Proc. Nati. Acad. Sci. USA Vol. 82, pp. 8153-8157, December 1985 Immunology Interleukin 1 can act as a B-cell growth and differentiation factor (cytokines/hapten-gelatin fractionation/T-cel-independent antigens/antibody-forming cell dones/enzyme-linked immunosorbent assay) BEVERLEY L. PIKE AND G. J. V. NOSSAL The Walter and Eliza Hall Institute of Medical Research, Post Office, Royal Melbourne Hospital, Victoria 3050, Australia Contributed by G. J. V. Nossal, July 23, 1985

ABSTRACT Splenic B lymphocytes specifically reactive to B cell, the bioactivities exhibited by IL-1 were identical to the hapten fluorescein (FLU) were prepared from nonimmune those in the filler cell-free system. adult mice by affinity fractionation on hapten-gelatin. These FLU-specific B cells were cultured as single cells or in small numbers in 10-ul wells either in the absence of any feeder, MATERIALS AND METHODS filler, or accessory cell or in the presence of 3T3 fibroblasts Mice and Preparation of Fluorescein (FLU)-Specific Splenic acting as filer cells. A selected batch ofa "T-cell-independent" B Cells. Specific-pathogen-free CBA/CaH/Wehi mice were antigen, FLU-Ficoll, which induces growth and differentiation used as spleen donors at 8-10 weeks of age. Hapten-specific only in the presence of lymphokines or cytokines acting as B-cell populations were prepared from spleen cell suspen- B-cell growth and differentiation factors (BGDF), was used as sions by fractionation on FLU-gelatin as described (15-17). the antigenic stimulus. It was found that murine interleukin 1 Adherent FLU-gelatin was removed from the recovered prepared by recombinant DNA technology was an effective, binding cells by collagenase. The binding population is 97% although weak, BGDF when acting with antigen on B cells B cells, -z70% FLU-binding, and -200-fold enriched for in cultured either under filler cell-free conditions or in the vitro reactivity to FLU conjugates (13, 16, 17). presence of3T3 cells. When the murine interleukin 1 was used Antigen. The hapten FLU was coupled onto aminoethyl- in combination with recombinant human interleukin 2, itself a carbonylmethylated Ficoll (AECM53-Ficoll), as described weak but effective BGDF in the system, an additive effect was (16, 17), and the FLU53-Ficoll was used as a final concen- observed. The results challenge the notion that interleukin 1 is tration of 0.1 ng/ml. exclusively or even primarily an activating cytokine. This EL4 Thymoma Cell-Derived BGDF. A lOx concentrate of system, in which pure factors are able to act with specific medium conditioned by concanavalin A-stimulated EL4 antigen on single hapten-specific B cells, will prove helpful for thymoma cells prepared as described (12) was used as a the further dissection of the respective roles of the various source ofT-cell-derived BGDF at a final concentration of5% factors that can act on B cells. (vol/vol). According to the recently proposed convention (18) this is termed EL-BGDF-pik. The important macrophage-derived growth-regulatory mol- Recombinant Murine IL-1 (r-mu-IL-1). The r-mu-IL-1 was ecule interleukin 1 (IL-1) (1) has recently been produced in prepared as described (2) and kindly provided by Hoff- pure form through recombinant DNA technology (2). Ever mann-La Roche (lot 11319-229-48). The concentration of since the description of IL-1 as a lymphocyte-activating r-mu-IL-1 is expressed as units/ml of IL-1 activity as deter- factor (3, 4), its bioactivity has been assumed to relate mined by the providers, using the thymocyte proliferation primarily if not exclusively to early events in the activation assay (19). of a resting, Go, cell. For example, IL-1 is seen as a cofactor Recombinant Human IL-2 (r-hu-IL-2). The r-hu-IL-2 was required for concanavalin A to exert its activating effects on prepared as described (20) and kindly provided by Cetus T lymphocytes (5) with