INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Human immune responses to specific antigens have been successfully studied with peripheral blood mononuclear cells (PBMC), which comprise B and T lymphocytes, monocytes, natural killer cells, and dendritic cells. Currently used anticoagulants include EDTA, heparin, and acid-citrate-dextrose (ACD). EDTA works as an anticoagulant by chelating metal ions; it is the reagent of choice, since it preserves cellular integrity (10), and is particularly recommended for achieving high yields of monocytes (2). However, blood stored in EDTA for more than a few hours yields PBMC contaminated with red cells and granulocytes (11). Heparinized blood, while avoiding this problem, nevertheless causes platelet aggregation and changes the white cell morphology (3). This platelet clumping in turn interferes with applications such as flow cytometric analysis of PBMC samples.

In our efforts to study antigen-specific T-cell responses, we have encountered a progressive loss of proliferative responses in PBMC obtained from blood stored in EDTA for increasing lengths of time. Storage of blood with EDTA for 2 days was reported to cause alterations in the proportion of lymphocyte subsets (11, 12). Earlier studies also described EDTA-mediated inhibition of phytohemagglutinin (PHA)-stimulated blast transformation of human PBMC (1). In these studies, the cells were continuously maintained in cultures in the presence of EDTA. A study of lymphocyte proliferative responses of whole blood and separated PBMC to different stimulants after overnight storage of blood in heparin, ACD, and citrate cell preparation tubes revealed varied responses, depending on the antigen used as well as the human immunodeficiency virus (HIV) status of the individual (23). In that study, however, responses to PHA were not significantly affected.

Here we report the irreversible loss of antigen-specific but not mitogen-induced responses in PBMC obtained from blood stored in EDTA for 24 h, although no EDTA was present during the 4 days of culturing with antigen. The paucity of metal ions following exposure to EDTA was not the cause of the defect, since EGTA did not do any damage to antigen-specific responses. In addition, metal ions added to the cultures failed to restore the functional competence of the affected cells. We show conclusively that EDTA damages antigen-presenting cells but not effector T cells.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Study population and experimental design. Ten milliliters of blood was drawn from volunteers after informed consent was obtained; 4 mM (0.15% [wt/vol]) EDTA or EGTA or 5 U of heparin per ml was used as an anticoagulant. Blood was either processed immediately or stored for up to 24 h at room temperature prior to isolation of PBMC.

Cell preparation. Blood was diluted with an equal volume of phosphate-buffered saline, layered over an equal volume of Ficoll-Hypaque (Pharmacia), and centrifuged in a table-top centrifuge with swing-out buckets at 900 × g for 30 min. The PBMC banding below the plasma were collected, washed multiple times with phosphate-buffered saline, and suspended in RPMI 1640 (Life Technologies-BRL) containing 2 mM L-glutamine, 10% human AB serum, 50 µg of streptomycin per ml, and 50 U of penicillin per ml. Cells were cultured at 0.5 × 10⁵ cells per well for M. tuberculosis antigen and 1 × 10⁵ cells per well for JEV antigen in 96-well flat-bottom plates (Costar). We routinely obtained 1 × 10⁶ to 2 × 10⁶ PBMC/ml of blood. Storing blood in the various anticoagulants for as long as 24 h resulted in no significant differences in the absolute numbers of viable lymphocytes obtained, as measured by trypan blue exclusion.

Antigen preparation and lymphocyte proliferation assays. Mycobacterial antigen was prepared by sonicating M. tuberculosis cells for a total of 15 min, followed by centrifugation for 30 min at 10,000 × g to remove insoluble material. The clear supernatant was filtered through a 0.2-µm-pore-size filter. The viral antigen used was glutaraldehyde-fixed lysates of JEV-infected Vero cells prepared as described previously (8).

Reconstitution of adherent antigen-presenting cells with Percoll-purified lymphocytes or T-cell lines. PBMC obtained from PPD-positive individuals by the Ficoll-Hypaque method of density gradient centrifugation were separated on a preformed continuous Percoll (Sigma) gradient in order to obtain separate populations of monocytes and lymphocytes as described previously (4). Briefly, the leukocytes (20 × 10⁶) were layered onto preformed Percoll gradients in 15-ml polycarbonate tubes and spun in swing-out buckets in a refrigerated centrifuge at 1,000 × g for 20 min. The cells from the two bands obtained were collected separately using sterile Pasteur pipettes and were found to be enriched for monocytes (upper band, 86%) and lymphocytes (lower band, 81%) by fluorescence-activated cell sorter analysis. Monocytes were seeded at 10⁵ cells/well, washed to remove nonadherent cells, and reconstituted with lymphocytes at 1.5 × 10⁵ cells/well, followed by the addition of mycobacterial antigen. The inability of either of these two populations to proliferate in response to ConA treatment indicated that the level of purity obtained was adequate for our experiments.

