Soluble anti-p primes for B cell differentiation
Eur. J. Immunol. 1992. 22: 1541-1545
Catherine Phillips.0 and Gerry G. B. Klaus National Institute for Medical Research, Mill Hill, London
1541
Soluble anti-p monoclonal antibodies prime resting B cells to secrete immunoglobulins in response to interleukins-4 and -5 u
v
Soluble anti-immunoglobulin (Ig) antibodies have been generally found to inhibit Ig secretion in B cells, via largely unknown mechanisms. To investigate this phenomenon further a two-step culture system was used in which Bcells are primed for 24-72 h with various soluble monoclonal or polyclonal anti-Ig antibodies: after washing the cells were placed in readout cultures with a combination of interleukin (IL)-5 and IL-4. Using this protocol B cells primed with (mitogenic or nonmitogenic) anti-p monoclonal antibodies differentiated into large numbers of IgM-secreting cells, comparable to responses to lipopolysaccharide. In contrast, priming with polyclonal rabbit anti-Ig or monoclonal anti-x antibodies, markedly inhibited Ig secretion induced by IL-4 IL-5. In addition, anti-p was markedly inhibitory if left in the readout cultures with the two 1ymphokines.These results, therefore, indicate that appropriate cross-linking of surface IgM receptors on B cells can prime the cells to secrete Ig when they are restimulated by T cell-derived lymphokines in the absence of anti-p. In contrast co-ligation of both surface IgM and surface IgD receptors apparently results in powerful inhibition of Ig secretion, which is not reversed by stimulation with IL-4 plus IL-5.
+
1 Introduction The use of anti-immunoglobulin antibodies (anti-Ig) as polyclonal analogues for antigen has been invaluable in the study of the role of sIg receptors in Bcell activation and proliferation [l,21. These studies have shown that appropriate forms of soluble anti-Ig [e.g. F(ab’)z fragments of rabbit anti-Ig], or of solid-phase (e.g. Sepharose-bound) anti-Ig, induce activation of virtually all mature murine B cells, and DNA synthesis in some of them (reviewed in [2]).Neither soluble or solid-phase anti-Ig per se induce B cells to secrete Ig. Instead, various studies have shown that soluble anti-Ig markedly inhibits Ig secretion induced by LPS [3-91. However, Sepharose-coupled anti-Ig will promote B cell differentiation to antibody-forming cells (AFC) in the presence of Tcell-derived cytokines [lo, 111. The mechanism responsible for the inhibitory effects of soluble anti-Ig are partially understood and appear to reflect postranscriptional modification of mRNA for secreted IgM [5-lo]. It is also known that the inhibition is not Fc mediated, as F(ab’)z fragments of anti-p inhibit the LPS response as effectively as intact antibodies [12]. The available evidence suggests that anti-Ig stimulation affords a polyclonal model for B cell activation by type-2 T-independent antigens (reviewed in [13]). The fact that in most studies soluble anti-Ig antibodies fail to prime B cells
to secrete Ig is, therefore, puzzling. We decided to reinvestigate the effects of a variety of soluble anti-Ig antibodies in terms of their capacity to activate Bcells to respond to T cell-derived lymphokines, most notably IL-4 and IL-5, which have both been implicated in the control of Bcell differentiation to AFC (reviewed in [14]). We report here preliminary results obtained with a model two-step culture system which separates the priming and expression phases of B cell differentiation. These experiments show that anti-p mAb prime B cells for lymphokine-induced Ig secretion, whereas polyclonal anti-Fab, or anti-x. mAb are markedly inhibitory.
2 Materials and methods 2.1 Experimental animals Male (CBA x C57BL/lO)F1 or CBA/Ca mice aged 4-5 months were obtained from the specific pathogen-free facilities at NIMR. Spleen cells were obtained from F1 mice in most experiments except for those in which mouse anti-Igh-5a (anti-6) monoclonal antibodies were used, where CBA/Ca spleen cells were used.
2.2 B cell activators [I 102681 ~~
~
Present address: Institute of Virology, University of Wiirzburg, Wiirzburg, FRG. 0 Supported by a fellowship from Glaxo Group Research. Correspondence: Gerry G. B. Klaus, Laboratory of Cellular Immunology, National Institute for Medical Research, Mill Hill, London NW7 l A A , Great Britain
0 VCH Verlagsgescllschaft mbH, D-6940 Weinheim, 1992
Goat (GaMIg) and rabbit (RaMIg) anti-mouse Fab antibodies were affinity purified on a column of mouse Fab, followed (in the case of RaMIg) by pepsin digestion and purification of F(ab’)* fragments as previously described [15]. Rat anti-mouse Cp2, b.7.6, (IgG1) [16, 171, anti-Cpl, (AK1: IgGz,) [12], anti-x (187.1; [IS]) mAb were purified by ion exchange chromatography on Bakerbond AbX. LPS from E. coli 055:B5 was obtained from Difco, Detroit, MI. 0014-2980/92/0606-1541$3.50+ .25/0
1542
Eur. J. Immunol. 1992. 22: 1541-1545
C. Phillips and G. G. B. Klaus
2.3 Lymphokines Supernatants of the plasmacytoma cell lines X63-Ag8-653 transfected with the BMGNeo vector containing either murine IL-4 or murine IL-5 cDNA [19] were titrated for activity on the HT-2 Tcell line [20] (for IL-4) or the BCLl B cell line [21] (IL-5). One unit is defined as the dilution of supernatant required to give 50% of maximal proliferation.
2.4 B cell preparation and culture Small resting B cells were prepared by first treating spleen cells with a monoclonal anti-Thy-1 antibody (NIMR-1 provided by R.M.E. Parkhouse, NIMR, Mill Hill, London) and guinea-pig complement, and then layering the cells on a stepwise gradient of 50% : 75% :85% Percoll (Pharmacia, Uppsala, Sweden) as described previously [22]. Gradients were centrifuged at 1300 x g for 15 min and cells from the top of the 85% layer were used for culture. The resulting preparations typically contained > 95% B cells. Cultures were set up in Costar plates (Cambridge, MA) at a density of 6 x 105/mlin the presence of anti-Ig antibodies or LPS as indicated, in RPMI 1640 medium (Flow, Irvine, Scotland), plus 5% FCS supplemented with 5 x lOP5M2-mercaptoethanol, 2 mML-glutamine, 1 mM sodium pyruvate, nonessential amino acids, penicillin and streptomycin. After 24-72 h these priming cultures were washed three times, counted and cells were replated at 2 x lo5viable celldm1in the presence of lymphokines (or LPS). The success of this culture system is critically dependent on the use of LPS-free reagents for the preparation and culture of B cells. The effects of LPS contamination can be minimized by the inclusion of 10 yg/ml polymyxin B during cell preparation,which blocks the mitogenic effects of LPS [23]. However, this was not done in the experiments described here.
sented as means (+ SEM, n = 3) of IgM AFC culture, normalized to lo6 recultured cells.
