Proc. Nati. Acad. Sci. USA Vol. 84, pp. 9140-9144, December 1987 Immunology Activation of resting human B cells by helper T-cell clone supernatant: Characterization of a human B-cell-activating factor (T-celi lymphokines/B-ceil stimulation) ANITA DIu*, MARIE-LISE GOUGEON*, JEAN-LOUIS MOREAU*, ELLIS L. REINHERZt, AND JACQUES THtZE* *Unitd d'Immunogdndtique Cellulaire, Institut Pasteur, 25, rue du Docteur Roux 75015 Paris, France; and tDivision of Immunobiology, Dana-Farber Cancer Institute, 44 Binney Street, Boston, MA 02115 Communicated by Baruj Benacerraf, August 17, 1987

ABSTRACT The effects of helper T-cell clone superna- (Pharmacia, Paris) density-gradient centrifugation of periph- tants on resting human B cells were investigated. Four different eral blood from healthy adult volunteers and were partially helper T-cell clones (two T4' and two T8+) were stimulated by depleted of adherent cells by adherence to plastic culture anti-T3 monoclonal antibodies on Sepharose beads or anti-T112 dishes. Nonadherent cells were submitted to rosette forma- plus anti-T113 monoclonal antibodies. The supernatants from tion with neuraminidase-treated sheep erythrocytes, and these activated clones induced the proliferation of highly nonrosetting cells were recovered after centrifugation on a purified resting B lymphocytes from the peripheral blood. The Lymphoprep gradient (Nyegaard, Oslo). The B cells were B cells exhibited a cell size and a surface-antigen pattern (4F2 then separated by using a discontinuous gradient of Percoll antigen and transferrin receptor) of phase Go B cells, and they (Pharmacia) as described (12), and cells banding at the were functionally resting. In response to T-cell supernatants a interface between Percoll concentrations 50o and 60% were large fraction of the B cells enlarged and expressed 4F2 collected. These preparations (referred to as dense B cells) antigens and transferrin receptors. In gel filtration, the cor- routinely contained <1% T cells, 2-3% monocytes, and responding activity migrated with an apparent Mr of 12,000- >85% B cells as determined by immunofluorescence staining 15,000. Our fwdings strongly support the existence ofa human using anti-T4 plus anti-T8, MO1, and anti-Bi mAbs, respec- B-cell-activating factor acting on resting B cells and causing tively (21, 22). The B-cell preparations showed no prolifer- them to enter phase G1 of the cell cycle. ative response to T-cell mitogens such as phytohemagglutinin (PHA), concanavalin A, or anti-T112 plus anti-T113 mAbs. In T-cell-dependent responses, specific T-cell-B-cell contact Dense B cells (5 x 104 per well) were cultured in 96-well is supposed to control B-cell activation, whereas later stages round-bottomed microtiter plates (Costar, Cambridge, MA) of B-cell maturation are driven by nonspecific T-cell-derived in 200-,lI final volume in RPMI 1640 medium (M.A. Bio- lymphokines (1, 2). In the human model, B-cell proliferation products, Walkersville, MD) containing 10% fetal calf serum is mediated by two B-cell growth factors (BCGF): a low-Mr (Flow Laboratories, Paris), 1 mM L-glutamine, 1% penicil- and a high-Mr BCGF (3-5). Human interleukin 4 (IL-4) also lin/streptomycin, and 10 mM Hepes. Thymidine uptake was acts as a B-cell proliferative factor (6). In vitro assays for the measured