CELLULAR IMMUNOLOGY 140, 158-172 (1992)

Differential Regulation of IgA Production by TGF-P and IL-5 TGF-/3 Induces Surface IgA-Positive Cells Bearing

IL-5 Receptor, Whereas IL-5 Promotes Their Survival and Maturation into IgA-Secreting Cells’

EIICHIRO SONODA,*,~ YASUMICHI HITOSHI,* NAOTO YAMAGUCHI,* TAKEHISA ISHII,$ AKIRA TOMINAGA,* SHUKURO AUKI,?

ANDKIYOSHI TAKATSU*,'

*Department of Biology, Institute for Medical Immunology, the TFirst Department of Medicine, Kumamoto University Medical School, 2-2-l Honjo, Kumamoto 860, Japan; and

$.Biosciences Laboratory Research Center, Mitsubishi-Kasei Co., Yokohama 227, Japan

Received November 6, 1991; accepted November 7, 1991

Transforming growth factor 6 (TGF-/3) and IL-5 have been shown to augment IgA production by LPS-stimulated murine B cells. We investigated the effect of TGF-fl on the expression of surface Ig-isotype and IL-5 receptor on LPS-stimulated B cells. TGF-fl increased the proportion of both surface IgA-positive (&A+) B cells and sIgG2b+ B cells and enhanced IgA and IgGrb production by LPS-stimulated B cells. TGF-fl synergized with IL-5 only for IgA production of the seven Ig-isotypes and in combination with IL-5 caused a significant increase in the proportion of s&A+ B cells up to 17.4%. In contrast, IL5 decreased the proportion of sIgGrb+ B cells and sIgGf B cells and inhibited the production of IgGzb and IgGr by LPS-stimulated B cells. About 50% of sIgA+ cells induced by TGF-@ expressed IL-5 receptor. They secreted peak levels of IgA and seemed to maintain long viability in the presence of IL5; whereas TGF-8 had the opposite effects on sIgA+ B cells and down-regulated the IL-5 receptor expression. These results indicate that TGF-0 increases the number of r&A+- and IL-5 receptor-positive B cells which respond to IL-5 giving rise to &A-secreting cells and also support the notions that TGF-8 preferentially induces switching to s&A+ B cells and IL5 induces the maturation of postswitch s&A+ B cells into IgA-secreting cells in a stepwise fashion. 0 1992 Academic PFS, hc.

INTRODUCTION

Maturation of activated B cells into immunoglobulin (Ig)-secreting cells is induced by various B cell differentiation factors derived from T cells or macrophages (1, 2). These differentiation factors can be divided into two groups with respect to their biological effects on B cells: One group is defined as class-switching factors (e.g., IL4 and IFN-7) which can regulate the expression of specific isotypes of LPS-stimulated B cells by enhancing the frequency of heavy chain class-switching to a certain isotype

’ This work was supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture and Science, Japan, by a Research Grant from Takeda Science Foundation, and by a Grant-in Aid from Tokyo Biochemical Research Foundation.

’ To whom correspondence should be addressed.

158

0008-8749/92 $3.00 Copyright 0 1992 by Academic press, Inc. AU rights of reproduction in any form reserved.

REGULATION OF IgA PRODUCTION BY TGFq3 AND IL-5 159

(3-6). Members of the other group function as maturation factors (e.g., IL-2, IL-5 and IL-6) which can induce growth and differentiation of B cells that have already switched to a certain isotype, probably in an isotype-nonspecific manner (7- 11). Using CH12.LX B cell lymphoma Kunimoto et al. (12) have shown that IL-4 induces dif- ferentiation of surface IgM-positive (sIgM+) cells to s&A+ cells and IL-5 enhances IgA secretion by the resulting sIgA+ cells in a sequential fashion. Their findings imply that the sequential stimulation with at least two differentiation factors are required for the non-IgM B cell differentiation.

Transforming growth factor p (TGF-P) was originally described for its ability to confer anchorage-independent growth to nonmalignant fibroblasts ( 13) and has sub- sequently been shown to have potent regulatory effects on a variety of cells in the immune system (14-17). TGF-@ inhibits cytotoxic T cell generation (17), NK cell function, T and B cell proliferation ( 14, 16), and antibody production ( 15). We and others have shown that TGF-0 specifically enhances IgA production by LPS-stimulated murine B cells and that IL-5 and IL-2 act with TGF-@ additively and synergistically, respectively ( 18-22). However, the mechanism of the synergistic effect of TGF-/3 with IL-2 or IL-5 on IgA production remains uncertain.

In a previous study (19), we have shown that TGF-@ enhances IgA production by LPS-stimulated splenic B cells, and IgA production was further augmented by the addition of IL-5. TGF-/3 acts on sIgA cells early in culture (by Day 2 in a 5-day culture) and IL-5 is required late in culture for IgA production. We interpreted this observation that TGF-/3 and IL-5 provide different signals to B cells in a stepwise manner to induce IgA production. In the present study, we extend our previous work to elucidate the mechanism(s) by which TGF-P and IL-5 act on sIgA B cells to enhance IgA production. The data suggest that TGF-/3 preferentially acts on sIgA B cells to induce heavy chain class-switching to a-chain in the presence of LPS, and that IL-5 propagates the proportion of sIgA+ B cells induced by TGF-P, maintains cell survival, and induces IgA secretion by postswitch s&A+ B cells.

MATERIALS AND METHODS

Animals

C57BL/6 mice (6 to 8 weeks) were originally purchased from the Japan SLC (Ha- mamatsu, Japan). All animals were maintained in the Laboratory Animal Facility of Kumamoto University.

Cell Lines

The Chinese hamster ovary (CHO) cell line transfected with cDNA for murine IL- 5 was provided by Dr. T. Nishihara (Suntory Central Research Laboratories, Osaka) and was maintained in our laboratory.