3 Results 3.1 Effects of priming with polyclonal or monoclonal anti-Ig antibodies for lymphokine-induced IgM secretion In the experiment shown in Fig. 1small resting B cells were primed for 72 h with various concentrations of RaMIg, rat anti-p (b.7.6), or anti-x (187.1).The cells were then washed thoroughly and recultured for a further 72 h with 30 U/ml IL-5 plus 10 U/ml IL-4, when IgM AFC were determined. B cells cultured in medium alone gave a background level of AFC, presumably as a result of activation via FCS. As expected, RaMIg markedly inhibited this response at concentrations as low as 0.1 pg/ml: the anti-x. mAb, 187.1, had similar effects. Priming with the anti-y mAb b.7.6, had quite different consequences: low concentrations (ca. 0.1 yg/ml) also inhibited the background response. However, concentrations of 10 yg/ml or greater substantially increased the number of IgM AFC.
3.2 Kinetics of priming and Ig secretion
To examine the conditions necessary both for priming and reculture, kinetic studies of both periods were done. B cells primed for 24 or 48 h with b.7.6 did not respond above background to lymphokines, whereas those cultured for 72 h did so (Fig. 2). Again, RaMIg markedly inhibited the background response at all time periods, illustrating the powerful (and seemingly irreversible) inhibition elicited by this form of anti-Ig. Fig. 3 demonstrates that maximal numbers of IgM AFC in b.7.6-primed cultures appeared after a 72-h reculture period in IL-4 plus IL-5.
2.5 Measurement of IgM AFC IgM-secreting cells in the readout cultures were enumerated after various times in a reverse-elispot assay (relispot) as adapted from Czerkinsky et al. [24]. Briefly, 24-well tissue culture plates (Costar) were coated with 10 pg/ml GaMIg in PBS overnight at 4°C and then blocked with 5% FCS for 1 h at 37 "C. Cultured cells were washed and plated out in triplicate at lo3 cells/well in medium plus 5% FCS. The plates were incubated for 2 h at 37"C, in 5% COz, and then washed x 3 in PBS containing 0.05% Tween 20 (Sigma, Saint Louis, MO), soaked for 5 min in the third wash and washed a further five times to detach cells 0 0.1 1 10 20 0 0.1 1 10 20 0 0.1 1 completely. The plates were then incubated with biotinyRaMlg lated goat anti-mouse IgM (Southern Biotechnology, Birmingham, AL), followed by streptavidin-conjugated alkaline phosphatase (Sigma). Plates were washed five times Figure 1. Effects of anti-p mAb priming of small resting B cells on and the substrate, 5-bromo-4-choro-3-indolyl-phosphatelymphokine-induced IgM secretion. Small resting B cells were (Sigma) was added in a cold 1.02 M 2-amino-2-methyl- primed with various concentration of anti-p (b.7.6), anti-x (187.1), or F(ab')z fragments of RaMIg for 72 h, washed and recultured 1-propanol buffer containing 0.74 mM MgC126 H2 0 , 1 5 mM with IL-5 (30 Ulrnl) plus IL-4 (10 U/ml) for a further 72 h. IgM NaN3 and 0.01% Triton X-405 [23]. Relispots developed in AFC were determined and results are presented as means k SEM 1to 3 h. The plates were washed in PBS and relispots were (n = 3) of AFC/106 viable recultured cells. Similar results were counted using an inverted microscope. Results are pre- obtained in three additional experiments.
Solublc anti-y primes for B cell differentiation
Eur. J. Immunol. 1992. 22: 1541-1545
1543
3.3 Effects of anti-p on LPS-induced Ig secretion
3.4 Effects of non-mitogenic anti-p mAb
In the experiment shown in Fig. 3 a portion of the control or anti-p-primed B cells were recultured with SO pg/ml LPS, and IgM AFC were assayed at various times thereafter. It is evident that priming with b.7.6 markedly enhanced the subsequent responses to both LPS and to the lymphokine mixture.The numbers of AFC elicited by the two secondary stimuli were similar at 48 and 72 h after reculture. By day 4 the numbers of lymphokine-induced IgM AFC were beginning to decline, whereas the response to LPS continued to increase until day 5, possibly due to the strong mitogenic properties of LPS. However, at no time did the numbers of AFC produced by anti-p-primed B cells recultured with lymphokines fall below those of unprimed cells recultured with LPS. Finally, unprimed cultures produced 20-fold fewer AFC than anti-p-primed cells when recultured with 1ymphokines.
The above experiments were performed with b.7.6,which is one of the few mitogenic anti-p mAb described [16]. To investigate if the effects observed were peculiar to this antibody, or if priming necessitated the use of a mitogenic antibody, further experiments were done with a nonmitogenic anti-p, AK1 [12].These two mAb also recognize different C, domains and are of different isotypes. As shown in Fig. 4, AK1 had similar effects to those of b.7.6: thus, low concentrations caused modest inhibition of background response, while > 10 pg/ml induced significant priming for IgM secretion. These results, therefore, indicate that the priming effects of anti-p are not related to the inherent mitogenicity of the antibodies.
We next assessed the relative importance of IL-4 and IL-S in the induction of Ig secretion (Fig. 5). It is clear that IL-S
Medium anti-p RaMlg
24 hrs
3.5 Requirements for IL-4 and IL-5 in Ig secretion
48 hrs
72 hrs
r"
Figure 2. Time course for priming with anti-y antibodies. Small resting B cells were incubated for 24,48 or 72 h with medium alone, anti-p (b.7.6; 20 p g h l ) , or RaMIg (10 yglml) washed and recultured for 3 days with IL-5 and IL-4 as in Fig. 1,when IgM AFC were assayed. 50
hi 0 0.1 1 1c
CLg/mI Figure 4. The effects of a non-mitogenic anti-p mAb (AK1) on priming for IgM secretion. Small resting B cells were incubated with the indicated concentrations of AK1 or medium for 72 h, washed, and subsequently recultured with 1L-4 plus IL-5 as in Fig. 1. IgM AFC were determined after 3 days of reculture. Results are means f SEM ( n = 3).
