Glucocorticosteroid-induced immunoglobulin production requires intimate contact between B cells and monocytes

CELLULARIMMUNOLOGY

112, 147-155 (1988)

Glucocorticosteroid-Induced lmmunoglobulin Production Requires Intimate Contact between B Cells and Monocytes’ FRANK M. ORSON* AND CURT A. AUZENNE Departments of Internal Medicine and Microbiology and Immunology, Baylor College ofMedicine, and VeteransAdministration Medical Center, Houston, Texas 77211 Received August 3, 1987; accepted October 12, 1987 Glucocorticosteroid (GCS)-induced immunoglobulin (Ig) production in vitro is dependent on the functions of T cells and monocytes. T cells produce a replacing factor (TRF-S) which, with monocytes and a broad spectrum of concentrations (both above and below the physiologic range) of GCS, stimulates B cells to synthesize Ig. TRF-S is produced by T cells in cultures of mononuclear cells in the absenceof stimulation over the initial 72 hr in culture. T cells, however, require the presence of monocytes or small quantities of interleukin 1 in order for the synthesis of TRF-S to occur. In addition to their role in stimulating TRF-S production, monocytes are also required in cultures of B cells responding to GCS and the cytokine. These experiments demonstrate that this monocyte function cannot be replaced by IL-l or crude supematants of monocyte cultures. Furthermore, exposure of TRF-S containing supematants to oxidizing conditions does not alter the dependence of the cytokine on monocytes or GCS. Coculture of B cells and monocytes separated by a permeable membrane demonstrated that the influence of monocytes on GCS-induced Ig production is unlikely to be mediated by stable soluble factors. Thus, GCS-induced Ig production requires intimate contact between monocytes and B cells in the form of surface contact or unstable soluble mediators. 0 1988 Academic press, IX.

INTRODUCTION Monocytes have been shown to perform essential functions in many facets of the immune response. Their roles in the presentation of antigen to both T cells and B cells (1) and in the production of some T-cell cytokines (2, 3) have been studied extensively. Other functions of monocytes in the development of immunoglobulin (Ig) secretion by B cells, however, have been investigated only more recently. Monocytes have been shown, for example, to releasea B-cell differentiation factor augmenting Ig production by Epstein-Barr virus (EBV)3-stimulated normal B cells and some lymphoblastoid B-cell lines (4). The monocyte requirement for effective lymphocyte ’ This work was supported by the Veterans Administration. ’ To whom correspondence should be addressed:Bldg. 2 11, Rm. 226, Veterans Administration Medical Center, 2002 Holcombe, Houston, TX 772 11. 3Abbreviations used: AET, 2-aminoisothiouronium bromide; BSS, balanced salt solution; Dex, dexamethasone; EBV, Epstein-Barr virus; FCS, fetal calf serum; GCS, glucocorticosteroids; HCTS, hydrocortisone; Ig, immunoglobulin; IgSC, immunoglobulin-secreting cells; IL-l, interleukin 1; MC, peripheral blood mononuclear cells; PWM, pokeweed mitogen; SEM, standard error of the mean; SPA, Staph protein A, SRBC, sheeperythrocytes; SN, supematant; TRF-S, T-cell replacing factor dependent on steroids. 147 0008-8749/88 $3.00 Copyright D 1988 by Academic Press,Inc. All rights of reproduction in any form reserved.

