Anticoagulants, and Five Cell Preparations'

Transcription

1 Published 1992 by Wiley-Liss, Inc. Cytometry 13:68-74 (1992) Flow Cytometric Analysis of Whole Blood Lysis, Three Anticoagulants, and Five Cell Preparations' Patricia H. Carter, Sandra Resto-Ruiz, Glennelle C. Washington, Steve Ethridge, Alessio Palini, Robert Vogt, Myron Waxdal, Thomas Fleisher, Philip D. Noguchi, and Gerald. E. Marti Laboratory of Cellular and Molecular Biology, Division Biochemistry and Biophysics, Center for Biologics Research and Evaluation, Food and Drug Administration, National Institutes of Health, Bethesda, Maryland 2892 (P.H.C., S.R.-R., G.C.W., P.D.N., G.E.M.); Division of Environmental Health Laboratory Sciences, Centers for Disease Control, Altanta, Georgia 3333 (S.E., R.V.); FAST Systems, Inc., Gaithersburg, Maryland 2877 (A.P., M.W.); Clinical Immunology Service, CPD, CC, NIH, Bethesda, Maryland 2893 (T.F.) Received for publication May 6, 1991; accepted August 1, 1991 We studied the effects of anticoagulants and cell preparation methods on lymphocyte forward-angle scatter (FSC), autofluorescence, and immunofluorescent staining for CD45, CD14, and CD13. Blood samples collected in ethylenediaminetetracetic acid (EDTA), heparin, and acid citrate dextrose (ACD) were processed by using conventional Hypaque-Ficoll (HF) separation and four whole blood (WB) lysis techniques: Immuno-lyse, Q-Prep, FACS Lyse, and Gen Trak Lysis. Lymphocytes prepared by using three of the four whole blood methods gave FCS values comparable to those isolated by HF, while one method (FACS Lyse) gave consistently lower values. Autofluorescence values were comparable by all methods except Immuno-lyse, which showed consistently higher values in blood stored for 24 h with any antico- agulant. Immunofluorescent values for CD45-stained cells were quite consistent across all methods, and among the whole blood methods, FACS Lyse and Q-Prep uniformly gave the highest purity of CD45-positive cells in the lymphocyte light scatter gates. Additionally, propidium iodide (PI) analyses of CD45-stained whole blood, and analyzed without lysis, confirmed that ACD and heparin were superior to EDTA for maintaining viable leucocytes overnight. Future studies should focus on other commonly used reagents, a wide variety of abnormal samples, and cell viability. Key terms: E thylenediaminetetracetic acid (EDTA), heparin, ACD, acid citrate dextrose, immunophenotyping, cell viability Whole blood (WB) lysis methods for flow cytometric immunophenotyping have essentially replaced Hypaque-Ficoll (HF) density gradient separation of peripheral blood mononuclear cells (PBMC) in clinical laboratories. Reports to date have focused primarily on hypotonic lysis compared with HF, while side-by-side comparisons of commercially available WB lysing methods have either not been available or have not been included in these studies (4,6). Although the results can be variable, the recent use of absolute CD4 counts in the management of AIDS patients has become common. We plan to immunophenotype thirdand fourth-generation family members in kindreds with familial cancer and wish to use WB lysis for a large number of individuals. Therefore, we have evaluated the effects of anticoagulants and cell preparation methods on the flow cytometric characterization and recovery of lymphocytes. This has entailed a side-byside comparison of four commercially available WB lysis methods together with conventionally prepared HFseparated PBMC. We collected sufficient blood from single normal blood bank donors by using three differ- 'This report was originally presented at an NCCLS Subcommittee meeting on Flow Cytometry, August 6-7,199, Philadelphia, PA, and in abstract form at the Fifth Annual Clinical Applications of Cytometry Meeting at Charleston, SC, September 199. Use of trade names is for identification only and does not constitute endorsement by the Food and Drug Administration, Public Health Service, or the U.S. Department of Health and Human Services.