Priming/Readout rnedium/medium o medium/lL-4 anti-phedium A anti-NIL-4
w 1.o
Figure 3. Kinetics of appearance of lymphokine-induced or LPSinduced IgM AFC in B cells primed for 3 days with anti-p. Small resting Bcells were incubated with either medium (square symbols), or 20 yglml b.7.6 (round symbols) for 72 h, washed and recultured with 50 pg/ml LPS (open symbols) or IL-4 plus IL-5 as in Fig. 1 (closed symbols). IgM AFC were assayed 1 to 6 days after reculture in each group (means f SEM, n = 3).
10
100
IL-5 U/ml
Figure 5. The requirements for JL-5 and IL-4 in the development of IgM AFC after priming with b.7.6. Small resting Bcells incubated for 72 h either with medium, or with 20 pg/ml b.7.6, were washed and recultured with various concentrations of IL-5, plus, or minus 10 U/ml IL-4. IgM AFC were determined 3 days after reculture (means f SEM, n = 3).
Eur. J. Immunol. 1992. 22: 1541-1545
C. Phillips and G. G. B. Klaus
1544
was an absolute requirement for IgM secretion as no AFC developed in its absence. Addition of IL-5 elicited a small, but significant dose-dependent increase in IgM AFC in both control and anti-y-primed cultures. Furthermore, although 10 U/ml IL-4 did not induce IgM secretion on its own, its inclusion in cultures of anti-y-primed cells with optimal concentrations of IL-5 increased the number of IgM AFC 10-fold. These findings are consistent with those of other workers [25, 261 who demonstrated that IL-4 and IL-5 synergize to promote Ig secretion. IL-4 only minimally increased the number of AFC developing from unprimed B cells recultured with IL-5.
3.6 Effects of continued presence of anti-p during the reculture period The stimulatory effects obtained with anti-y are in apparent conflict with those of earlier investigations on the effects of anti-Ig on Ig secretion [4-91. We, therefore, designed further experiments to attempt to resolve these discrepancies. In the experiment summarized in Fig. 6 some anti-y (or RaM1g)-primed cultures were not washed before the addition of lymphokines on day 3 and assay on day 6. The continued presence of b.7.6 throughout the culture period abolished the priming effect and indeed reduced the background significantly.To rule out the possibility that this effect was due to exhaustion of the medium, additional primed cultures were washed on day 3 and b.7.6 or RaMIg were added back along with the lymphokines during the 3-day reculture period. This again abrogated the priming effect. Finally, it is noteworthy that RaMIg was powerfully inhibitory under all these experimental conditions. This experiment, therefore, confirms that the continued presence of anti-y inhibits lymphokine-induced Ig secretion in activated B cells. Gp: Washed Read-out A
B
+
{
IL-4 IL-5
+
Priming:
anti-lg
Medium anti-p RaMlg
I 0
I 50
I 100
IgM AFC x 10-3/1 O6 Cells Figure 6. The effects of the continued presence of anti-p during the readout cultures. Small resting B cells were incubated either with medium, 20 pg/ml b.7.6, or RaMIg (10 pg/ml) for 72 h. Cells in group A were washed and recultured with IL-5/IL-4 for 3 days as usual. Group B was also washed, but recultured with IL-5/IL-4 in the presence of the same concentrations of antibodies as those used in the priming culture. Group C was not washed, but IL-5 and IL-4 were added to the primary culture on day 3. IgM AFC were determined 3 days after reculture. (means f SEM, n = 3).
4 Discussion The results of these experiments indicate that soluble anti-y mAb effectively prime quiescent B cells for both lymphokine- and LPS-induced IgM secretion, provided that they are washed out of the readout cultures. This effect is not demonstrably dependent on the isotype of the mAb, on the domain of the y chain recognized, nor on the inherent mitogenicity of the mAb. Effective priming requires prolonged exposure (72 h) to relatively high concentrations of the mAb (10-50 yg/ml), while lower concentrations (< 1 yg/ml) tend to be inhibitory (Figs. 1and 4). Both IL-4 and IL-5 were required in the readout cultures to generate optimal responses (Fig. 5). These effects are in complete contrast to those of anti-?t mAb, or polyclonal F(ab’)z rabbit anti-Fab which both cause powerful (and possibly irreversible) inhibition of lymphokine-induced IgM secretion. The lymphokine-induced IgM production by anti-p-primed B cells compared very favorably with that induced by LPS (Fig. 3). Until day 2 after reculture of anti-y-primed B cells, the numbers of AFC induced by IL-5 plus IL-4 by primed B cells were identical to those induced by LPS. By day 4, the lymphokine-induced response began to decline, whereas the LPS-induced response increased until day 5. The differences in kinetics and magnitude between the response to LPS (Fig. 3), versus that to IL-5 plus IL-4 are probably due to the greater inherent mitogenicity of LPS. However, the capacity of IL-4 to promote isotype switching from IgM, to IgGl and IgE may also be relevant. This is currently under investigation and will be reported elsewhere (C. Phillips and G. G. B. Klaus in preparation). It is also possible that additional stimuli (lymphokines?) may be able to enhance the lymphokine-induced response further. These results demonstrate that small resting B cells can be induced to secrete IgM after appropriate triggering of sIgM receptors, followed by restimulation with IL-5 and IL-4. This response does not apparently require B cell membrane contact with T helper cells: this implies that prolonged ligation of sIgM receptors followed by these two lymphokines are sufficient to induce T cell differentiation to AFC. This may be relevant to the capacity of T-independent antigens to induced Ig secretion, a process in which sIgM apparently plays a greater role than that of sIgD [27, 281. The results with anti-% and RaMIg further suggest that cross-linking both sIgM and sIgD receptors generates a more powerful inhibitory signal in resting B cells, which is not overcome by washing the cells and restimulating them with lymphokines. Further experiments have indeed shown that anti-6 mAb do not prime B cells for Ig secretion, but instead inhibit priming by anti-y (C. Philipps and G. G. B. Klaus in preparation). These findings shed some further light on what determines whether anti-Ig enhances or inhibits Ig secretion in activated B cells. As illustrated in Fig. 6 if anti-y is left in the cultures during their stimulation with lymphokines it is inhibitory. This observation is, therefore, in line with those of various earlier studies, using either anti-y mAb, or polyclonal anti-Ig antibodies plus LPS [4-9].This inhibition has been shown to be largely due to lack of posttranscriptional modification of mRNA of the y chain to the secreted form, as the membrane form remains unaffected
