CELLULAR
IMMUNOLOGY
144,417-428( 1992)
Transforming Growth Factor-P inhibits the Production of IgG, IgM, and IgA in Human Lymphocyte Cultures A.WARMOLDL.VANDENWALLBAKE,KELLEY P.BLACK, ROSE KULHAVY,JIRI MESTECKY,ANDSUSANJACKSON Department of Microbiology, University of Alabama at Birmingham, Birmingham, Alabama 35294 Received April 9, 1992; acceptedJuly 6, 1992 Transforming growth factor p (TGFO) has potent immunoregulatory effects acting on both T and B cells. It strongly inhibits secretion of IgG and IgM in human and murine B cell cultures, but has been shown to have an enhancing effect on IgA production in the mouse. We have studied the effect of TGFfl on the production of IgA in human lymphocyte cultures. The addition of TGFP to pokeweed-stimulated peripheral blood lymphocytes resulted in a suppression of IgA production of both subclasses,similar in magnitude to the suppressionof IgG and IgM production. Membrane IgA expression was not increased by culturing tonsillar lymphocytes with TGFP. In conclusion, we find no evidence for a selective enhancing effect of TGF@ on IgA synthesis in humans, in contrast to the findings reported in mice. 0 1992 Academic Pm, Inc.
INTRODUCTION Transforming growth factor p (TGFP) is a member of a family of cytokines synthesized by a large variety of cell types and has a wide spectrum of biological effects (l-3). Three members of this family, designated TGF@, TGF/32, and TGFP3, have been identified in mammals. TGFP 1, a 25kDa homodimer, is now recognized as an important mediator of several cellular and physiologic processes,modulating growth and differentiation of many cell types (3). Its amino acid sequence is identical in various mammalian species,including humans and pigs (4). Recent studies have made it increasingly clear that TGFP 1 has potent immunoregulatory actions. It is produced by both T and B lymphocytes and is the most potent endogenous suppressant of lymphocyte proliferation known. TGFPl inhibits the proliferation of T cells in response to IL-2 (5) and suppressesthe development of cytotoxic T cells (6). With respect to B cells, TGFPl also inhibits factor-dependent proliferation, as well as secretion of IgG and IgM in humans (5,7,8). In the mouse, TGFfll also suppressesthe production of IgG and IgM in lipopolysaccharide-stimulated B cells, but leads to a selective increase in the production of IgA (9- 1l), possibly by acting as a switch factor ( 12, 13). However, in mice this enhancing effect on IgA production depends on the type of B cell stimulation used: basal and IL-5-induced IgA production by mesenteric lymph node cells is profoundly suppressedby TGFP 1, whereaslipopolysaccharide-induced IgA secretion is increased (14). In humans the effects of TGFPl on IgA synthesis have not yet been clearly defined, although indirect evidence suggestsa suppressive effect similar to that observed for IgG and IgM (15). Others have recently found evidence for a direct 417 0008-8749/92$5.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
418
VAN
DEN
WALL
BAKE
ET AL.
switching effect to the IgA isotype (16). In view of the potentially important role of TGF/3 1 as an isotype-specific regulation factor, we studied the effect of TGFP 1 on IgA synthesis by human peripheral blood and tonsillar lymphocytes. MATERIALS AND METHODS Cell Culture Peripheral blood mononuclear cells (PBMC) were isolated from heparinized venous blood from healthy white adults ( 17 males, 3 females,ages25-63 years) using a standard Ficoll density gradient (Lymphocyte Separation Media, Organon Teknika Corp., Durham, NC). Mononuclear cells from tonsils were prepared from surgical specimens provided by The Children’s Hospital (Birmingham, AL). Tonsils were removed for tonsillitis in a 6-year-old black female and a 6-year-old white male, and were carefully dissected in tissue culture media. Mononuclear cells were isolated from the cell suspensions by density gradient centrifugation using Lymphocyte Separation Media. Isolated mononuclear cells were washed twice in cold Dulbecco’s phosphate-buffered saline (PBS; Mediatech, Washington, DC), and contaminating red cells lysed in 0.82% ammonium chloride solution (pH 7.4). The cells were washedtwice more in Dulbecco’s PBS, and once in complete tissue culture media consisting of RPM1 1640 (Mediatech) supplemented with 2 mM Igh.ttamine (GIBCO BRL, Grand Island, NY), penicillin (100 U/ml), and streptomycin (100 pg/ml) (GIBCO), and 10% heat-inactivated fetal bovine serum (Flow-ICN, Costa Mesa, CA). PBMC cultures were performed in 24-well tissue culture plates (Costar, Cambridge, MA). Each well contained 1 X lo6 cells in 1 ml complete tissue culture medium. All cell cultures were performed in duplicate. Cells were cultured for 7 days at 37°C in humidified air containing 5% COZ. At the end of the culture period, the cells were spun down, and the supernatants stored at -20°C until the measurement of immunoglobulin content. The cells were harvestedby vigorous pipetting with cold Dulbecco’s PBS, washed twice in PBS, and once in complete tissue culture media before enumeration of immunoglobulin-producing cells in the ELISPOT assay.The number of viable cells for each type of culture was established by trypan blue exclusion in a hemocytometer. Tonsillar cells were cultured at a concentration of 1 X 106/ml in a 10 ml volume in 25cm2 tissue culture flasks (Corning Glass Works, Coming, NY). After 7 days of culture, cells were harvested and culture supematants were saved for ELISA analysis. Growth Factors All growth factors were added at the beginning of the 7-day culture period. Pokeweed mitogen (GIBCO) was used at a 1:lOO dilution. Recombinant human interleukin 2 (rIL-2) with a purity >95% and specific activity of 2.5 X lo6 BRMP units/mg (Genzyme Research Products, Cambridge, MA) was added to the culture medium at a concentration of 50 U/ml. Highly purified TGFP 1, isolated from human and porcine platelets, was obtained from R & D Systems(Minneapolis, MN). Basedon our own observations and results reported in the literature, porcine TGF/31 was used at a concentration of 1 rig/ml except where noted. Fluorescence-Activated Cell Sorter (FACS) Analysis An aliquot (50 X 1O6cells) of the tonsillar mononuclear cell preparation was stained overnight on ice with FITC-labeled anti-human IgA and.biotin-conjugated anti-human
TGFP INHIBITS
PRODUCTION
OF IgG, IgM, AND
IgA
419
IgM (Becton-Dickinson Immunocytometry Systems, San Jose, CA). The cells were then washed with cold Dulbecco’s PBS and were stained with PE-labeled streptavidin (Becton-Dickinson). After a final wash in Dulbecco’s PBS, the cells were resuspended in culture media and were sorted on a FACStar analyzer (Becton-Dickinson) for cells expressing surface (s) sIgM but not sIgA (s&M’, sIgA cells). sIgM+, sIgA cells were cultured for 48 hr in culture media with or without 1 rig/ml of TGFP 1. At the end of the culture period, the cells were harvested, restained as above, and were analyzed on the FACStar for any shift of sIgM+, sIgA- cells to sIgM+, s&A+ cells. Enzyme-Linked Immunoassay (ELBA) The concentrations of immunoglobulin isotypes in the culture supematants were measuredby ELISA. Polyvinylchloride 96well plates (Dynatech laboratories, Chantilly, VA) were coated overnight at room temperature with 100 &well of the primary antibody appropriately diluted in PBS. For the measurement of IgA we used affinitypurified goat F(ab), anti-human IgA (Pel-Freeze,Rogers, AR). The coating antibodies for the IgG and IgM determinations were heavy chain-specific affinity-purified goat F(ab)Z fragments (Jackson, West Grove, PA). The plates were washed three times with PBS containing 0.05% Tween 20 (Sigma) (PBS-Tween) between all subsequent steps in the ELISA. The wells were blocked for 1 hr at 37°C with PBS-Tween containing 1% bovine serum albumin (BSA) (Sigma; PBS-Tween-BSA). This buffer was also used as a diluent in all subsequent steps of the ELISA. Appropriate serial dilutions of the tissue culture supernatants were added to duplicate wells. As a standard we used seven serial twofold dilutions of a pooled human serum standard containing known concentrations of IgG, IgA, and IgM, added to duplicate wells. After an overnight incubation at 4°C bound immunoglobulins were detected with heavy chain-specific, affinity-purified goat F(ab)* fragments, coupled to biotin (Tago Inc., Burlingame, CA). The incubation with the biotinylated antibodies was performed at 37°C for 2 hr, and followed by avidin-conjugated peroxidase (Sigma) for 1 hr at 37°C. Conversion of the substrate (2,2’-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (Sigma), containing 0.0075% H202 was measuredat 4 14 nm on a I’,, kinetic microplate reader (Molecular Devices Co., Palo Alto, CA). OD readings of blanks, obtained in wells containing no serum, were subtracted from the OD values in the wells containing standards and unknowns. The concentrations of the various isotypes were obtained by interpolation on the standard curves using a four-parameter modeling procedure (Deltasoft, BioMetallics Inc., Princeton, NJ). The final concentration in eachsamplewas calculated as the mean of the results of duplicate cultures. Enzyme-Linked Immunospot Assay (ELISPOT) Immunoglobulin-producing cells of the various isotypes were enumerated by the ELISPOT technique ( 17). The wells of nitrocellulose-bottomed 96-well Millititer HA plates (Millipore, Bedford, MA) were coated overnight at 4°C with 100 &well of the primary antibody appropriately diluted in PBS. The coating antibodies for the enumeration of IgG-, IgA-, and IgM-producing cells were the same as described for the ELISA. For the IgA subclasseswe used murine monoclonal antibodies (a kind gift from Drs. J. Radl and J. J. Haaijman) specific for IgAl (clone 69-l 1.4) and IgA2 (clone H5-5 12) (18). The plates were washed three times in PBS, and allowed to soak for 5 min. Blocking was achieved by incubating the plates for 30 min at 37°C with
420
VAN
DEN
WALL
BAKE
ET AL.