148

ORSON AND AUZENNE

responsesto polyclonal B-cell activators has been known for many years (5), but the precise functions of monocytes in such cultures have not been well defined. Among the stimulants of polyclonal Ig production of particular interest to our laboratory are glucocorticosteroids (GCS), which also depend on monocytes for their stimulatory effectson B cells (3,6-8). Exposure of human peripheral blood mononuclear cells (MC) in vitro to a broad spectrum of GCS concentrations, both above and below the physiologic range, results in the development of immunoglobulin-secreting cells (IgSC) (6). Both T cells and monocytes are required for this responseto occur. T cells produce a soluble cytokine T-cell replacing factor dependent on steroids (TRF-S) (7), while monocytes have a dual role: (i) they are necessary to provide small quantities of interleukin 1 (IL-l) required by T cells for production of the cytokine, and (ii) the presenceof monocytes is necessaryfor B cells stimulated with TRF-S and steroids to develop into IgSC (3). This paper reports the results of experiments investigating the influence of monocytes on the response of B cells to TRF-S and GCS. Monocytes could not be replaced by IL- 1, monocyte supernatants, or exposure of TRF-S containing supernatants to oxidative conditions. It is unlikely that any stable soluble factors mediate this response since B cells separated from GCS-stimulated MC in culture by filters permeable to molecules but not cells cannot be stimulated to produce Ig by GCS and TRFS. Thus, B cells require intimate contact with monocytes for GCS-induced Ig production to occur. MATERIALS AND METHODS Preparation of cell populations. Heparinized venous blood was obtained from healthy young adult volunteers, and the mononuclear cell (MC) population was prepared by standard density gradient centrifugation on a mixture of sodium diatrizoate and Ficoll (LSM, Bionetics, Kensington, MD), followed by extensive washing with modified Hanks’ balanced salt solution (BSS, phosphate instead of carbonate buffered, GIBCO, Grand Island, NY). To separate T cells and B cells, MC suspensions were mixed with 2-aminoethylisothiouronium bromide (AET, Sigma Chemical Co., St. Louis, MO)-treated sheep erythrocytes (SRBC) at a ratio of 1:150 and were incubated for 2 hr at 4°C. The erythrocyte-lymphocyte mixture was then gently resuspended and the rosetting cells were separated from the nonrosetting population on another LSM gradient. The SRBC were removed from the rosetting cells by lysis with an NH&l-lysing buffer. The rosetting population was demonstrated to contain 97% rosette-positive cells and will be referred to as the T-cell population. The interface of nonrosetting cells contained B cells, monocytes, “null” cells, and 3% rosettepositive cells, and will be referred to as “non-T cells.” Monocyte-enriched cells were obtained by suspending 5 X 10’ MC in 10 ml of culture medium supplemented with 20% fetal calf serum and pipetting this suspension into a 250-ml flask (Costar 3275, Data Packaging, Cambridge, MA) which was preincubated for 1 hr with 5 ml of FCS. After 60-90 min at 37°C nonadherent cells were removed by aspiration and washing of the adherence surface with BSS.Adherent cells were collected by incubation of the flask for 15 min with 15 ml of 1:5000 Versene (GIBCO), followed by vigorous washing of the adherence surface by repeated pipetting of the Versene. After washing twice with BSS, more than 90% of the resulting cells contained nonspecific esterase(9). Purified B cells were obtained by depleting non-T cells of monocytes with iron car-