Antibodies

Anti-Thy1.2 mAb was purchased from Serotec, Ltd. (England). Anti-CD4 mAb RL172.4 (IgM class) and Anti-CD8 mAb 3.155 (IgM class) were kindly provided as hybridoma cells by Drs. R. Nagasawa (Saitama Medical School, Kawagoe, Japan) and 0. Kanagawa (Washington University, St. Louis, MO). Anti-mouse cr chain mAb (7 l- 14) was kindly provided as hybridoma cells by Dr. F. Kroese (Rijksuniversiteit Gron-

160 SONODA ET AL.

ingen, Groningen, the Netherlands). Affinity-purified goat anti-mouse IgE antibody and standard IgE were kindly provided by Dr. H. Miyata (Kissei Central Research Laboratories, Matsumoto, Japan). Affinity-purified goat anti-mouse isotype-specific antibodies, except anti-&E, were purchased from Southern Biotechnology Associates, Inc. (Birmingham, AL). Anti-mouse IL-5 receptor, H7 mAb and anti-mouse IL-5, TB 13 mAb were prepared and purified as described (23-25). Rabbit polyclonal anti- TGF-/3 antibody was obtained from R & D Systems, Inc. (Minneapolis, MN).

Lymphokines and Reagents

Human recombinant TGF-PI which had been previously described (19, 26) was a kind gift of Genentech, Inc. (South San Francisco, CA). Murine IL-5 was purified from culture supematant of CHO cells which had been transfected with murine IL-5 cDNA by an immuno-affinity column coupled with anti-IL5 mAb (23). Specific ac- tivity of purified IL-5 was 2.2 X lo7 units/mg protein. Human recombinant IL-2, murine recombinant IL-4, IL-6 was kindly provided by Dr. T. Kakinuma (Takeda Pharmaceutical Chemicals), Dr. K. Hama (Ono Pharmaceutical Co., Ltd.), and Drs. T. Hirano and T. Kishimoto (Osaka University, Osaka, Japan), respectively. Murine recombinant TNF-(U was purchased from Genzyme Corp. (Cambridge, MA). t3H]Thymidine ([3H]TdR) (sp act, 5 Ci/mmol) and [35S]methionine (sp act, 1000 Ci/ mmol) were obtained from Daiichi Radio Chemical Co. Ltd. (Tokyo, Japan). Bio- synthetically [35S]methionine-labeled IL5 was prepared by the methods previously described (27).

Cell Preparations

T cell-depleted spleen cell suspension was prepared as described (19) and used as B cell sources. Briefly, spleen cells were treated with a cocktail of anti-Thy 1, anti-CD4, and anti-CD8 mAbs, followed by incubation with nontoxic rabbit complement. sIgA+ B cells were depleted by using a magnetic cell sorter (MACS) (Miltenyi Biotec GmbH, FRG) according to the procedures described by Abts et al. (28). Briefly, splenic B cells were incubated with biotinylated goat anti-mouse IgA for 25 min at 4°C washed, and labeled with FITC-streptavidin for 15 min at 4°C. The cells were then washed and incubated for 5 min at 4°C with l/ 100 diluted biotinylated superpara- magnetic beads. After the magnetic staining, the cells were loaded onto a separation column filled with stainless steel wool in the magnetic field of the MACS. The effluent cells were collected and used as a s&A- population. In certain experiments, adherent accessory cells were removed from spleen cell suspension by passage through a Sephadex G- 10 column (Pharmacia Fine Chemicals, Uppsala, Sweden). Adherent accessory cells were prepared as follows. Briefly, spleen cell suspensions were plated on a glass petri dish for 2 hr at 37°C after which nonadherent cells were removed. Adherent cells were collected with a rubber policeman and treated with anti-Thy 1 mAb and complement. These adherent cells were X-irradiated (4000 rad) and added (2 X 104/well) to an adherent cell-depleted fraction.

Cell Culture

Cells were cultured in round-bottomed, 96-well plates (No. 25850, Coming) in RPM1 1640 medium (Sigma Chemical Co., St. Louis, MO) supplemented with 5 X

REGULATION OF IgA PRODUCTION BY TGF-/3 AND IL-5 161

10e5 M 2-ME, penicillin (100 U/ml), streptomycin (100 p/ml), and 10% FCS (Lot No. 855, Flow Laboratories, McLean, VA). A total of 1 X lo5 spleen B cells was cultured with LPS (10 pg/ml; Difco Laboratory, Detroit, MI). Recombinant cytokines to be tested were added at the beginning of the culture. Cultures were set up in triplicates. After 8 days culture supernatants were pooled and assayed for Ig production by ELISA using isotype-specific antibodies.

Measurement of Zg Production

First, 96-well flexible assay plates (Becton Dickinson Labware, Oxnard, CA) were coated with isotype-specific goat anti-mouse Ig antibodies and incubated overnight at 4°C. Plates were then washed with PBS containing 0.05% Tween-20 (Bio-Rad Lab- oratories, Richmond, CA) and incubated with PBS containing 2% BSA (Sigma Chem- ical Co.) for 2 hr at room temperature followed by washing. Culture supematants were then added to the wells and incubated at room temperature for another 2 hr followed by washing. Biotinylated isotype-specific goat anti-mouse Ig antibodies were then added to the wells and incubated for 2 hr at room temperature. After washing the plates, horseradish peroxidase-conjugated streptavidin (Zymed Laboratories, Inc., South San Francisco, CA) was added to each well, incubated for 30 min at room temperature followed by washing. A loo-p1 aliquot of substrate, o-phenylenediamine (diluted to 0.4 mg/ml) and hydrogen peroxide (final 0.015%) dissolved in 0.1 Mcitrate buffer (pH 5.0) were added to each well. After 10 min incubation, the enzyme reaction was stopped by adding 2 M sulfuric acid and optical density at 495 nm was read by V-max Kinetic Micro Plate Reader (Molecular Devices Corp., Palo Alto, CA). Using myeloma proteins (Miles Scientific, Naperville, IL) standard curves were generated for each isotype and nanograms of Ig per milliliter of supematants were determined with the computer program Delta-Soft (BioMetallics Inc., Princeton, NJ). The spec- ificity of the ELISA was tested by assaying all isotype-specific antibodies against a panel of myeloma proteins of each of the six other isotypes.

