Enumeration of IgE secreting B cells

Journal oflmmunologicalMethods, 132 (1990) 37-43 Elsevier

37

JIM05653

Enumeration of IgE secreting B cells A filter spot-ELISA Christopher L. King 1, Georgos Thyphronitis 2 and Thomas B. N u t m a n 1 I Laboratory of Parasitic Diseases, National Institutes of Health, Bethesda, MD, U.S.A., and e Department of Medicine, Uniformed Services University of the Health Sciences, Bethesda, MD, U.S.A.

(Received 20 February 1990, revised received 26 April 1990, accepted 26 April 1990)

To assess the frequency of IgE producing cells in humans a filter immunoplaque assay has been developed to detect IgE secretion from individual B lymphocytes in unfractionated peripheral blood mononuclear cells (PBMC). PBMC were incubated in microfilter plates containing nitrocellulose membranes coated with polyclonal anti-human IgE antibody, and the IgE production by a single cell was detected using a specific anti-human IgE monoclonal antibody followed by enzymatic development. The products of the enzymatic reaction were visualized as blue plaques on the membranes. The assay was both sensitive and specific as determined by: (1) a near 1 : 1 correlation between direct cell counts of an IgE producing myeloma cell line (U266) and the number of plaques in the filter immunoassay; and (2) the absence of detectable plaques generated by human B cells transformed by Epstein-Barr Virus (EBV) and producing only IgG or IgM. The presence of other cell types in PBMC did not affect the ability to detect IgE secreting cells. Replicate cultures of highly purified B lymphocytes, first transformed with EBV and then stimulated with recombinant human interleukin-4, produced IgE levels in culture supernatants that correlated closely with the number of IgE producing cells (r = 0.93; P < 0.001). Furthermore, using the same transformed cells, the number of IgE secreting cells assessed by the immunoplaque assay was significantly correlated (r --- 0.94; P = 0.002) with the number of IgE producing cells assessed by immunofluorescence staining of intracytoplasmic IgE. This assay provides a simple and direct method to assess the frequency of IgE producing lymphocytes in humans. Key words: lgE; ELISA, filter spot-; B lymphocyte

Introduction IgE antibody levels in human sera and in cell cultures are very much lower than those of other immunoglobulin isotypes, and they are frequently difficult to detect even with highly sensitive ira-

Correspondence to: C.L. King, National Institutes of Health, Bldg 4, Room 126. Bethesda, MD 20892, U.S.A.

munoassays (Leung and Geha, 1986). This low level of immunoglobulin production is associated with a correspondingly low number of IgE secreting B lymphocytes, estimated to range between 1/1000-1/100000 lymphocytes in normal individuals (Saxon et al., 1980; Lanzavecchia, 1983; Stein et al., 1986). In disorders associated with elevated serum IgE levels (e.g., atopic disorders, the hyper-IgE recurrent infection syndrome (HIE), or parasitic helminth infections) it is presumed

0022-1759/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)

38 that the frequency of IgE producing B lymphocytes is increased. However, no systematic study has actually assessed numbers of IgE secreting cells in these conditions, and studies on the frequency of IgE producing lymphocytes has, in part, been hampered by the lack of a simple and direct method to measure IgE producing cells. Current methods for such measurements of IgE producing cells have relied on either limiting dilution analysis of EBV transformed B cells or immunofluorescent staining of IgE producing lymphocytes (MacKenzie and Dosch, 1989; Thyphronitis et al., 1989). Both methods, however, have serious limitations. Limiting dilution analysis only provides an indirect measure of IgE producing ceils, is extremely variable and requires multiple replicates; moreover, it is absolutely dependent on the number of B cells actually infected and transformed by EBV. The alternative of direct counting of immunofluorescence-stained IgE in B cells can be laborious and inaccurate, especially in preparations of peripheral blood mononuclear cells where such cells are at very low concentrations. Recently a technique has been described by which individual immunoglobulin secreting lymphocytes can be enumerated by filter hybridization using highly specific monoclonal and polyclonal antibodies (Czerkinsky et al., 1983; Sedgwick and Holt, 1983; Mrller and Borrebaeck, 1985). The present study adapts these techniques to measure the number of IgE secreting lymphocytes in preparations of PBMC and evaluates certain technical issues important in such estimation; namely the sensitivity, specificity, and reproducibility of the assay and the effects of cryopreservation of peripheral blood cell populations prior to assessing the numbers of IgE-secreting B cells.

