Thomas A. Fleisher and Raul C. Braylan
Flow cytometry is a technology used routinely in most hematology laboratories. Its entry into the mainstream of clinical laboratory analysis has been aided by the increasing availability of monoclonal antibodies that define cell surface and intracellular proteins as markers of cell lineage, differentiation, activation, and other biologic properties. Instrument design advances have yielded benchtop cytometers with fixed optics that, when linked with new developments in fluorochrome chemistry, enable a wide range of clinical applications. In addition, proficiency testing is now available in support of these clinical applications through the College of American Pathologists as mandated by the Clinical Laboratory Improvement Amendment of 1988. The major advantage that flow cytometry provides is its capacity to assess multiple measurements on large numbers of individual cells. Flow cytometric studies have extended the understanding of hematopoietic cell development, differentiation, activation, and apoptosis. In addition, they have provided important information regarding hematologic malignancies, insight into reconstitution after stem cell transplantation, and understanding of cell abnormalities that result in immune or hematologic deficiencies. As such, flow cytometry plays an important role in the diagnosis, characterization, and monitoring of a number of hematologic disorders.
The basic design of a flow cytometer involves four major elements: optics, fluidics, electronics, and a computer equipped with specialized software.1,2 The optical system utilizes one or more light sources, typically one or several lasers that produce monochromatic light and serve as the excitation beams. At the opposite side of the optical bench, light generated from the cells that have intersected the excitation beam is filtered, reflected by dichroic mirrors set in fixed locations, and finally collected by linked photodetectors to allow quantitation of the emitted light at specific wavelengths. To ensure that all cells analyzed experience consistent exposure to the excitation beam, the fluidic system must maintain the cells in a consistent location as they move sequentially through the beam. To accomplish this, the cell suspension is injected into a flowing stream of sheath fluid that hydrodynamically focuses the inner stream of cells within the outer sheath fluid stream.1 The intersection of the cells with the excitation light beam(s) produces characteristic light scatter (nonfluorescent) signals; additional fluorescent signals are generated by fluorochromes that typically are linked to specific reagents that bind antigens present on or within the cells of interest. The various light signals (parameters) are collected by the optical bench, while instrument design determines the number of parameters collected per cell. The two reagent-independent (nonfluorescent) parameters are forward-angle light scatter, as a marker of cell size, and side-angle light scatter, as an index of cellular regularity/ granularity. The combination of these two parameters allows for an approximate discrimination among the three major types of leukocytes as well as evaluation of red blood cells and platelets in whole blood samples.3
The fluorescent data collected by a flow cytometer are the result of either cell surface or intracellular binding of antibodies or other specific ligands conjugated directly to fluorochromes or detected with secondary reagents conjugated to fluorochromes, as well as reagents that are inherently fluorescent. Fluorochromes are excited by light of a defined wavelength and emit light of lower energy (longer wavelength). There are currently many different fluorochromes used in clinical flow cytometry, including fluorescein isothiocyanate, phycoerythrin, peridin chlorophyll protein, and allophycocyanin. More recently, combinations of two fluorochromes linked to each other have been developed; they depend on the transfer of energy from the first fluorochrome to excite the second fluorochrome. These tandem fluorochromes extend the range of emission wavelengths available from one excitation beam. The availability of multiple fluorochromes that absorb light of the same wavelength but emit light at different wavelengths means that multiple reagents can be used simultaneously with a single light source to yield a multicolor (polychromatic) study. One or more additional light sources are present in most current clinical instruments to extend the range of multicolor studies. Extended multicolor studies require complex color compensation and data management processes that typically involve sequential data evaluation.
The clinical application of flow cytometry in hematology saw its earliest use as a supplement to the morphologic classification of leukemias and lymphomas, affording not only lineage information but also the state of differentiation and/or maturation,4,5 cell growth and apoptosis.6 In addition, flow cytometry provided the best prognosticator in human immunodeficiency virus (HIV) infection based on absolute CD4 T cell numbers.7 More recently, flow cytometry has proven important in characterizing hematopoietic stem cells, detecting minimal residual neoplastic disease, defining immune deficiencies, identifying certain red blood cell related disorders, measuring the number of contaminating leukocytes in plasma and red blood cell transfusions, evaluating platelets and characterizing other blood cells.8-12 Flow cytometry also can be used to look inside the cell as well as to evaluate cell surface characteristics. Fixation and permeabilization facilitate intracellular entry of reagents to determine the presence of specific proteins and to assess functional characteristics.13 This chapter is directed at basic concepts of flow cytometry including data presentation and interpretation followed by a brief review of applications for hematologists.