concomitant expression of IL-2 Immune (Palo Alto, CA) (lot LP-222). Activity is expressed receptors, after which IL-2 is seen as the only growth factor as units/ml of r-hu-IL-2 as determined by using a murine required for further proliferation ofthe T-cell clone. IL-1 has CTL-L line as described (21). also been implicated in the activation of B lymphocytes B-Cell Cloning Systems. FLU-specific B cells were cultured (6-10), though in this instance, some work has suggested that in 60-well Terasaki trays at a mean of0.4-12 cells per well in it acts later than another postulated activating factor, B-cell- 10 ,ud of RPMI 1640 medium supplemented with 5% (vol/vol) stimulating factor 1 (BSF-1) (6). fetal calf serum and 100 AM 2-mercaptoethanol, either in the We have reported (11-14) a system in which normal murine absence of any added filler cells or in the presence of 300 B lymphocytes, preselected on the basis of their antigen BALB/c 3T3 fibroblast cells as described (13, 14, 17). For specificity, are able to be stimulated to proliferate and some purposes-e.g., the delineation of factor concentra- differentiate to immunoglobulin secretion status when cul- tion-response characteristics-an oligoclonal approach using tured singly in the presence of antigen and a source of B-cell a mean of 10 B cells per well was used as described (14). The growth and differentiation factors (BGDF). Under these additional presence of filler cells markedly improves both conditions, the B cell itself is the only possible target for the cloning efficiency and clone size (13, 14, 17). Routinely, 300 action of antigen and factors. In the belief that this system 3T3 cells were dispensed into the culture wells in 5 Al of was ideal for testing various purified factors for bioactivity on medium prior to the addition of the B cells to allow time for B cells, we have used it to assess the bioactivity of murine adherence. To avoid intertray variance, B cells are dispensed IL-1, prepared by recombinant DNA technology (r-mu-IL-1). into the wells of all trays in 5 1.d of medium containing twice The results show IL-1 to have a combination ofbioactivities, the required concentration of antigen. For filler cell-free including the promotion ofboth growth and differentiation of B cells. Moreover, when filler cells were added to the single Abbreviations: FLU, fluorescein; AFC, antibody-forming cell; BSF- 1, B-cell-stimulating factor 1; BGDF, B-cell growth and differenti- ation factor(s); BCGF, B-cell growth factor; IL-1 and IL-2, inter- The publication costs ofthis article were defrayed in part by page charge leukin 1 and 2; r-mu-IL-1, recombinant murine IL-1; r-hu-IL-2, payment. This article must therefore be hereby marked "advertisement" recombinant human IL-2; SAM, sheep anti-murine immunoglobulin in accordance with 18 U.S.C. §1734 solely to indicate this fact. antibody. 8153 Downloaded by guest on September 27, 2021 8154 Immunology: Pike and Nossal Proc. Natl. Acad Sci. USA 82 (1985)

conditions, a further 5 ul of medium alone was added. Number of B cells per well Factors were added in a further 1 /.l of medium at lix the 0.4 0.6 0.8 1.0 1.2 1 required final concentration. Assessment of Clonal Proliferation. Before assay for anti- body formation, culture wells were examined, by using an inverted phase microscope as described (12, 13, 17), for the presence or absence of a proliferating B-cell clone. In the presence of 3T3 cells, the proliferating B cells could usually be distinguished, due to the adherence to plastic, the different morphology, and the substantially larger size ofthe 3T3 cells. Assessment of Antibody Formation. Antibody formation was assessed by using a sensitive enzyme-linked immuno-