Materials and methods

Cells An IgE myeloma cell line, U266, was obtained from the American Type Culture Collection (Rockville, MD) and cloned by limiting dilution to ensure that all cells were high IgE producers. PBMC from either normal donors or individuals with HIE were obtained by Ficoll-diatrizoate

gradient centrifugation (Bryum, 1968), and highly purified B cells were obtained by negative selection using a cocktail of monoclonal antibodies directed against mononuclear cells as described in detail elsewhere (Thyphronitis et al., 1989). EBV infection of B cells was performed by 2 h incubation of 1 x 105 cells with supernatants of the B95-8 EBV-producing marmoset line containing 3 x 105 infectious units of EBV. Cultures were maintained at 5% CO 2 at 37°C in Iscove's medium with L-glutamine, gentamicin (0.4 mg/ml) and 10% fetal calf serum; half of the culture media was replaced weekly. EBV infected cells were subsequently cloned and supernatants assayed for production of either IgM or IgG by immunoassay (Nutman et al., 1984).

Intracytoplasmic staining of cells for IgE Cultured cells were separated on Ficoll-diatrizoate, then mixed with formalin-fixed chicken erythrocytes at a 100 : 1 ratio and centrifuged onto a glass microscope slide with a cytocentrifuge. Cells were air-dried, fixed for 1 h in methanol at 4°C, dried, stained for 30 min with fluoresceinlabeled mouse monoclonal antibody to IgE (Becton Dickinson, Mountain View, CA), washed, mounted in glycerol, and examined with a fluorescence microscope. The percentage of cells staining for IgE was determined by counting under UV light the number of lymphoid cells that displayed green cytoplasmic fluorescence in a field that contained a total of 100 chicken erythrocytes (and, therefore, 10000 lymphocytes) (Finkelman et al., 1987). Chicken erythrocytes were identified by their oval shape and orange nuclear fluorescence.

Reagents Plates were coated with an affinity-purified rabbit anti-human IgE whose specificity has been characterized previously (Hamilton et al., 1981) and further documented by the spot-ELISA described below. The second antibody was a highly specific anti-human IgE monoclonal antibody, 4.15 (obtained from Dr. Andrew Saxon, and described previously (Macy et al., 1988)). An alkaline phosphatase (AP)-conjugated rabbit anti-mouse IgG antibody was purchased commercially (Jackson ImmunoResearch, West Grove, PA). The AP chromogen substrate solution, consisting of 5-

39 bromo-4-chloro-3-indolyl phosphate toluidine salt and p-nitroblue tetrazolium chloride ( B C I P / N B T ) was prepared according to the manufacturer's instructions (Kirkegaard & Perry, Gaithersburg, MD). The enzyme substrate solutions could be kept in the dark at 4°C indefinitely and was successfully kept at room temperature for 2 weeks without adverse effects. Recombinant human IL-4 was purified from supernatants of transfected C127 mouse mammary tumor cells as described previously (Thyphronitis et al., 1989).