after 3 days of culture with a 16-hr exposure to 1 detection of these lymphokines involve submitogenic con- ,uCi of [3H]thymidine per well (1 Ci = 37 GBq). centrations of anti-, heavy chain immunoglobulin (anti-,) T-Cell Clones. P2 and N2 (T8+) were cloned twice by antibodies or Staphylococcus aureus Cowan A strain (SAC) limiting dilution methods, in which T cells from a healthy that are supposed to polyclonally activate resting B cells. volunteer (0.3 cell per well) were plated in the presence of Interleukin 2 (IL-2) also induces proliferation and, in some irradiated autologous feeder layers in IL-2-containing medi- cases, differentiation of SAC or anti-,-activated human B um with 0.25 ,ug of PHA per ml (Wellcome). Tetanus cells (7). In addition, a direct effect ofvery high doses ofIL-2 toxoid-specific T-cell clones TT1 and TT9 (T4+) were ob- on unstimulated B cells (8) was also reported. tained in the same way from an antigen-specific cell line In the mouse model, we have shown that the T cells can be raised from T cells ofa healthy volunteer using tetanus toxoid replaced by their supernatant for mouse resting B-cell acti- at 0.3 floculation unit/ml (Massachusetts State Biological vation, proliferation, and differentiation (9-11). We have Laboratories, Jamaica Plain, MA). T cells were cultured in hypothesized that a new lymphokine called B-cell-activating IL-2-containing medium. Restimulation was done every 4 factor (BCAF) was responsible for resting B-cell activation, weeks in the presence of irradiated feeder cells, PHA (0.25 and we now have evidence that this factor is distinct from ,tg/ml), and IL-2. The T-cell cultures tested negatively for other mouse lymphokines. In this report, we extend these mycoplasma at least twice during the course of this work. observations to the human model. Using helper T-cell clones Production ofT-Cell Supernatants. T-cell clones were taken stimulated by anti-T3 or anti-T11 monoclonal antibodies 15 days after restimulation, washed once, and incubated in 5 (mAbs), we have produced supernatants that could polyclon- ml of polypropylene tubes (Falcon) at 106 cells/ml for 6 hr or ally activate highly purified resting human B cells from the 24 hr at 37°C in medium alone, or in the presence of anti-T3 peripheral blood. Our findings strongly support the existence mAbs on Sepharose beads (1/50) (13), or a combination of of a human BCAF with an apparent Mr of 12,000- anti-T112 plus anti-T113 mAbs (1/200 dilution of ascitic fluid) 15,000. (14). The supernatants were then collected, filtered through a 0.22-,um Millipore membrane, and used immediately or MATERIALS AND METHODS stored at -20°C. B-Cell Preparation and Proliferation Assay. Peripheral blood mononuclear cells were obtained by Ficoll/Hypaque Abbreviations: anti-A, anti-, heavy chain immunoglobulin antibod- ies; BCAF, B-cell-activating factor; BCGF, B-cell growth factor; IL-2, interleukin 2; IL-4, interleukin 4; mAb, monoclonal antibody; The publication costs of this article were defrayed in part by page charge PHA, phytohemagglutinin; P2(T11)-SN, P2(T11) supernatant; FITC, payment. This article must therefore be hereby marked "advertisement" fluorescein isothiocyanate; SAC, Staphylococcus aureus Cowan A in accordance with 18 U.S.C. §1734 solely to indicate this fact. strain.