200 &well of complete culture media. The cells obtained from the harvested cultures were added to duplicate wells at varying cell densities in 100 &well of complete culture media, and were incubated in the tissue culture incubator for 3 hr. Subsequently, the cells were removed by three washings with PBS. Between all subsequent stepsthe plates were washed three times in PBS-Tween and were allowed to soak in this buffer for 5 min. The spots were developed by overnight incubation at 4°C with the same biotinylated antibodies as described for the ELISA, diluted in PBS-Tween-BSA. After a 1-hr incubation with 0.5 pg/ml extravidin conjugated to alkaline phosphatase(Sigma) diluted in PBS-Tween, the color was developed with the chromogen 5-bromo-4-chloro3-indolyl phosphate toluidine salt and p-nitroblue tetrazolium chloride (Sigma) (17). Spots were counted under low magnification and were expressed as the number of spot-forming cells (SFC) per IO6 cells added to the wells. The final result for each culture was calculated as the mean of duplicate wells.
Statistical Analysis Both the amounts of immunoglobulin produced and the numbers of SFC proved to be not normally distributed, and were consequently transformed to their logarithms prior to statistical analysis. Comparisons between the various types of culture were performed by analysis of variance for repeated measures,with a P value < 0.0 1 considered as statistically significant, using Fisher’s least significant difference (LSD) method. Data are represented as geometric means with the 95% confidence interval (CI). RESULTS
Dose-ResponseRelationship of TGFBl on Immunoglobulin Production To determine the optimal concentration of porcine TGF@1, we studied the effect of increasing concentrations of TGF@l added to PWM-stimulated PBMC from three healthy volunteers. As shown in Fig. 1, there was a dose-dependent decreasein the induction of SFC of all three isotypes. The suppressive effect of TGFPl reached a plateau at concentrations above 1 @ml. Consequently, this concentration of TGF/31 was used in the subsequent experiments.
Inhibition of P WA&InducedImmunoglobulin Secretionby TGF@l Figure 2 shows the quantities of IgA, IgG, and IgM produced in PBMC cultures from 20 healthy individuals. PWM stimulation led to significant increases in the amounts produced for all three isotypes, as compared to media controls (P < 0.01). The addition of 1 rig/ml porcine TGFPl to the PWM-stimulated cultures significantly suppressedthe quantities secreted of not only IgG and IgM, but also IgA (P < 0.01 for all three isotypes). The extent of suppression for the various isotypes induced by TGF/3 1 is given in Table 1. The total viable cell numbers at the end of the culture period were not decreased by the addition of TGF/?l . In control cultures containing media only cell survival was 52.3% (95% CI 43.4-63.0). In PWM-stimulated cultures cell survival was 83.6% (63. l110.7), and in cultures containing PWM and TGF@l 107.7% (79.6-145.9).
TGFP INHIBITS PRODUCTION
421
OF IgG, IgM, AND IgA
80
0
I@ W
fl
0
3.067
0.2
1.8
0.6
TGF concentration
IgM
5.4
(nglml)
FIG. 1. Dose-response relationship of increasing quantities of porcine TGFP 1 added to PWM-stimulated PBMC cultures (n = 3). The number of SFC of each isotype is given as a percentage of the number found in the absenceof TGFpl, which is set at 100%.Data are the geometric means.
For tonsil mononuclear cells, the amounts of immunoglobulin produced are given in Table 2. Similar to the results obtained using PBMC, the secretion of all three isotypes was inhibited by adding TGFP 1 to the PWM-stimulated cultures. Efect of TGFPl on the Production of IgAl and IgA2
We studied the effect of TGFPl on the induction of SFC of the two subclassesof IgA to determine if there was any differential effect on the subclasses.The numbers of IgA 1 and IgA2 SFC induced in unstimulated PBMC cultures and cultures stimulated with PWM or PWM plus TGFPl are shown in Figs. 3 and 4. PWM stimulation led to significant increases of both IgA 1 and IgA2 SFC compared to media controls (P < 0.0 1). Similar to the effect observed on total IgA secretedin culture, 1 rig/ml of TGFP 1 caused a significant decreasein the numbers of IgAl and IgA2 SFC when added to PWM-stimulated cultures (P < 0.0 1). The geometric mean percentage of suppression induced by TGFPl was 63% for IgAl and 64% for IgA2. As shown in Table 3, the IgA subclassdistribution, calculated as the number of IgAl SFC divided by the sum of IgAl and IgA2 SFC, was not appreciably influenced by the type of stimulation used. This demonstrates that TGFPl did not have a preferential effect on either of the two IgA subclasses. Comparison of porcine and human TGFPl
In a subgroup of six healthy individuals we compared the suppressive effect of porcine and human TGFP 1 on PWM-induced immunoglobulin production in PBMC
422
VAN DEN WALL BAKE ET AL.
p< 0.01
pc 0.01 I
IL-
q IgA IgG El
PWM stimulation
IgM
PWM+TGF
FIG. 2. ImmunogIobulin produced in cultures is shown. Quantities of IgA, I@, and IgM were measured in the supernatants of PBMC cultures (n = 20), using media only, PWM, and PWM plus 1 rig/ml of porcine TGFfll. Data are the geometric means, with vertical bars representing the 95% confidence intervals.
cultures. Porcine and human TGFPl appeared to have very similar inhibitory effects on the production of IgG, IgM, and IgA (Table 4) and on the induction of SFC of both IgA subclasses(Table 5). Using another type of stimulation, the combination of rIL-2 (50 U/ml) and human TGFfi I ( 1 rig/ml), we were equally unable to demonstrate any selective enhancement of IgA production, as shown in Tables 4 and 5. Similar to the effect on PWM stimulation, TGFPl inhibited immunoglobulin synthesis induced by IL-2. Induction of Surface IgA Expression on Tonsillar MNC by TGFPl Sorted sIgM+, sIgK tonsillar MNC were cultured for two days to study the induction of surface IgA expression. The addition of 1 rig/ml of TGFfll to the culture media TABLE I Percentage of Inhibition of PWM-Induced Immunoglobulin Synthesis in PBMC Cultures by Porcine TGFPI Isotype
% Inhibition
IgA
65.0 (46.0-77.3) 56.9 (35.4-71.2) 83.5 (73.8-89.6)
w W
Note. Values are geometric means (n = 20) of percentage of inhibition by the addition of 1 rig/ml porcine TGFflI to PWM-stimulated PBMC cultures. Cells were cultured at 106/ml. Values in parenthesesare the 95% confidence intervals.
