) or CD34/CD38/CD19 (b). Leukemic cells frequently fall outside these normal boundaries and occupy different positions in the “white area.” In a and b, 53% to 59% of ALL cases can be regarded as aberrant, but other combinations exist, and most cases of ALL can be followed with one or another combination staining. The percentage values are the proportions of ALL cases that are positioned as aberrant. For example, 10.4% of cases show much higher CD10 display on ALL blasts than on normal CD20+ B-cell forms (35). These tests for MRD are regarded to be useful if they can identify fewer then one leukemic cells per 104 normal cells. The three- and four-color combinations used for T- and B-lineage ALL and for acute myeloid leukemia at the St. Jude Children's Hospital, Memphis, Tennessee, USA and at the University Hospital, Salamanca, Spain are published in detail (37). In clinical practice, the children are studied at the end of remission induction by conventional morphologic methods. If no abnormality is seen, the bone marrow is retested by flow cytometry for MRD because this technique is more sensitive to detect residual leukemia. The prognostic significance of these diagnostic results has been proven in a large cohort by observing the relapses that developed during subsequent years. The patients were grouped into four categories: (a) non-detectable MRD, (b) MRD detectable but less than 0.1%, (c) MRD between 0.1% and 1%, and (d) MRD greater than 1%. Five years later, the relapse rates were 10%, 23%, 43%, and 72%, indicating the clinical use of the assay (36, 38). These tests for MRD are complementary to molecular techniques. Both of these serve as surrogate endpoints in the clinical testing of the new therapeutic approaches. The studies described above already gave the impression that the marker expression on normal leucocytes was regular and that the flow cytometers provided reliable quantitative results. These premises can be studied further by methods that quantitate antibody-binding capacity (ABC) expressed as molecules per cell (39). Of these, the indirect immunofluorescence tests, such as the Quantitative Indirect ImmunoFluorescence (QIFI) test (40), is the most reliable because the first antibody used to bind to the membrane antigens at saturating concentrations remains unconjugated and therefore unmodified; the fluorochrome-labeled second antibody, used at saturating level, is then a universal reagent used in every assay. Thus, the results observed are comparable and allow the construction of “league tables” for the expression of different membrane antigens on different cell types (39, 40). For example, the CD45 antigens on lymphocytes are expressed at four times higher levels (at 200 × 103 molecules/cell) than CD4 antigens on “helper-type” T cells (50 × 103 molecules/cell) and 10 times higher than CD19 antigens on B lymphocytes (20 × 103 molecules/cell; Fig. 7a). Expressions of CD38, CD7, CD19, CD4, and CD45 antigens on lymphocytes from different donors (a) and the preservation of these antigens on stabilized cell preparations prepared from fresh cord blood cells (b). The ABC was determined by using the indirect quantitative immunofluorescence method referred to as the QIFI test (40). The expressions of CD4 and CD45 molecules is particularly regular and robust, so these moieties can be used as controls for quantitative flow cytometry (41). Once the ABC values are known, these help to interpret the staining of normal lymphocytes (c) and aberrant populations (d) in a quantitative manner. Hypothesis 5: When tested by the QIFI test, some markers will prove to be regularly maintained at their characteristic known levels of expression in different people, to be used as acceptable biological controls; further, these molecules might be stabilized by fixation to provide cellular standards of long shelf-life (41). The ABC values for the expression of various CD antigens on lympho-, mono-, and granulocytes have been documented using the QIFI assay (40). The hierarchy of expression is known and the individual personal variations for CD45 and CD4 antigen expression are less than 10% (Fig. 7a). These antigens, also tested on cord blood cells for CD45, CD4, and CD38, are maintained at a stable high level on blood preparations stabilized for long-term use (Fig. 7b). These preparations of long shelf-life (42) serve as standards in the stabilized cellular immunofluorescence assay (41). With the help of these standards during flow cytometry, the values of mean fluorescence intensity can be replaced by values of ABC (molecules per cell). In normal blood, CD4+ T cells uniformly express 50 × 103 ABCs/cell (Fig. 7c). CD8+ cells are heterogeneous: “proper” CD3+,CD8++ T cells exhibit 130 × 103 molecules, whereas the CD3−,CD8+ natural killer cells display 10 to 100 × 103 CD8 