9140 Downloaded by guest on September 26, 2021 Immunology: Diu et al. Proc. Natl. Acad. Sci. USA 84 (1987) 9141 RESULTS IL-2 Is not Involved in the B-Cell Response to T-Cell Supernatant. Because high doses of IL-2 were able to induce Supernatants from Activated T-Cell Clones Aire Sufficient to some proliferation of resting B cells and because T-cell Induce the Proliferation of Unstimulated B Ccells. Two T4' supernatants did contain some IL-2 (as measured on an T-cell clones (TT1 and TT9) and two T8' clone1s (P2 and N2) IL-2-dependent T-cell line), it was of interest to evaluate the were activated for 24 hr by anti-T3 mAbs 4on Sepharose contribution of IL-2 to our results. The final concentration of beads, or by a combination of anti-T112 and aniti-T113 mAbs. IL-2 from P2(Tll)-SN in the B-cell assays (between 0.6 and The culture supernatants were collected and as&sayed in B-cell 1.25 ng/ml) was too low to account by itselffor the observed cultures. As shown in Fig. 1, dense B cells are driven to response (Table 1). We further excluded the participation of proliferate in response to these helper T-cell suipernatants, in IL-2 in the effects ofthe T-cell supernatant by either blocking the absence of any other added stimulus, andI this effect is the IL-2 receptor or neutralizing the IL-2 activity with dose dependent. The release of this proliferrative activity appropriate mAbs. As shown in Table 2, when resting B cells needs stimulation of the T-cell clones becauwse little or no were cultured with anti-pt and recombinant IL-2, both 33B3-1 activity is detectable in supernatants from uinstimulated T (anti-IL-2R IgG2a, ref. 15) and DMS-1 (anti-IL-2 IgG1, ref. cells. In addition, the proliferation of B cellIs induced by 16) mAbs were able to inhibit the B-cell response to IL-2. On T-cell supernatant is not major histocompatiboility complex- the contrary, the response to P2(T11)-SN was not signifi- iit comf cantly affected by either mAb. As a control, we checked that restricted. Supernatants obtained after stimul o ploex 33B3-1 and DMS-1 mAbs did not affect the BCGF-induced P2 with anti-T112 plus anti-T113 mAbs [P2(Ta11tSN'11)-SN] weree response of anti-,u activated B cells (data not shown). studied in greater detail. Size Analysis of Resting B Cells Stimulated with P2(Tll)-SN. The Responding B-Cell Population Consists c sf Functionally The change in size of small B cells was analyzed at different Resting B Cells. The B cells used in the culturels consisted of times after stimulation with P2(T11)-SN. Fig. 2 shows that at dense B cells with a mean volume correspondling to cells in initiation of the culture, Percoll gradient-purified B cells phase Go of the cell cycle (Fig. 2A). To verify Ithat they were exhibited a mean cell volume ranging between 140 and 160 also functionally resting and to exclude that our results could ,1Mm3. During culture with P2(T11)-SN (Fig. 2B), the size be explained by a contamination of in vivo-actiivated B cells, distribution gradually increased; blast cells accounted for we assayed low-Mr BCGF, IL-4, and IL-2 tihat have been 40% ofthe population by 96 hr. Fig. 2A shows that culture of reported to act only on preactivated B cells (3,, 6, 7). Table 1 small B cells in medium alone did not significantly alter their shows that when these lymphokines were assayed alone on size distribution. small B cells, no significant proliferative respcrnse was seen. Effects of P2(Tll)-SN on Resting B-Cell Expression of 4F2 However, when anti-IL antibodies coupled to Sepharose Antigen and Transferrin Receptor. The mAbs 4F2 and 5E9 beads were added to the cultures, the same coells responded (antitransferrin receptor) recognize cell surface glycopro- to all three lymphokines. We also verified tha.t interleukin-1 teins that appear to be related to lymphocyte activation and and y-interferon do not induce any proliferatiion of small B proliferation (17). Fig. 3 shows that at initiation ofthe culture, cells (data not shown). small B cells were totally negative for transferrin-receptor and 4F2-antigen staining. During culture with P2(T11)-SN, 15 A 15 B the expression of both antigens gradually increased, and by 72 hr, the percentage of highly positive cells was 48% for 4F2 antigen and 57% for transferrin receptor, respectively. In 10 10 addition, 30% of the cells were shown to express the IL-2 receptor after 3 days ofstimulation with P2(T11)-SN (data not shown). Culture medium by itself did not induce any detect- 0 5-I- 5 able change in 4F2-antigen or transferrin-receptor expression I- (data not shown). -0 0 The Proliferative Activity Migrates with an Apparent Mr of 5 10 5 10 12,000-15,000. P2(Tll)-SN was separated by molecular u0 sieve fast protein liquid chromatography. Fig. 4 shows one 10- IG c D out of four separate experiments that gave similar results. z I Proliferation of resting B cells was tested for each fraction, C010 and the proliferative activity for resting B cells was identified I in the Mr range of 12,000-15,000. Some B-cell proliferation I.; / /IsO was seen with one or two fractions corresponding to a Mr of I >500,000 and probably represented aggregates or material not retained on the column. The Superose column did not separate the IL-2 activity from the B-cell proliferative activity. However, under the 5b10 experimental conditions described in Table 2 both 33B3-1 5 10 (anti-IL-2R) and DMS-1 (anti-IL-2) mAbs did not inhibit the SUPERNATANT DILUTION (%v v) proliferative response of resting B cells to column fractions 58-62 (data not shown). FIG. 1. Induction of resting B-cell proliferation by supernatants derived from activated T-cell clones. Dense B cells were cultured with supernatant from the T-cell clones P2 (A), 1711 (B), N2 (C), and DISCUSSION TT9 (D). The supernatants were obtained after a 24-hr culture of the These experiments clearly demonstrate that upon activation, T cells in complete medium alone (o), with anti-T3 mAbs coupled to T-cell clones release beads or with a helper factors that can by themselves Sepharose (e), combination of anti-T112 plus activate unstimulated B anti-T113 mAbs (A). Results are expressed as the mean cpm from polyclonally cells. Our initial obser- duplicate cultures. In such experiments the proliferative response of vation was established with four different T-cell clones. the B-cell preparations to optimal concentrations of anti-T112 plus Although two of them had a T4' phenotype and two had a anti-T113 mAbs or to anti-T3 Sepharose beads alone ranged between T8+ phenotype, they all exhibited similar helper/inducer 1000 and 2000 cpm. properties when cocultured with B cells. Helper T-cell clones Downloaded by guest on September 26, 2021 9142 Immunology: Diu et al. Proc. Natl. Acad. Sci. USA 84 (1987)