2 WHOLE BLOOD LSIS 69 IMM- FACS a HIFGEN 4 IMM- FACS Q HIF GEN Ann. I c- > p" 2 ro a B : IMM- FACS a H/F GEN WSE LSE PREP TRAK IMM- FACS a HIF GEN FIG. 1. Summary of percentage lymphocytes (upper panels) and mean channel values (lower panels) for cells analyzed immediately (d, left panels) and the following day (d 1, right panels) after overnight storage at room temperature. For each plot, the lysing sys- terns are: 1 = Immuno-lyse; 2 = FACS Lyse; 3 = Q-Prep; 4 = Hypaque- Ficoll PBMC; 5 = Gen Trak. The three anticoagulants are from left to right for each lysis method EDTA (solid black bar), heparin (middleslashed bar), and ACD (solid gray bar). ent anticoagulants in order to permit the preparation of five different samples for each anticoagulant. Additional experiments were directed at assessing cell viability and quantitation of the fluorescence intensity (FI) of both autofluorescence and CD45 expression. EXPERIMENTAL PROCEDURES Blood Samples Whole blood specimens from normal blood bank donors (Dept. Transfusion Medicine, Clinical Center, National Institutes of Health, NIH) were collected by using EDTA, heparin, and acid citrate dextrose (ACD) as anticoagulants (Vacutainer tubes, Becton-Dickinson, San Jose, CA). Three to six tubes were drawn on individual donors, and a portion of the blood was processed immediately. The remainder was processed after overnight storage in a laminar flow hood at room temperature (RT). Blood Preparation HF samples were prepared after diluting a 3 to 5 ml aliquot of blood 1:2 with phosphate-buffered saline (PBS). We prepared WB lysis samples with Immunolyse and Q-Prep (Coulter Immunology, Hialeah, FL), FACS Lyse (BDIS, San Jose, CAI, and Gen Trak Lysis (Gen Trak, Inc., Plymouth, PA) by following the manufacturers' recommended procedures. Of these methods only the Q-Prep is an automated procedure that incor- porates three reagents: (A) formic acid (lysing agent); (B) sodium carbonate, sodium chloride and sodium sulfate (stabilizers); (C) paraformaldehyde (fixative). These reagents are added sequentially in the order listed by the machine (Q-PREP) over a 35 s time period. The other procedures are all manual and timed. The final preparations were washed twice with PBS and were resuspended in 1 ml of 1% paraformaldehyde- PBS. Some preparations were left unstained to measure autofluorescence, and others were stained with a mixture of CD45-FITC and CD14-phycoerytherin (PE) (Leucogate [BDIS], San Jose, CAI or Leucogate plus CD13-PE (My 7 RD1 [Coulter Immunology, Hialeah, FLI). Flow Cytometry A FACScan flow cytometer calibrated with Autocomp (BDIS) and QuickCal (Flow Cytometry Standards, Incorp, Research Triangle, NC) was used. We acquired and analyzed data using FACScan Research Software (BDIS). Forward-angle scatter (FCS), side scatter (SSC), log fluorescence green (FLl), and log fluorescence red (FL2) were collected for 1,- 2, events in list mode files. In some experiments, both linear and log SSC parameters were collected. Preparations were evaluated for lymphocyte recovery, light scatter properties, the percentage of positive and fluorescence intensity of unstained cells, and + CD45 CD14- lymphocytes.

