T cells and control of C3 gcne expression
Eur. J. Immunol. 1992. 22: 3103-3109
Margaret B. GoldmanAo, Mary Ann KnovichA and John N. Go1dman.O Department of MedicineA and Department of Microbiology and Immunology", The Pennsylvania State University College of Medicine, Hershey
3103
T lymphocytes mediate immunologic control of C3 gene expression* Immunologic control of C3 gene expression by tissue macrophages can be accomplished by treatment of spleen fragments with anti-C3 antibody. We now demonstrate that suppression of C3 requires participation of T lymphocytes of both the CD4+ and CD8+ phenotypes. Pretreatment of splenic tissue with anti-Thy-1.2 monoclonal antibody blocks the ability of the antLC3 antibody to induce C3 suppression. Reduction in either the CD4+ or CD8+ subpopulations of T lymphocytes also abrogates C3 suppression demonstrating that both T cell subsets are required in addition to the inducing antibody. Artificially elevating intracellular levels of CAMP with cholera toxin can partially substitute for the effects mediated by T cells in this reaction. Therefore, normal expression of the C3 gene can be suppressed by a regulatory network that requires the presence of a specific inducing antibody and Tlymphocytes of both the CD4+ and CD8+ subsets.This regulatory network has many similarities to regulatory networks that have been well documented in suppression of specific murine immunoglobulin allotypes.
1 Introduction Administration of antibody directed against individual proteins to newborn animals or to cells in tissue culture may result in suppression of synthesis and secretion of those proteins. This has been accomplished with immunoglobulins [I-41 and with individual complement components [S,61. Several laboratories have demonstrated that the suppression of production of immunoglobulin allotypes [7-101 or idiotypes [111 can be the result of regulatory networks that actively block the normal expression of those proteins. Our laboratory has previously demonstrated that in vivo treatment of newborn mice with anti-C5 antibody can suppress CS production in hepatocytes for months [5, 61. I n vitro treatment of guinea pig spleen cells with antiLC4 antibody can block C4 biosynthesis by macrophages [12-171. Similar regulation of mouse C3 can be accomplished by in vitro treatment of mouse spleen cells with antiX3 antibody (181. Although C4 and C3 biosynthesis gradually resumes, the total amount produced (intracellular plus extracellular) never returns to normal.
C3 protein synthesis and C3 mRNA levels in macrophages were compared during the suppression phase and the recovery phase for spleen cell populations treated with anti-C3 antibody or nonimmune serum. These studies demonstrated that the rate determining steps of immunologic control of C3 gene expression were posttranscriptional
[HI. Immediately after antibody treatment, C3 mRNA levels were either unchanged or were elevated compared to controls even though C3 protein was not detectable as metabolically labeled molecules or as antigen by sensitive immunoassays for C3 protein. Therefore, posttranscriptional mechanisms must account for the block in C3 gene expression. Two days after antibody treatment C3 mRNA levels were reduced SO % . Although this reproducible change in C3 mRNA was demonstrated, it was too small to account for the total disappearance of C3 protein. During the recovery phase, C3 mRNA increased as C3 biosynthesis resumed. The overall conclusion of these studies is that immunologic control of C3 gene expression resulted mainly from impaired translation. With suppression of each of the the three complement components, data were obtained that demonstrated a requirement for lymphocytes in addition to the inducing antibody. There are also many other parallels between in vivo suppression of mouse C5 and in vivo suppression of mouse immunoglobulin allotypes where suppression is actively induced and maintained with T lymphocytes. We, therefore, undertook the present studies to verify that T lymphocytes are important for immunologic control of C3 gene expression and to determine whether either CD4+ or CD8+ lymphocytes or both are required for the regulatory network.