A B

N N 0 0 I.- OH 11 oH 3J if w 5. %w z 72H zcc

-u V- -1 24H 0 48H *t_-72 96H

100 300 500 700 900 100 300 500 700

CELL VOLUME (pm3) CELL VOLUME (#im3)

FIG. 2. Cell-volume analysis of B cells during 4 days of stimulation with P2(T11)-SN. Resting B cells were cultured for various lengths of time in complete medium alone (A) or in the presence of P2(T11)-SN (10%o vol/vol) (B) and were then subjected to cell-volume analysis using a Coulter Counter and a Channelyser (Coultronics, Margency, France). H, hr.

of the T81 phenotype have already been described (18). The from the cultures. The factors found in the supernatant must, T-cell clones could be activated through either the T3-T-cell therefore, be secreted by T cells themselves. receptor complex (13), or more efficiently through the Til The responding cell population consisted ofhighly purified alternative pathway (14). Activation of T cells with mAbs B lymphocytes and did not show any response to T-cell was done at a time where all feeder cells had been eliminated mitogens. In addition, the proportion of cells bearing B-cell markers remained constant over the 3-day culture period with Table 1. Resting B-cell response to P2(T11)-SN, IL-2, T-cell supernatant (data not shown). In all experiments, IL-4, and low-Mr BCGF resting B cells were separated according to cell density on a Percoll gradient (12), and dense B cells were used in the Addition(s) to culture [3H]Thymidine uptake, cpm cultures. Different criteria were used to verify the state of P2(T11)-SN activation of the responding B-cell population. The size 10% 12,122 profile and the surface antigen pattern (transferrin receptor 5% 9,112 and 4F2 antigen) of the dense B cells were consistent with 0.5% 3,657 cells in phase Go ofthe cell cycle (17). In addition, the B cells 0 450 were functionally resting because they failed to proliferate in rIL-2 response to low-Mr BCGF, IL-4, and IL-2. Several controls 100 ng/ml 2,298 10 ng/ml 1,235 Table 2. Effect of 33B3-1 (anti-IL-2 receptor) and DMS-1 2.5 ng/ml 1,949 (anti-IL-2) mAbs on B-cell proliferation 1 ng/ml 986 [3H]Thymidine uptake, cpm 0 450 anti-,u 4,811 anti-jI + P2(T11)-SN, anti-c + IL-2 (2.5 ng/ml) 14,632 mAb (jzg/ml) IL-2 (1.25 (2.5%, rIL-4 added to culture Control ng/ml) vol/vol) 1% 1,355 33B3-1 0.5% 1,057 0 1092 8634 22,666 0.25% 1,035 3 1053 1700 (100%) 19,004 (16%) 0 1,271 0.3 4536 (601%) 20,122 (12%) anti-c 1,536 0.03 8886 (0%) 22,790 (0%) anti-p + IL-4 (0.5%) 8,169 DMS-1 Low-M, BCGF 0 384 6655 5,014 20%o 1,394 200 588 3109 (76%) 6,244 (0%) 10%0 1,143 50 5194 (31%) 6,835 (0%) 5% 700 25 5909 (16%) 6,462 (0%1o) 0 1,117 anti-si 718 Small B cells were cultured in the presence ofP2(T11)-SN or anti-/A beads (10 ug/ml) plus recombinant (r) IL-2 at indicated concentra- anti-st + BCGF (5%) 5,857 tions. DMS-1 mAb was added 1 hr before the cellular addition. Small B cells were cultured in the presence of P2(T11)-SN, Percentage of inhibition over background was calculated by the recombinant (r) IL-2, low-M, BCGF (Cellular Products, Buffalo, equation: 1 - [(cpm - background cpm)/(maximum cpm - back- NY), recombinant IL-4, anti-. beads (Bio-Rad) (10 .g/ml), or anti-Is ground cpm)], and percentage is bracketed. Proliferation with anti-js beads plus lymphokines. Thymidine uptake was measured after 3 beads alone (1803 cpm in experiment with mAb 33B3-1 and 1986 cpm days of culture with a 16-hr exposure to 1 ACi [3H]thymidine per well in experiment with mAb DMS-1) was used as background cpm for (1 Ci = 37 GBq). Results are expressed as the mean cpm of duplicate calculating the percentage of inhibition in cultures containing anti-/A cultures. plus IL-2. -, Not done. Downloaded by guest on September 26, 2021 Immunology: Diu et al. Proc. Natl. Acad. Sci. USA 84 (1987) 9143

Us-LJ C.)