3 7 CARTER ET AL. U 6 > 8 If u z i i 2 2 IMM FACS Q H/F GEN LSE LSEPREP TRAK r IMMFACS HIF GEN LSELSE PREP TRAK I I r 4 > 4 s U U o y IMM FACS H/F GEN IMM FACS 4 H/F 5 GEN r i IMM FACS Q H/F GEN IMM FACS Q H/F GEN FIG. 2. Summary of mean channel (MC) fluorescence intensity data for unstained, cell autofluorescence, and CD45+CD14 cells analyzed immediately (left panels) and the following day after overnight storage at room temperature (right panels), using EDTA, heparin, and ACD. The solid black bar in each plot is the MC for unstained cells, while the other bars are for CD45'CD14- cells stained with either Leucogate (middle slashed bar) or Leucogate plus My 7 RD1 (solid gray bar). The lysing systems are: 1 = Immuno-; 2 = FACS Lyse; 3 = Q- Prep; 4 = Hypaque-Ficoll PBMC; 5 = Gen Trak. The upper panels are for the anticoagulant EDTA, the middle panels are for heparin, and the lower panels are for ACD. Whole Blood Propidium Iodide (PI) Uptake In a separate series of experiments, viability and light scatter parameters were analyzed on an EPICS 742 standardized for fluorescence intensity and light scatter with microspheres. Whole blood from a single normal blood bank donor (Clinical Center, NIH Bethesda, MD) was collected in EDTA, heparin, and ACD (two Vacutainer tubes each). One tube was processed immediately while the other tube was held overnight unopened at RT. WB (5 pl) was stained with CD45 (HLel, FITC, BDIS) at RT and washed once with PBS and.5% bovine serum. Propidium iodide (PI) (2 pl; 5 pg/pl) was added, and the non-lysed sample was analyzed for PI uptake by CD45-FITC-positive cells. RESULTS Initial Observations Our initial experiments focused on incomplete red blood cell (RBC) lysis. Macroscopic evidence of incomplete hemolysis with the manual methods may be due

4 WHOLE BLOOD LSIS 71 Table 1 Summary of Percentage Positive CD45+ CD14 Lymphocytes Method of lysis HPaque Immunolyse FACS Lyse Q-Prep ficoll Gen Trak EDTA Leucogate Leucogate + My Heparin Leucogate Leucogate + My ACD Leucogate Leucogate + My to insufficient time for hemolysis, while this problem with the automated method could be due to instrument malfunction. Incomplete hemolysis was not a serious problem with any of the methods used in this study. We also addressed whether the combination of linear FSC and linear SSC vs. linear FSC and log SSC had any effect on determining the percentage of lymphocytes in a given preparation. This comparison (linear vs. log amplification for SSC) demonstrated no difference with regard to leukocyte population resolution. Consequently, all experiments were performed by using linear amplification for FSC and log amplification for SSC. This observation is consistent with the findings of Dawson and co-workers (3). Whole Blood Lysis A checker-board comparison of three anticoagulants and five cell preparation methods conducted immediately and after overnight storage resulted in a set of 3 two-parameter light scatter histograms. There were three analyses for each condition: unstained cells, cells stained with CD45/CD14, and cells stained with CD451 CD14 plus CD13. Each was analyzed for the percentage of lymphoid cells, FSC and SSC mean channel values, and the FI of unstained cells and CD45'CD14- cells in terms of mean channel values as well as the percentage positive CD45+CD14- cells. Selected data summarized in Figure 1 (upper panels) show the percentage of lymphocytes as a function of the method of cell preparation in which cells are prepared immediately and stored overnight at RT. The results were similar for cells prepared at both time points. We have not observed any significant difference in the percentage of lymphocytes as a function of the method of WB lysis used. Obviously a larger proportion of lymphoid cells is in the HF PBMC preparation. Also shown in Figure 1 (lower panels) are the mean channel FSC values for cells found in the lymphocyte gates. As can be seen, adding FACS Lyse solution results in a decrease in the FSC mean channel value at both time points. Changes in the SSC mean channel values were variable. These changes were not noted for the other methods of cell prepara- tion. The FIs for unstained cells and CD45+CD14- cells are summarized in Figure 2. As can be seen, there is some variation in the autofluorescence as a function of anticoagulant and method of WB lysis. Immuno-lyse showed higher autofluorescence