Statistical analysis. In the experiment with convalescent JEV patients, positive responses to viral antigen were scored based on the criteria that (i) the SI was equal to or greater than 2.4, since the maximum value of the proliferative response induced in PBMC of control individuals plus 1.96 times the standard deviation of the mean was 2.388, and (ii) the total counts per minute observed on stimulation with viral antigen measured 1,500 or more. The calculation of the SEM for M. tuberculosis and JEV antigens has already been described above. The SIs for both bacterial and viral antigens were logarithmically transformed (21), and significance values were calculated using Student's t test.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Effect of EDTA exposure on antigen-specific responses of PBMC. In our initial experiments, we observed a near complete loss of mycobacterial antigen-specific lymphoproliferation in PBMC obtained from the blood of nine PPD-positive individuals and stored in 4 mM EDTA as an anticoagulant for more than 20 h. The proliferative responses of PBMC from six representative individuals after various periods of exposure to EDTA are shown in Fig. 1. There was a dramatic and significant reduction in the SI, ranging from 86 to 94% in the individuals tested (P < 0.01). The results are plotted as the mean ± SEM of the SIs obtained from three independent experiments.

View larger version (37K): [in this window] [in a new window]   FIG. 1.   Progressive loss of antigen-specific proliferative responses in PBMC obtained from blood stored in EDTA for increasing periods of time. Proliferation was measured as outlined in Materials and Methods. V1 to V6 indicate the six PPD-positive volunteers participating in the study. Values in parentheses indicate the mean ± SEM counts per minute incorporated by control PBMC. The SI indicated is the mean value obtained from three independent experiments. Bars indicate standard errors.

Metal ions fail to reverse EDTA-mediated damage. We attempted the restoration of antigen-specific lymphoproliferation by adding Ca²⁺, Mg²⁺, or Zn²⁺ to the blood samples during storage in EDTA. The effect of Ca²⁺ could not be studied, as Ca²⁺ resulted in clotting of the blood samples. Mg²⁺ did not show any beneficial effect. We observed only a marginal reversal of the EDTA effect when 0.16 mM Zn²⁺ was added to whole blood during the 24-h exposure to EDTA (data not shown). As an alternative means of providing depleted ions, we next carried out the lymphoproliferation assay with the metal ions added to the PBMC in cultures during the 4-day antigen exposure period. Chloride salts of Ca²⁺, Mg²⁺, Zn²⁺, and Fe²⁺ were added singly or in combination to the PBMC along with antigen. None of these metal ions helped restore the proliferative response of PBMC (Fig. 2). In view of the fact that EDTA was removed from the cells by extensive washing and the presence of adequate quantities of several divalent cations in culture medium formulations, the concentrations of the added divalent cations achieved in the cultures were perhaps in excess of the optimal levels required, as suggested by the observed suppressive effect on the SI when some of the metal ions or combinations thereof were added.

View larger version (19K): [in this window] [in a new window]   FIG. 2.   Inability of metal ions to reverse EDTA-mediated suppression of antigen-specific lymphocyte proliferation. Responses were assayed as described in the legend to Fig. 1 with various metal ions added either singly or in combinations as indicated during the 4-day exposure to antigen. The concentrations of the metal ions used were as follows: Ca2+, 2 mM; Mg2+, 0.8 mM; Zn2+, 0.1 mM; and Fe2+, 0.2 mM.

EDTA-exposed PBMC respond maximally to IL-2 and ConA. In order to determine the overall ability of EDTA-exposed lymphocytes to proliferate in response to nonspecific stimuli, we treated the PBMC with IL-2 or ConA in place of antigen or, in the case of IL-2, in combination with antigen as well. Each of these agents was able to elicit maximum proliferation of the treated lymphocytes in an antigen-independent manner. Use of ConA or IL-2 brought about maximum levels of proliferation regardless of the period of exposure to EDTA (Table 1). The addition of IL-2 to PBMC from blood stored in EDTA as late as 6 days after they were seeded into culture wells still elicited maximum levels of thymidine incorporation (data not shown). Our results clearly indicated the presence of a fully functional DNA-synthesizing machinery in the EDTA-exposed lymphocytes. A similar differential response to mitogens as opposed to specific antigens was also reported for HIV patients (23). These results suggested to us that the effector T-cell population in the PBMC was perhaps not the target of the EDTA-mediated damage and suggested the possibility that EDTA affected the antigen-presenting cells in PBMC.