Filter spot-ELISA assay The assay consists of six steps: (1) preparation of a solid-phase immunoadsorbent, (2) incubation of the cell suspension, (3) addition of the unconjugated monoclonal antibody, (4) addition of APconjugated antibody, (5) addition of AP chromogen substrates which yield insoluble blue spots, and (6) enumeration of the spots. Solid-phase immunoadsorbent. Individual wells of nitrocellulose bottomed 96-well Millititer HA plates (Millipore Co., Bedford, MA) were filled with 0.05-0.075 ml of PBS pH 9.6 containing 2 /~g/ml of rabbit anti-human IgE and incubated overnight at 4°C to coat the nitrocellulose membranes. (Coated plates could be stored for 1-2 weeks at 4°C.) Unabsorbed protein was removed by washing six times with PBS containing 0.025% Tween 20 (PBS-T; Sigma, St. Louis, MO). In order to saturate remaining protein binding sites, 0.1 ml of PBS with 3% BSA was added to individual wells and incubated at 37°C for at least 1 h. Care was taken to dry the bottom of the nitrocellulose membrane with absorbent paper towels to prevent leaking. Cell incubation. Microtiter wells were filled with 0.1 ml of cell culture medium containing various concentrations of cells and incubated undisturbed either for 4 h or overnight at 37°C in a humidified CO 2 incubator. Routinely, samples were run in quadruplicate and serially diluted (two-fold dilutions) for at least four dilutions. The serial dilutions enabled more precise enumeration of ceils in that an optimal dilution could be selected for counting cells. Monoclonal antibody and immunoenzyme stage. At the completion of the cell incubation stage, plates were washed six times in PBST, flicked

manually to remove any remaining wash buffer, and blotted dry; 0.075 ml of PBS-T containing 1% bovine serum albumin (PBS-BT; sigma, St. Louis, MO) with 2 / ~ g / m l of 4.15 anti-human IgE monoclonal was added to each well. Plates were incubated for 6 h at room temperature or overnight at 4°C. Plates were again washed six times with PBS-T, and blotted dry before 0.075 ml of PBS-BT containing 1 /~g/ml of AP-conjugated rabbit anti-mouse IgG was added to each well; the plate was then incubated at 37°C for 2 h and washed six times in PBS prior to development. Deoelopment. To each well, 0.075 ml of B C I P / N B T substrate solution was added, and the wells were examined for the appearance of blue spots. Spots first appeared in 5-10 rain; the reaction was allowed to proceed for an additional 5-10 rain after which the wells were washed with PBS and allowed to dry. Careful monitoring of plates was necessary after adding substrate to prevent excessive background staining. Enumeration of spots. Developed plates were dried and the blue spots were enumerated under low magnification ( x 10- × 20) using a dissecting microscope equipped with a vertical light source. Positive reactions could be easily identified as circular, densely granulated foci with decreased granulation radiating from the center, so that the spots had a 'halo'-like appearance. Their diameters ranged from 0.05 mm to 0.2 mm. Occasionally, small dense spots were observed that were either slightly irregular in shape or did not have a granular appearance; these were considered as background and could be distinguished from true spots.

Fluorescein activated cell staining The proportion of B cells in PBMC or enriched B cell preparations was determined by staining cells with phycoerythrin-conjugated (PE) anti-CD 19 antibody (Becton Dickinson, Mountain View, CA), and analyzed in an EPICS fluorescenceactivated cell sorter (Coulter, Haileleha, FL). Data analysis Data are expressed as m e a n _ normally distributed after natural tion. Means were compared using and correlations of groups of

SEM and were log transformaStudent's t test data were de-

40 termined by simple linear regression of transformed data.

Results

TABLE I THE PROPORTION OF IgE SECRETING B LYMPHOCYTE PREPARATIONS OF PBMC COMPARED TO HIGHLY PURIFIED B CELLS Diagnosis

Sensitivity and specificity The sensitivity of this spot-ELISA was determined using IgE-producing myeloma cells (U266) in which direct counts of these cells were compared with the number of IgE secreting cells detected as 'spots'. There was a near 1 : 1 correlation between the number of cells added and the number of plaques in the filter immunoassay at dilutions of 125, 62, and 31 cells per well (corresponding to counts of 132, 65 and 38 spots/well, as seen in Fig. 1). The assay was also highly specific, as no detectable plaques were observed using EBV transformed B cells producing only IgG or IgM (Fig. 1).