DATA PRESENTATION AND INTERPRETATION
Current flow cytometers generally yield graphic displays of the cell frequency versus the light intensity for one or more parameters by means of specialized computer software. Figure 28.1, shows a single-parameter histogram that reflects the quantitative distribution of cells (y-axis) versus light signal strength (x-axis). Alternatively, the signal intensity of two correlated parameters can be plotted versus cell frequency, the latter displayed as dot density (dot plot, Fig. 28.2A) or a series of concentric lines (contour plot, Fig. 28.2B). When measuring multiple parameters (as colors), the data are typically evaluated by using dual-parameter displays. Typically, 10,000 to 20,000 events are collected to provide sufficient numbers of cells for meaningful data relative to the subpopulations of interest. However, when the cell or cells of interest are infrequent, such as evaluating hematopoietic stem cells (CD34+) in peripheral blood or detecting minimal residual disease (MRD) in leukemia, substantially larger total numbers of cells must be collected.8,14
Distinguishing a positive signal is usually based on defining background signals by evaluating either unstained cells (no monoclonal antibody added) or cells that have been incubated with a fluorochromeconjugated irrelevant antibody. By convention, the negative–positive discriminator is defined by the intensity of a signal that includes 99% (or 98%) of all cells based on one of the background conditions described above; cells that emit a signal above this discriminator are scored as positive for binding by the specific reagent(s) added to the cell suspension. This approach applies to well-defined populations with homogeneous antigen expression, but modifications or alternative interpretations may be necessary when the cell populations analyzed are heterogeneous or express dim fluorescence, which can lead to rather arbitrary calculations of percentages of positive cells.
The data generated by the computer are only as good as the instrument capabilities, settings, reagents, and cell preparation used. To prevent reporting invalid data, certain standards must be met.15 First, optimal instrument performance is integral and depends on a quality control program utilizing specialized software and methods. The use of validated reagents is also part of good laboratory practices, while the quality of the cell preparation can be assessed using the nonfluorescent parameters, forward- and side-angle light scatter, to confirm the presence of the population of interest. Each major blood cell type has distinctive features in this scatter plot. Platelets are obviously smaller than all other blood cells and heterogeneous in size, characteristics that can be confirmed in comparison to red blood cells. Erythrocytes have a characteristic appearance based on forward- and side-angle light scatter (Fig. 28.3A), which partially overlaps with that of lymphocytes. However, because of the large difference in frequency of circulating erythrocytes and leukocytes, there is no practical concern of lymphocytes contaminating erythrocytes (a collection of 10,000 erythrocytes normally would include less than 20 lymphocytes). In contrast, the presence of erythrocytes makes evaluation of lymphocytes virtually impossible. In part because of this reason, the study of lymphocytes in a whole blood sample or other specimens containing a significant amount of blood typically involves a red blood cell lysis step to eliminate erythrocytes (Fig. 28.3B). Following successful red cell lysis, a three-part differential is observed in peripheral blood with normal lymphocytes representing the smallest (forward-angle light scatter) and most regular/ agranular (side-angle scatter) cells, while granulocytes are slightly larger (higher forward-angle scatter) and show substantial granularity (high side-angle scatter), and monocytes fall between these two cell types (Fig. 28.3B). The two less prevalent granulocyte types differ in their location on a scatter plot, with eosinophils typically falling within the granulocyte population while basophils overlap with lymphocytes. Hematopoietic stem cells are normally found in the lymphocyte area of the scatter plot. It is important to recognize that the cell relationships noted above do not necessarily apply in cases of hematopoietic malignancy since neoplastic cells may exhibit altered light scatter properties or appear as a distinct population, separate from normal elements. Current standard practice for identifying the major leukocyte types, and particularly lymphocytes, includes using the pan-leukocyte monoclonal antibody CD45, alone or in combination with the monocyte-specific antibody CD14. CD45 is usually included in each staining combination as a specific lymphocyte identifier (characteristically bright staining) when red cell lysis is inadequate (Fig. 28.4) or substantial numbers of non-lymphocytes or debris contaminate the lymphocyte gate.16 Malignant cells also may differ from their normal counterparts in their staining characteristics with a variety of reagents, including CD45 expression (Figs. 28.5 and 28.6). The correlated analysis of side scatter and CD45 with or without appropriate sequential gating of subpopulations of interest is extremely helpful in recognizing hemopoietic and lymphoid neoplasia.17
FIGURE 28.1 Single-parameter histogram, a distribution plot of CD3 fluorescence intensity (x-axis) versus number of events/cells (y-axis) evaluating lymphocytes.