Eur. J. Immunol. 1992. 22: 1541-1545
[29, 301. It is not clear why only soluble anti-Ig has this effect, whereas Sepharose-bound anti-Ig does not [12]. Further studies are required to evaluate the effects of anti-p in T cell-dependent B cell activation. It is now well established that contact-mediated signals are involved in the initial phase of T-B interaction (reviewed in [31]). In line with this B cells can be induced to secrete Ig by contact with histoincompatible activated T cells [32-341 possibly through contact with surface molecules other than Ia [35, 361 plus lymphokines. Indeed, membranes of activated T cells induce substantial proliferation of small resting B cells, but only stimulate Ig secretion in the presence of lymphokines [36]. Stimulation with anti-Ig can enhance proliferation [35] and differentiation of B cells [37] in response to activated T cells, but the enhancement disappears at high T : B cell ratios or in the presence of more “efficient” Tcells [38]. Thus, membrane contact with appropriate numbers of activated T cells and activation via sIgM cross-linking may represent alternative mechanisms by which B cells can become competent to secrete Ig in response to lymphokines. It appears that these mechanisms can function either alone or in tandem. Other workers have shown that membrane contact with activated T cells is sufficient to induce this competency in small resting B cells. However, this is the first report that soluble anti-p mAb can also induce this competency independent of T cell membrane contact. We thank Mary Holrnan for her invaluable assistance with some of’ these experiments, and Anneliese Schirnpl for helpful discussion. We are also grateful to E Melchers, Chris Heusser and David Scott,for gift.7 of cell lines. Received January 13, 1992; in revised form February 24, 1992
5 References 1 Parker, D. C., Nature 1975. 258: 361. 2 DeFranco, A. L., Raveche, E. S., Asofsky, R. and Pau1.W. E . , J. Exp. Med. 1982. 155: 1523. 3 Anderson, J., Bullock, W. W. and Melchers. F., Eur. J. Irnrnunol. 1974. 4: 715. 4 Kearney, J. F., Klem, J., Bockman. D. E., Cooper, M. D. and Lawton, A. R., J. Irnrnunol. 1978. 120: 158. 5 Flahart, R. E. and Lawton, A. R., J. Exp. Med. 1987.166: 864. 6 Lawton, A. R.,Summerlin,V. S. and Dooley, J. S., J. Irnrnunol. 1990. 145: 3177. 7 Flahart. R. E. and Lawton, A. R., Mol. Cell. Irnrnunol. 1987. 3: 61. 8 Yuan, D., Mol. Cell. Irnrnunol. 1987. 3: 133. 9 Chen, U., Gene 1988. 72: 209.
Soluble anti-k primes for B cell differentiation
1545
10 Parker, D. C., Irnrnunol. Rev. 1980. 52: 115. 11 Birkeland, M. L., Simpson, L., Isakson, P. C. and Pure, E., I . Exp. Med. 1987. 166: 506. 12 Leptin, M., Eur. .I. Irnrnunol. 1985. 15: 131. 13 DeFranco, A. L., Kung, J. T. and Paul, W. E., Irnrnunol. Rev. 1982. 64: 161. 14 Klaus, G . G. B., Hawrylowicz, C. M. and Carter, C. J., Irnrnunol. 1985. 55: 411. 15 Hawrylowicz, C. M., Keeler, K. D. and Klaus, G. G. B., Eur. J. Irnrnunol. 1984. 14: 244. 16 Julius, M. H., Heusser, C. H . and Hartmann, K.-U., Eur. J. Irnrnunol. 1984. 14: 753. 17 Leptin, M., Potash, M. J., Grutzmann, R., Heusser, C., Shulman, M., Kohler, G. and Melchers, F., Eur. J. Irnrnunol. 1984. 14: 534. 18 Yelton, D. E., Desaymard, C. and Scharff, M. D., Hyhridorna 1981. I : 5 . 19 Karasuyama, H. and Melchers, F., Eur. J. Irnrnunol. 1988. 18: 97. 20 Smith, C. A. and Rennick, D. M., Proc. Natl. Acad. Sci. USA 1986. 83: 1857. 21 Kinashi,T., Harada, N., Severinson, E.,Tanabe,T., Sideras, P., Konishi, M., Azuma, C., Tominaga, A., Bergstedt-Lindqvist, S., Takahashi. M., Fumihiko, M.. Yaoita. Y , Takatsu, K. and Honjo, T., Nature 1986. 324: 70. 22 Klaus, G . G . B., Hawrylowicz, C. M. and Carter, C. M., Irnrnunology 1985. 55: 411. 23 Jacobs, D. M. andMorrison, D. C.,J. Exp. Med. 1975.14: 1453. 24 Czerkinsky, C. C.,Tarkowski, A., Nilsson, L.-A., Ouchertony, O., Nygren, H . and Gretzer, C., J. Irnrnunol. Methods 1984. 72: 489. 25 Murray, l? D., McKenzie, D.T., Swain, S. L. and Kagnoff, M. F., J. Irnrnunol. 1987. 130: 2669. 26 Coffman, R. L., Seymour, B. W. l?, Lebman, D. A., Hiraki, D. D., Christiansen, J. A., Shrader, B., Cherwinski, H. M., Savelkoul, H . F. J., Finkelman, F. D., Bond, M. W. and Mosmann,T. R., Irnrnunol. Rev. 1988. 102: 5. 27 Cambier, J. C., Ligler, F. S., Uhr, J. W., Kettman, J. R. and Vitetta, E. S., Proc. Natl. Acad. Sci. USA 1978. 75: 432. 28 Zitron, I. M., Mosier, D. E. and Paul,W. E., J. Exp. Med. 1977. 146: 1707. 29 Chen-Bettecken, U.,Wecker, E. and Schimpl, A., Proc. Natl. Acad. Sci. U S A 1985. 82: 7384. 30 Hogbom, E., Martensson, I.-L. and Leanderson, T., Proc. Natl. Acad. Sci. U S A 1987. 84: 9135. 31 Vitetta, E . S., Fernandez-Botran, R., Meyers, C. D. and Sander,V. M., Adv. Irnrnunol. 1989. 45: 1. 32 Riedel, C., Owens, T. and Nossal, G. J. V., Eur. J. Irnrnunol. 1988. IS: 403. 33 Tite, J. l?, Kaye, J. and Jones, B., Eur. J. Irnrnunol. 1984. 14: 553. 34 Julius, M. H. and Rammensee, H.-G., Eur. J. Irnrnunol. 1988. 18: 375. 35 Whalen, B. J. ,Tony, H. P. and Parker, D. C., J. Irnrnunol. 1988. 141: 2230. 36 Hodgkin, l? D.,Yamashita, L. C., Coffman, R. L. and Kehry, M. R., J. Irnrnunol. 1990. 145: 2025. 37 Julius, M. H.,Von Boehmer, 13. and Sidman, C. L.. Proc. Natl. Acad. Sci. USA 1982. 79: 1989. 38 Pereira, l?, Forsgren, S., Portnoi, D., Bandcira, A , , MartinezA., C:. and Coutinho, A . , Eur. J. Imrnunol. 1986. 16: 355.