FACS Analysis and Cell Sorting

Cell suspensions were prepared in PBS containing 1% FCS and 0.0 1% sodium azide. Cells were stained with biotinylated isotype-specific goat anti-mouse Ig, followed by FITC-labeled streptavidin (Bethesda Research Laboratories Life Technologies, Inc., Gaithersburg, MD). For two-color immunofluorescence staining, cells (1 X 106) were first incubated with isotype matched mAbs, anti-Lyt-2 (53-6.7; rat IgG*a) or anti-lyt- 1 (53-7.3; rat IgGza), for 10 min at 4°C to avoid nonspecific binding of labeled anti- bodies. Cells were then incubated at 4°C for 25 min with biotinylated anti-IL-5 receptor mAb (H7, rat IgGza), washed, and incubated at 4°C for 15 min with phycoerythrin- streptavidin (Becton Dickinson, Mountain View, CA) followed by washing and in- cubation at 4°C for 25 min with FITC-labeled anti-a-chain mAb (71-14, rat IgG*a). After the last incubation, cells were washed and resuspended in PBS/FCS containing 2 &ml of 7-amino-actinomycin D (7AAD, Sigma Chemical Co.). Cells were analyzed on a FACScan (Becton Dickinson). For sorting study, sIgA spleen cells were stimulated for 4 days with TGF-P (0.5 rig/ml) in the presence of LPS (10 &ml). On Day 4 cells were harvested and stained with FITC-labeled anti-a-chain mAb at 4°C for 20 min. Before sorting, dead cells were excluded by staining with unlabeled superparamagnetic

162 SONODA ET AL.

beads followed by passage through a MACS. Cells thus obtained were sorted on a FACStar (Becton Dickinson).

Binding Assay for IL-5

The IL-5 binding assay was carried out as described (27) using biosynthetically 3sS- labeled IL-5. Nonspecific binding was determined in the presence of loo-fold excess of unlabeled IL-5.

RESULTS

Selective Enhancing Efect of TGF-p and IL-5 on IgA Production by LPS-Stimulated B Cells

As we reported (19), TGF-/3 induced a substantial amount of IgA production by LPS-stimulated splenic B cells from BALB/c mice when added on Day 1 of a 7-day culture. Addition of IL5 together with TGF-@ further enhanced the IgA production. To elucidate the enhancing effect of IL5 on polyclonal IgA production by LPS- and TGF-P-stimulated B cells in detail, we examined the expression of IL-5 receptor on spleen cells from various strains of mice by staining with biotinylated H7 (anti-IL-5 receptor) mAb and phycoerythrin-streptavidin. The results revealed that splenic B cell population from C57BL/6 mice showed the most abundant IL-5 receptor expres- sion of those we had tested (data not shown). Furthermore, LPS-stimulated splenic B cells from C57BL/6 mice responded to TGF-/3 and IL-5 for the IgA production better than those from BALB/c mice. The addition of TGF-@ from the beginning of the culture rather than on Day 1 in an g-day culture, as described under Materials and Methods, induced much more IgA response in LPS-stimulated B cells (Table 1). In this system, the addition of IL-5 with TGF-/3 further augmented the IgA production. IL-4, IL-6, or TNF-a alone or in combination with TGF-fl had marginal effect on IgA production (Table 2). Depletion of adherent accessory cells from the responding

TABLE 1

TGF-/3 Exerted Its Activity Early in the Culture for IgA Production

Cytokines (rig/ml) added on

Cytokines TGF-0 Day 0 Day 1 Day 0 Day 1

Medium

IL-5

IL-2

Exp. 1 Exp. 2

268 242 92 84 1871 393 1760 420 876 595 444 206

9515 1480 9500 1375 819 714 287 326

1420 1230 1049 1540

Note. Splenic B cells of C57BL/6 mice were stimulated with LPS (10 &ml) and cultured for 8 days. TGF-p (1 t&ml) alone or in combination with IL5 (4 @ml) or IL-2 (4 rig/ml) was added on either Day 0 or Day 1. Culture supematants were harvested and assayed for IgA production by ELISA. Values are the mean of triplicate cultures.

REGULATION OF IgA PRODUCTION BY TGF-0 AND IL-5 163

TABLE 2

Effect of Cytokines on Ig Production by LPS-Stimulated B Cells

Polyclonal Ig response

Medium

IL-4

IL-5

IL-6

TNF-cu

- 35.2 2550 12 + 8.2 36 1277 - 22.2 10.880 41 + 5.8 288 1990 - 84.6 2223 425 + 38.2 12 8084 - 30.5 2209 120 + 7.2 36 2004 - 33.2 1408 176 + 8.8 27 1751

Note. Splenic B cells of C57BL/6 mice were stimulated with LPS (10 pg/ml) and cultured for 8 days. Various cytokines IL-4 (10 U/ml), IL5 (4 @ml), IL-6 (100 U/ml), TNF-(Y (100 U/ml), alone or in com- bination with TGF-p (1 rig/ml), were added on Day 0. Culture supematants were harvested and assayed for IgA production by ELISA. Values are the mean of triplicate cultures.

B cell population caused a significant decrease of IL-5-induced IgA production and addition of X-irradiated adherent cells into the adherent cell-depleted B cell culture restored the IgA production (Table 3). In contrast, TGF-B-induced IgA production was marginally affected, suggesting that TGF-@-induced IgA production is relatively accessory cell independent. We adopted the culture system in which splenic B cells from C57BL/6 mice are stimulated with LPS and TGF-P on Day 0 in subsequent experiments.