FIGURE 28.2 A. Dot plot of two-color (CD4 and CD8) staining evaluating lymphocytes. Frequency of events is reflected by the number of dots. B. Contour plot of two-color (CD4 and CD8) staining evaluating lymphocytes. Frequency of events is reflected by the contour levels.
FIGURE 28.3 (A) Dot plot of forward scatter (x-axis) versus side scatter (y-axis) on nonlysed whole blood sample (B) dot plot of forward scatter (x-axis) versus side scatter (y-axis) on lysed whole blood sample demonstrating a three-part leukocyte differential: lymphocytes, monocytes, and granulocytes.
FIGURE 28.4 Side scatter and CD45 analysis of normal peripheral blood demonstrates distinct clusters of granulocytes, monocytes, lymphocytes, and residual red cells.
The interpretation of fluorescent data based on antibody binding reflects the biology of the particular cognate cell surface protein. When the monoclonal reagent identifies exclusively one cell population, data interpretation is unambiguous, as for the pan-T-cell marker CD3 shown in Figure 28.1. In this example, when the evaluation is confined to lymphocytes using a lymphocyte gate, there clearly are two populations, CD3 negative cells, including B and natural killer (NK) lymphocytes, and CD3 positive T cells. In other situations, biologic variability in surface protein expression impacts data interpretation; examples are shown in Figures 28.7 and 28.8. In both histograms, there are at least three cell populations: cells negative for the marker, cells showing intermediate fluorescence, and cells that have bright fluorescence. In Figure 28.7, the CD8 intermediate cells are predominantly NK cells, while the bright staining cells are primarily CD8 positive T cells. In Figure 28.8A, the CD4 intermediate staining cells are monocytes, while the bright staining cells are T cells, with the former being present only in very small numbers with proper gating on lymphocytes (Fig. 28.8B). The finding of low-density CD4 expression on monocytes helps to explain HIV infection of this cell lineage.
Many monoclonal antibodies, individually or in combination, can serve to distinguish cells of a specific lineage (Table 28.1), and characteristic binding features can be used to direct a flow cytometry study to a specific cell population. As mentioned above, nonfluorescent parameters of forward-angle and side-angle scatter help distinguish among lymphocytes, monocytes, granulocytes, and platelets.3 Within the granulocyte population, neutrophils and eosinophils can be discriminated by the differential expression of the complement receptor CD16: neutrophils stain for CD16 while eosinophils are negative.11 Cells of the erythroid lineage can be identified based on the expression of glycophorin. Within the lymphocyte population, lineage-specific antibodies differentiate various populations and subpopulations. Hematopoietic stem cells can be identified by the expression of the cell surface protein, CD34, a valuable marker that has enabled evaluation and ex vivo isolation of bone marrow or mobilized circulating hematopoietic stem cells for transplantation.8
FIGURE 28.5 Analysis of side scatter and CD45 of peripheral blood from a patient with acute myeloid leukemia demonstrates charcteristic blasts (myeloblasts).
FIGURE 28.6 (A) Side scatter and CD45 analysis of peripheral blood from a patient with chronic lymphocytic leukemia shows increased lymphocytes. A lymphocyte gate (oval selection) is used for further analysis (B) that demonstrates that the majority of the lymphocytes abnormally coexpress CD19 and CD5, which is a typical feature of chronic lymphocytic leukemia (CLL).