Proportion of IgE secreting cells in PBMC compared to B cell enriched subpopulations Although the use of purified B lymphocytes would be optimal in the spot-ELISA assay since it

No. of Cells added

HIE HIE Normal

Number of IgE secreting cells/10 5 B lymphocytes PBMC

Purified B lymphocytes

1593 +85 193 _+13 0.7+ 0.2

1607 _56 157 _+ 4 1.1 + 0.3 a

a p = 0.04. would avoid any potential effect of other cell types in PBMC on estimates of the number of IgE-secreting cells, adding PBMC directly to the well would greatly simplify the assay. In order to determine whether estimates of the numbers of IgE-secreting cells in PBMC and purified B lymphocytes are comparable, aliquots of PBMCs obtained from two patients with markedly elevated serum IgE levels and from a normal individual were purified into highly enriched B lymphocytes (91-94%) and assayed in parallel with total PBMC

Clones of E~ cell lines producing IgE

IgG

IgM

125

63

31

Fig. l. Sensitivity and specificity of the IgE spot-ELISA. Spot-EL1SA assay to which decreasing numbers of cloned B cells producing IgE, IgG, or IgM were added.

41 1000

.,,J

100 n

_.=~ 1,-o

O~

10

u.I "'

1

04

10 5

!

10 6

i

10 7

PBMC/WELL

Fig. 2. Effect of initial PBMC concentration on the n u m b e r of IgE secreting B cells in two patients. The n u m b e r of lgE secreting B cells as a function of initial plating density is plotted. Data are expressed as mean + SEM of quadruplicate cultures.

(Table I). In two of the three individuals estimates of the number of IgE secreting cells were the same for both methods of assessment, and in the other individual there was a significant, but small (19%), difference in approximation of the number of IgE secreting lymphocytes. Of note, the proportion of IgE secreting cells was much greater in the patients with HIE (0.2-1.6%) compared to the normal individual (0.001%).

Effect of PBMC concentration on the number of IgE secreting B lymphocytes PBMCs were isolated from two HIE patients, added to microtiter wells at increasing concentrations (5 x 10 4 t o 2 x 10 6 eells/well) and the number of IgE secreting cells determined (Fig. 2). Estimates of the numbers of IgE secreting cells were not significantly affected by concentrations of 1 x 10 6 PBMC/well and lower. However, at 2 x 10 6 PBMCs/well the numbers of IgE secreting cells detected were significantly reduced (by 43% and 66%) in both patients. This effect was independent of the proportion of IgE-secreting cells because the patients, one with a high, and the other with a comparatively low proportion of IgE secreting cells both demonstrated lower estimates of IgE secreting cells at 2 X 10 6 cells/well.

Effect of cryopreservation of PBMC on estimates of lgE secreting cells Cryopreservation of cells significantly reduced estimates of the number of IgE secreting cells when compared to fresh cells obtained concurrently (by 26%, 50% and 56%) in three HIE patients so studied. The lower estimates of IgE secreting cells in cryopreserved cells appeared to result from a proportionately greater loss of IgE producing cells in the freezing and thawing process than other types of B lymphocytes since the proportion B cells in cryopreserved cells was actually higher than in the fresh cells for all three patients studied (data not shown). Reproducibility Although there was a significant difference in estimates of the frequency of IgE secreting cells between fresh and cryopreserved obtained from the same sample, estimates of the number of IgE secreting cells from aliquots of the same cryopreserved sample performed in different experiments was highly reproducible. The degree of reproducibility of the assay was demonstrated when PBMC from a single individual were aliquoted, cryopreserved, and subsequently assayed in four separate experiments, yielding estimates of 9.3, 6.5, 6.0, and 6.9 IgE secreting cells/100,000 B lymphocytes. Estimates of IgE secreting B-cell frequency: spotELISA vs: mtracytoplasmic immunofluorescence staining The number of IgE secreting cells detected by filter spot-ELISA was compared to immunofluorescence staining of intracytoplasmic IgE in lymphocytes from the same cultures. The cultures of IgE producing lymphocytes were generated by EBV-transforming highly purified B lymphocytes (96%) obtained from a normal individual in the presence of 2 x 103 U of IL-4 for 26 days. This procedure generated a series of cultures each of which produced different amounts of IgE. Cells from each culture were counted and standardized to 1 x 105 lymphocytes; the 'old' media replaced with fresh media and incubated for an additional 3 days. There was a wide range of IgE detected in different culture supernatants, but the IgE concentrations correlated closely with the proportion of IgE secreting B lymphocytes ( r = 0.93, P <