TGFP INHIBITS PRODUCTION
423
OF IgG, IgM, AND IgA
TABLE 2 Production of Immunoglobulin in Tonsil Cultures Stimulation
IgA
w
kM
IS1 280 89
147 390 191
193 342
ND” ND ND
Experiment 1 Media PWM PWM + human TGFfil
43 101 47 Experiment 2
Media PWM PWM + human TCFPl
137 159 130
189
Note. Immunoglobulin production (t&ml) in 7-day tonsil cultures containing lo6 mononuclear cells. a ND, not detectable.
did not lead to an increased frequency of surfaceIgA-expressing B cells when compared to medium alone. In the first experiment, the percentage of sIgM+, s&A+ cells was 14.6 for the TGFPl stimulation, and 12.8 for media alone. In the second experiment the percentages were 2.6 and 2.2 respectively. The results from one representative experiment are shown in Fig. 5. 100000
p-z 0.01
p< 0.01
r----II----l
. .
10000-
. l .**
.
.
9.
. .
.
Y
.
. d 0
.
1000 -
l
f
.
.
0-0 l
.
:
Ei
z .
0 v-
.
.*
.
.
l * .* --*
loo-
.
.
a** .
3
%
.
l
.
iE
.
.
.
.
10
media
PWM
PWM+TGF
stimulation FIG. 3. Number of IgAl SFC per lo6 cells, induced in PBMC cultures (n = 1S), using media only, PWM, and PWM plus 1 rig/ml of porcine TGFPl. Data points are the results of the individual cultures, with vertical bars representing the geometric means and 95% confidence intervals.
424
VAN DEN WALL BAKE ET AL.
p< 0.01
p< 0.01
. .* ..: ..I
II 10000
!!! d 0
1000
.
s 0 v2 ih
..
.
.. ..r: ... $
100
.
. .
.** . l* lz . . J
. . .
. . 10
media
PWM
PWM+TGF
stimulation FIG. 4. Number of IgA2 SFC per I O6cells, induced in PBMC cultures (n = 17), using media only, PWM, and PWM plus I rig/ml of porcine TGFfll. Data points are the results of the individual cultures, with vertical bars representing the geometric means and 95% confidence intervals.
DISCUSSION In this study, we have examined the effect of TGFBl on the production of IgA and its subclassesby peripheral blood and tonsil B cells. The study was prompted by analogous studies in mice showing a selective increase in the production of IgA, with concurrent suppression of the secretion of IgG and IgM (9- 13). The results from these murine studies have suggestedthat TGFPl may operate as an &A-specific switch factor (I 2, 13). If such a role for TGF/3 1 could be confirmed in humans, this would have important consequencesfor our understanding of the regulation of the IgA immune response,both in health and in diseaseswhere a dysregulation of the IgA immune response appears to be involved in the pathogenesis. TABLE 3 IgA SubclassDistribution in PBMC Cultures Stimulation Media PWM PWM + porcine TGF@l (1 rig/ml)
% IgAl SFC 69.0 (64.3-74.1) 64.4 (58.0-71.5) 66.4 (61.6-71.6)
Note. Values are geometric means (n = 18) of the percentagesof IgA spot forming cells in PBMC cultures that produce IgA 1. Values in parenthesesare the 95% confidence intervals.
TGFP INHIBITS PRODUCTION
425
OF IgG, IgM, AND IgA
TABLE 4 Production of Immunoglobulin in PBMC Cultures Stimulation Media PWM PWM + porcine TGFPl PWM + human TGFfil IL-2 IL-2 + human TGF@
74 (34-161) 607 (315-1171) 234 (92-597) 215 (90-51 I) 173 (85-354) 62 (29-133)
75 (46-124) 769 (430-I 376) 46 1 (256-828) 417 (232-750) 400 (188-853) 86 (62-l 18)
13 (7-24) 1294 (621-2698) 275 (89-852) 286 (93-882) 598 (191-1869) 29 ( 19-44)
Note. Values are geometric means (n = 6) of nanograms of immunoglobulin produced in PBMC cultures containing lo6 cells. Values in parenthesesare 95% confidence intervals.
The results of our study do not support the hypothesis that TGFPl selectively enhances the production of IgA in humans. In contrast, we found equal or greater inhibition of the secretion of IgA and its subclassesas compared to IgG and IgM. This significant suppression of IgA synthesis was found both in PBMC cultures and in cultures of mucosally derived lymphocytes. Apart from the obvious conclusion that the differences in our findings compared to those reported in mice are the consequence of a difference between the species, other factors may be responsible for the discrepancy. One possible explanation may be in the different type of stimulation in the various studies. Most of the results obtained in mice were the result of stimulation with a combination of lipopolysaccharide (LPS) and TGFP 1 (9- 13). In mice, the type of lymphocyte stimulation has been reported to be an important determinant for the effect of TGF@l. When IL-5 was used as the stimulus for IgA production, a profound inhibitory effect of TGFPl was observed (14). Similarly, the samestudy showed a suppressiveeffect on the basal IgA production by mesenteric lymph node lymphocytes. However, when murine B cells are preactivated with LPS, IL-5 and TGFPl act additively in stimulating IgA production (10). LPS is not an effective stimulus for immunoglobulin production in humans, and we consequently chose other types of stimulation for our study. It is possible that the situation in which TGFPl is used in combination with PWM, as in our study, is more similar
TABLE 5 Spot Forming Cells Induced in PBMC Cultures Stimulation
IsAl
IgA2
Media PWM PWM + porcine TGFfll PWM + human TGFPI IL-2 IL-2 + human TGF@l
28 1 (I 54-509) 3908 (18 13-8424) 1372 (455-4135) 1520 (587-3936) 922 (433-1960) 208 (112-385)
144 (69-303) 2115 (1107-4043) 676 (162-2815) 834 (212-3285) 707 (374-l 337) 211 (155-289)
Note. Values are geometric means (n = 6) of number of SFC per IO6 cells induced in PBMC cultures. Values in parenthesesare 95% confidence intervals.
426
VAN DEN WALL BAKE ET AL.