Eur. J. Immunol. 1992. 22: 1541-1545
Catherine Phillips.0 and Gerry G. B. Klaus National Institute for Medical Research, Mill Hill, London
1541
Soluble anti-p monoclonal antibodies prime resting B cells to secrete immunoglobulins in response to interleukins-4 and -5 u
v
Soluble anti-immunoglobulin (Ig) antibodies have been generally found to inhibit Ig secretion in B cells, via largely unknown mechanisms. To investigate this phenomenon further a two-step culture system was used in which Bcells are primed for 24-72 h with various soluble monoclonal or polyclonal anti-Ig antibodies: after washing the cells were placed in readout cultures with a combination of interleukin (IL)-5 and IL-4. Using this protocol B cells primed with (mitogenic or nonmitogenic) anti-p monoclonal antibodies differentiated into large numbers of IgM-secreting cells, comparable to responses to lipopolysaccharide. In contrast, priming with polyclonal rabbit anti-Ig or monoclonal anti-x antibodies, markedly inhibited Ig secretion induced by IL-4 IL-5. In addition, anti-p was markedly inhibitory if left in the readout cultures with the two 1ymphokines.These results, therefore, indicate that appropriate cross-linking of surface IgM receptors on B cells can prime the cells to secrete Ig when they are restimulated by T cell-derived lymphokines in the absence of anti-p. In contrast co-ligation of both surface IgM and surface IgD receptors apparently results in powerful inhibition of Ig secretion, which is not reversed by stimulation with IL-4 plus IL-5.
+
1 Introduction The use of anti-immunoglobulin antibodies (anti-Ig) as polyclonal analogues for antigen has been invaluable in the study of the role of sIg receptors in Bcell activation and proliferation [l,21. These studies have shown that appropriate forms of soluble anti-Ig [e.g. F(ab’)z fragments of rabbit anti-Ig], or of solid-phase (e.g. Sepharose-bound) anti-Ig, induce activation of virtually all mature murine B cells, and DNA synthesis in some of them (reviewed in [2]).Neither soluble or solid-phase anti-Ig per se induce B cells to secrete Ig. Instead, various studies have shown that soluble anti-Ig markedly inhibits Ig secretion induced by LPS [3-91. However, Sepharose-coupled anti-Ig will promote B cell differentiation to antibody-forming cells (AFC) in the presence of Tcell-derived cytokines [lo, 111. The mechanism responsible for the inhibitory effects of soluble anti-Ig are partially understood and appear to reflect postranscriptional modification of mRNA for secreted IgM [5-lo]. It is also known that the inhibition is not Fc mediated, as F(ab’)z fragments of anti-p inhibit the LPS response as effectively as intact antibodies [12]. The available evidence suggests that anti-Ig stimulation affords a polyclonal model for B cell activation by type-2 T-independent antigens (reviewed in [13]). The fact that in most studies soluble anti-Ig antibodies fail to prime B cells
to secrete Ig is, therefore, puzzling. We decided to reinvestigate the effects of a variety of soluble anti-Ig antibodies in terms of their capacity to activate Bcells to respond to T cell-derived lymphokines, most notably IL-4 and IL-5, which have both been implicated in the control of Bcell differentiation to AFC (reviewed in [14]). We report here preliminary results obtained with a model two-step culture system which separates the priming and expression phases of B cell differentiation. These experiments show that anti-p mAb prime B cells for lymphokine-induced Ig secretion, whereas polyclonal anti-Fab, or anti-x. mAb are markedly inhibitory.
2 Materials and methods 2.1 Experimental animals Male (CBA x C57BL/lO)F1 or CBA/Ca mice aged 4-5 months were obtained from the specific pathogen-free facilities at NIMR. Spleen cells were obtained from F1 mice in most experiments except for those in which mouse anti-Igh-5a (anti-6) monoclonal antibodies were used, where CBA/Ca spleen cells were used.
2.2 B cell activators [I 102681 ~~
~
Present address: Institute of Virology, University of Wiirzburg, Wiirzburg, FRG. 0 Supported by a fellowship from Glaxo Group Research. Correspondence: Gerry G. B. Klaus, Laboratory of Cellular Immunology, National Institute for Medical Research, Mill Hill, London NW7 l A A , Great Britain
0 VCH Verlagsgescllschaft mbH, D-6940 Weinheim, 1992
Goat (GaMIg) and rabbit (RaMIg) anti-mouse Fab antibodies were affinity purified on a column of mouse Fab, followed (in the case of RaMIg) by pepsin digestion and purification of F(ab’)* fragments as previously described [15]. Rat anti-mouse Cp2, b.7.6, (IgG1) [16, 171, anti-Cpl, (AK1: IgGz,) [12], anti-x (187.1; [IS]) mAb were purified by ion exchange chromatography on Bakerbond AbX. LPS from E. coli 055:B5 was obtained from Difco, Detroit, MI. 0014-2980/92/0606-1541$3.50+ .25/0
1542
Eur. J. Immunol. 1992. 22: 1541-1545
C. Phillips and G. G. B. Klaus
2.3 Lymphokines Supernatants of the plasmacytoma cell lines X63-Ag8-653 transfected with the BMGNeo vector containing either murine IL-4 or murine IL-5 cDNA [19] were titrated for activity on the HT-2 Tcell line [20] (for IL-4) or the BCLl B cell line [21] (IL-5). One unit is defined as the dilution of supernatant required to give 50% of maximal proliferation.
2.4 B cell preparation and culture Small resting B cells were prepared by first treating spleen cells with a monoclonal anti-Thy-1 antibody (NIMR-1 provided by R.M.E. Parkhouse, NIMR, Mill Hill, London) and guinea-pig complement, and then layering the cells on a stepwise gradient of 50% : 75% :85% Percoll (Pharmacia, Uppsala, Sweden) as described previously [22]. Gradients were centrifuged at 1300 x g for 15 min and cells from the top of the 85% layer were used for culture. The resulting preparations typically contained > 95% B cells. Cultures were set up in Costar plates (Cambridge, MA) at a density of 6 x 105/mlin the presence of anti-Ig antibodies or LPS as indicated, in RPMI 1640 medium (Flow, Irvine, Scotland), plus 5% FCS supplemented with 5 x lOP5M2-mercaptoethanol, 2 mML-glutamine, 1 mM sodium pyruvate, nonessential amino acids, penicillin and streptomycin. After 24-72 h these priming cultures were washed three times, counted and cells were replated at 2 x lo5viable celldm1in the presence of lymphokines (or LPS). The success of this culture system is critically dependent on the use of LPS-free reagents for the preparation and culture of B cells. The effects of LPS contamination can be minimized by the inclusion of 10 yg/ml polymyxin B during cell preparation,which blocks the mitogenic effects of LPS [23]. However, this was not done in the experiments described here.
sented as means (+ SEM, n = 3) of IgM AFC culture, normalized to lo6 recultured cells.