Figure 1 shows the effect on five antibody isotype production of TGF-/3 alone, or in combination with IL-5. TGF-@ induced IgA production by LPS-stimulated B cells in a dose-dependent manner, whereas it suppressed Igh4 and IgGr production. TGF- /I also augmented IgGzb and to a lesser extent IgG3 production, whereas higher con- centrations of TGF-@ (> 1 rig/ml) suppressed both IgG2b and IgG3 production. TGF- /3 caused slight inhibition of the IgGza and IgE response (data not shown). The addition of both TGF-/? and IL-5 induced much more striking IgA production than that induced by either cytokines, whereas it showed no synergistic or additive effect on six other isotype production. No significant IgA production was observed in the absence of LPS (data not shown). IL-5 alone augmented Igh4 and to a lesser extent IgA production, but caused inhibition of IgG2b and IgG3 production in the presence of LPS. Similar results were obtained in three separate experiments.

Efect of TGF-/3 and IL-5 on the Isotype Expression on LPS-Stimulated B Cells

It was reported that TGF-fi induces an increase in the number of s&A+ cells in LPS-stimulated splenic sIgK B cell culture (20) and IL-5 acts on sIgA+ B cells in Peyer’s patches as a maturation factor to induce differentiation into &A-secreting cells (7,9-l 1,29). We therefore asked whether the significant IgA production induced by TGF-fl and IL-5 correlated with the increase in the total number of sIgA+ B cells.

164 SONODA ET AL.

!ib(k*( lo-'to0 10' 102

i;E2 iih

0 10~10-210~'100 10' 102

I

a

TGF-/3 (rig/ml) FIG. 1. Effect of TGF-@ and IL-5 on Ig production by LPS-stimulated splenic B cells. A total of 1 X 10’

splenic B cells was stimulated with LPS (10 &ml). Various concentrations of TGF-p alone (0) or in combination with IL5 (4 rig/ml; 0) were added on Day 0. Supematants from triplicate cultures were pooled on Day 8 and assayed for isotype.-specific antibody by ELISA. Each point represents the mean value from triplicate cultures. Parallel results were obtained in three other experiments.

Splenic B cells were cultured with or without cytokines in the presence of LPS, harvested every day from Day 2 to Day 6, and analyzed for their surface isotype expression. Figure 2A shows one of the representative kinetics data of the isotype expression. When the cells were stimulated with LPS alone, an increase in the proportion of sIgG:, sIgG2b+, and sIgG: B cells was obvious. The addition of TGF-8 together with LPS caused a substantial increase in the proportion of sIgA+ and sIgG2b+ B cells (Fig. 2A). In IL-5 containing cultures, there were few sIgA+ B cells on Day 6 (Figs. 2A and 2B). In contrast, TGF-/3 caused a slight increase (2 to 6%) in the proportion of sIgA+ B cells and a dramatic decrease in sIgM+ cells. Furthermore, the addition of TGF-P in combination with IL-5 caused a marked increase in the proportion of s&A+ B cells during culture. s&A+ cells became up to 17.4% of the total cells on Day 6 (Fig. 2B). IL-5 decreased the proportion of sIgG,b+ and sIgGf B cells irrespective to the existence of TGF-P in the LPS-stimulated B cell culture. The proportion of sIg+ cells appeared to correspond to the levels of antibody production in each isotype (Fig. 1).

To estimate the induction of s&A+ B cells by TGF-/3 and IL5 in detail, splenic B cells were cultured with or without cytokines in the presence of LPS and harvested every day from Day 3 to Day 7. Cells were then stained with FITC-coupled anti-a- chain mAb and analyzed on a FACScan. It is evident from Fig. 3 that TGF-P alone, or in combination with IL-5, caused an increase in the number of sIgA+ B cells by Day 4. These s&A+ B cells decreased in number by subsequent culture with TGF-8 alone, whereas in the presence of IL-5 these s&A+ B cells did not decrease until Day 6 and seemed to survive for longer periods of time.

A 100

‘ii90 6

G@J =

870

m 60

+r 60

$ 40

30 o-01 0 2 4 6 0 2 4 6 0 2 4 6

REGULATION OF IgA PRODUCTION BY TGF-/3 AND IL-5 165

0 2 4 6 0 2 4 6

Days in Culture

Log Fluorescence

FIG. 2. Effect of TGF-fl and IL-5 on the isotype expression of LPS-stimulated B cells. Splenic B cells were cultured with medium (m), IL-5 (100 U/ml; q ), TGF-fi (I rig/ml; l ), or TGF-/3 and IL-5 (0) in the presence of LPS (10 r&ml) and harvested. (A) Cells were stained with isotype-specific goat anti-mouse Ig plus FITC- streptavidin and analyzed on a FACScan. Before analysis cells were stained with 7AAD and dead cells were excluded using FL 3 channel. The proportions of sIg+ cells were calculated from FACS profiles. (B) Cells were harvested on Day 6 and stained with FITC-labeled anti-a-chain mAb. The gates for s&A+ cells are indicated. The percentage of s&A+ cells within the gated region is shown in the lower right-hand corner.

IL-S Receptor Expression on TGF-fi-Induced sIgA+ Cells

It is worthwhile to evaluate whether sIgA+ cells can differentiate into IgA-secreting cells in response to TGF-p or IL-5 To deplete the proliferative B cells which have already expressed sIgA on their surface at the beginning of the culture, we separated

166 SONODA ET AL.

0’.3 4 5 6 7 Days in Culture

FIG. 3. Kinetics of rIgA+ B cell induction. Splenic B cells were cultured with medium (B), IL5 (100 U/ ml; Cl), TGF-/3 (1 @ml; l ), or TGF-P and IL5 (0) in the presence of LPS (10 &ml) and harvested. Cells were stained with FITC-labeled anti-a-chain mAb and analyzed on a FACScan. Before analysis cells were stained with 7AAD and dead cells were excluded using FL 3 channel. The number of sIgA+ cell was calculated from the proportion to the total viable cells.