-loo
ICI
toa
10s
104
-xl0
10’
loa
10’
1oJ
FIG. 5. FACS analysis of tonsi B lymphocytes stained with FITC-labeled anti-human IgA (horizontal axis) and PE-labeledanti-human IgM (vertical axis) after incubation with media alone (A) and after incubation with TGF/31 (B).
to the combination of IL-5 with TGFfll, which leads to inhibition of IgA synthesis in mice. However, this explanation seemsless likely in view of the fact that other forms of costimulation also did not seemto enhance IgA production in our study of human lymphocytes. The combination of rIL-2 and TGFP 1 did not stimulate IgA production in PBMC and led to decreasedproduction of all three isotypes compared to stimulation with IL-2 alone (Tables 4 and 5). We have also examined the effect of killed Stuphylococcus aweus Cowan strain I (0.01%) in combination with rIL-2 (100 U/ml) on purified B cells from human tonsil and appendix. Costimulation with human TGFP 1 (1 rig/ml) did not lead to an increase in IgA synthesis in either of these tissues (results not shown). An alternative explanation is that TGFPl has a mixed effect on cultures containing both B and T cells (such as our PBMC cultures), in contrast to the direct effect on purified B cells used in the murine studies (9- 13). The mixed effect could consist of, on the one hand, an induction of a switch to membrane IgA expression in B cells but, on the other hand, a strong suppressive effect on the accompanying T cells, which are necessary for the subsequent proliferation and differentiation to immunoglobulinproducing cells. This mixed effectcould be responsible for the inhibition of IgA secretion that we found, using the T-cell-dependent stimulation by PWM. However, the negative results in the B cell cultures stimulated with SAC and IL-2 argue against this explanation, as does the absence of any stimulator-y effect of TGF/31 on the expression of surface IgA observed in tonsillar B cells. BecauseTGF@l also has a strong inhibitory effect on the proliferation of activated B cells, this in itself could also be responsible for the suppression of immunoglobulin secretion (7, 19). In the concentrations used, TGF/31 does not appear to be toxic to lymphocytes, and does not lead to a decreasedviability at the end of the 7-day culture period ( 19). Recent indirect evidence also supports our conclusion of a suppressive effect of TGFBI on PWM-induced IgA synthesis in human PBMC. Kekow et al. reported a strong inhibitory effect of PBMC culture supernatants from HIV-infected individuals on the production of IgA, IgG, and IgM by healthy donor PWM-stimulated PBMC cultures ( 15). The suppressive effect of these supernatants could be reversed by the
TGF(3 INHIBITS
PRODUCTION
OF IgG, IgM, AND
I&4
427
addition of anti-TGF/3 antibody, demonstrating that the suppression of immunoglobulin synthesis in vitro was caused by TGFP. 1n vim, there is no evidence of suppressed immunoglobulin production in HIV-infected patients. In contrast, such patients often manifest polyclonal B cell activation, reflected in increased serum levels of IgG, IgA, and IgM (20, 2 1). Possibly, the effectsof increased IL-6 production and Epstein-Barr virus infection override the suppressive effects of TGFP in these patients (15). In summary, we find no evidence of a selective &A-inducing effect of TGFPl in human lymphocyte cultures. Recently, the induction of germ-line transcripts of both IgA subclasseshas been described in human splenic B cells cultured with TGFP 1, as an event preceeding an IgA switch (22). Similarly, human tonsillar sIgD+ B cells could be induced to synthesize IgA using a combination of TGFP, IL- 10, and cross-linking of CD40 (23). In contrast, the IgA secretion by sIgD- B cells was suppressedby TGFP. Another important factor determining the effect of TGFP on IgA production by human B cells may be the duration of exposure to the cytokine. Removal of the TGFP after 24 hr led to increased IgA production, whereas continued presenceof the cytokine in human splenic B cell cultures led to profound suppression of IgA synthesis ( 16). Although an early direct effect on B cells consisting of switching to the IgA isotype seems possible, the net effect of TGFP 1 on PWM-stimulated human lymphocyte cultures is a suppression of IgA synthesis, similar in magnitude to that reported for IgG and IgM. ACKNOWLEDGMENTS The researchof Dr. A. W. L. van den Wall Bake has been made possible by a career development award from the Royal Netherlands Academy of Arts and Sciences.The work was supported by U.S. PHS Grants DK 40117, AI 23952, and AI 18745, and the Nephrology Research and Training Center. We appreciate accessto the Flow Cytometry Core Facility of the UAB AIDS Center.
REFERENCES 1. 2. 3. 4. 5.
Roberts, A. B., Frolik, C. A., Anzano, M. A., and Sporn, M. B., Fed. Proc. 42,2621, 1983. Sporn, M. B., Roberts, A. B., Wakefield, L. M., and de Crombrugghe, B., J. CeN.Biol. 105, 1039, 1987. Massague,J., Annu. Rev. Cell. Biol. 6, 597, 1990. Sporn, M. B., and Roberts, A. B., J. Am. Med. Assoc. 262,938, 1989. Kehrl, J. H., Wakefield, L. M., Roberts, A. B., Jakowlew, S., Alvarez-Mon, M., Derynck, R., Sport-r, M. B., and Fauci, A. S., J. Exp. Med. 163, 1037, 1986. 6. Ranges,G. E., Figari, I. S., Espevik, T., and Palladino, M. A., J. Exp. Med. 166, 991, 1987. 7. Kehrl, J. H., Roberts, A. B., Wakefield, L. M., Jakowlew, S., Sporn, M. B., and Fauci, A. S.. J. Immunol. 137, 3855, 1986.
8. Kehrl, J. H.. Thevenin, C., Rieckmann, P., and Fauci, A. S., J. Immunol. 146,4016, 1991. 9. Coffman, R. L., Lebman, D. A., and Shrader, B., J. Exp. Med. 170, 1039, 1989. IO. Sonoda, E., Matsumoto, R., Hitoshi, Y., Ishii, T., Sugimoto, M., Araki, S., Tominaga, A., Yamaguchi, N., and Takatsu, K., J. Exp. Med. 170, 1415, 1989. 11. Kim, P.-H., and Kagnoff, M. F.. J. Zmmunol. 144, 3411, 1990. 12. Kim, P.-H., and Kagnoff, M. F., J. Immunol. 145, 3773, 1990. 13. Lebman, D. A., Nomura, D. Y., Coffman, R. L., and Lee, F. D.. Proc. Nutl. Acud. Sci. USA 87, 3962, 1990. 14. Chen, S.-S., and Li, Q., Cell Immunol. 128, 353, 1990. 15. Kekow, J., Wachsman, W., McCutchan, J. A.. Gross, W. L.. Zacharia, M., Carson, D. A.. and Lotz, M.. J. Clin. Invest. 87, 1010, 1991. 16. Van Vlasselaer, P., Punnonen, J.. and De Vries, J. E., J. Immunol. 148, 2062. 1992. 17. Czerkinsky, C., Moldoveanu, Z., Mestecky, J., Nilsson, L.-A.. and Ouchterlony, 0.. J. Immunol. Methods 115, 31, 1988.
428
VAN DEN WALL BAKE ET AL.
18. Valentijn, R. M., Radl, J., Haaijman, J. J., Vermeer, B. J., Weening, J. J., Kauffman, R. H., Daha, M. R., and Van Es, L. A., Kidney ht. 26,760, 1984. 19. Kehrl, J. H., Taylor, A. S., Delsing, G. A., Roberts, A. B., Sporn, M. B., and Fauci, A. S., J. Immunol. 143, 1868, 1989. 20. Lane, H. C., Masur, H., Edgar, L. C., Whalen, G., Rook, A. H., and Fauci, A. S., N. Engl. J. Med. 309, 453, 1983. 2 1. Seligmann, M., Chess,L., Fahey, J. L., Fauci, A. S., Lachmann, P. J., L’Age-Stehr, J., Ngu, J., Pinching, A. J., Rosen, F. S., Spira, T. J., and Wybran, J., N. Engl. J. Med. 311, 1286, 1984. 22. Islam, K. B., Nilsson, L., Sideras, P., Hammarstrom, L., and Smith, C. I. E., ht. Immunol. 3, 1099, 1991. 23. Defiance, T., Vanbervliet, B., Briere, F., Durand, I., Rousset, F., and Banchereau, J., J. Exp. Med. 175, 671, 1992.