sorbent assay (ELISA) procedure as described (14). As the In target B cells were selected for specificity for FLU by a) prefractionation on hapten-gelatin, a sheep anti-murine 0) immunoglobulin antibody (SAM) was used as the polyvalent GJ- antigen in the ELISA assay rather than to FLU coupled cmC bovine serum albumin. This change was to measure antibody a) production per se, regardless of its affinity, as a result of particular stimuli. Briefly, the supernatant fluid of each culture well was individually transferred into the wells of a 96-well flexible U-bottomed polyvinyl plate (Dynatech, Alexandria, VA) precoated with affinity-purified SAM (Silenus Laboratories, Dandenong, Australia) at 3 ,ug/ml and containing 50 /.l of0.3% skim milk powder and 0.05% Tween 20 in phosphate-buffered saline. After a period of >2 hr at room temperature, the plates were washed in PBS-tween 20 and horseradish peroxidase-coupled SAM was added for a period of >4 hr. After washing, the substrate 2,2'-azinobis(3- ethylbenzthiazolinesulfonic acid) (ABTS) (Sigma) was added and the absorbance ofthe fluid in the wells was read 1 hr later with a Titertek Multiscan ML (Flow Laboratories) using dual wavelengths (414 and 492 nm). A well was considered positive if its absorbance exceeded the mean ± 3 SEM of the FIG. 1. Limiting dilution analysis ofthe response ofFLU-specific background as calculated on the basis of a large number of B cells as assessed by AFC clone formation by ELISA. (A) Response replicates of supernatants from wells lacking B-cell input. to FLU-Ficoll and EL-BGDF-pik (x-x) (frequency value, 46.3% ± Control wells lacking antibody consistently gave an absorb- 12.5%). (B) Response to FLU-Ficoll alone (x-x) (frequency value, 2.76% ± 0.6%) or FLU-Ficoll plus r-mu-IL-1 at 100 units/ml (a-0) ance value of <0.010 unit. The frequency of antibody- (frequency value, 9.75% ± 1.54%). The 95% confidence limits are forming cell (AFC) clone precursors was determined by indicated by broken lines. Poisson analysis as described (12-15, 22). Data for superna- tants from wells of oligoclonal cultures are expressed as the mean ± SEM of the absorbance of 12 replicate wells. definite BGDF activity, its addition raising the cloning efficiency and final antibody amount above background, but RESULTS only to about one-quarter the level of EL-BGDF-pik. The relative effects are further illustrated by Table 2, which Synergy of IL-1 and Antigen in Promotion of AFC Clone shows the pooled results of eight experiments in which Formation by Single Hapten-Specific B Cells in the Presence of antibody formation with FLU-Ficoll alone, FLU-Ficoll plus 3T3 Filler Cells. Fig. 1 shows a typical limiting dilution r-mu-IL-1 at 100 units/ml, and FLU-Ficoll plus EL-BGDF- analysis in which various numbers of FLU-specific B cells pik were directly compared. The results confirm the data of were stimulated with FLU-Ficoll plus EL-BGDF-pik (Fig. Table 1 and add the further point that the amount of antibody lA) or FLU-Ficoll in the absence and presence of r-mu-IL-1 formed in the presence of factors is greater than that of the at 100 units/ml (Fig. 1B). While r-mu-IL-1 was clearly not as occasional clones arising in their absence. strong a BGDF as the mixture of lymphokines contained within EL-BGDF-pik, a significant elevation of AFC clone Table 1. Dose-response of BGDF activity of r-mu-IL-1 acting frequency was seen when both antigen and IL-1 were with FLU-Ficoll on FLU-specific B cells in the presence present, as opposed to antigen alone. Furthermore, the of 3T3 cells response followed the expected Poisson distribution, indi- cating that only the responder B cells are limiting in the r-mu-IL-1, Absorbance x 103/ system. Thus IL-1 can synergize with antigen to promote units/ml % AFC clones* 10 B cells AFC clone development from single B cells. Proliferation as 0 4.02 49 observed by microscopic visualization correlated with AFC 1 12.9 211 clone formation (data not shown). The dose-response char- 10 12.9 155 acteristics of the BGDF activity of r-mu-IL-1 