STEROID-INDUCED

IMMUNOGLOBULIN

PRODUCTION

149

bony1 filings (Lymphocyte Separator Reagent, Technicon Instruments Corp., Tarrytown, NY). Suspensions of non-T cells were added in a 5-ml volume to 4 ml of the filings suspension and 1 ml of fetal calf serum. After incubation for 90 min at 37°C nonadherent cells were decanted and a magnet was used to remove the iron filings and iron-containing phagocytes from the suspension. The remaining cells were subjected to another LSM gradient to remove any residual iron-containing cells. The resulting B cells had fewer than 1%monocytes by nonspecific esterasestaining. Lymphocyte cultures. Cultures of MC utilized 96-well microtiter round-bottomed plates (Costar 3799) with a volume of 0.2 ml/well. MC cultures contained 3 X lo5 cells, whereas non-T-cell cultures contained 10’ cells. Hydrocortisone (HCTS, 1Om6 M, Sigma), dexamethasone (Dex, 1O-6A4, Sigma), pokeweed mitogen (PWM, titered to the maximal response for the lot used, a 1:200 final dilution, GIBCO), and EBV (10% by volume of the supernatant of the marmoset cell line B95-8) were used as stimulants for polyclonal Ig production as previously described (8). Fetal calf serum (FCS, Reheis, Kankakee, IL) was used at a concentration of 10% in culture. Dualchamber cultures were constructed of a plastic tube (1.3-cm external and 0.6-cm internal diameters) sealed on one end with a l-pm pore Nucleopore filter (Nucleopore Corp., Pleasanton, CA) which was placed in a 50-ml plastic culture tube (Falcon 2070, Beckton-Dickinson Labware, Oxnard, CA). The resulting culture surfaces were separated by 5 mm. After gas sterilization, cell suspensions containing lo6 B cells, non-T cells, or 2 X lo6 MC were placed in either chamber in a final total volume of 1.5 ml. Supernatant (SN) preparations. MC were cultured for 3 days in 50-ml flasks (Costar 25 100, Data Packaging, Cambridge, MA) in a volume of 5 ml at an initial cell concentration of 5 X lo6 cells/ml in RPM1 1640 (GIBCO) supplemented with 10% FCS, 2 mM L-glutamine (GIBCO), and 10 pg/ml gentamicin (Sigma). T-cell and monocyte supernatants were similarly prepared except for using 3 X lo6 or 1 X lo6 cells/ml, respectively. Supernatants were harvested by sedimenting the cells and filtering the decanted fluid through 0.45-pm filters (Acrodiscs, Gelman, Ann Arbor, MI). TRF-Sassay. Supernatants were tested for TRF-S activity by utilizing steroid-stimulated cultures of non-T cells. Each microtiter well contained lo5 non-T cells, 10e6 M Dex, and dilutions of supernatant. After a culture period of 8-10 days at 37°C in a humidified COZincubator, the cultures were assayedfor the presenceof Ig-secreting cells by reverse hemolytic plaque assay. Treatment of SN with H202 (Sigma) was accomplished by adding a sufficient amount of an aliquot of SN to provide a final concentration of low4 M, and incubating at 37°C for 1 hr before adding the SN to assaycultures. Reverse hemolytic plaque assay. After 8- 10 days in culture, the cells were washed in the microtiter plate three times with Hanks’ balanced salt solution (GIBCO, buffered with sodium phosphate). The cells were then resuspended in 0.125 ml of BSS in the plate, and a mixture of Staph protein A (SPA, Genzyme, Boston, MA)coated SRBC, guinea pig serum (Cappel Laboratories, Westchester, PA) as a source of complement, and rabbit antisera to human Ig was added in dilutions appropriate to maximize plaque development. The antiserum reacts with human y-, c+, p-, X-, and K-Ig chains, and thus detects Ig secretion of all major isotypes. The mixture for each culture was then pipetted into a loo-p1 glass chamber (Air-Lot Slides, Spiral Scientific, Cincinnati, OH). The chambers were sealed with hot wax and then incubated for 3 hr at 37°C. At the end of the incubation period, the number of IgSC

150

ORSON AND AUZENNE

. . . .

hbd

GCS

PWM MC

. .

EBV

Med

GCS

PWM

Mf8 Depleted

EBV

MC

Med

GCS

T Depleted

PWM

EBV

MC

FIG. 1. GCS-induced Ig synthesis depends on T cells and monocytes. Each culture contained 3 x lo5 mononuclear cells, 3 X 1O5MC depleted of monocytes, or 1 X lo5 T-depleted MC. Optimal concentrations of stimulants or medium alone was added to give a final volume of 0.2 ml. Each data point represents the geometric mean of three cultures in an individual experiment, and each bar is the geometric mean of the data points for the particular experimental condition.

was enumerated (f 10%) with a video-based automatic plaque counter (Optomax, Hollis, NH). Cytokines. Highly purified IL-l was obtained commercially (Genzyme) and was previously shown to have IL- 1 activity in a thymocyte costimulation assay(4). RESULTS Monocytes are requiredfor GCS-induced Igproduction. The monocyte dependence of some polyclonal B-cell activators for stimulation of Ig production was not initially recognized because only small numbers of monocytes were sufficient to support responses(5). As with pokeweed mitogen, GCS stimulation of Ig production is inhibited only with stringent monocyte depletion (6). Figure 1 demonstrates that intact mononuclear cells can respond to various stimulants with large increasesin the number of IgSC (left set of columns). These concentrations of Dex and HCTS are in the middle of the very broad range of concentrations capable of inducing maximal numbers of IgSC (6). Although Dex binds much lessto cortisol-binding globulin and albumin than HCTS, this distinction has little apparent influence on their capacity to induce IgSC (6), except perhaps at very low concentrations. Depletion of either monocytes (center columns) or T cells (right columns) eliminates the response to both GCS and PWM. In contrast, stimulation with EBV is dependent only on the presence of B cells which are activated to produce Ig by direct infection with the virion ( 10). Since GCS-induced Ig production does not involve mitogens or antigens,

STEROID-INDUCED

IMMUNOGLOBULIN

151

PRODUCTION

.

.

?!