sIgA cells from &A’ cells with a MACS according to procedures described under MuteriuZs and Methods. Contamination of the sIgA- population with sIgA+ cells was less than 0.3% upon reanalysis. These purified sIgA B cells were stimulated with TGF-@ in the presence of LPS for 3 days, stained with anti-a-chain mAb and anti- IL-5 receptor mAb, and were analyzed on a FACScan. It was clear that TGF-P again did induce the increase in sIgA+ B cells (Fig. 4A). Each of s&A+ and sIgA cell pop-

slgA

IL-5Fl

FIG. 4. Flow cytometric analysis of B cells stimulated with TGF-@. T cell-depleted sIgA cells were cultured with LPS (10 rglml) and TGF-j3 (0.5 @ml). On Day 3 cells were harvested and stained with biotinylated anti-IL5 receptor mAb plus streptavidin-PE followed by FITC-labeled anti-a-chain mAb. Before analysis cells were stained with 7AAD and dead cells were excluded using FL 3 channel. (A) The box in dual color profile illustrates the sIgA+ B cell population induced by TGF-fi. (B) sIgA and sIgA+ B cells were analyzed for their IL-5 receptor expression. Cells were stained with streptavidin-PE alone (-). with anti-IL-5 receptor plus streptavidin-PE (.........), or with anti-IL5 receptor mAb and streptavidin-PE in the presence of a 170-fold molar excess of IL5 (. . . .).

REGULATION OF IgA PRODUCTION BY TGF-fl AND IL-5 167

TABLE 3

Effect of Adherent Cells on Cytokine-Induced IgA Production

IgA production (rig/ml)

Cytokines Undepleted Adherent cells depleted Adherent cells added back

Medium 42 47 103 IL-5 210 53 283 TGF-/3 672 396 447

Note. Before T cell depletion, spleen cells were passed through Sephadex G-10 columns (adherent cells depleted) or not passed (undepleted). Enriched adherent cells were added to adherent cell-depleted fraction (adherent cells added back). Each cell population was stimulated with LPS (10 &ml) and cultured for 8 days. TGF-P (1 @ml) or IL-S (4 r&ml) was added on Day 1. Culture supematants were harvested and assayed for IgA production by ELISA. Values are the mean of triplicate cultures.

ulation displayed in Fig. 4A was examined for IL-5 receptor expression. About 50% of &A+ B cell population expressed detectable levels of IL-5 receptor on their surfaces (Fig. 4B). The sIgA B cell population also contained IL-5 receptor-positive cells, although the proportion with respect to the total cells was less than that in the sIgA+ population. The binding of H7 mAb to each B cell population was markedly inhibited in the presence of an excess amount of IL-5 (Fig. 4B).

To evaluate the possibility that postswitch sIgA+ B cells respond to TGF-/3 for IgA production, we cultured splenic sIgA B cells with TGF-/3 for 4 days and then separated sIgA+ B cells from sIgA B cells by sorting. Each cell population and unsorted B cell population was cultured with various stimulants. sIgA B cells produced marginal but

TABLE 4

Effect of IL-5 and TGF-@ on IgA Secretion by sIgA+ B Cells Induced by TGF-fl

IgA production (q/ml)

Stimulant Unsorted s1a.K sIti+

None 443 23 4905 IL-5 2169 266 62,558 TGF-P 156 10 2424 LPS 1619 117 17,932 LPS + IL5 3312 571 97,424 LPS + TGF-fl 448 63 3780

Note. Splenic B cells were stained with biotinylated goat anti-mouse IgA antibody. sIgA cells were separated by MACS as described under Materials and Methods. These s1g.K cells (1 X 106/mI) were cultured with TGF-fi (1 rig/ml) in the presence of LPS (10 &ml) and harvested on Day 4. Then cells were stained with FITC-labeled anti-a-chain mAb and separated into sIpA+ and s&A populations by a FACStar or stained and not sorted (unsorted). Each cell population was cultured with IL-5 (8 rig/ml) or TGF-8 (1 rig/ml) in the presence or absence of LPS (10 g/ml) at 5 X lo4 cells/well for an additional 5 days, i.e., Day 9 from the initiation of cultures. Culture supematants were harvested and assayed for IgA secretion by ELISA. AI1 cultures were set up in duplicate and values are the mean of duplicates. These data are representative of two separate experiments.

168 SONODA ET AL.

much less IgA than unsorted or sIgA+ B cells, irrelevant to the existence of stimulants. The sorted s&A+ B cells secreted a large amount of IgA in response to IL-5 (Table 4), although they produced marginal amounts of IgA in the culture with LPS or medium alone. The maximal IgA production was always seen in the presence of IG5. In contrast, TGF-@ inhibited &A-secretion (Table 4) and DNA synthesis (data not shown) by these sIgA+ B cells. Similar results were obtained in two separate experiments. Unsorted B cells responded to both IL-5 and LPS for the enhanced IgA production, whereas TGF- fi had the opposite effect. These results suggest that IL-5, but not TGF-/3, is the principal maturation factor of postswitch s&A+ cells and that TGF-/3 inhibits maturation and growth of these cells.

Efect of Anti-TGF-fi Antibody on Isotype Production

Thus TGF-@ increases the number of sIgA+ B cells early in culture, while it inhibits the IgA secretion late in culture. To dissect the above two activities of TGF-/3 on B cells, we added anti-TGF-@ antibody to the LPS-stimulated B cell culture on Day 0 or on Day 2. Results are summarized in Table 5. Addition of anti-TGF-/3 antibody on Day 0 to the culture containing TGF-/I alone, or in combination with IL-5, sig- nificantly inhibited the enhancement of IgA and IgG2b production and recovered from the suppressed production of IgM, IgG, , and IgG3. In contrast, the addition of anti-TGF-/I antibody to the culture containing TGF-/3 on Day 2 could abrogate neither the enhancement of IgA and IgG2b production nor the suppression of IgM production. However, it could abrogate the suppression of IgG, and IgG, production. Anti-TGF- /3 antibody added on Day 2 to the culture containing TGF-/3 and IG5 caused maximum IgA production. Interestingly, in the absence of cytokines, anti-TGF-P antibody caused a substantial inhibition of IgGzb and IgG, production by LPS-stimulated B cells. These results suggest that TGF-/3 acts on B cells by Day 2 to enhance IgA and IgG2b