IMMUNOLOGY
144,417-428( 1992)
Transforming Growth Factor-P inhibits the Production of IgG, IgM, and IgA in Human Lymphocyte Cultures A.WARMOLDL.VANDENWALLBAKE,KELLEY P.BLACK, ROSE KULHAVY,JIRI MESTECKY,ANDSUSANJACKSON Department of Microbiology, University of Alabama at Birmingham, Birmingham, Alabama 35294 Received April 9, 1992; acceptedJuly 6, 1992 Transforming growth factor p (TGFO) has potent immunoregulatory effects acting on both T and B cells. It strongly inhibits secretion of IgG and IgM in human and murine B cell cultures, but has been shown to have an enhancing effect on IgA production in the mouse. We have studied the effect of TGFfl on the production of IgA in human lymphocyte cultures. The addition of TGFP to pokeweed-stimulated peripheral blood lymphocytes resulted in a suppression of IgA production of both subclasses,similar in magnitude to the suppressionof IgG and IgM production. Membrane IgA expression was not increased by culturing tonsillar lymphocytes with TGFP. In conclusion, we find no evidence for a selective enhancing effect of TGF@ on IgA synthesis in humans, in contrast to the findings reported in mice. 0 1992 Academic Pm, Inc.
INTRODUCTION Transforming growth factor p (TGFP) is a member of a family of cytokines synthesized by a large variety of cell types and has a wide spectrum of biological effects (l-3). Three members of this family, designated TGF@, TGF/32, and TGFP3, have been identified in mammals. TGFP 1, a 25kDa homodimer, is now recognized as an important mediator of several cellular and physiologic processes,modulating growth and differentiation of many cell types (3). Its amino acid sequence is identical in various mammalian species,including humans and pigs (4). Recent studies have made it increasingly clear that TGFP 1 has potent immunoregulatory actions. It is produced by both T and B lymphocytes and is the most potent endogenous suppressant of lymphocyte proliferation known. TGFPl inhibits the proliferation of T cells in response to IL-2 (5) and suppressesthe development of cytotoxic T cells (6). With respect to B cells, TGFPl also inhibits factor-dependent proliferation, as well as secretion of IgG and IgM in humans (5,7,8). In the mouse, TGFfll also suppressesthe production of IgG and IgM in lipopolysaccharide-stimulated B cells, but leads to a selective increase in the production of IgA (9- 1l), possibly by acting as a switch factor ( 12, 13). However, in mice this enhancing effect on IgA production depends on the type of B cell stimulation used: basal and IL-5-induced IgA production by mesenteric lymph node cells is profoundly suppressedby TGFP 1, whereaslipopolysaccharide-induced IgA secretion is increased (14). In humans the effects of TGFPl on IgA synthesis have not yet been clearly defined, although indirect evidence suggestsa suppressive effect similar to that observed for IgG and IgM (15). Others have recently found evidence for a direct 417 0008-8749/92$5.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
418
VAN
DEN
WALL
BAKE
ET AL.
switching effect to the IgA isotype (16). In view of the potentially important role of TGF/3 1 as an isotype-specific regulation factor, we studied the effect of TGFP 1 on IgA synthesis by human peripheral blood and tonsillar lymphocytes. MATERIALS AND METHODS Cell Culture Peripheral blood mononuclear cells (PBMC) were isolated from heparinized venous blood from healthy white adults ( 17 males, 3 females,ages25-63 years) using a standard Ficoll density gradient (Lymphocyte Separation Media, Organon Teknika Corp., Durham, NC). Mononuclear cells from tonsils were prepared from surgical specimens provided by The Children’s Hospital (Birmingham, AL). Tonsils were removed for tonsillitis in a 6-year-old black female and a 6-year-old white male, and were carefully dissected in tissue culture media. Mononuclear cells were isolated from the cell suspensions by density gradient centrifugation using Lymphocyte Separation Media. Isolated mononuclear cells were washed twice in cold Dulbecco’s phosphate-buffered saline (PBS; Mediatech, Washington, DC), and contaminating red cells lysed in 0.82% ammonium chloride solution (pH 7.4). The cells were washedtwice more in Dulbecco’s PBS, and once in complete tissue culture media consisting of RPM1 1640 (Mediatech) supplemented with 2 mM Igh.ttamine (GIBCO BRL, Grand Island, NY), penicillin (100 U/ml), and streptomycin (100 pg/ml) (GIBCO), and 10% heat-inactivated fetal bovine serum (Flow-ICN, Costa Mesa, CA). PBMC cultures were performed in 24-well tissue culture plates (Costar, Cambridge, MA). Each well contained 1 X lo6 cells in 1 ml complete tissue culture medium. All cell cultures were performed in duplicate. Cells were cultured for 7 days at 37°C in humidified air containing 5% COZ. At the end of the culture period, the cells were spun down, and the supernatants stored at -20°C until the measurement of immunoglobulin content. The cells were harvestedby vigorous pipetting with cold Dulbecco’s PBS, washed twice in PBS, and once in complete tissue culture media before enumeration of immunoglobulin-producing cells in the ELISPOT assay.The number of viable cells for each type of culture was established by trypan blue exclusion in a hemocytometer. Tonsillar cells were cultured at a concentration of 1 X 106/ml in a 10 ml volume in 25cm2 tissue culture flasks (Corning Glass Works, Coming, NY). After 7 days of culture, cells were harvested and culture supematants were saved for ELISA analysis. Growth Factors All growth factors were added at the beginning of the 7-day culture period. Pokeweed mitogen (GIBCO) was used at a 1:lOO dilution. Recombinant human interleukin 2 (rIL-2) with a purity >95% and specific activity of 2.5 X lo6 BRMP units/mg (Genzyme Research Products, Cambridge, MA) was added to the culture medium at a concentration of 50 U/ml. Highly purified TGFP 1, isolated from human and porcine platelets, was obtained from R & D Systems(Minneapolis, MN). Basedon our own observations and results reported in the literature, porcine TGF/31 was used at a concentration of 1 rig/ml except where noted. Fluorescence-Activated Cell Sorter (FACS) Analysis An aliquot (50 X 1O6cells) of the tonsillar mononuclear cell preparation was stained overnight on ice with FITC-labeled anti-human IgA and.biotin-conjugated anti-human
TGFP INHIBITS
PRODUCTION
OF IgG, IgM, AND
IgA
419
IgM (Becton-Dickinson Immunocytometry Systems, San Jose, CA). The cells were then washed with cold Dulbecco’s PBS and were stained with PE-labeled streptavidin (Becton-Dickinson). After a final wash in Dulbecco’s PBS, the cells were resuspended in culture media and were sorted on a FACStar analyzer (Becton-Dickinson) for cells expressing surface (s) sIgM but not sIgA (s&M’, sIgA cells). sIgM+, sIgA cells were cultured for 48 hr in culture media with or without 1 rig/ml of TGFP 1. At the end of the culture period, the cells were harvested, restained as above, and were analyzed on the FACStar for any shift of sIgM+, sIgA- cells to sIgM+, s&A+ cells. Enzyme-Linked Immunoassay (ELBA) The concentrations of immunoglobulin isotypes in the culture supematants were measuredby ELISA. Polyvinylchloride 96well plates (Dynatech laboratories, Chantilly, VA) were coated overnight at room temperature with 100 &well of the primary antibody appropriately diluted in PBS. For the measurement of IgA we used affinitypurified goat F(ab), anti-human IgA (Pel-Freeze,Rogers, AR). The coating antibodies for the IgG and IgM determinations were heavy chain-specific affinity-purified goat F(ab)Z fragments (Jackson, West Grove, PA). The plates were washed three times with PBS containing 0.05% Tween 20 (Sigma) (PBS-Tween) between all subsequent steps in the ELISA. The wells were blocked for 1 hr at 37°C with PBS-Tween containing 1% bovine serum albumin (BSA) (Sigma; PBS-Tween-BSA). This buffer was also used as a diluent in all subsequent steps of the ELISA. Appropriate serial dilutions of the tissue culture supernatants were added to duplicate wells. As a standard we used seven serial twofold dilutions of a pooled human serum standard containing known concentrations of IgG, IgA, and IgM, added to duplicate wells. After an overnight incubation at 4°C bound immunoglobulins were detected with heavy chain-specific, affinity-purified goat F(ab)* fragments, coupled to biotin (Tago Inc., Burlingame, CA). The incubation with the biotinylated antibodies was performed at 37°C for 2 hr, and followed by avidin-conjugated peroxidase (Sigma) for 1 hr at 37°C. Conversion of the substrate (2,2’-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (Sigma), containing 0.0075% H202 was measuredat 4 14 nm on a I’,, kinetic microplate reader (Molecular Devices Co., Palo Alto, CA). OD readings of blanks, obtained in wells containing no serum, were subtracted from the OD values in the wells containing standards and unknowns. The concentrations of the various isotypes were obtained by interpolation on the standard curves using a four-parameter modeling procedure (Deltasoft, BioMetallics Inc., Princeton, NJ). The final concentration in eachsamplewas calculated as the mean of the results of duplicate cultures. Enzyme-Linked Immunospot Assay (ELISPOT) Immunoglobulin-producing cells of the various isotypes were enumerated by the ELISPOT technique ( 17). The wells of nitrocellulose-bottomed 96-well Millititer HA plates (Millipore, Bedford, MA) were coated overnight at 4°C with 100 &well of the primary antibody appropriately diluted in PBS. The coating antibodies for the enumeration of IgG-, IgA-, and IgM-producing cells were the same as described for the ELISA. For the IgA subclasseswe used murine monoclonal antibodies (a kind gift from Drs. J. Radl and J. J. Haaijman) specific for IgAl (clone 69-l 1.4) and IgA2 (clone H5-5 12) (18). The plates were washed three times in PBS, and allowed to soak for 5 min. Blocking was achieved by incubating the plates for 30 min at 37°C with
420
VAN
DEN
WALL
BAKE
ET AL.