1000

: G ii -0

500

d 200

+Med

+TRF-S

+TRF-S

+TRF-S +M@

SN

+ IL-1

PIG. 2. Monocyte SN and IL-l cannot substitute for monocytes in GCS-induced Ig production. Mononuclear cells were depleted of T cells by SRBC rosetting, and isolated B cells were prepared by subsequent iron carbonyl treatment of nonadherent T-cell-depleted populations. Cultures contained 1 X IO5 non-T cells or purified B cells in a final volume of 0.2 ml/well. Supematant concentrations were SO%,dexamethasone concentration was lo-” h4, and EBV was used at 10% (v/v). The monocyte supematant concentration was 25%, and a final concentration of 1 unit/ml was used in those cultures containing IL-l. Each point represents the mean of triplicate cultures, and the bars represent the geometric mean of the individual experiments.

the function of monocytes in this system was not immediately apparent, so further experiments were performed to study the role of monocytes. Monocytes are required for TRF-S activity. The dependence of GCS-induced Ig production on T cells was previously shown to be mediated by a soluble cytokine, TRF-S (3, 7), and, like some other T-cell cytokines, its production depended on the monocyte product IL- 1. Monocytes clearly had another function, however, since they not only induce the synthesis of TRF-S but also are required for the expression of TRF-S activity. B cells were isolated from T cells and monocytes by sequential depletions using AET-SRBC rosetting and iron carbonyl particles. The resulting populations contained fewer than 1% monocytes. Figure 2 shows that such isolated B cells (open bars) were unable to respond to stimulation with TRF-S and GCS, in contrast to B cell cultures containing monocytes (striped bars). Furthermore, this lack of responsiveness could not be overcome by the addition of either monocyte supernatants or highly purified IL-l (third and fourth pairs of bars). Control cultures stimulated with EBV, however, demonstrated that B cells purified by these procedures were capable of producing Ig in the absence of monocytes since there was no difference in the number of IgSC induced by EBV with or without added monocytes (far right pair of bars).

152

ORSON AND AUZENNE

In some immune responses, cytokines produced by T cells are modified by monocytes through an oxidative effect (11). Soluble immune response suppressor (SIRS), for example, can be converted from an inactive form by treatment with hydrogen peroxide. To determine whether monocytes may activate TRF-S in a similar fashion, we treated TRF-S-containing supernatants with H202 and compared the response of isolated B cells with that of B cells and monocytes to these SN with GCS. Isolated B cells failed to develop an increase in the number of IgSC different from treatment with medium alone (~200 IgSC/culture), whereas the number of IgSC in the cultures of B cells with monocytes increased appropriately with this same SN (1030/culture). Longer exposure periods or higher concentrations of H202 resulted in a marked loss of TRF-S activity. Similarly, incubation of TRF-S-containing SN with monocytes for varying periods of time failed to enable the SN to subsequently induce IgSC in isolated B cells treated with GCS (data not shown). Thus, neither IL-l, monocyte SN, nor treatment with peroxide was able to substitute for the role of monocytes in the expression of TRF-S activity in GCS-induced Ig production. Intimate contact between B cells and monocytes is required for GCS-induced Ig production. To determine whether soluble factors which could mediate monocyte function in TRF-S activity might be detectable, we cultured isolated B cells in chambers that were separated by a filter from mononuclear cells stimulated with GCS. The culture vessels consisted of a plastic tube sealed on the bottom with a filter and a culture tube into which the other was placed. Thus, the two chambers were separated by a 5-mm vertical space of culture medium. The filters had pores of 1 pm and thus could freely pass soluble molecules, but prevented migration of cells from one population to another. Whether the isolated B cells were placed in the upper or lower chamber did not alter the outcome, and so the data presented in Fig. 3 consist entirely of cultures in which the B cells were in the upper chamber (chamber 1) and the MC were in the lower chamber (chamber 2). B cells alone (far left bar), or B cells with added monocytes (second bar from left), as expected, did not respond to GCS stimulation when cultured alone. The third and fourth pairs of bars show the number of IgSC/culture when B cells alone or B cells and monocytes were cultured in the upper chamber and MC were cultured in the lower chamber and stimulated with GCS. Although large increases in IgSC developed in the MC cultures, isolated B cells had no more IgSC than the control cultures. In contrast, the B cells with monocytes developed a large increase in IgSC, indicating that TRF-S passed through the filter effectively. Additional control cultures stimulated with EBV (bars on the right) for both isolated B cells and B cells with monocytes showed each population to be viable and capable of developing IgSC. These results demonstrate the transmission of soluble cytokines between the two chambers can occur (TRF-S had to cross the filter for the cultures of B cells with monocytes to respond to GCS stimulation). However, monocytes appear not to produce such soluble signals, since no IgSC developed in the isolated B cells, despite a vigorous response in the MC population 5 mm away from them. This suggests that an intimate interaction such as surface contact between the monocyte and B cell may be required for GCS-induced Ig production to occur, although there are other potential explanations as discussed below. DISCUSSION Glucocorticosteroids have profound influences on many aspects of immune system function ( 12- 15). Short-term administration of GCS in vivo results in a gradual