TABLE 5

Effect of Anti-TGF-j3 Antibody on Isotype Eroducion by LPS-Stimulated B Cells

Addition Supematants level (rig/ml)

Cytokines Antibodies IgA IiN If@, Wzb k%

None

TGF-fl

TGF-P + IL-5

Control Ab (Day 0) 28 55,200 6732 2766 8558 Anti-TGF-8 (Day 0) 22 57,000 5872 724 3406 Control Ab (Day 0) 944 10,800 549 9811 3998 Anti-TGF-8 (Day 0) 245 32,600 4753 5241 15,361 Anti-TGF-,!I (Day 2) 624 24,900 4591 7065 13,767 Control Ab (Day 0) 1556 49,800 595 1978 1451 Anti-TGF-8 (Day 0) 4635 90,000 8722 2065 3496 Anti-TGF-/3 (Day 2) 16,429 87,800 4925 3221 4143

Note. Splenic B cells (1 X lO’/well) were stimulated with LPS (10 &ml). TGF-fl (1 rig/ml) alone, or in combination with IL-5 (4 r&ml), was added on Day 0. Anti-TGF-0 antibody (30 &ml) or rabbit control antibody was added on either Day 0 or Day 2 as indicated in parentheses. Culture supematants were harvested on Day 8 and assayed for isotype-speck antibody by ELISA. Ig production (@ml) in culture with IL-5 alone was IgA, 602; IgM, 107,300; IgG,, 8105; IgGrb, 1290, IgG3, 3519. Values are the mean of triplicate cultures. These data are representative of three separate experiments.

REGULATION OF IgA PRODUCTION BY TGF-j3 AND IL-5 169

production and to suppress Igh4 production and that it acts after Day 3 to suppress the IgA, IgG, , and IgG, production. The latter effect may be mediated, in part, through the down-regulation of IL-5 receptor expression. In fact, preculturing of LPS-stimulated B cells with TGF-P decreased the specific binding of 35S-labeled IL-5 (Table 6) and decreased the binding of anti-IL-5 receptor antibody (data not shown).

DISCUSSION

In this study, we attempted to clarify the mechanisms of the synergistic effect of IL-5 with TGF-8 on IgA production. In our culture system employed in the present experiments, TGF-P alone, or in combination with IL-5 induced more IgA production by LPS-stimulated B cells than in the culture system which has already been described (18-20, 22). This may be, in part, due to the source of responding B cells and the timing of the addition of TGF-p. We used B cells from C57BL/6 mice and other investigators mainly used B cells from BALB/c mice ( 18-20, 22). We also found that the addition of TGF-fl alone, or in combination with IL-5 at the beginning of the culture rather than on Day 1 was more effective for IgA production. Besides, anti-IL- 5 receptor mAb enabled us to define the IL-S-responsive cells, at a single cell level.

We found that TGF-P did induce the generation of sIgA+ B cells from a sIgA cell fraction, confirming the observation reported by Lebman et al. (20). They also reported that IL-2 acts synergistically with TGF-B to augment IgA secretion without affecting the total number of sIgA+ cells and speculated that IL-2 further enhances IgA secretion by a post-transcriptional mechanism (20). In contrast, the addition of IL-5 to LPS- stimulated splenic B cell culture with TGF-@ caused a significant increase in the number of s&A+ B cells, which are the precursors of &!-secreting cells. About 50% of sIgA+ B cells generated from sIgA B cells during the culture with TGF-fl also expressed the IL-5 receptor on their surface and secreted peak IgA levels in response to IL-5, but not in response to TGF-/3 (Fig. 4 and Table 4). We would speculate at this moment based on these results that IL-5 synergizes with TGF-P for IgA production by different mechanisms from that of IL-2.

Although TGF-p induced the generation of sIgA+ B cells from LPS-stimulated sIgA B cells, TGF-/3 inhibited the growth and differentiation of resulting sIgA+ B cells into &$-secreting cells. It had previously been demonstrated that TGF-fl inhibited IgA secretion by sIgA+ Peyer’s patch cells (18). This finding suggests that the sIgA+ pop- ulation in the Peyer’s patch may not behave differently from spleen cells induced to

TABLE 6

Effect of TGF-fl on the Expression of “S-labeled IL-5 Binding Site of LPS-Stimulated B Cells

Preculture with 35S-Labeled IL-5 bound to target cells

LPS 1196 (557) LPS + TGF-8 597 (414)

Note. Splenic B cells were stimulated with LPS (10 &ml) and cultured with or without TGF-8 ( 1 ng/ ml) for 2 days prior to the binding assay. On Day 3 cells were harvested and each group of 8.5 X lo6 cells were incubated for 10 min at 37°C with 300 pM 3sS-labeled IL-5. Values are the mean cpm from duplicate cultures. Parentheses represent the value of cpm obtained in the presence of a lOO-fold molar excess of unlabeled IL-5.

170 SONODA ET AL.

switch sIgA expression. In contrast, IL5 induced IgA production by these s&A+ B cells. Taken together, TGF-@ acts specifically as an isotype-switching factor to induce switching of sIgA B cells to s&A+ B cells and IL-5 acts as a maturation factor to induce the differentiation of resulting sIgA+ B cells into &$-secreting cells. It is clear therefore that TGF-P and IL-5 enhance IgA secretion by different mechanisms and exert their activity on a B cell at the different stages of differentiation in a stepwise fashion.