Antigen-presenting cells but not effector T cells are damaged by EDTA. In order to determine whether effector or presenting cells were affected by EDTA, we resorted to the preparation of pure populations of these two subsets of cells and then reconstituted them for the antigen stimulation assay. Percoll gradients were used to obtain monocyte and lymphocyte populations. The data from a representative PPD-positive volunteer (Table 2) showed that while monocytes from fresh blood successfully reconstituted the antigen-specific response of lymphocytes obtained from both fresh and EDTA-stored blood, monocytes from EDTA-exposed blood were incapable of eliciting the same effect.

Exposure of blood to heparin or EGTA does not damage the antigen-specific responses of PBMC. Anticoagulants other than EDTA were examined for their ability to preserve the antigen responsiveness of PBMC from stored blood. Heparin, which is a reagent commonly used for this purpose, as well as EGTA, also a metal ion chelator, were examined. Heparin allowed unaltered mycobacterial antigen-specific proliferation of lymphocytes from blood stored for up to 24 h for two of the four PPD-positive individuals tested (Fig. 3A). We did observe a significant reduction (>50%) in proliferation brought about by storage in heparin for the remaining two volunteers. An earlier report described deterioration in PHA-induced blastogenic responses in PBMC obtained from blood processed after 24 h of storage in heparin (7). In our studies, however, the antigen-specific proliferative response of PBMC from heparinized blood samples was vastly superior to that of PBMC from EDTA-treated blood. Surprisingly, in contrast to the results obtained with EDTA, lymphocytes from blood stored in EGTA for up to 24 h were relatively more proficient in antigen-specific proliferation (Fig. 3B). Two of the volunteers tested, V1 and V2, did, however, show a 40 to 50% reduction in antigen-specific responses. PBMC from heparinized or EGTA-treated blood also responded maximally to both ConA and IL-2 (data not shown). We have not tested cells obtained from blood stored in EGTA or heparin for periods longer than 24 h.

View larger version (46K): [in this window] [in a new window]   FIG. 3.   Heparin and EGTA do not adversely affect antigen-specific responses of PBMC from stored blood. Proliferation was assayed as described in Materials and Methods. (A) Blood stored in heparin. (B) Blood stored in EGTA. V1 to V4 indicate the four PPD-positive volunteers participating in the study. Values in parentheses indicate the mean ± SEM counts per minute incorporated by control PBMC. The SI indicated is the mean value obtained from three independent experiments. Bars indicate standard errors.

Lymphoproliferative responses to JEV antigen are not impaired on storage of blood in EGTA. EGTA was also found to be very efficient in preserving JEV antigen-specific lymphoproliferative responses on storage of blood obtained from convalescent JEV patients for as long as 24 h after collection (Fig. 4). Blood samples stored in EGTA for 24 h (n = 20) gave results similar to those of blood samples processed at the site of collection (n = 10) (P > 0.05). All 30 samples had SIs above 2.4, the cutoff value chosen as described in Materials and Methods. The loss in response upon storage in EDTA for 16 h (n = 17) was extremely significant (P < 0.001) compared to the results for fresh samples. All 17 individuals had SIs below 2.4 on stimulation with JEV antigen, although their ConA responses were comparable to those of the other 30 volunteers (data not shown).

View larger version (13K): [in this window] [in a new window]   FIG. 4.   EGTA preserves virus-specific proliferative responses in convalescent JEV patients. Blood samples obtained from convalescent JEV patients were processed at the time of collection (n = 10) or stored in 4 mM EDTA (n = 17) or EGTA (n = 20) for 16 to 24 h at room temperature prior to processing. Samples were scored as positive if the SI was >2.4.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The analysis of lymphocyte subsets in peripheral blood often serves as a useful marker of disease progression and prognosis, especially in infections such as HIV type 1. The choice of anticoagulant is of major importance when blood samples are to be used for techniques such as immunophenotyping by flow cytometry or for analysis of antigen-specific cellular responses. The three most commonly used anticoagulants, EDTA, heparin, and ACD, have been compared by several investigators for their ability to preserve the proportion of lymphocyte subsets and their morphological features, surface markers, and ability to proliferate in response to various antigens (11, 12, 14, 19, 20, 23). Although EDTA is the anticoagulant of choice for hematology (10), it was nevertheless reported to cause alterations in the proportions of T- and B-cell subsets when blood was processed 2 days after collection compared to fresh blood samples (11, 12). In addition, PHA-induced lymphocyte transformation was also reported to be inhibited by the continuous presence of EDTA in cultures (1). The choice of the anticoagulant used has also been found to have profound effects on the surface expression of the activation markers HLA-DR and CD11b on peripheral leukocytes (19). Studies of the proliferative ability of lymphocytes from HIV-infected patients have shown that 24 h of storage of blood in heparin or ACD does affect antigen-specific responses, while mitogen-specific responses are better preserved (23). Of the anticoagulants tested, heparin preserved only cytomegalovirus responses better than ACD, while samples stored in citrate cell preparation tubes gave better results than those stored in ACD with all antigens tested. None of the anticoagulants tested was able to provide results comparable to those obtained with fresh blood.