ABCs/cell (Fig. 7c). This picture is complicated by the existence of populations that can be or in cases can aberrant T cells be seen in higher than In such a (Fig. the cells are CD8+ T cells with (20 × CD4 display chronic T chronic lymphoid These quantitative on normal blood standards and on stabilized blood have a in characterizing the aberrant display of various antigens in leukemia and The of CD4+ is the in the of The level of these cells in the blood is therefore the most for disease associated with the first on immunodeficiency flow cytometry has remained the for CD4 in Recently, however, this has been into The introduction of in has the for but the of CD4 tests has been to as high as to For these CD4 methods have been to the standards relevant to be First, physicians in may a full of analysis and natural killer whereas in with a large might only CD4 Second, the of recently has been established when with the on morphologic a of antibody reactivity be used when CD4+ lymphocytes are defined by (Fig. The for this to a more is that markers such as CD4 and CD45 are maintained in at the time when It is to count CD4+ T lymphocytes using a CD4 antibody labeled with a of fluorochromes such as or using CD4 intensity on the and on the is the small lymphocytes are from the CD4+ The tests using a CD4 antibody provide CD4 when with other that use more reagents such as the method on CD4 after the are not as reliable as flow cytometry because it is as documented by the CD4 values with the flow cytometric CD4 The CD4 cases show high CD4 This CD4 might from activated and/or in patients with infectious disease such as Hypothesis flow cytometry tests for CD4 can be and by using a CD4 with the current these will be more then the more methods Hypothesis The of CD4 counting by flow is that this quantitative method between lymphocytes and CD4+ methods are to this additional for and to use when CD4+ is associated with such as and diseases. CD4 staining on the flow cytometer between T cells of 50 × 103 CD4+ of × 103 of the fluorochromes used (Fig. The assay method such as however, CD4 antigen in blood after cells have been normal when most CD4 antigens from CD4+ the and CD4 antigens are into the background during the construction of the in patients with and this and with CD4+ T cells might be referred to as T are detectable that do not to when tested by flow cytometry (Fig. This is on in diseases to at CD4 above the but these might have been by CD4 molecules It is therefore to the levels of from the currently CD4 when these are on the of a and CD4 counting method the for patients during the of new information on genes and the of genomics and proteomics are to be in an to the functional of cells, cellular systems, and It is therefore an development that the discipline of with its technical driving force by flow cytometry, image analysis, and modern microscopy, has been recently to to a new phase of scientific development (6, 7). is an approach cells and systems using the available information in a functional The of this approach is by the that the using flow cytometry as the technology, have been, from their early of development to the present hypothesis (Fig. The concept of cytomics includes the three of proteomics, and cytometry to cellular and functions and to biosciences with their medical is a hypothesis-driven as illustrated in this In this review, the have been primarily to document two aspects of the hypotheses (a) their to the science to diagnostic medicine to provide for the that cytomics a new interface between the biosciences and clinical and (b) to provide practical in to scientific In the field by flow cytometry is than suggested The of science that is by the of Leucocyte Typing Workshops by cellular antigens, through the monoclonal antibody technology and their CD the achievements of the proteomics discipline and is more to to its essentially functional The of the monoclonal antibody on flow cytometry vice is It is to see these two of science CD and can in this review, the modern of flow cytometry to develop of analysis by lasers, optical and of antibodies with new of fluorochromes have not been these investigations are to sort the of in such as and in and human such as the of the newly regulatory CD4+ T cells the emerging new platforms show the of the diagnostic systems with flow cytometric analysis of cell including CD4+ This is the concept which is more than its This development is a example to two (a) the current highly multiarray systems, with contributions from will soon clinical in a and (b) flow cytometry is an part of the current technical in the of multiarray to Greaves London, Jude Children's Hospital, Memphis, of the of for Howard Shapiro David Hospital, and Hospital, for also to of for to his (Fig.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.