200 &well of complete culture media. The cells obtained from the harvested cultures were added to duplicate wells at varying cell densities in 100 &well of complete culture media, and were incubated in the tissue culture incubator for 3 hr. Subsequently, the cells were removed by three washings with PBS. Between all subsequent stepsthe plates were washed three times in PBS-Tween and were allowed to soak in this buffer for 5 min. The spots were developed by overnight incubation at 4°C with the same biotinylated antibodies as described for the ELISA, diluted in PBS-Tween-BSA. After a 1-hr incubation with 0.5 pg/ml extravidin conjugated to alkaline phosphatase(Sigma) diluted in PBS-Tween, the color was developed with the chromogen 5-bromo-4-chloro3-indolyl phosphate toluidine salt and p-nitroblue tetrazolium chloride (Sigma) (17). Spots were counted under low magnification and were expressed as the number of spot-forming cells (SFC) per IO6 cells added to the wells. The final result for each culture was calculated as the mean of duplicate wells.
Statistical Analysis Both the amounts of immunoglobulin produced and the numbers of SFC proved to be not normally distributed, and were consequently transformed to their logarithms prior to statistical analysis. Comparisons between the various types of culture were performed by analysis of variance for repeated measures,with a P value < 0.0 1 considered as statistically significant, using Fisher’s least significant difference (LSD) method. Data are represented as geometric means with the 95% confidence interval (CI). RESULTS
Dose-ResponseRelationship of TGFBl on Immunoglobulin Production To determine the optimal concentration of porcine TGF@1, we studied the effect of increasing concentrations of TGF@l added to PWM-stimulated PBMC from three healthy volunteers. As shown in Fig. 1, there was a dose-dependent decreasein the induction of SFC of all three isotypes. The suppressive effect of TGFPl reached a plateau at concentrations above 1 @ml. Consequently, this concentration of TGF/31 was used in the subsequent experiments.
Inhibition of P WA&InducedImmunoglobulin Secretionby TGF@l Figure 2 shows the quantities of IgA, IgG, and IgM produced in PBMC cultures from 20 healthy individuals. PWM stimulation led to significant increases in the amounts produced for all three isotypes, as compared to media controls (P < 0.01). The addition of 1 rig/ml porcine TGFPl to the PWM-stimulated cultures significantly suppressedthe quantities secreted of not only IgG and IgM, but also IgA (P < 0.01 for all three isotypes). The extent of suppression for the various isotypes induced by TGF/3 1 is given in Table 1. The total viable cell numbers at the end of the culture period were not decreased by the addition of TGF/?l . In control cultures containing media only cell survival was 52.3% (95% CI 43.4-63.0). In PWM-stimulated cultures cell survival was 83.6% (63. l110.7), and in cultures containing PWM and TGF@l 107.7% (79.6-145.9).
TGFP INHIBITS PRODUCTION
421
OF IgG, IgM, AND IgA
80
0
I@ W
fl
0
3.067
0.2
1.8
0.6
TGF concentration
IgM
5.4
(nglml)
FIG. 1. Dose-response relationship of increasing quantities of porcine TGFP 1 added to PWM-stimulated PBMC cultures (n = 3). The number of SFC of each isotype is given as a percentage of the number found in the absenceof TGFpl, which is set at 100%.Data are the geometric means.
For tonsil mononuclear cells, the amounts of immunoglobulin produced are given in Table 2. Similar to the results obtained using PBMC, the secretion of all three isotypes was inhibited by adding TGFP 1 to the PWM-stimulated cultures. Efect of TGFPl on the Production of IgAl and IgA2
We studied the effect of TGFPl on the induction of SFC of the two subclassesof IgA to determine if there was any differential effect on the subclasses.The numbers of IgA 1 and IgA2 SFC induced in unstimulated PBMC cultures and cultures stimulated with PWM or PWM plus TGFPl are shown in Figs. 3 and 4. PWM stimulation led to significant increases of both IgA 1 and IgA2 SFC compared to media controls (P < 0.0 1). Similar to the effect observed on total IgA secretedin culture, 1 rig/ml of TGFP 1 caused a significant decreasein the numbers of IgAl and IgA2 SFC when added to PWM-stimulated cultures (P < 0.0 1). The geometric mean percentage of suppression induced by TGFPl was 63% for IgAl and 64% for IgA2. As shown in Table 3, the IgA subclassdistribution, calculated as the number of IgAl SFC divided by the sum of IgAl and IgA2 SFC, was not appreciably influenced by the type of stimulation used. This demonstrates that TGFPl did not have a preferential effect on either of the two IgA subclasses. Comparison of porcine and human TGFPl
In a subgroup of six healthy individuals we compared the suppressive effect of porcine and human TGFP 1 on PWM-induced immunoglobulin production in PBMC
422
VAN DEN WALL BAKE ET AL.
p< 0.01
pc 0.01 I
IL-
q IgA IgG El
PWM stimulation
IgM
PWM+TGF
FIG. 2. ImmunogIobulin produced in cultures is shown. Quantities of IgA, I@, and IgM were measured in the supernatants of PBMC cultures (n = 20), using media only, PWM, and PWM plus 1 rig/ml of porcine TGFfll. Data are the geometric means, with vertical bars representing the 95% confidence intervals.
cultures. Porcine and human TGFPl appeared to have very similar inhibitory effects on the production of IgG, IgM, and IgA (Table 4) and on the induction of SFC of both IgA subclasses(Table 5). Using another type of stimulation, the combination of rIL-2 (50 U/ml) and human TGFfi I ( 1 rig/ml), we were equally unable to demonstrate any selective enhancement of IgA production, as shown in Tables 4 and 5. Similar to the effect on PWM stimulation, TGFPl inhibited immunoglobulin synthesis induced by IL-2. Induction of Surface IgA Expression on Tonsillar MNC by TGFPl Sorted sIgM+, sIgK tonsillar MNC were cultured for two days to study the induction of surface IgA expression. The addition of 1 rig/ml of TGFfll to the culture media TABLE I Percentage of Inhibition of PWM-Induced Immunoglobulin Synthesis in PBMC Cultures by Porcine TGFPI Isotype
% Inhibition
IgA
65.0 (46.0-77.3) 56.9 (35.4-71.2) 83.5 (73.8-89.6)
w W
Note. Values are geometric means (n = 20) of percentage of inhibition by the addition of 1 rig/ml porcine TGFflI to PWM-stimulated PBMC cultures. Cells were cultured at 106/ml. Values in parenthesesare the 95% confidence intervals.