3 Results 3.1 Effects of priming with polyclonal or monoclonal anti-Ig antibodies for lymphokine-induced IgM secretion In the experiment shown in Fig. 1small resting B cells were primed for 72 h with various concentrations of RaMIg, rat anti-p (b.7.6), or anti-x (187.1).The cells were then washed thoroughly and recultured for a further 72 h with 30 U/ml IL-5 plus 10 U/ml IL-4, when IgM AFC were determined. B cells cultured in medium alone gave a background level of AFC, presumably as a result of activation via FCS. As expected, RaMIg markedly inhibited this response at concentrations as low as 0.1 pg/ml: the anti-x. mAb, 187.1, had similar effects. Priming with the anti-y mAb b.7.6, had quite different consequences: low concentrations (ca. 0.1 yg/ml) also inhibited the background response. However, concentrations of 10 yg/ml or greater substantially increased the number of IgM AFC.
3.2 Kinetics of priming and Ig secretion
To examine the conditions necessary both for priming and reculture, kinetic studies of both periods were done. B cells primed for 24 or 48 h with b.7.6 did not respond above background to lymphokines, whereas those cultured for 72 h did so (Fig. 2). Again, RaMIg markedly inhibited the background response at all time periods, illustrating the powerful (and seemingly irreversible) inhibition elicited by this form of anti-Ig. Fig. 3 demonstrates that maximal numbers of IgM AFC in b.7.6-primed cultures appeared after a 72-h reculture period in IL-4 plus IL-5.
2.5 Measurement of IgM AFC IgM-secreting cells in the readout cultures were enumerated after various times in a reverse-elispot assay (relispot) as adapted from Czerkinsky et al. [24]. Briefly, 24-well tissue culture plates (Costar) were coated with 10 pg/ml GaMIg in PBS overnight at 4°C and then blocked with 5% FCS for 1 h at 37 "C. Cultured cells were washed and plated out in triplicate at lo3 cells/well in medium plus 5% FCS. The plates were incubated for 2 h at 37"C, in 5% COz, and then washed x 3 in PBS containing 0.05% Tween 20 (Sigma, Saint Louis, MO), soaked for 5 min in the third wash and washed a further five times to detach cells 0 0.1 1 10 20 0 0.1 1 10 20 0 0.1 1 completely. The plates were then incubated with biotinyRaMlg lated goat anti-mouse IgM (Southern Biotechnology, Birmingham, AL), followed by streptavidin-conjugated alkaline phosphatase (Sigma). Plates were washed five times Figure 1. Effects of anti-p mAb priming of small resting B cells on and the substrate, 5-bromo-4-choro-3-indolyl-phosphatelymphokine-induced IgM secretion. Small resting B cells were (Sigma) was added in a cold 1.02 M 2-amino-2-methyl- primed with various concentration of anti-p (b.7.6), anti-x (187.1), or F(ab')z fragments of RaMIg for 72 h, washed and recultured 1-propanol buffer containing 0.74 mM MgC126 H2 0 , 1 5 mM with IL-5 (30 Ulrnl) plus IL-4 (10 U/ml) for a further 72 h. IgM NaN3 and 0.01% Triton X-405 [23]. Relispots developed in AFC were determined and results are presented as means k SEM 1to 3 h. The plates were washed in PBS and relispots were (n = 3) of AFC/106 viable recultured cells. Similar results were counted using an inverted microscope. Results are pre- obtained in three additional experiments.
Solublc anti-y primes for B cell differentiation
Eur. J. Immunol. 1992. 22: 1541-1545
1543
3.3 Effects of anti-p on LPS-induced Ig secretion
3.4 Effects of non-mitogenic anti-p mAb
In the experiment shown in Fig. 3 a portion of the control or anti-p-primed B cells were recultured with SO pg/ml LPS, and IgM AFC were assayed at various times thereafter. It is evident that priming with b.7.6 markedly enhanced the subsequent responses to both LPS and to the lymphokine mixture.The numbers of AFC elicited by the two secondary stimuli were similar at 48 and 72 h after reculture. By day 4 the numbers of lymphokine-induced IgM AFC were beginning to decline, whereas the response to LPS continued to increase until day 5, possibly due to the strong mitogenic properties of LPS. However, at no time did the numbers of AFC produced by anti-p-primed B cells recultured with lymphokines fall below those of unprimed cells recultured with LPS. Finally, unprimed cultures produced 20-fold fewer AFC than anti-p-primed cells when recultured with 1ymphokines.
The above experiments were performed with b.7.6,which is one of the few mitogenic anti-p mAb described [16]. To investigate if the effects observed were peculiar to this antibody, or if priming necessitated the use of a mitogenic antibody, further experiments were done with a nonmitogenic anti-p, AK1 [12].These two mAb also recognize different C, domains and are of different isotypes. As shown in Fig. 4, AK1 had similar effects to those of b.7.6: thus, low concentrations caused modest inhibition of background response, while > 10 pg/ml induced significant priming for IgM secretion. These results, therefore, indicate that the priming effects of anti-p are not related to the inherent mitogenicity of the antibodies.
We next assessed the relative importance of IL-4 and IL-S in the induction of Ig secretion (Fig. 5). It is clear that IL-S
Medium anti-p RaMlg
24 hrs
3.5 Requirements for IL-4 and IL-5 in Ig secretion
48 hrs
72 hrs
r"
Figure 2. Time course for priming with anti-y antibodies. Small resting B cells were incubated for 24,48 or 72 h with medium alone, anti-p (b.7.6; 20 p g h l ) , or RaMIg (10 yglml) washed and recultured for 3 days with IL-5 and IL-4 as in Fig. 1,when IgM AFC were assayed. 50
hi 0 0.1 1 1c
CLg/mI Figure 4. The effects of a non-mitogenic anti-p mAb (AK1) on priming for IgM secretion. Small resting B cells were incubated with the indicated concentrations of AK1 or medium for 72 h, washed, and subsequently recultured with 1L-4 plus IL-5 as in Fig. 1. IgM AFC were determined after 3 days of reculture. Results are means f SEM ( n = 3).
Priming/Readout rnedium/medium o medium/lL-4 anti-phedium A anti-NIL-4
w 1.o
Figure 3. Kinetics of appearance of lymphokine-induced or LPSinduced IgM AFC in B cells primed for 3 days with anti-p. Small resting Bcells were incubated with either medium (square symbols), or 20 yglml b.7.6 (round symbols) for 72 h, washed and recultured with 50 pg/ml LPS (open symbols) or IL-4 plus IL-5 as in Fig. 1 (closed symbols). IgM AFC were assayed 1 to 6 days after reculture in each group (means f SEM, n = 3).