FIGURE 28.7 CD8 histogram evaluating lymphocytes.
Many of the monoclonal reagents used to evaluate hematopoietic elements detect antigens that are not exclusively expressed on one specific cell type, and interpretation of data must incorporate knowledge of different surface protein expression patterns. A combination of additional antibodies often clarifies the relative expression of different antigens on specific cell populations. Cell surface proteins may be altered under different circumstances during the life cycle of a cell, including preferential expression early and/or late during differentiation, expression in response to cell activation and/or in various states of cell-specific function. Protein upregulation implies a range of expression that could include cells transforming from negative to clearly positive, depending on the temporal pattern of expression. For example, the α chain of the interleukin-2 receptor (CD25) shows such a pattern on T cells (Fig. 28.9), but the interpretation of CD25 expression as an activation marker has been complicated by the identification of T regulatory cells among CD25 expressing CD4+ T cells.18 When the interpretation of positive and negative is visually less clear, consistent interpretation criteria are crucial for valid comparison of data between different studies. In some circumstances, isoforms of a specific protein are differentially expressed, and cells may express one or the other isoform or both (Fig. 28.10).
As mentioned above, sometimes the use of percentage positive for a specific marker is misleading, as shown in Figure 28.11, where the histogram for the unstained cells overlaps significantly with that of stained cells. The histogram overlay demonstrates that there is a shift in the stained cells that would not be adequately reflected by simply scoring cells as positive or negative. Currently, many laboratories typically note the geometric mean channel (GMC) fluorescence of the unstained and stained cells and then report the cells to be positive for the specific marker with an increased fluorescence of x-fold over background (based on the quotient of the GMC-stained cells divided by the GMC of unstained or irrelevant antibody-treated cells). These considerations are particularly relevant for many markers used to evaluate malignant cells. In fact, consensus groups have repeatedly emphasized that reporting percentage values when interpreting results in hematopoietic malignancies is generally unsatisfactory.19,20 These values may not be sufficiently informative to allow the detection of neoplastic cells and cannot adequately describe the phenotype of the malignant cells. For this reason, it is recommended that interpretation of flow cytometric results in hematopoietic and lymphoid malignancies be based on the visual examination of the plots for each of the antibodies used, and that the results be primarily descriptive, in a manner similar to the microscopic evaluation of cells or tissues. Numerical values are only used to indicate the fraction of neoplastic cells or other well-defined cell populations present in the sample.
Flow cytometry has also been applied to investigate intracellular characteristics and, specifically, the presence of proteins which may only be detected intracellularly13 or as well as those that ultimately are also expressed on the cell surface. In addition, there are a series of reagents that bind to DNA and allow assessment of cell-cycle status.21 More recently, intracellular flow cytometry has been applied to measure some functional properties of cells, including the detection of intracellular cytokines following cell stimulation and cell-activation-specific processes such as calcium flux, pH changes, and phosphorylation of intracellular signaling proteins,13 but these applications currently have limited applications in routine clinical laboratory practice. Certain intracellular proteins that are not expressed on the cell surface serve diagnostic or prognostic purposes in malignant conditions and are often used clinically; these include terminal deoxynucleotidyl transferase,22 bcl-2,23-25 and ZAP-70.26,27
FIGURE 28.8 A. CD4 histogram evaluating mononuclear cells (lymphocytes and monocytes). B. CD4 histogram evaluating only lymphocytes.