TGFP INHIBITS PRODUCTION
423
OF IgG, IgM, AND IgA
TABLE 2 Production of Immunoglobulin in Tonsil Cultures Stimulation
IgA
w
kM
IS1 280 89
147 390 191
193 342
ND” ND ND
Experiment 1 Media PWM PWM + human TGFfil
43 101 47 Experiment 2
Media PWM PWM + human TCFPl
137 159 130
189
Note. Immunoglobulin production (t&ml) in 7-day tonsil cultures containing lo6 mononuclear cells. a ND, not detectable.
did not lead to an increased frequency of surfaceIgA-expressing B cells when compared to medium alone. In the first experiment, the percentage of sIgM+, s&A+ cells was 14.6 for the TGFPl stimulation, and 12.8 for media alone. In the second experiment the percentages were 2.6 and 2.2 respectively. The results from one representative experiment are shown in Fig. 5. 100000
p-z 0.01
p< 0.01
r----II----l
. .
10000-
. l .**
.
.
9.
. .
.
Y
.
. d 0
.
1000 -
l
f
.
.
0-0 l
.
:
Ei
z .
0 v-
.
.*
.
.
l * .* --*
loo-
.
.
a** .
3
%
.
l
.
iE
.
.
.
.
10
media
PWM
PWM+TGF
stimulation FIG. 3. Number of IgAl SFC per lo6 cells, induced in PBMC cultures (n = 1S), using media only, PWM, and PWM plus 1 rig/ml of porcine TGFPl. Data points are the results of the individual cultures, with vertical bars representing the geometric means and 95% confidence intervals.
424
VAN DEN WALL BAKE ET AL.
p< 0.01
p< 0.01
. .* ..: ..I
II 10000
!!! d 0
1000
.
s 0 v2 ih
..
.
.. ..r: ... $
100
.
. .
.** . l* lz . . J
. . .
. . 10
media
PWM
PWM+TGF
stimulation FIG. 4. Number of IgA2 SFC per I O6cells, induced in PBMC cultures (n = 17), using media only, PWM, and PWM plus I rig/ml of porcine TGFfll. Data points are the results of the individual cultures, with vertical bars representing the geometric means and 95% confidence intervals.
DISCUSSION In this study, we have examined the effect of TGFBl on the production of IgA and its subclassesby peripheral blood and tonsil B cells. The study was prompted by analogous studies in mice showing a selective increase in the production of IgA, with concurrent suppression of the secretion of IgG and IgM (9- 13). The results from these murine studies have suggestedthat TGFPl may operate as an &A-specific switch factor (I 2, 13). If such a role for TGF/3 1 could be confirmed in humans, this would have important consequencesfor our understanding of the regulation of the IgA immune response,both in health and in diseaseswhere a dysregulation of the IgA immune response appears to be involved in the pathogenesis. TABLE 3 IgA SubclassDistribution in PBMC Cultures Stimulation Media PWM PWM + porcine TGF@l (1 rig/ml)
% IgAl SFC 69.0 (64.3-74.1) 64.4 (58.0-71.5) 66.4 (61.6-71.6)
Note. Values are geometric means (n = 18) of the percentagesof IgA spot forming cells in PBMC cultures that produce IgA 1. Values in parenthesesare the 95% confidence intervals.
TGFP INHIBITS PRODUCTION
425
OF IgG, IgM, AND IgA
TABLE 4 Production of Immunoglobulin in PBMC Cultures Stimulation Media PWM PWM + porcine TGFPl PWM + human TGFfil IL-2 IL-2 + human TGF@
74 (34-161) 607 (315-1171) 234 (92-597) 215 (90-51 I) 173 (85-354) 62 (29-133)
75 (46-124) 769 (430-I 376) 46 1 (256-828) 417 (232-750) 400 (188-853) 86 (62-l 18)
13 (7-24) 1294 (621-2698) 275 (89-852) 286 (93-882) 598 (191-1869) 29 ( 19-44)
Note. Values are geometric means (n = 6) of nanograms of immunoglobulin produced in PBMC cultures containing lo6 cells. Values in parenthesesare 95% confidence intervals.
The results of our study do not support the hypothesis that TGFPl selectively enhances the production of IgA in humans. In contrast, we found equal or greater inhibition of the secretion of IgA and its subclassesas compared to IgG and IgM. This significant suppression of IgA synthesis was found both in PBMC cultures and in cultures of mucosally derived lymphocytes. Apart from the obvious conclusion that the differences in our findings compared to those reported in mice are the consequence of a difference between the species, other factors may be responsible for the discrepancy. One possible explanation may be in the different type of stimulation in the various studies. Most of the results obtained in mice were the result of stimulation with a combination of lipopolysaccharide (LPS) and TGFP 1 (9- 13). In mice, the type of lymphocyte stimulation has been reported to be an important determinant for the effect of TGF@l. When IL-5 was used as the stimulus for IgA production, a profound inhibitory effect of TGFPl was observed (14). Similarly, the samestudy showed a suppressiveeffect on the basal IgA production by mesenteric lymph node lymphocytes. However, when murine B cells are preactivated with LPS, IL-5 and TGFPl act additively in stimulating IgA production (10). LPS is not an effective stimulus for immunoglobulin production in humans, and we consequently chose other types of stimulation for our study. It is possible that the situation in which TGFPl is used in combination with PWM, as in our study, is more similar
TABLE 5 Spot Forming Cells Induced in PBMC Cultures Stimulation
IsAl
IgA2
Media PWM PWM + porcine TGFfll PWM + human TGFPI IL-2 IL-2 + human TGF@l
28 1 (I 54-509) 3908 (18 13-8424) 1372 (455-4135) 1520 (587-3936) 922 (433-1960) 208 (112-385)
144 (69-303) 2115 (1107-4043) 676 (162-2815) 834 (212-3285) 707 (374-l 337) 211 (155-289)
Note. Values are geometric means (n = 6) of number of SFC per IO6 cells induced in PBMC cultures. Values in parenthesesare 95% confidence intervals.
426
VAN DEN WALL BAKE ET AL.