are shown in 100 12.4 440 Table 1. Preliminary screening was performed under oligoclonal conditions, using an average of 10 FLU-specific EL-BGDF-pik 53.1 1470 B cells per well, then the findings were confirmed by clonal FLU-specific B cells were stimulated for 5 days as indicated in the analysis. With clonal analysis, a flat dose-response relation- first column prior to analysis of individual culture supernatants for ship was noted, whereas oligoclonal analysis showed a the presence of AFC clones on SAM by an ELISA. modest increase in the amount of antibody with r-mu-IL-1 at *Values shown represent AFC clone frequencies as determined from 100 units/ml. In both systems, r-mu-IL-1 exhibited weak but two directly comparative experiments. Downloaded by guest on September 27, 2021 Immunology: Pike and Nossal Proc. Natl. Acad. Sci. USA 82 (1985) 8155 Table 2. BGDF activity of r-mu-IL-1 when acting with antigen Table 3. BGDF activity of r-mu-IL-1 acting with antigen on on single FLU-specific B cells in the presence of 3T3 filler cells single FLU-specific B cells in the absence of fillers Factor(s) % AFC Absorbance x 103 r-mu-IL-1, % AFC Absorbance x 103/ added clones P* Per 10 B cells Per clone units/ml clones P* 10 B cells None 2.65 ± 0.53 - 56 ± 17 245 ± 79 0 0.99±0.17 7± 2 r-mu-IL-1 10.6 ± 1.99 <0.0005 457 ± 67 464 ± 42 1 1.41 ± 0.22 NS 12 ± 4 EL-BGDF-pik 37.2 ± 5.7 <0.0005 1550 ± 233 482 ± 67 10 2.80 ± 0.63 <0.025 30 ± 10 100 3.59 ± 0.54 <0.0025 98 ± 30 r-mu-IL-1 was 100 units/ml when present. Values represent mean 1000 2.35 ± 0.98 NS 29 ± 8 ± SEM ofeight directly comparative experiments. Highly significant elevation of the response to antigen alone was seen by the further addition of either IL-1 or EL-BGDF-pik. The response to antigen EL-BGDF-pik 13.7 ± 3.20 <0.01 198 ± 35 plus EL-BGDF-pik was significantly greater than that ofantigen plus FLU-specific B cells were stimulated with FLU-Ficoll at 0.1 ng/ml r-mu-IL-1 (P < 0.025). in the presence of r-mu-IL-1 or EL-BGDF-pik as indicated, under *Determined by Student's t test and representing difference from filler cell-free conditions. Values are mean ± SEM from four or five response with antigen alone. experiments. *Determined by Student's t test and representing difference from BGDF Activity of r-mu-IL-l Acting with Antigen on Single response with antigen alone. NS, not significant. Hapten-Specific B Cells in the Absence of Filler Cells. The ability of IL-1 to act as a BGDF on single hapten-specific B delayed for up to 36 hr (data not shown). Both the number of cells stimulated with specific antigen in the absence offillers clones stimulated and the total amount of antibody synthe- or any other cell type was investigated. FLU-specific B cells sized per input B cell were considerably smaller with IL-1 or were stimulated with FLU-Ficoll and various concentrations IL-2, alone or together, than with EL-BGDF-pik. of r-mu-IL-i. Both oligoclonal and clonal analyses were performed (Table 3). As expected, far fewer AFC clones developed in the DISCUSSION absence offillers, and the absorbance values were also lower. The results presented in this paper have clearly shown that a Dose-response analysis showed that 10 or 100 units/ml had soluble macrophage-derived factor, IL-1, prepared by significant BGDF activity. One unit/ml was clearly recombinant DNA technology, can act in synergy with suboptimal and 1000 units/ml was clearly supra-optimal, in specific antigen to promote growth and differentiation contrast with the rather flat dose-response relationship in the amongst single, isolated hapten-specific B cells and can do so presence of fillers (Table 1). Proliferation as scored micro- in the absence ofany other lymphokine or any other cell. This scopically correlated almost exactly with antibody forma- result stands in contrast to the concept of IL-1 as purely an tion, all proliferating