APPLICATIONS OF FLOW CYTOMETRY IN HEMATOLOGY
Flow cytometry is of value in the evaluation of numerous hematologic conditions, but no other diseases have benefited more from the use of this technology than hematopoietic and lymphoid neoplasias. Flow cytometry has revolutionized the manner in which we diagnose, classify, and monitor acute leukemia or lymphoproliferative disorders, and it is rare now that patients with these disorders are treated without including the flow cytometric data. The technology is rapid, quantitative, and can analyze simultaneously multiple antigens in a large number of cells, allowing the easy detection, characterization, and enumeration of malignant cells, even when admixed with normal elements. The ability of flow cytometry to recognize malignancy is based on its capacity to distinguish differences in antigen expression between normal and neoplastic cells. Normal hematopoietic cells originate from a stem cell in the marrow that subsequently gives rise to a progeny of different cell lineages. These cell progenitors traverse various developmental stages and ultimately evolve into mature elements in the circulation and other peripheral organs. As the hematopoietic cells develop and differentiate, they undergo changes in their surface or intracellular antigenic profile that is characteristic of their lineage and differentiation stage. Hematopoietic malignancies are clonal cell populations that express similar antigens to those of their non-neoplastic counterparts but usually with a different expression pattern that is unique for each type of neoplasia. This antigen expression can be increased, decreased, absent, asynchronous, or may be of a different cell lineage. Thus, knowledge of the immunophenotype of a cell together with its physical properties revealed by light scatter signals allows determination of not only their lineage and developmental stage but also, in most instances, their normal or neoplastic nature. Furthermore, in T- or B-cell lymphoproliferative disorders, the clonal nature of the lymphocytes can be established by recognizing the restriction of expression of immunoglobulin light chains28 or T-cell receptor beta chains.29 The identification of monoclonal lymphoid expansions simultaneously with other informative antigens has proven extremely useful in the differential diagnosis between benign and malignant lymphoid disorders30 not only in marrow and blood31 but also in lymph nodes32 and extranodal lymphoid sites.33 In conjunction with cytologic examination, this technology is especially helpful in samples obtained from fine-needle aspirates.34-36
Table 28.1 Commonly Used Leukocyte Antigens Used in Clinical Flow Cytometry Based on Cluster of Differentiation Designation
CD1a: Cortical thymocytes, dendritic cells, langerhans cells
CD2: T cells, thymocytes, NK-cell subset
CD3: T cells, thymocytes
CD4: T-cell subset, thymocyte subset, monocytes/macrophages
CD5: T cells, B-cell subset
CD7: Thymocytes,T cells, NK cells, early myeloid cells
CD8: T-cell subset, thymocyte subset, NK-cell subset
CD10: Early B cell, neutrophils, bone marrow stromal cells
CD11b: Monocytes, granulocytes, NK cells
CD11c: Myeloid cells, monocytes
CD13: Myelomonocytic cells
CD14: Monocytes, myelomonocytic cells
CD15: Granulocytes, monocytes, endothelial cells
CD16: NK cells, granulocytes, macrophages
CD19: B cells (from pre-B-cell stage to plasma cells)
CD20: Mature B cells
CD21: Mature B cells, follicular dendritic cells
CD22: Mature B cells
CD23: Activated B cells
CD25: Activated T cells, activated B cells, regulatory T cells
CD27: Memory B cells
CD30: Activated T, B, NK cells, monocytes, Reed Sternberg cells
CD33: Myeloid cells, myeloid progenitor cells, monocytes
CD34: Hematopoietic precursor cells, capillary endothelium
CD38: Most thymocytes, activated T cells, B-cell precursors, germinal center B cells, plasma cells, myeloid cells, monocytes, NK cells
CD36: Platelets, monocytes/macrophages
CD41: Megakaryocytes, platelets
CD42b: Megakaryocytes, platelets
CD45: Leukocytes
CD45RA:T-cell (naïve) subsets, B cells, monocytes
CD45RO:T-cell (memory) subsets, B-cell subsets, monocytes/macrophages
CD56: NK cells, NK T cells
CD57: NK cells, T-cell subsets, B cells, monocytes
CD61: Megakaryocyte platelets, megakaryocytes, macrophages
CD64: Mature neutrophils, monocytes
CD71: Erythroid precursors, proliferative cells
CD79a: B cells
CD95 (Fas): Lymphocytes (upregulated after activation), monocytes, neutrophils
CD103: Intestinal epithelial T lymphocytes
CD117: Myeloid blast cells, mast cells
CD138: Epithelial cells, plasma cells
CD, cluster of differentiation; NK, natural killer.
FIGURE 28.9 Single-parameter histogram of CD25 expression on lymphocytes (left panel) and contour plot of CD25 (y-axis) and CD3 (x-axis) expression.