STEROID-INDUCED

IMMUNOGLOBULIN

153

PRODUCTION

.

Chamber

B

Med

B + M#

i

2

1

+ GCS

tad

2 + GCS

BMC

1 + GCS

I

2

.I.49 MC 1

2 + GCS

B

Med

1

2 + EBV

Med

B+M 1

2

+ EBV

FIG. 3. Mononuclear cells responding to GCS release TRF-S, but not monocyte factors. Mononuclear cells, T-cell-depleted MC, and isolated B cells were cultured in dual chambers separated by a Nucleopore filter with a 1.O-pm pore size as described under Materials and Methods. Chamber 1 refers to the upper chamber and chamber 2 to the lower. Each chamber contained a culture of 1 X lo6 B cells or B cells with monocytes, 2 X lo6 mononuclear cells, or culture medium alone. Stimulation was with HCTS ( 10m6M final concentration) or EBV (10% v/v). Each point represents the mean number of IgSC for each culture.

decline in serum Ig levels for the major isotypes ( 16). Despite this, specific antibody responses to immunization were not diminished in humans ( 17, 18), or were even enhanced (in primary, but not secondary responses) ( 19). To reconcile these observations, we suggest that the overall decline in serum Ig levels may represent primarily an antiproliferative effect on the lymphocytes of pharmacologic corticosteroid doses (20). The proliferation inhibition, which most strongly affects T cells, essential for the regulation of B-cell responses, may prevent sustained responses while enhancing the early phases of the antibody response. The effects of GCS in vitro on immunoglobulin synthesis are quite variable, and highly dependent on the conditions used by particular investigators. The stimulation of Ig production in vitro has been demonstrated in various laboratories (3,6, 14,2 1, 22). The T-cell contribution to this response has been well established; T4+ cells synthesize the soluble cytokine TRF-S which is produced following exposure to small quantities of IL-1 (3). The cytokine is produced spontaneously in MC cultures, but it requires the presence of both GCS and monocytes to induce Ig synthesis by B cells. Monocytes, therefore, make two contributions to GCS-induced Ig production. First, in cultures of unseparated cells, they produce small quantities of IL- 1 essential to induce T cells to synthesize TRF-S. Second, as investigated in this study, a close cell-to-cell interaction of monocytes with B cells is essential for TRF-S and GCS to

154

ORSON AND AUZENNE

exert their stimulatory effect. This second function cannot be replaced by IL- 1, crude monocyte SN, pretreatment of TRF-S SN with monocytes in culture, or any soluble factors that remain stable in media supporting GCS-induced Ig production. The most compelling demonstration that close cell-to-cell contact is required for monocytes to carry out this role was produced by the experiments using dual-chamber cultures. Despite the sharing of culture medium through a filter, monocytes in one chamber could not supply what was needed by purified B cells in the other chamber to induce an IgSC response in the presence of adequate concentrations of TRF-S and GCS. Thus it seems that close contact between the monocytes and the B cells is required, possibly because monocytes produce a highly unstable factor needed for B cells to respond to GCS and TRF-S, or because molecules closely associated with the monocyte surface must contact the B-cell surface directly. The roles of GCS-induced Ig production in in vivo responses remain poorly understood, and require further study. It is recognized that GCS stimulate the maturation of many cell types (23). In the rabbit splenic fragment system, steroids were found to be absolutely required for antibody production (24). Thus, steroids may be required for, or may enhance, the differentiation of one or more subsets of B cells. They could affect any or all of the events which are necessary for synthesis of Ig, from gene rearrangements to secretion of fully assembled molecules via the Golgi apparatus. Although this maturational effect of GCS is unlikely to be antigen specific, specificity may be conferred in physiologic responses by local factors, e.g., the clonal expansion of both T and B cells triggered by antigen. Indeed, most other systems in which Bcell growth and differentiation factors act, there is likewise no antigen specificity (25). Apart from the contrived conditions in the laboratory, the major role of glucocorticosteroids may be in the adaption of the body to stress. In the earliest immune responses to antigenic challenge, antibodies with degenerate specificity are produced that bind with low avidity to more than one antigen (26). These less specific but quickly produced antibodies may be particularly useful early in infections as an attempt to limit invasion by microorganisms. GCS could thus also play a role in the development of hypergammaglobulinemia associated with many chronic infections and inflammatory diseases (27). REFERENCES 1. 2. 3. 4.