42 TABLE II ESTIMATES OF IgE SECRETING B CELL FREQUENCY: SPOT ELISA VS. IMMUNOFLUORESCENCE STAINING IN EBV TRANSFORMED AND IL-4 STIMULATED NORMAL B CELLS ELISA IgE (ng/ml)

Spot E L I S A IgE Secreting/ 104 B cells

Immunofluorescence IgE Staining/ 104 B cells

249.0 195.0 178.0 120.0 97.0 36.4 13.9 11.5 5.9 0.4 0.0

210.0 77.0 87.0 61.0 51.0 21.0 3.6 3.2 2.1 0.9 0.0

189.0 139.0 nd a 62.0 nd 10.0 nd nd 1.7 1.3 0.0

a Not done.

0.001) (Table II). Estimates of the number of IgE producing cells from most of the same cultures, as determined by immunofluorescence staining of intracytoplasmic IgE, also correlated closely with the frequency of IgE secreting ceils estimated by the spot-ELISA (r = 0.94; P = 0.002) (Table II).

Discussion A filter spot enzyme-linked immunoassay is described for detection of individual IgE secreting lymphocytes based on the use of highly specific polyclonal and monoclonal antibodies to human IgE. The procedure follows the general principles of the E L I S P O T assay initially described for use in polystyrene plates with alkaline phosphatase (Sedgwick and Holt, 1983) or horseradish peroxidase (Czerkinsky et al., 1983) and later modified for use on nitrocellulose plates (MiSller and Borrebaeck, 1985; Czerkinsky et al., 1988). The assay described here makes use of alkaline phosphatase-labelled antiglobulin that in the presence of B C I P / N B T chromogen substrate produces an insoluble precipitate visualized as a blue plaque. The immunoplaque assay was sufficiently sensitive to detect all the IgE secreting ceils added to each well up to 125 IgE producing cells per well.

Furthermore, the assay was specific in that no spots were detected from EBV transformed B cells secreting only I g G or IgM. Therefore, this technique should be able to assess levels of IgE production lower than that estimated in culture supernatants because one or more cells secreting IgE can be visualized in up to 1 × 106 cells, while the same number of cells may produce insufficient protein to be measured in the culture supernatant (unpublished observations). This assay has the advantage of being able to estimate IgE secreting cells directly from PBMC populations. Although such estimation requires determining the proportion of B cells in the PBMC by immunofluorescence staining of cells, use of PBMC preparations was much simpler than preparing purified B cells for study. Our results demonstrate that estimates of IgE secreting cells using PBMC are the same as using purified B cells and indicate that in this assay the presence of other cell types does not affect IgE secretion. The disadvantage of using PBMC was that adding more than 1 × 106 PBMC per microtiter well leads to underestimates of the number of IgE secreting cells. Whether this underestimate results from inhibition of IgE secretion by other cell types or from the physical limitations caused by large numbers of cells in a microtiter well is unclear, but we have found that this problem can be rectified by using wells with a larger surface area of nitrocellulose (unpublished observations). The number of IgE secreting cells estimated by the spot-ELISA was almost identical to the number of IgE committed B lymphocytes determined by intracytoplasmic staining of IgE in cells obtained from the same culture. While intracytoplasmic staining of IgE in normal B lymphocytes may not indicate actual secretion of the immunoglobulin, in this experiment cells were EBV transformed and therefore activated to secrete immunoglobulin; thus the presence of intracytoplasmic IgE should indicate a plasma cell. In addition, the amount of IgE produced in the same culture supernatants was also closely correlated with the number of IgE producing cells. This finding indicates that the amount of IgE produced in lymphocyte cultures is primarily a function of the number of lgE producing plasma cells rather than a difference in the amount of immunoglobu-

43

lin produced by an individual cell. Whether this relationship holds for cultures of lymphocytes from individuals with different levels of serum IgE remains to be determined. Our studies demonstrate that this filter immunoassay is both highly sensitive and specific and provides a simple and direct method for detecting IgE producing lymphocytes in human PBMC. This technique should provide a useful alternative to currently available assays such as limiting dilution analysis and plaque-inhibition assays. Other potential applications of the spotELISA technique that are being explored include measurement of both total and antigen-specific IgE secreting cell frequencies in various diseases associated with elevated serum IgE levels. Furthermore, this assay has the capacity to measure the efficacy of therapy for various atopic or parasitic conditions by assessing antigen-specific IgE production that is not easily detected by current methods.