-loo
ICI
toa
10s
104
-xl0
10’
loa
10’
1oJ
FIG. 5. FACS analysis of tonsi B lymphocytes stained with FITC-labeled anti-human IgA (horizontal axis) and PE-labeledanti-human IgM (vertical axis) after incubation with media alone (A) and after incubation with TGF/31 (B).
to the combination of IL-5 with TGFfll, which leads to inhibition of IgA synthesis in mice. However, this explanation seemsless likely in view of the fact that other forms of costimulation also did not seemto enhance IgA production in our study of human lymphocytes. The combination of rIL-2 and TGFP 1 did not stimulate IgA production in PBMC and led to decreasedproduction of all three isotypes compared to stimulation with IL-2 alone (Tables 4 and 5). We have also examined the effect of killed Stuphylococcus aweus Cowan strain I (0.01%) in combination with rIL-2 (100 U/ml) on purified B cells from human tonsil and appendix. Costimulation with human TGFP 1 (1 rig/ml) did not lead to an increase in IgA synthesis in either of these tissues (results not shown). An alternative explanation is that TGFPl has a mixed effect on cultures containing both B and T cells (such as our PBMC cultures), in contrast to the direct effect on purified B cells used in the murine studies (9- 13). The mixed effect could consist of, on the one hand, an induction of a switch to membrane IgA expression in B cells but, on the other hand, a strong suppressive effect on the accompanying T cells, which are necessary for the subsequent proliferation and differentiation to immunoglobulinproducing cells. This mixed effectcould be responsible for the inhibition of IgA secretion that we found, using the T-cell-dependent stimulation by PWM. However, the negative results in the B cell cultures stimulated with SAC and IL-2 argue against this explanation, as does the absence of any stimulator-y effect of TGF/31 on the expression of surface IgA observed in tonsillar B cells. BecauseTGF@l also has a strong inhibitory effect on the proliferation of activated B cells, this in itself could also be responsible for the suppression of immunoglobulin secretion (7, 19). In the concentrations used, TGF/31 does not appear to be toxic to lymphocytes, and does not lead to a decreasedviability at the end of the 7-day culture period ( 19). Recent indirect evidence also supports our conclusion of a suppressive effect of TGFBI on PWM-induced IgA synthesis in human PBMC. Kekow et al. reported a strong inhibitory effect of PBMC culture supernatants from HIV-infected individuals on the production of IgA, IgG, and IgM by healthy donor PWM-stimulated PBMC cultures ( 15). The suppressive effect of these supernatants could be reversed by the
TGF(3 INHIBITS
PRODUCTION
OF IgG, IgM, AND
I&4
427
addition of anti-TGF/3 antibody, demonstrating that the suppression of immunoglobulin synthesis in vitro was caused by TGFP. 1n vim, there is no evidence of suppressed immunoglobulin production in HIV-infected patients. In contrast, such patients often manifest polyclonal B cell activation, reflected in increased serum levels of IgG, IgA, and IgM (20, 2 1). Possibly, the effectsof increased IL-6 production and Epstein-Barr virus infection override the suppressive effects of TGFP in these patients (15). In summary, we find no evidence of a selective &A-inducing effect of TGFPl in human lymphocyte cultures. Recently, the induction of germ-line transcripts of both IgA subclasseshas been described in human splenic B cells cultured with TGFP 1, as an event preceeding an IgA switch (22). Similarly, human tonsillar sIgD+ B cells could be induced to synthesize IgA using a combination of TGFP, IL- 10, and cross-linking of CD40 (23). In contrast, the IgA secretion by sIgD- B cells was suppressedby TGFP. Another important factor determining the effect of TGFP on IgA production by human B cells may be the duration of exposure to the cytokine. Removal of the TGFP after 24 hr led to increased IgA production, whereas continued presenceof the cytokine in human splenic B cell cultures led to profound suppression of IgA synthesis ( 16). Although an early direct effect on B cells consisting of switching to the IgA isotype seems possible, the net effect of TGFP 1 on PWM-stimulated human lymphocyte cultures is a suppression of IgA synthesis, similar in magnitude to that reported for IgG and IgM. ACKNOWLEDGMENTS The researchof Dr. A. W. L. van den Wall Bake has been made possible by a career development award from the Royal Netherlands Academy of Arts and Sciences.The work was supported by U.S. PHS Grants DK 40117, AI 23952, and AI 18745, and the Nephrology Research and Training Center. We appreciate accessto the Flow Cytometry Core Facility of the UAB AIDS Center.
REFERENCES 1. 2. 3. 4. 5.
Roberts, A. B., Frolik, C. A., Anzano, M. A., and Sporn, M. B., Fed. Proc. 42,2621, 1983. Sporn, M. B., Roberts, A. B., Wakefield, L. M., and de Crombrugghe, B., J. CeN.Biol. 105, 1039, 1987. Massague,J., Annu. Rev. Cell. Biol. 6, 597, 1990. Sporn, M. B., and Roberts, A. B., J. Am. Med. Assoc. 262,938, 1989. Kehrl, J. H., Wakefield, L. M., Roberts, A. B., Jakowlew, S., Alvarez-Mon, M., Derynck, R., Sport-r, M. B., and Fauci, A. S., J. Exp. Med. 163, 1037, 1986. 6. Ranges,G. E., Figari, I. S., Espevik, T., and Palladino, M. A., J. Exp. Med. 166, 991, 1987. 7. Kehrl, J. H., Roberts, A. B., Wakefield, L. M., Jakowlew, S., Sporn, M. B., and Fauci, A. S.. J. Immunol. 137, 3855, 1986.
8. Kehrl, J. H.. Thevenin, C., Rieckmann, P., and Fauci, A. S., J. Immunol. 146,4016, 1991. 9. Coffman, R. L., Lebman, D. A., and Shrader, B., J. Exp. Med. 170, 1039, 1989. IO. Sonoda, E., Matsumoto, R., Hitoshi, Y., Ishii, T., Sugimoto, M., Araki, S., Tominaga, A., Yamaguchi, N., and Takatsu, K., J. Exp. Med. 170, 1415, 1989. 11. Kim, P.-H., and Kagnoff, M. F.. J. Zmmunol. 144, 3411, 1990. 12. Kim, P.-H., and Kagnoff, M. F., J. Immunol. 145, 3773, 1990. 13. Lebman, D. A., Nomura, D. Y., Coffman, R. L., and Lee, F. D.. Proc. Nutl. Acud. Sci. USA 87, 3962, 1990. 14. Chen, S.-S., and Li, Q., Cell Immunol. 128, 353, 1990. 15. Kekow, J., Wachsman, W., McCutchan, J. A.. Gross, W. L.. Zacharia, M., Carson, D. A.. and Lotz, M.. J. Clin. Invest. 87, 1010, 1991. 16. Van Vlasselaer, P., Punnonen, J.. and De Vries, J. E., J. Immunol. 148, 2062. 1992. 17. Czerkinsky, C., Moldoveanu, Z., Mestecky, J., Nilsson, L.-A.. and Ouchterlony, 0.. J. Immunol. Methods 115, 31, 1988.
428
VAN DEN WALL BAKE ET AL.
18. Valentijn, R. M., Radl, J., Haaijman, J. J., Vermeer, B. J., Weening, J. J., Kauffman, R. H., Daha, M. R., and Van Es, L. A., Kidney ht. 26,760, 1984. 19. Kehrl, J. H., Taylor, A. S., Delsing, G. A., Roberts, A. B., Sporn, M. B., and Fauci, A. S., J. Immunol. 143, 1868, 1989. 20. Lane, H. C., Masur, H., Edgar, L. C., Whalen, G., Rook, A. H., and Fauci, A. S., N. Engl. J. Med. 309, 453, 1983. 2 1. Seligmann, M., Chess,L., Fahey, J. L., Fauci, A. S., Lachmann, P. J., L’Age-Stehr, J., Ngu, J., Pinching, A. J., Rosen, F. S., Spira, T. J., and Wybran, J., N. Engl. J. Med. 311, 1286, 1984. 22. Islam, K. B., Nilsson, L., Sideras, P., Hammarstrom, L., and Smith, C. I. E., ht. Immunol. 3, 1099, 1991. 23. Defiance, T., Vanbervliet, B., Briere, F., Durand, I., Rousset, F., and Banchereau, J., J. Exp. Med. 175, 671, 1992.