The utilization of flow cytometry has recently extended to the assessment of myelodysplasia37 and has also been found to be useful in plasma cell disorders.38 In the latter, this analysis has relevance in the differential diagnosis between myeloma and other plasma cell disorders, and also provides prognostic utility based on the identification of high-risk monoclonal gammopathy of undetermined significance and smoldering myeloma.39
FIGURE 28.10 Contour plot of CD45RA and CD45RO expression on CD4+ T cells.
FIGURE 28.11 Overlapping control and positively stained histograms.
Flow cytometry is a useful tool for the detection of MRD. Broader capabilities provided by technical advances, a larger selection of antibodies and fluorochromes, and new analytical approaches have resulted in an increased sensitivity in the identification of malignant cells. The presence of MRD has prognostic value for predicting acute leukemia relapse prior to or following curative therapy attempts such as allogeneic hematopoietic stem cell transplants,40-44 and is now a critical component of many therapeutic protocols. While genetic abnormalities in acute leukemic cells can be recognized by molecular genetic techniques, flow cytometry is an excellent alternative in cases where genetic markers are absent. In addition, in an era of curative attempts, MRD detection by flow cytometry has become a useful tool in chronic lymphoid leukemias.31
Many applications in nonmalignant diseases are also routinely performed in the clinical laboratory. This technology remains a critical tool in monitoring disease progression and therapy in HIV infection.7There are specific lymphocyte findings characteristic of primary immunodeficiencies, including loss of cell populations or subpopulations, the absence of specific cell surface or intracellular proteins, and changes in normal immunologic processes that can be detected by flow cytometry (e.g., the development of memory T cells and/or B cells).13 Assessment of specific lymphocyte populations and subpopulations is being studied in a variety of disorders characterized by inflammation, with particular attention to expression of activation markers. Reconstitution of the immune system following intensive chemotherapy and hematopoietic stem cell transplantation can be monitored by flow cytometry, an approach that is even more significant due to the recent focus on immunotherapy and vaccines in experimental treatment protocols for malignancies.
Assessment of leukocytes other than lymphocytes is emerging in the clinical laboratory. Monocytes can be evaluated to define deficiencies associated with defective monocyte surface receptor expression.13Granulocyte expression of critical adhesion molecules and their capacity to generate reactive oxygen species can be assessed by flow cytometry.13 In addition, granulocyte-specific autoantibodies can be detected. Flow cytometric methods are being used to characterize eosinophils in settings of increased production such as the hypereosinophilia syndrome,16 and basophils have been studied for intracellular cytokine production as well as activation of antigen expression ex vivofollowing specific antigen exposure.
Hematopoietic stem cell identification is dependent on flow cytometry, which is routinely used to characterize and quantitate stem cells in the transplantation setting, generally based on CD34 expression together with other cell surface markers.8 Separation techniques to purify stem cells from either bone marrow or peripheral blood typically utilize CD34 selection methods, and patients are followed by flow cytometric testing post transplantation in order to assess donor cell engraftment and in certain settings, donor–host cell chimerism.
The evaluation of erythrocytes by flow cytometry has been applied to cell surface proteins, autoantibodies in hemolytic anemia, and the detection of F-cells in fetomaternal hemorrhage and sickle cell anemia.9,45 The detection of glycosylphosphatidylinositol-anchored proteins on erythrocytes and leukocytes by flow cytometry is now the favored method to accurately diagnose paroxysmal nocturnal hemoglobinuria.46
Flow cytometric evaluation of platelets has been described as a method to study these cells in whole blood, thus eliminating the need for platelet isolation and minimizing cell manipulation.10 This approach allows for the detection of platelet-associated immunoglobulin, assessment for states of platelet activation and aggregation, and detection of reticulated platelets. Flow cytometry is also a rapid and useful method to detect platelet defects in structural or functional glycoproteins, such as the abnormal expression of gpIIb/IIIa in Glanzmann thrombasthenia and gpIb in Bernard-Soulier disease.10
SUMMARY
Flow cytometry has evolved as integral in the laboratory assessment of many hematologic disorders. This technology provides a powerful tool to assess simultaneously cell surface and intracellular characteristics. The increasing range of reagents and expanded understanding of cell biology mean that flow cytometry will play an even larger role in the clinical evaluation of various cellular components of the hematologic system, both in benign and malignant conditions.
References