clones forming some antibody (data not "activating" cytokine. shown). While no consensus exists about the respective roles ofthe AFC Clone Development in the Presence of Both IL-1 and various growth factors capable of affecting B lymphocytes IL-2. Our recent studies have shown that IL-2 can act on (6-10, 23-31), much of the literature on "T-independent" single B cells to promote growth and differentiation (14, 21) B-cell activation (e.g., ref. 8) supports a scheme shown in when acting under either filler cell-free (21) or 3T3 cell- Fig. 2A. A small, resting B cell is seen as being activated from supported conditions (14). It has been reported that exposure its Go state by the conjoint action of multivalent antigen (or to IL-1 is a prerequisite for T lymphocytes to be able to anti-p heavy chain antibody) and an activation factor, seen respond to IL-2 (5). Furthermore, it has been proposed that by some as an IL-1-like molecule (8) or by others as a resting B cells require a similar signal from IL-1 before they particular T-cell-derived lymphokine [e.g., BSF-1 (27, 29)]. are able effectively to respond to BGDF (7, 8). The possibility This activated B cell displays an increased amount of surface that IL-1 and IL-2 may exhibit some synergy in their action Ia antigens (30-32), an increased amount ofRNA and protein on B lymphocytes was therefore investigated. FLU-specific synthesis (30, 32, 33), and a slight increase in size (30, 33). Its B cells were stimulated with antigen in the absence of added activated state is referred to as the Go or GlA (33) growth factors, or with 100 units/ml ofeither r-mu-IL-1 or r-hu-IL-2, state. At this stage, the cell is postulated to display receptors or both, and the AFC clone frequency was determined by for one or more B-cell growth factors, which drive the cell ELISA after 5-6 days. Rather than a synergistic effect, an into active cycle. After an undefined number ofdivisions, the additive effect was achieved by the addition ofboth cytokines cells are postulated to display receptors for one or more (Table 4). Furthermore, and somewhat surprisingly, the differentiation factors, which encourage differentiation amount of antibody formed per clone was not significantly towards active secretory status (6, 8, 23-26, 29). greater when both lymphokines were present. No difference While much effort has gone into the purification of factors in the response was noted when the addition of the IL-2 was based on assay systems that seek to force each factor into one

Table 4. BGDF activity of IL-1 and IL-2 acting with FLU-Ficoll on FLU-specific B cells Filler cell-free* 3T3 fillerst Factor(s) added % AFC clones Absorbance x 103/10 B cells % AFC clones Absorbance X 103/10 B cells None 0.46 7 2.61 ± 0.57 95 ± 46 r-mu-IL-1 1.75 21 9.34 ± 1.79 486 ± 110 r-hu-IL-2 1.30 20 7.75 ± 1.28 330 ± 126 IL-1 + IL-2 3.24 64 20.4 ± 4.31 1060 ± 230 EL-BGDF-pik NT 164 29.2 ± 6.14 1950 ± 510 FLU-specific B cells were stimulated with FLU-Ficoll in the presence or absence of 100 units/ml of the factors as indicated, either under filler cell-free conditions or in the presence of 300 3T3 filler cells. *Results of a typical experiment. NT, not tested. tMean ± SEM of three experiments. Downloaded by guest on September 27, 2021 8156 Immunology: Pike and Nossal Proc. Natl. Acad Sci. USA 82 (1985)

The current paradigm of B lymphocyte activation A

Small Go Early GI Activated B blast Antibodyforming B cell B cell B blas cone tL:e

LI'~~~~~~ Antigen plus BCGF receptors appear BCDF receptors appear 'Activation'cytokine BCGF acts to promote growth BCDF acts to promote differentiation to secretory status An alternative model of B lymphocyte activation B

Small Go Early GI Activated B blast Antibody-forming B cell B cell B blast clone clone.

FIG. 2. Alternative views on the roles of monokines and lymphokines in B-lymphocyte activation.