values after overnight storage for all three anticoagulants. The FI of stained cells was remarkably constant for cells prepared both immediately and the following day, however. We next evaluated the percentage of CD45+CD14- cells in each of the various preparations. In viewing the representative values (Table l), we found a variable degree of dim to negative CD45 events that were more prevalent after overnight storage. This was found with varying combinations of anticoagulant and WB lysis methods. These findings are important because they can result in a decrease in the percentage of CD45+ lymphocytes present in the lymphocyte gates. A percentage less than 9-95 of lymphocytes in the total gated population may result in the technical rejection of the sample or, alternatively, require correction of results involving other lymphoid markers after an explanation for this loss is found. Adding CD13 to CD45 and CD14 increases the fluorescence of the myeloid components and moves them into the double-positive quadrant. Back-gating (PAINT-A-GATE analysis, BDIS) of these data suggests that these events are due primarily to the inclusion of debris in the lymphocyte gate (see Fig. 3 for selected examples). To a lesser extent, however, CD14- monocytes and dim CD45 myeloid cells (basophils) may also be present (5). The debris may consist of microparticles (microspherocytes) derived from a variety of cells including platelet aggregates as a result of the lysis procedure itself (2). The PI uptake of unlysed, WB samples immediately (day ) and on the following (day 1) is summarized in Figure 4 There were no significant differences using the three anticoagulants at day. On day 1, however, there was a significant increase in the number of granulocytes positive for PI (dead) in the EDTA anticoagulated blood. This cell death was less evident but present nonetheless with ACD and it was negligible with heparin.

5 72 CARTER ET AL. FIG. 3. Selected examples of back gating (Paint-A-Gate, BDIS). Left column contains dot plots for FSC vs. SSC. Right column contains dot plots for FL1 vs FL2. The anticoagulants and the method of lysis used were EDTA and Gen Trak for A-D, while EDTA and FACS Lyse were used for E-H. In B and F, whole blood was stained with Leucogate, while in D and H, whole blood was stained with Leucogate plus My 7 RD1. The red-orange cluster is designated as lymphocytes, while the green and magenta clusters are monocytes and granulocytes, respectively. Debris is blue. Note the shift in FSC using FACS Lyse (E, G) and the shift in debris with the addition of My 7 RD1 to Leucogate (D, H). DISCUSSION The clinical use of flow cytometry will undoubtedly continue to increase as the diagnostic and therapeutic values of monoclonal antibodies are validated. Developments in flow cytometry software will further refine the routine clinical application of this method. Rapid sample preparation methods are evolving to speed processing of the sample load for flow cytometric analysis. We have also seen a progression of these procedures from the original HF method through hypotonic lysis methods up to the present commercially available, pro- prietary (non-hypotonic) methods. The practice of WB staining followed by lysis (rather than lysis followed by WB staining) has been evaluated and is now a common practice (1). With no previous comparative reports in the literature, we performed a side-by-side comparison of four commercially available methods for WB lysis in order to compare them to conventional HF-separated PBMC. Although not included in this report, hypotonic lysis using NH,C1 after staining gave results similar to the results from other WB lysis methods (personal observation, Marti).

6 WHOLE BLOOD LSIS , I - LOG SIDE SCATTERDAy LOG RED FLUORESCENCE Anticoagulant: ACD LOG SIDE SCATTER LOG RED FLUORESCENCE DA 1 Anticoagulant: ACD LOG SIDE SCATTER LOG RED FLUORESCENCE DA Anticoaguiont: HEPARIN LOG SIDE SCATTER LOG RED FLUORESCENCE DA 1 Anticoagulant: HEPARIN LOG SIDE SCATTER LOG RED FLUORESCENCE LOG SIDE SCATTER LOG RED FLUORESCENCE DA DA 1 Anticoagulant: EDTA Anticoagulant: EDTA FIG. 4. Summary of PI uptake of unlysed whole blood in ACD, EDTA, and heparin (top to bottom) on gated FITC-CD45+ cells. The left panels are for the day blood sample, and the right panels are for the day 1 sample. See text (Experimental Procedures section) for method.