Unanue, E. R., Annu. Rev. Immunol. 2,395, 1984. Mizel, S. B., Immunol. Rev. 63,51, 1982. Orson, F. M., Flange, F. P., and Cashaw, J. L., J. Zmmunol. 137,578, 1986. Rossen, R. D., Laughter, A. L., Orson, F. M., Flagge, F. P., Cashaw, J. L., and Sumaya, C., J. Immunol. 135,3289,1985. 5. Waldmann, T. A., and Broder, S., Adv. Immunol. 32, 1, 1982. 6. Grayson, J., Dooley, N. J., Koski, I. R., and Blaese, R. M., J. C/in. Invest. 68, 1539, 198 I. 7. Orson, F. M., Grayson, J., Pike, S., De Seau, V., and Blaese, R. M., J. Exp. Med. 158, 1473, 1983. 8. Orson, F. M., De Seau, V., Pike, S., and Blaese, R. M., J. Immunol. 133,208, 1983. 9. Koski, I. R., Poplack, D. G., and Blaese, R. M., In “Zn Vitro Methods in Cell-Mediated and Tumor Immunity” (R. R. Bloom and J. R. David, Eds.), pp. 358-362. Academic Press, New York, 1976. 10. Yarchoan, R., Tosato, G., Blaese, R. M., Simon, R. M., and Nelson, D. L., J. Exp. Med. 157, 1, 1983. 11, Aune, T. M., and Pierce, C. W., Proc. Natl. Acad. Sci. USA 78,5099, 198 1. 12. Claman, H. N., N. Engl. J. Med. 287,388, 1972. 13. Fauci, A. S., In “Glucocorticosteroid Hormone Action” (J. D. Baxter and G. G. Rousseau, Eds.), pp. 449-466. Springer-Verlag, Berlin/New York, 1979. 14. Cupps, T. R., and Fauci, A. S., Immunol. Rev. 65,133,1982.

STEROID-INDUCED 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

IMMUNOGLOBULIN

PRODUCTION

155

Parillo, J. E., and Fauci, A. S., Annu. Rev. Pharmacol. Toxicol. 19, 179, 1979. Butler, W. T., and Rossen, R. D., J. Clin. Invest. 52,2629, 1973. Butler, W. T., Transplant. Proc. I, 49, 1975. Claman, H. N., N. Engl. J. Med. 287,388, 1972. Tuchinda, M., Newcombe, R. W., and DeVald, B. J., Int. Arch. Allergy42,533, 1972. Cupps, T. R., and Fauci, A. S., Immunol. Rev. 65,133,1982. Goodwin, J. S., and Atluru, D., J. Immunol. 136,3455, 1986. Jelinek, D. F., and Lipsky, P. E., J. Immunol. 130,2597, 1983. Ballard, P., In “Glucocorticosteroid Hormone Action” (J. D. Baxter and G. G. Rousseau, Eds.), pp. 494-5 15, Springer-Verlag, New York, NY, 1979. Ambrose, C. T., J. Exp. Med. 119, 1027, 1964. Muraguchi, A., Kehrl, J. H., Butler, J. L., and Fauci, A. S., J. Clin. Immunol. 4,337, 1984. Kan, E. A. R., andsiskind, G. W., J. Immunol. 124, 1758, 1980. Stites, D. P., In “Basic and Clinical Immunology” (D. P. Stites, J. D. Stobo, H. H. Fudenberg, and J. V. Wells, Eds.), pp. 325-365, Lange Medical Publications, Los Altos, CA, 1982.