ofthese three conceptual categories, each acting at a specific suggest the possibility that any one of several factors/ stage in the differentiation pathway, the model has not gone cytokines can promote the full sequence (though perhaps unchallenged. Thus, it has recently been claimed that BSF-1 only on particular subsets of B cells). This is clear for IL-2 (formerly termed BSF-pl and before that BCGF-I) is not (21), IL-1 (present paper, ref. 9), gel filtration and chromato- primarily a growth factor but rather is an activation factor (27, graphic fractions of EL-BGDF-pik (12), and also highly 29). Furthermore, recent evidence (34) suggests that B-cell purified human BCGF-II and B-cell differentiation factor growth factor II (BCGF-II) (28) can promote both growth and (BCDF) (reviewed in refs. 24 and 25) kindly provided by T. secretion among activated B cells. Similarly, the believed Kishimoto (data not shown). A possible heterogeneity among differentiation factor from the hybridoma B151K12 (26) now B cells is suggested by the findings that the frequency of appears to have characteristics very similar to BCGF-II (35). clones generated with optimal amounts of either IL-1 or IL-2 Our own studies using single hapten-specific B cells, is lower than that found when both were used (Table 4), in reinforced by the present study, do not support the current which case an additive effect was noted. More direct inves- scheme as shown in Fig. 2A. Rather, the data demonstrate tigation of B-cell heterogeneity in factor responsiveness will that a single factor has all three of the claimed activities: soon be feasible. This will require purification of all relevant promotion of the Go to G1 transition, promotion of cell B cell-active factors, perhaps five or more; identification of division, and promotion of differentiation to secretory status respective cellular receptors, and development of anti-re- (12-14, 17, 21). We therefore propose an alternative scheme ceptor monoclonal antibodies, combined with multiparame- of "T-independent" activation as outlined in Fig. 2B. We ter cell sorter analysis and appropriate clonal assays. Downloaded by guest on September 27, 2021 Immunology: Pike and Nossal Proc. Natl. Acad. Sci. USA 82 (1985) 8157 In contrast to claims for T-cell activation (5), preliminary 7. Corbel, C. & Melchers, F. (1984) Immunol. Rev. 78, 51-74. kinetic studies did not favor the view that prior exposure of 8. Melchers, F. & Andersson, J. (1984) Cell 37, 715-720. cells to antigen plus IL-1 increases the response to IL-2 in the 9. Booth, R. J. & Watson, J. D. (1984) J. Immunol. 133, system. The of increased IL-2 receptor 1346-1349. single-cell question 10. Liebsen, H. J., Marrack, P. & Kappler, J. W. (1982) J. Immu- expression remains to be explored. nol. 129, 1398-1402. If IL-1 were predominantly an activating factor helping 11. Vaux, D. L., Pike, B. L. & Nossal, G. J. V. (1981) Proc. Natl. B cells to move from Go to G1, one might anticipate that it Acad. Sci. USA 78, 7702-7706. would exert its maximum effects on small B lymphocytes and 12. Pike, B. L., Vaux, D. L., Clark-Lewis, I., Schrader, J. W. & be less effective on large B cells. We have reported (13) that Nossal, G. J. V. (1982) Proc. Natl. Acad. Sci. USA 79, B cells slightly larger than median yield the highest cloning 6350-6354. frequency when EL-BGDF-pik is used as a factor source. 13. Pike, B. L., Vaux, D. L. & Nossal, G. J. V. (1983) J. Immu- This poor responsiveness of small B cells could be due to the nol. 131, 544-560. 14. Pike, B. L. & Nossal, G. J. V. (1985) Proc. Natl. Acad. Sci. absence of an IL-1-like activating factor in the EL-BGDF- USA 82, 3395-3399. pik. Preliminary data (not shown) argue against this expla- 15. Nossal, G. J. V. & Pike, B. L. (1976) Immunology 30, nation. The addition of optimal amounts of IL-1 does not 189-202. augment the EL-BGDF-pik response. Furthermore, the re- 16. Nossal, G. J. V., Pike, B. L. & Battye, F. L. (1978) Eur. J. sponse oflarge cells to FLU-Ficoll and IL-1 far exceeded that Immunol. 