7 74 CARTER ET AL. Essentially all of the methods gave equivalent results in terms of the absolute number and proportion of lymphocytes for analysis. Qualitatively, granulocyte numbers appeared to be comparable on immediately prepared blood samples, while losses of monocytes were more variable. Decreases in FSC were noted only for the FACS Lyse preparation, but this did not appear to interfere with lymphocyte enumeration. In fact, all of the laboratories involved in this study used a different methods of WB lysis. We are unable to choose or recommend any one WB lysis method. This decision is made based on individual laboratory preference. Variation in the percentage of CD45+CD14- lymphocytes was of some concern. It did not seem to be related to any one method of lysis and it was most likely related to inadequate exclusion of WB lysis debris by light scatter gating. The recent report of back gating by Loken and co-workers offers an excellent means of correcting this problem on normal samples; however, before it can be widely recommended, more experience with abnormal patient samples is required (5). We were also concerned about the effect WB lysis methods might have on the level of FI. Autofluorescence seemed to vary as a function of the WB lysis method and, to a lesser extent, of the anticoagulant. The relationship of autofluorescence to the background fluorescence of a blank bead is a convenient method of checking instrumentiassay sensitivity and is a definite requirement in measuring some of the low-density markers associated with leukemic cells. In the case of CD45 FI, we saw little or no variation in normal blood samples. FI measurements for other commonly used markers such as CD4 and CD8 should be evaluated as a function of the anticoagulant and method of lysis. Viability was a primary concern when we evaluated the anticoagulants. We realized that in most settings, a laboratory can set up a sample within 24 h, so we did not conduct longer time studies. On the basis of these study results, we advise that samples anticoagulated with EDTA be analyzed immediately rather than stored overnight due to the significant decrease in granulocyte viability. In an earlier study, potassium EDTA was associated with loss of function in lympho- cyte mitogen stimulation assays (7). Because of this, we continue to prefer heparin if the sample cannot be analyzed immediately. ACD may also be a useful alternative to heparin since the results were comparable for lymphocytes and there was only a minimal decrease in granulocyte viability over 24 h. ACD may, however, result in platelet aggregation in samples from patients with high platelet counts (personal observation, Marti). Future studies should include commonly used reagents applied to a population of normal adult individuals as well as to a variety of patient populations. The issues to be addressed should include examination of light scatter parameters, methods of gating, and FI of the various reagents used. In situations where samples are routinely processed after h, viability testing should be considered; this may be facilitated by the recently describe method of using an ultra-violet, lightactivated dye for viability assessment followed by fixation (8). LITERATURE CITED 1. Ashmore LM, Shopp GM, Edwards BS: Lymphocyte subset analysis by flow cytometry. Comparison of three different staining techniques and effects of blood strage.j Immunol Methods 188:29-215, Ault KA: Flow cytometric measurement of platelet-associated immunoglobulin. Pathol Immunopathol Res 7:395-48, Dawson CD, Scheppler JA, Nicholson JKA, Holman RC: Enumeration of antigen sites on cells by flow cytometry. Pathobiol J Immunopathol Mol Cell Biol 59:57-61, Hoffman RA, Kung PC, Hansen WP, Goldstein G: Simple and rapid measurement of human T lymphocytes and their subclasses in peripheral blood. Proc Natl Acad Sci USA 77: , Loken MR, Brosnan JM, Bach BA, Ault KA: Establishing optimal lymphocyte gates for immunophenotyping by flow cytometry. Cytometry 11(4): , Muirhead KA, Wallace PK, Schmitt TC, Frescatore RL, Franco JA, Horan PK: Methodological considerations for implementation of lymphocyte subset analysis in a clinical reference laboratory. Ann N Acad Sci 468: , Patrick CW, Swartz S, Keller RH: A comparative study of heparin and potassium EDTA as anticoagulants for cell surface markers and immunologic functional analysis. [Meeting abstract.] Am J Clin Pathol 81(6):797, Riedy MC, Muirhead KA, Jensen CP, Stewart CC: Use of a photolabeling technique to identify nonviable cells in fixed homologous or heterologous cell populations. Cytometry 12(2): , 1991.