10
100
IL-5 U/ml
Figure 5. The requirements for JL-5 and IL-4 in the development of IgM AFC after priming with b.7.6. Small resting Bcells incubated for 72 h either with medium, or with 20 pg/ml b.7.6, were washed and recultured with various concentrations of IL-5, plus, or minus 10 U/ml IL-4. IgM AFC were determined 3 days after reculture (means f SEM, n = 3).
Eur. J. Immunol. 1992. 22: 1541-1545
C. Phillips and G. G. B. Klaus
1544
was an absolute requirement for IgM secretion as no AFC developed in its absence. Addition of IL-5 elicited a small, but significant dose-dependent increase in IgM AFC in both control and anti-y-primed cultures. Furthermore, although 10 U/ml IL-4 did not induce IgM secretion on its own, its inclusion in cultures of anti-y-primed cells with optimal concentrations of IL-5 increased the number of IgM AFC 10-fold. These findings are consistent with those of other workers [25, 261 who demonstrated that IL-4 and IL-5 synergize to promote Ig secretion. IL-4 only minimally increased the number of AFC developing from unprimed B cells recultured with IL-5.
3.6 Effects of continued presence of anti-p during the reculture period The stimulatory effects obtained with anti-y are in apparent conflict with those of earlier investigations on the effects of anti-Ig on Ig secretion [4-91. We, therefore, designed further experiments to attempt to resolve these discrepancies. In the experiment summarized in Fig. 6 some anti-y (or RaM1g)-primed cultures were not washed before the addition of lymphokines on day 3 and assay on day 6. The continued presence of b.7.6 throughout the culture period abolished the priming effect and indeed reduced the background significantly.To rule out the possibility that this effect was due to exhaustion of the medium, additional primed cultures were washed on day 3 and b.7.6 or RaMIg were added back along with the lymphokines during the 3-day reculture period. This again abrogated the priming effect. Finally, it is noteworthy that RaMIg was powerfully inhibitory under all these experimental conditions. This experiment, therefore, confirms that the continued presence of anti-y inhibits lymphokine-induced Ig secretion in activated B cells. Gp: Washed Read-out A
B
+
{
IL-4 IL-5
+
Priming:
anti-lg
Medium anti-p RaMlg
I 0
I 50
I 100
IgM AFC x 10-3/1 O6 Cells Figure 6. The effects of the continued presence of anti-p during the readout cultures. Small resting B cells were incubated either with medium, 20 pg/ml b.7.6, or RaMIg (10 pg/ml) for 72 h. Cells in group A were washed and recultured with IL-5/IL-4 for 3 days as usual. Group B was also washed, but recultured with IL-5/IL-4 in the presence of the same concentrations of antibodies as those used in the priming culture. Group C was not washed, but IL-5 and IL-4 were added to the primary culture on day 3. IgM AFC were determined 3 days after reculture. (means f SEM, n = 3).
4 Discussion The results of these experiments indicate that soluble anti-y mAb effectively prime quiescent B cells for both lymphokine- and LPS-induced IgM secretion, provided that they are washed out of the readout cultures. This effect is not demonstrably dependent on the isotype of the mAb, on the domain of the y chain recognized, nor on the inherent mitogenicity of the mAb. Effective priming requires prolonged exposure (72 h) to relatively high concentrations of the mAb (10-50 yg/ml), while lower concentrations (< 1 yg/ml) tend to be inhibitory (Figs. 1and 4). Both IL-4 and IL-5 were required in the readout cultures to generate optimal responses (Fig. 5). These effects are in complete contrast to those of anti-?t mAb, or polyclonal F(ab’)z rabbit anti-Fab which both cause powerful (and possibly irreversible) inhibition of lymphokine-induced IgM secretion. The lymphokine-induced IgM production by anti-p-primed B cells compared very favorably with that induced by LPS (Fig. 3). Until day 2 after reculture of anti-y-primed B cells, the numbers of AFC induced by IL-5 plus IL-4 by primed B cells were identical to those induced by LPS. By day 4, the lymphokine-induced response began to decline, whereas the LPS-induced response increased until day 5. The differences in kinetics and magnitude between the response to LPS (Fig. 3), versus that to IL-5 plus IL-4 are probably due to the greater inherent mitogenicity of LPS. However, the capacity of IL-4 to promote isotype switching from IgM, to IgGl and IgE may also be relevant. This is currently under investigation and will be reported elsewhere (C. Phillips and G. G. B. Klaus in preparation). It is also possible that additional stimuli (lymphokines?) may be able to enhance the lymphokine-induced response further. These results demonstrate that small resting B cells can be induced to secrete IgM after appropriate triggering of sIgM receptors, followed by restimulation with IL-5 and IL-4. This response does not apparently require B cell membrane contact with T helper cells: this implies that prolonged ligation of sIgM receptors followed by these two lymphokines are sufficient to induce T cell differentiation to AFC. This may be relevant to the capacity of T-independent antigens to induced Ig secretion, a process in which sIgM apparently plays a greater role than that of sIgD [27, 281. The results with anti-% and RaMIg further suggest that cross-linking both sIgM and sIgD receptors generates a more powerful inhibitory signal in resting B cells, which is not overcome by washing the cells and restimulating them with lymphokines. Further experiments have indeed shown that anti-6 mAb do not prime B cells for Ig secretion, but instead inhibit priming by anti-y (C. Philipps and G. G. B. Klaus in preparation). These findings shed some further light on what determines whether anti-Ig enhances or inhibits Ig secretion in activated B cells. As illustrated in Fig. 6 if anti-y is left in the cultures during their stimulation with lymphokines it is inhibitory. This observation is, therefore, in line with those of various earlier studies, using either anti-y mAb, or polyclonal anti-Ig antibodies plus LPS [4-9].This inhibition has been shown to be largely due to lack of posttranscriptional modification of mRNA of the y chain to the secreted form, as the membrane form remains unaffected