8, 151-157. of small cells, although the proportionate increase above the 17. Pike, B. L. & Nossal, G. J. V. (1984) J. Immunol. 132, "antigen alone" background was somewhat greater for the 1687-1695. small cells. The capacity ofIL-1 to enhance the performance 18. Paul, W. E. (1983) Eur. J. Immunol. 13, 956. of large cells certainly substantiates that it is more than just 19. Mizel, S. B., Dukovich, M. & Rothstein, J. (1983) J. Immunol. 131, 1834-1837. a cycle-initiating monokine. 20. Rosenberg, S. A., Grimm, E. A., McGrogan, M., Doyle, M., The finding that IL-1 can act with antigen to promote Kawasaki, W., Koths, K. & Mark, D. F. (1984) Science 223, activation, division, and differentiation of B cells was sur- 1412-1415. prising, and it suggests that all claims for factor synergies 21. Pike, B. L., Raubitschek, A. & Nossal, G. J. V. (1984) Proc. should really be confirmed in a system in which a single B cell Natl. Acad. Sci. USA 81, 7917-7921. is the unequivocal target of factor action. So far, no mixture 22. Good, M. F., Boyd, A. W. & Nossal, G. J. V. (1983) J. of pure lymphokines has been as effective in our hands as Immunol. 130, 2046-2055. EL-BGDF-pik. It thus remains a challenge to explore pos- 23. Howard, M., Nakanishi, K. & Paul, W. E. (1984) Immunol. sible synergistic interactions between factors, which can be Rev. 78, 185-210. 24. Kishimoto, T. (1985) Annu. Rev. Immunol. 3, 133-157. done at the level of the single B cell as more pure molecules 25. Kishimoto, T., Yoshizaki, K., Kimoto, M., Okada, M., become available. Kuritani, T., Kikutani, H., Shimizo, K., Nakagawa, T., Nakagawa, T., Miki, Y., Kishi, H., Fukunaga, K., Yoshibuko, The r-mu-IL-1 used in these studies was a generous gift of T. & Taga, T. (1984) Immunol. Rev. 78, 97-118. Hoffmann-La Roche, obtained through the courtesy of Drs. P. T. 26. Takatsu, K., Tanaka, K., Tominaga, A., Kumahar, Y. & Lomedico, A. Stem, and S. Mizel. The r-hu-IL-2 used was a Hamaoka, T. (1980) J. Immunol. 125, 2646-2653. generous gift ofCetus Immune (Palo Alto, CA), obtained through the 27. Rubin, E. M., Ohara, J. & Paul, W. E. (1985) Proc. Natl. courtesy of Dr. A. Raubitschek. The excellent technical assistance Acad. Sci. USA 82, 2935-2939. of Maureen Zanoni and Mandy Ludford is gratefully acknowledged. 28. Swain, S. L., Howard, M., Kappler, J., Marrack, P., Watson, This work was supported by the National Health and Medical J., Booth, R., Wetzel, G. D. & Dutton, R. W. (1983) J. Exp. Research Council, Canberra, Australia; by Grant AI-03958 from the Med. 158, 822-835. U.S. National Institute of Allergy and Infectious Diseases; and by 29. Oliver, K., Noelle, R. J., Uhr, J. W., Krammer, P. H. & the generosity of a number ofprivate donors to the Walter and Eliza Vitetta, E. S. (1985) Proc. Natl. Acad. Sci. USA 82, Hall Institute. 2465-2467. 30. Roehm, N. W., Liebson, H. J., Zlotnik, A., Kappler, J., 1. Mizel, S. B. & Mizel, D. (1981) J. Immunol. 126, 834-836. Marrack, P. & Cambier, J. C. (1984) J. Exp. Med. 160, 2. Lomedico, P. T., Gubler, U., Hellman, C. P., Dukovich, M., 679-694. Gini, J. G., Yu-Ching, P., Collier, K., Semionow, R., Oshua, & A. & Mizel, S. B. (1984) Nature (London) 312, 458-462. 31. Noelle, R., Krammer, P. H., Ohara, J., Uhr, J. Vitetta, 3. Gery, I., Gershon, R. K. & Waksman, B. H. (1972) J. Exp. E. S. (1984) Proc. Natl. Acad. Sci. USA 81, 6149-6153. Med. 136, 128-138. 32. Monroe, J. G. & Cambier, J. C. (1983) J. Immunol. 130, 4. Mizel, S. B., Oppenheim, J. J. & Rosenstreich, D. L. (1978) J. 626-631. Immunol. 120, 1497-1503. 33. Wetzel, G. D., Swain, S. L., Dutton, R. W. & Kettman, J. R. 5. Larsson, E.-L., Iscove, N. N. & Coutinho, A. (1980) Nature (1984) J. Immunol. 133, 2327-2332. (London) 283, 664-666. 34. Swain, S. L. (1985) J. Immunol. 134, 3934-3943. 6. Howard, M. & Paul, W. E. (1983) Annu. Rev. Immunol. 1, 35. Harada, N., Kikuchi, Y., Tominaga, A., Takaki, S.-& Takatsu, 307-333. K. (1985) J. Immunol. 134, 3944-3951. Downloaded by guest on September 27, 2021