Acknowledgements The authors thank Dr. Eric Ottesen for his support throughout this project and critical review of the manuscript.

References B/Syum, A. (1968) Isolation of leukocytes from blood and bone marrow. Scand. J. Clin. Lab. Invest. 21, 97. Czerkinsky, C., Nilsson, L.-A., Nygren, H., Ouchterlony, O. and Tarkowski, A. (1983) A solid-phase enzyme-linked immunospot (ELISPOT) assay for enumeration of specific antibody-secreting cells. J. lmmunol. Methods 65, 109. Czerkinsky, C., Moldoveanu, Z., Mestecky, J., Nilsson, L.-A. and Ouchterlony, O. (1988) A novel two colour ELISPOT

assay. I. Simultaneous detection of distinct types of antibody-secreting cells. J. Immunol. Methods 115, 31. Finkelman, F.D., Snapper, C.M., Mountz, J.D. and Katona, I. (1987) Polyclonal activation of the murine immune system by a goat antibody to mouse IgE. IX. Induction of a polyclonal IgE response. J. Immunol. 138, 2826. Hamilton, R.G., Hussain, R., Ottesen, E.A. and Adkinson, Jr., N.F. (1981) The quantitation of parasite-specific human IgG and IgE in sera: evaluation of solid-phase RIA and ELISA methodology. J. Immunol. Methods 44, 101. Lanzavecchia, A. (1983) One out of five peripheral blood B lymphocytes is activated to high-rate Ig production by human alloreactive T cell clone. Eur. J. lmmunol. 13, 820. Leung, D.Y.M. and Geha, R.S. (1986) Control lgE synthesis in man. J. Clin. Immunol. 6, 273. MacKenzie, T. and Dosch, H.-M. (1989) Clonal and molecular characteristics of the human IgE-committed B cell subset. J. Exp. Med. 169, 407. Macy, E., Kemeny, M. and Saxon, A. (1988) Enhanced ELISA: how to measure less than 10 picograms of a specific protein (immunoglobulin) in less than 8 h. FASEB 2, 3003. MSller, S.A. and Borrebaeck, C.A.K. (1985) A filter immunoplaque assay for the detection of antibody-secreting cells in vitro. J. Immunol. Methods 79, 195. Nutman, T.B., Ottesen, E.A., Fauci, A.A. and Volkman, D.J. (1984) Parasite antigen-specific T cell lines and clones: MHC restriction and B cell helper function. J. Clin. Invest. 73, 1178. Saxon, A., Morrow, C. and Stevens, R.H. (1980) Subpopulations of circulating B cells and regulatory T cells involved in vitro immunoglobulin E production in atopic patients with elevated serum immunoglobulin E. J. Clin. Invest. 65, 1457. Sedgwick, J.D. and Holt, P.G. (1983) A solid-phase immunoenzymatic technique for the enumeration of specific antibody-secreting cells. J. Immunol. Methods 57, 301. Stein, L.D., Chan, M.A., Toshifumi, H. and Dosch, H.-M. (1986) Epstein-Barr virus-induced IgE production in limiting dilution cultures of normal human B cells. Eur. J. lmmunol. 16, 1167. Thyphronitis, G., Tsokos, G.C., June, C.H., Levine, A.D. and Finkelman, F.D. (1989) IgE secretion by Epstein°Barr virus-infected purified human B lymphocytes is stimulated by interleukin 4 and suppressed by interferon-T. Proc. Natl. Acad. Sci. U.S.A. 86, 5580.