-J A. z 0Uo 104 lo4 RELATIVE FLUORESCENCE INTENSITY FIG. 3. Surface expression of4F2 antigen and transferrin receptor on B cells during culture with P2(T11)-SN. Resting B cells were cultured for various lengths of time in the presence of 10%o (vol/vol) P2(T11)-SN. The cells were stained by means of indirect fluorescence with the following: (A) 4F2 mAb (1/600 dilution of ascitic fluid) plus fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse immunoglobulins (GAM-FITC, 2.5 ug/ml, Coulter Clone, Margency, France); (B) 5E9 mAb (1/1000 dilution of ascitic fluid) plus GAM-FITC. Background fluorescence was determined by staining the cells with GAM-FITC alone and was superimposable to Fig. 3B time 0 hr. Cells were analyzed on a FACS-Can analyzer (Becton Dickinson).

in the experiments also excluded an activation of B cells due eration of resting B cells (Table 1 and refs. 3, 6), but we to culture conditions (Fig. 2A and data not shown). cannot exclude their participation in the observed prolifera- Our main observation is that supernatants from T-cell tive response. IL-2 at moderate concentrations only acts on clones act on resting B-cell populations independently of preactivated B cells (7), but IL-2 was reported to have a direct additional stimuli, such as anti-,a or SAC. Such supernatants effect on resting B cells at high concentrations (>100 ng/ml) induce surface-activation antigens, blastogenesis, and thy- (8). The use of two different types of mAbs, one directed at midine uptake in the culture. By using parameters that can be the IL-2 receptor (33B3-1) and the other one directed at measured on each cell individually, we saw that at least 50o soluble IL-2 (DMS-1) allowed us to exclude any participation of the B cells undergo marked changes in size or in surface- of IL-2 in our results. antigen expression in response to T-cell factors. Our results strongly support the existence of a T-cell- The resting B-cell proliferative activity was recovered after derived soluble factor that acts early in resting B-cell acti- gel filtration in the 12,000-15,000 Mr range. We can therefore vation. This BCAF could either induce the proliferation of exclude a requirement for high-Mr BCGF in our results but resting B cells or prepare the B cells to respond to other not of low-Mr BCGF or IL-4, both of which have a reported lymphokines such as BCGF or IL-4. We have already Mr of -15,000. These two factors do not induce the prolif- described a factor for mouse resting B lymphocytes with