The fact that TGF+ induces switching to sIgA+ has been confirmed by the ap- pearance of germ-line Ca transcripts (20, 30), by the formation of IgA-specific switch circular DNA (3 l), and by the frequency analysis of &A-secreting cells (2 1). However, it still remains unclear whether TGF-fl induces only IgA isotype switching. Whitemore et al. (32) reported that TGF-/I increased the frequency of switching to IgA and IgG*b in CH12.LX, a clone of IgM producing Ly-1 B cell which is capable of spontaneous isotype switching to IgA, IgGzb, IgG, , and IgGs. Their results suggested that TGF-fi acts as a trigger or inducing signal for isotype switching, but does not direct the class switch to any particular isotype. In this study, we described that TGF-fl enhanced not only IgA but also IgGzb, and to a lesser extent IgG3, production by LPS-stimulated B cells at optimal concentrations. Moreover, the addition of anti-TGF-@ antibody to LPS-stimulated B cell culture decreased the IgG2b and IgG3 production in the absence of exogenous TGF-@ These are the major isotypes induced by the stimulation of LPS. It is unclear, however, whether TGF-/!I can determine the isotype to which a B cell will switch, or can only induce switching, to isotypes to which the cells are already committed in LPS-stimulated B cell culture.

It has been reported that TGF-P can be produced by transformed cells as well as normal cells (26), including activated T cells (14) and macrophages (33). Since splenic B cells used in many experiments contain substantial numbers of accessory cells, TGF- fi which may be secreted during culture by LPS-stimulated accessory cells such as macrophages could affect B cell differentiation. The depletion of adherent cells from splenic B cells caused a moderate loss of IgA enhancement induced by TGF-P (Table 3), suggesting that adherent cells, or factors secreted by them might affect the IgA production by LPS-stimulated B cells. One of the critical issues is why TGF-fl alone can induce IgA production of LPS-stimulated sIgA B cells, although it inhibits mat- uration of s&A+ B cells into IgA-secreting cells. IL-6 was not effective in our culture system, although it was reported that IL-6 induces IgA production (8) (Table 2). We speculate that LPS may simply act as a maturation factor or that LPS and/or TGF-P may induce the production of unidentified cytokines other than IL-6 by B cells or adherent cells, and one of such cytokines may act on sIgA+ B cells for IgA production.

Several investigators reported that the mechanism by which IL-5 augmented IgA- secretion by sIgA+ B cells is through the induction of IgA-secretion, but not through the proliferation of sIgA+ cells (7, 10,29). In this study, we showed that IL-5 increased the number of sIgA+ B cell in combination with TGF-fl and LPS. IL-5 slightly enhanced the [3H]TdR incorporation by sorted sIgA+ B cells compared to control cultures, and cell recoveries in IG5-containing cultures were always better than those in control cultures irrespective of the presence of LPS (data not shown). It is not clear, however, whether IL-5 simply supports the survival of sIgA+ B cells, induces the proliferation of sIgA+ B cells, or increases the frequency of switching of sIgA to s&A+. The mainte- nance of sIgA+ B cell survival may be one of the mechanisms of enhanced IgA pro- duction induced by IL-5.

REGULATION OF IgA PRODUCTION BY TGF-@ AND IL-5 171

Interestingly, the addition of IL-5 to LPS-stimulated B cell culture caused the suppression of IgGab and IgG, production. These are the major isotypes induced by LPS stimulation (Fig. 1). We do not think that IL-5 decreases the proportion of sIgG2b+ and sIgG: cells by conversion into Ig-secreting cells or that IL-5 inhibits their growth in an isotype-specific manner. We favor the possibility that IL-5 may suppress both IgGzb and IgG3 production by decreasing the frequency of isotype switching of LPS- stimulated B cells. This suggests that the induction of maturation of sIgM+ B cells by IL-5 may be related to the concomitant inhibition of isotype switching of sIgM+ B cells to other isotypes. We are currently studying the mechanism of inhibitory effect of IL-5 on isotype switching.

It has been reported that most Ly-l+ B cells in the peritoneal cavity and a small population of splenic B cells have detectable IL-5 receptors determined by flow cy- tometric analysis with anti-IL-5 receptor mAbs (25,34, 35) and seem to be the major IL-5 responsive B cell population in vivo. Germinal center cells in peripheral lymph nodes from immune mice were also reported to be a major IL-S-responsive B cell population (36). It is of a particular interest to elucidate whether the IL-5-responsive splenic B cell population is derived from germinal centers. If germinal center cells that have already switched to non-IgM isotypes are activated with natural antigens such as bacterial flora, IL-5 may act to support their survival and to induce their maturation into Ig-secreting cells. IL-5 may be one of the important factors for the maturation and maintenance of postswitch memory cells in vivo. Immunohistochem- ical staining using anti-IL-5 receptor mAb should facilitate clarification of this issue.

In conclusion, TGF-P and IL5 induce IgA production through distinct mechanisms and act on B cells in a stepwise fashion. TGF-/3 acts as an isotype-switching factor to induce switching of sIgA B cells to sIgA+ cells and inhibit the maturation of sIgA+ B cells into Q&secreting cells. In contrast, IL-5 induces the maturation of postswitch s&A+ B cells and supports their survival. The experimental system described here should provide feasible tools in further analysis of generation, maturation, and selection of IgA memory B cells.

ACKNOWLEDGMENTS

We thank Dr. T. Nishihara, Dr. F. Kroese, and Genentech Inc. (South San Francisco, CA) for providing IL-5-cDNA-transfected CHO cells, hybridoma (7 1 - 14) and recombinant TGF-/3 1, respectively. We also thank Drs. S. Mita, R. Matsumoto, and Y. Kikuchi for their helpful advice throughout this study; Drs. T. Hattori and K. Takatsuki for allowing us to use a FACStar; Dr. T. Murakami for valuable advice for operating a FACStar; and Dr. E. Barsoumian for his critical review of this manuscript.