Eur. J. Immunol. 1992. 22: 1541-1545
[29, 301. It is not clear why only soluble anti-Ig has this effect, whereas Sepharose-bound anti-Ig does not [12]. Further studies are required to evaluate the effects of anti-p in T cell-dependent B cell activation. It is now well established that contact-mediated signals are involved in the initial phase of T-B interaction (reviewed in [31]). In line with this B cells can be induced to secrete Ig by contact with histoincompatible activated T cells [32-341 possibly through contact with surface molecules other than Ia [35, 361 plus lymphokines. Indeed, membranes of activated T cells induce substantial proliferation of small resting B cells, but only stimulate Ig secretion in the presence of lymphokines [36]. Stimulation with anti-Ig can enhance proliferation [35] and differentiation of B cells [37] in response to activated T cells, but the enhancement disappears at high T : B cell ratios or in the presence of more “efficient” Tcells [38]. Thus, membrane contact with appropriate numbers of activated T cells and activation via sIgM cross-linking may represent alternative mechanisms by which B cells can become competent to secrete Ig in response to lymphokines. It appears that these mechanisms can function either alone or in tandem. Other workers have shown that membrane contact with activated T cells is sufficient to induce this competency in small resting B cells. However, this is the first report that soluble anti-p mAb can also induce this competency independent of T cell membrane contact. We thank Mary Holrnan for her invaluable assistance with some of’ these experiments, and Anneliese Schirnpl for helpful discussion. We are also grateful to E Melchers, Chris Heusser and David Scott,for gift.7 of cell lines. Received January 13, 1992; in revised form February 24, 1992
5 References 1 Parker, D. C., Nature 1975. 258: 361. 2 DeFranco, A. L., Raveche, E. S., Asofsky, R. and Pau1.W. E . , J. Exp. Med. 1982. 155: 1523. 3 Anderson, J., Bullock, W. W. and Melchers. F., Eur. J. Irnrnunol. 1974. 4: 715. 4 Kearney, J. F., Klem, J., Bockman. D. E., Cooper, M. D. and Lawton, A. R., J. Irnrnunol. 1978. 120: 158. 5 Flahart, R. E. and Lawton, A. R., J. Exp. Med. 1987.166: 864. 6 Lawton, A. R.,Summerlin,V. S. and Dooley, J. S., J. Irnrnunol. 1990. 145: 3177. 7 Flahart. R. E. and Lawton, A. R., Mol. Cell. Irnrnunol. 1987. 3: 61. 8 Yuan, D., Mol. Cell. Irnrnunol. 1987. 3: 133. 9 Chen, U., Gene 1988. 72: 209.
Soluble anti-k primes for B cell differentiation
1545
10 Parker, D. C., Irnrnunol. Rev. 1980. 52: 115. 11 Birkeland, M. L., Simpson, L., Isakson, P. C. and Pure, E., I . Exp. Med. 1987. 166: 506. 12 Leptin, M., Eur. .I. Irnrnunol. 1985. 15: 131. 13 DeFranco, A. L., Kung, J. T. and Paul, W. E., Irnrnunol. Rev. 1982. 64: 161. 14 Klaus, G . G. B., Hawrylowicz, C. M. and Carter, C. J., Irnrnunol. 1985. 55: 411. 15 Hawrylowicz, C. M., Keeler, K. D. and Klaus, G. G. B., Eur. J. Irnrnunol. 1984. 14: 244. 16 Julius, M. H., Heusser, C. H . and Hartmann, K.-U., Eur. J. Irnrnunol. 1984. 14: 753. 17 Leptin, M., Potash, M. J., Grutzmann, R., Heusser, C., Shulman, M., Kohler, G. and Melchers, F., Eur. J. Irnrnunol. 1984. 14: 534. 18 Yelton, D. E., Desaymard, C. and Scharff, M. D., Hyhridorna 1981. I : 5 . 19 Karasuyama, H. and Melchers, F., Eur. J. Irnrnunol. 1988. 18: 97. 20 Smith, C. A. and Rennick, D. M., Proc. Natl. Acad. Sci. USA 1986. 83: 1857. 21 Kinashi,T., Harada, N., Severinson, E.,Tanabe,T., Sideras, P., Konishi, M., Azuma, C., Tominaga, A., Bergstedt-Lindqvist, S., Takahashi. M., Fumihiko, M.. Yaoita. Y , Takatsu, K. and Honjo, T., Nature 1986. 324: 70. 22 Klaus, G . G . B., Hawrylowicz, C. M. and Carter, C. M., Irnrnunology 1985. 55: 411. 23 Jacobs, D. M. andMorrison, D. C.,J. Exp. Med. 1975.14: 1453. 24 Czerkinsky, C. C.,Tarkowski, A., Nilsson, L.-A., Ouchertony, O., Nygren, H . and Gretzer, C., J. Irnrnunol. Methods 1984. 72: 489. 25 Murray, l? D., McKenzie, D.T., Swain, S. L. and Kagnoff, M. F., J. Irnrnunol. 1987. 130: 2669. 26 Coffman, R. L., Seymour, B. W. l?, Lebman, D. A., Hiraki, D. D., Christiansen, J. A., Shrader, B., Cherwinski, H. M., Savelkoul, H . F. J., Finkelman, F. D., Bond, M. W. and Mosmann,T. R., Irnrnunol. Rev. 1988. 102: 5. 27 Cambier, J. C., Ligler, F. S., Uhr, J. W., Kettman, J. R. and Vitetta, E. S., Proc. Natl. Acad. Sci. USA 1978. 75: 432. 28 Zitron, I. M., Mosier, D. E. and Paul,W. E., J. Exp. Med. 1977. 146: 1707. 29 Chen-Bettecken, U.,Wecker, E. and Schimpl, A., Proc. Natl. Acad. Sci. U S A 1985. 82: 7384. 30 Hogbom, E., Martensson, I.-L. and Leanderson, T., Proc. Natl. Acad. Sci. U S A 1987. 84: 9135. 31 Vitetta, E . S., Fernandez-Botran, R., Meyers, C. D. and Sander,V. M., Adv. Irnrnunol. 1989. 45: 1. 32 Riedel, C., Owens, T. and Nossal, G. J. V., Eur. J. Irnrnunol. 1988. IS: 403. 33 Tite, J. l?, Kaye, J. and Jones, B., Eur. J. Irnrnunol. 1984. 14: 553. 34 Julius, M. H. and Rammensee, H.-G., Eur. J. Irnrnunol. 1988. 18: 375. 35 Whalen, B. J. ,Tony, H. P. and Parker, D. C., J. Irnrnunol. 1988. 141: 2230. 36 Hodgkin, l? D.,Yamashita, L. C., Coffman, R. L. and Kehry, M. R., J. Irnrnunol. 1990. 145: 2025. 37 Julius, M. H.,Von Boehmer, 13. and Sidman, C. L.. Proc. Natl. Acad. Sci. USA 1982. 79: 1989. 38 Pereira, l?, Forsgren, S., Portnoi, D., Bandcira, A , , MartinezA., C:. and Coutinho, A . , Eur. J. Imrnunol. 1986. 16: 355.