500DOod 50OACAd 15.oood 5,oood

0 Irl5l K I.-I 4 E E 2.5. M0 3z I-. A o5E

COLUMN FRACTION NUMBER FIG. 4. M, determination of the BCAF activity by gel filtration column analysis. P2(T11)-SN (2 ml) was applied to a Superose 12 column (Pharmacia, Paris) and eluted with 0.15 M NaCl/25 mM Tris/0.1% betarne buffer, pH 7.8 on a fast protein liquid chromatography system (Kontron, Trappes, France) at a flow rate of0.5 ml/min. One-milliliter fractions were collected and assayed on resting B cells at a final dilution of 25% (vol/vol) in 200-pul final volume. [3H]Thymidine incorporation is expressed as specific cpm above background (500 cpm). d, Dalton. Downloaded by guest on September 26, 2021 9144 Immunology: Diu et al. Proc. Natl. Acad. Sci. USA 84 (1987) similar properties (9-11) and comparable Mr (C. Roth, Fauci, A. S. (1984) Immunol. Rev. 78, 75-96. J.-L.M., M. Komer, and J.T., unpublished results). Several 3. Mehta, S. R., Conrad, D., Sandler, R., Morgan, J., Montagna, observations made in the human model with soluble T-cell- R. & Maizel, A. L. (1982) J. Immunol. 135, 3298-3302. 4. Ambrus, J. L., Jurgensen, C. H., Brown, E. J. & Fauci, A. S. derived factors agree with our results. Peer cell line super- (1985) J. Exp. Med. 162, 1319-1335. natant (19) and supernatant of a T-cell hybridoma (20) induce 5. Dugas, B., Vazquez, A., Gerard, J. P., Richard, Y., Auf- the proliferation of tonsillar B cells. These observations, as fredou, M. T., Delfraissy, J. F., Fradelizzi, D. & Galanaud, P. well as our results in the human and mouse system, all (1985) J. Immunol. 135, 333-338. suggest the existence of a BCAF. This remains an hypothesis 6. Yokota, T., Otsuka, T., Mosmann, T., Banchereau, J., De- until complete biochemical characterization of such a factor. france, T., Blanchard, D., De Vries, J. E., Lee, F. & Arai, K. We cannot exclude that a combination of already known (1986) Proc. Natl. Acad. Sci. USA 83, 5894-5898. lymphokines may account for our results, although this has 7. Jelinek, D. F., Splawski, J. B. & Lipsky, P. E. (1986) Eur. J. never been reported. Immunol. 16, 925-932. 8. Bich-Thuy, L. & Fauci, A. S. (1985) Eur. J. Immunol. 15, BCAF-containing supernatant acts in the absence of T- 1075-1079. cell-B-cell contact and activates B cells independently of 9. Leclercq, L., Bismuth, G. & Theze, J. (1984) Proc. NatI. their major histocompatibility complex antigens and speci- Acad. Sci USA 81, 6491-6495. ficity. This activation suggests that resting B lymphocytes 10. Leclercq, L., Cambier, J. C., Mishal, Z., Julius, M. H. & can respond to lymphokines independently of surface immu- Theze, J. (1986) J. Immunol. 136, 539-545. noglobulin cross-linking or Ia recognition by helper T cells. 11. Diu, A., Leclercq, L., Dautry-Varsat, A. & Theze, J. (1987) It is important to examine the consequences of such an Cell. Immunol. 107, 471-478. activation pathway on the specificity of the B-cell immune 12. Kuritani, T. & Cooper, M. D. (1983) J. Immunol. 131, 1306- response. Because helper T cells can only secrete this factor 1311. 13. Meuer, S. C., Hodgdon, J. C., Hussey, R. E., Protentis, J. P., upon activation, one can imagine that B cells acting as Schlossmann, S. F. & Reinherz, E. L. (1983) J. Exp. Med. antigen-presenting cells would be the first target to T-cell 158, 988-993. lymphokines. The target specificity thus would be maintained 14. Meuer, S. C., Hussey, R. E., Fabbi, M., Fox, D., Acuto, O., by the proximity of the antigen-specific T and B cells. Fitzgerald, K. A., Hodgdon, J. C., Protentis, J. P., Schloss- However, the existence ofa BCAF could explain the recruit- mann, S. F. & Reinherz, E. L. (1984) Cell 36, 897-906. ment of bystander, unspecific B cells during an antigen- 15. Olive, D., Raymond, J., Dubreuil, P., Charmot, D., Jacques, specific immune response. Purification of this lymphokine Y. & Mawas, C. (1986) Eur. J. Immunol. 16, 611-616. will permit a closer examination of its participation in the 16. Smith, K. A., Favata, M. F. & Oroszlan, S. (1983) J. Immu- humoral immune nol. 131, 1808-1815. response. 17. Kehrl, J. H., Muragushi, A. & Fauci, A. S. (1984) J. Immunol. The authors thank Drs. L. M. Nadler, A. S. Freedman, Y. 132, 2857-2861. Jacques, K. Smith, A. Dautry-Varsat, J. Banchereau, and Cetus 18. Suciu-Foca, N. & King, D. W. (1986) Transplant. Proc. 18, Corporation for the gift of mAbs and interleukins. We especially 230-234. thank Drs. C. Grillot-Courvalin and S. Longacre for helpful advice 19. Bowen, D. L., Ambrus, J. L. & Fauci, A. S. (1986) J. Immu- and Mrs. F. Georges for expert secretarial assistance. This work was nol. 136, 2158-2163. supported by grants from Institut National de la Sante et de la 20. Mayer, L., Fu, S. M. & Kunkel, H. G. (1984) Immunol. Rev. Recherche Mddicale (INSERM) and Centre National de la Re- 78, 119-135. cherche 21. Todd, R. F., Nadler, L. M. & Schlossman, S. F. (1981) J. Scientifique (CNRS). Immunol. 126, 1435-1442. 1. Kishimoto, T. (1985) Annu. Rev. Immunol. 3, 133-157. 22. Stashenko, P., Nadler, L. M., Hardy, R. & Schlossman, S. F. 2. Kehrl, J. H., Muragushi, A., Butler, J. L., Falkoff, R. J. M. & (1980) J. Immunol. 125, 1678-1685. Downloaded by guest on September 26, 2021