REFERENCES

1. Howard, M., and Paul, W. E., Annu. Rev. Immunol. 1, 307, 1983. 2. Kishimoto, T., Annu. Rev. Immunol. 3, 133, 1985. 3. Vitetta, E. S., Brooks, K., Chen, Y.-W., Isakson, P., Jones, S., Layton, J., Mishra, G. C., Pure, E., Weiss,

E., Word, C., Yuan, D., Tucker, P., Uhr, J. W., and Krammer, P. H., Immunol. Rev. 78, 137, 1984. 4. Yuan, D., and Vitetta, E. S., Mol. Immunol. 20, 367, 1983. 5. Snapper, C. M., and Paul, W. E., Science 236,944, 1987. 6. Paul, W. E., and Ohara, J., Annu. Rev. Immunol. 5, 429, 1987. 7. Harriman, G. R., Kunimoto, D. Y., Elliott, J. F., Paetkau, V., and Strober, W., J. Immunol. 140, 3033,

1988. 8. Beagley, K. W., Eldridge, J. H., Lee, F., Kiyono, H., Everson, M. P., Koopman, W. J., Hirano, T.,

Kishimoto, T., and McGhee, J. R., J. Exp. Med. 169,2 133, 1989.

172 SONODA ET AL.

9. Matsumoto, R., Matsumoto, M., Mita, S., Hitoshi, Y., Ando, M., Araki, S., Yamaguchi, N., Tominaga, A., and Takatsu, K., Immunology 66,32, 1989.

10. Schoenbeck, S., McKenzie, D. T., and Kagnoff, M. F., Eur. J. Immunol. 19, 965, 1989. 11. Kunimoto, D. Y., Nordan, R. P., and Strober, W., J. Zmmunol. 143,2230, 1989. 12. Kunimoto, D. Y., Harriman, G. R., and Strober, W., J. Zmmunol. 141, 713, 1988. 13. Spom, M. B., Roberts, A. B., Wakefield, L. M., and Assoian, R. K., Science 233, 532, 1986. 14. Kehrl, J. H., Wakefield, L. M., Roberts, A. B., Jakowlew, S., Alvarez-Mon, M., Derynck, R., Spom,

M. B., and Fauci, A. S., J. Exp. Med. 163, 1037, 1986. 15. Kehrl, J. H., Roberts, A. B., Wakefield, L. M., Jakowlew, S., Spom, M. B., and Fauci, A. S., J. Immunol.

137,3855, 1986. 16. Lee, G., Ellingsworth, L. R., Gillis, S., Wall, R., and Kincade, P. W., J. Exp. Med. 166, 1290, 1987. 17. Jin, B., Scott, J. L., Vadas, M. A., and Bums, G. F., Immunology66, 570, 1989. 18. Coffman, R. L., Lebman, D. A., and Shrader, B., J. Exp. Med. 179, 1039, 1989. 19. Sonoda, E., Matsumoto, R., Hitoshi, Y., Ishii, T., Sugimoto, M., Araki, S., Tominaga, A., Yamaguchi,

N., and Takatsu, K., .I. Exp. Med. 170, 1415, 1989. 20. Lebman, D. A., Lee, F. D., and Coffman, R. L., J. Immunol. 144,952, 1990. 21. Kim, P.-H., and Kagnoff, M. F., J. Immunol. 145, 3773, 1990. 22. Chen, S.-S., and Li, Q., Cell. Immunol. 128, 353, 1990. 23. Harada, N., Takahashi, T., Matsumoto, M., Kim&i, T., Ohara, J., Kikuchi, Y., Koyama, N., Sever&on,

E., Yaoita, Y., Honjo, T., Yamaguchi, N., Tominaga, A., and Takatsu, K., Proc. Natl. Acad. Sci. CJSA84,4581, 1987.

24. Yamaguchi, N., Hitoshi, Y., Mita, S., Hosoya, Y., Murata, Y., Kikuchi, Y., Tominaga, A., and Takatsu, K., Int. Immunol. 2, 181, 1990.

25. Hitoshi, Y., Yamaguchi, N., Mita, S., Sonoda, E., Takaki, S., Tominaga, A., and Takatsu, K., J. Immunol. 144,4218, 1990.

26. Derynck, R., Jarrett, J. A., Chen, E. Y., Eaton, D. H., Bell, J. R., Assoian, R. K., Roberts, A. B., Spom, M. B., and Goeddel, D. V., Nature (London) 316, 701, 1985.

27. Mita, S., Harada, N., Naomi, S., Hitoshi, Y., Sakamoto, K., Akagi, M., Tominaga, A., and Takatsu, K., J. Exp. Med. 168, 863, 1988.

28. Abts, H., Emmerich, M., Miltenyi, S., Radbruch, A., and Tesch, H., J. Immunol. Methods 125, 19, 1989.

29. Beagley, K. W., Eldridge, J. H., Kiyono, H., Everson, M. P., Koopman, W. J., Honjo, T., and McGhee, J. R., J. Zmmunol. 141,2035, 1988.

30. Lebman, D. A., Nomura, D. Y., Coffman, R. L., and Lee, F. D., Proc. Natl. Acad. Sci. USA 87, 3962, 1990.

3 1. Matsuoka, M., Yoshida, K., Maeda, T., Usuda, S., and Sakano, H., Cell 62, 135, 1990. 32. Whitmore, A. C., Prowse, D. M., Hat&ton, G., and Arnold, L. W., Int. Immunol. 3, 95, 1991. 33. Assoian, R. K., Fleurdelys, B. E., Stevenson, H. C., Miller, P. J., Madtes, D. K., Raines, E. W., Ross,

R., and Spom, M. B., Proc. Natl. Acad. Sci. USA 84,6020, 1987. 34. Rolink, A. G., Melchers, F., and Palacios, R., J. Exp. Med. 169, 1693, 1989. 35. Rolink, A. G., Thalmann, P., Kikuchi, Y., and Erdei, A., Eur. J. Immunol. 20, 1949, 1990. 36. Rabinowitz, J. L., Tsiagbe, V. K., Nicknam, M. H., and Thorbecke, G. J., J. Zmmunol. 145, 2440,

1990.