The analysis of human antigen-specific T-cell responses in populations living in remote areas in which a pathogen is endemic often entails storage of blood samples in anticoagulants for as long as 24 h during transportation before the samples reach the laboratory for processing. Because blood is the only source of immune cells that can be obtained from human volunteers, PBMC obtained from such stored samples have to be used for lymphoproliferative as well as cytotoxic responses, the effectors for which may often be present in low proportions in peripheral blood.

Calcium is a divalent cation vital for downstream transduction of signals in lymphocytes following receptor interaction with major histocompatibility complex-bound antigen and appeared to be the most likely candidate whose loss may have been responsible for the results that we observed. Ca²⁺ has also been shown to restore EDTA- and citrate-mediated inhibition of PHA-induced proliferation of lymphocytes when these metal ion chelators were added to cultures (1). We were, however, unable to detect any restoration by metal ions such as Ca²⁺, Mg²⁺, Fe²⁺, and Zn²⁺, over a range of concentrations, of EDTA-induced reduction of antigen-specific lymphoproliferation. The vital factor depleted by EDTA exposure has yet to be identified.

A striking aspect of our results was the absence of proliferation inhibition when blood was stored in EGTA as an anticoagulant. In fact, such differential effects of EDTA and EGTA were also reported for the inhibition of DNA synthesis in cultured chick embryo cells (18). The affinity of EGTA for Ca²⁺, Zn²⁺, and Cu²⁺ is very similar to that of EDTA for these three divalent cations (18). For Ni²⁺, Co²⁺, and Mn²⁺, EGTA has a lower affinity than EDTA (1, 5). The benign nature of EGTA is perhaps due to the low affinity of EGTA for some trace ions vital for the functional integrity of antigen-presenting cells.

A 24-h delay in the processing of heparinized blood was reported earlier to cause more than a 50% decrease in PHA-stimulated DNA synthesis (7). In our studies, too, heparin caused up to 70% inhibition of proliferative responses in some individuals. Moreover, we observed a small but distinct enhancement of total thymidine incorporation in lymphocytes when blood was stored in EGTA as opposed to heparin for about half of the individuals tested, a result which translated to increased fold stimulation and therefore was more like the responses obtained for fresh blood. EGTA was able to preserve virus-specific responses in convalescent JEV patients at levels comparable to those in blood processed immediately following collection (P > 0.05). Our results suggest that EGTA can successfully replace EDTA as an anticoagulant in instances where prolonged storage of unclotted blood may become necessary before the samples can be transported to a laboratory for obtaining lymphocytes.

The observation that antigen-presenting cells were the victims of EDTA was unexpected. The ability of PBMC stored in EDTA to respond to IL-2 or ConA was the initial indication of this possibility. Similarly, PHA-specific responses in HIV patients were found not to be altered on storage in heparin and ACD (23), indicating that PHA-specific proliferation differed significantly from antigen-specific proliferation. Reconstitution of antigen-specific proliferation of either untreated or EDTA-exposed T-cell lines by fresh untreated monocytes but not by EDTA-treated monocytes established beyond a doubt that presenting cells developed a defect following prolonged exposure to EDTA. We also observed that treatment of these affected monocytes with bacterial lipopolysaccharide for 24 h failed to make them proficient in presenting antigen. The EDTA effect appeared unlikely to be brought about by a loss of major histocompatibility complex class II molecules on the antigen-presenting cells of the PBMC population, since it was shown that exposure of mouse tissues to 10% EDTA (269 mM) for as long as 14 days to achieve demineralization for histological examination did not cause any loss of class II molecules (6). The blastogenic response of human or guinea pig T cells to mitogens such as PHA or ConA has been conclusively shown to require the presence of monocytes (9, 17). It is therefore surprising that despite damage to monocytes by EDTA, the response of whole PBMC to the mitogen ConA or to IL-2 was wholly unaffected by EDTA.