2 Materials and methods 2.1 Animals [I 104671 *' This work was supported by National Institutes of Health Grant
ROI A1 25775 and a grant from the Eleanor Naylor Dana Charitablc Trust. Correspondence: Margaret B. Goldman, Division of lnfectious Diseases and Epidemiology, The Milton S . Hershey Mcdical Center, SO0 University Drive, Hershey, PA 17033, USA Abbreviations:
SpF: Splenic fragments CT: Cholera toxin
0 VCH Vcrlagsgesellschaft mbH. D-6940 Weinheim, 1992
DBA/l, CS7BL/6, BALB/c, BALB/cBy female mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Har: athymic nude mice were obtained from Harlan Sprague Dawley, Inc. (Madison, WI). All animals were housed in a facility that is approved by the American Association for the Accreditation of Laboratory Care and were maintained according to standards outlined in Guide to Care and Use ojlahoratoryAnimals as established by the National Institutes of Health (Bethesda, MD). OO14-29SO/92/1212-3103$3.SO + .25/O
3104
Eur. J. Immunol. 1992. 22: 3103-3109
M. B. Goldman, M. A. Knovich and J. N. Goldman
cose DMEM (Gibco-BRL) supplemented with 25 mM Hepes (Sigma, St. Louis, MO), L-asparagine (36 mg/liter), L-arginine (116 mg/liter), folic acid (6 mg/liter), 1% Lglutamine, 100 U/ml penicillin G (Squibbmarsham, Princeton, NJ), 100 pg/ml streptomycin sulfate (Pfizer Laboratories, New York, NY) and 10 YO heat-inactivated horse or fetal calf serum (Gibco-BRL). Fetal calf serum was used with HO-13-4, 2B6, GK1.5, and 2.43. For 3.155 and T24 the final serum concentration was increased to 20 '70.Horse serum (10 YO)was used for R17.217.1.3.2-ME was added to a final concentration of 0.05 mM for HO-13-4, 2B6, and R17.217.1.3. MEM nonessential amino acids (Gibco-BRL) were added to a final concentration of 0.001 mM for HO-13-4 and 2B6. Media for 2B6 was also supplemented with 1 mM sodium pyruvate. Ascites fluids were prepared according to published procedures [24] by injection of athymic nude mice with either R17.217.1.3, 3.155, or 2B6 cells and by injection of BALB/cBy mice with HO-13-4 cells. IgG monoclonal antibodies were purified from tissue culture supernatants by passage over an ImmunoPure Plus Immobilized Protein G column (Pierce) following the manufacturer's instructions. IgM monoclonal antibodies were purified from ascites fluids by euglobulin precipitation [25]. The specificity of the reagents prepared in our own laboratory was verified by FACS analysis of thymic lymphocytes. The immunofluorescent titer of each monoclonal antibody was determined against mouse thymocytes by indirect staining with a second antibody to IgM that was conjugated to FITC. Each antibody had the following titers: HO-13-4 1:800, 2B6 1:100, and 3.155 1:100. The irrelevant antibody, R17.217 did not stain mouse thymocytes.
2.2 Antisera Rabbit anti-mouse C3 antibody was prepared in our laboratory as previously described [ 181. Goat anti-mouse C3 antibody was purchased from Organon Teknika Cappel (Durham, NC). The specificity of the commercial antibody was identical to that of our own rabbit antibody [lS]. Specific anti-C3 antibodies were purified from the antisera by affinity chromatography and their protein concentration determined in a bicinchoninic acid assay (Pierce, Rockford, IL).The undiluted goat antiserum contained 1.8 mg/ml of specific antLC3 antibody. The undiluted rabbit serum contained 3.4 mg/ml of specific antibody.
2.3 Cell culture Using a Mcllwain tissue chopper (Sybron Brinkmann, Westbury, NY), mouse spleens were sectioned into fragments (1 mm x 1 mm x 1 spleen thickness) and placed in tissue culture with four fragments per well. Visual assessment was used to ensure that each culture well had an equal mass of spleen tissue. Spleen fragments (SpF) were maintained according to published procedures that are standard for our laboratory [18]. Briefly, experimental cultures were treated several days with heat-inactivated (56" for 30 min) anti-C3 antisera at a 1:10 dilution and control cultures were treated with a 1:10 dilution of heat-inactivated nonimmune serum. When the hyperimmune and normal sera were derived from rabbits, treatment lasted 4 days. When the hyperimmune and normal serum were derived from goats, the treatment lasted 7 days. After antibody treatment, fragments were washed three times and maintained thereafter in media containing 2 YO agammaglobulinemic horse serum (Gibco-BRL, Grand Island, NY). At various times after removal of anti-C3 antibody, SpF were placed for 4-18 h in 0.5 ml DMEM without L-methionine (GibcoBRL) to which was added 100 pCi of [35S]L-methionine (New England Nuclear Research Products, Boston, MA) and 2 % agammaglobulinemic horse serum. Individual radiolabeled proteins were isolated by immunoprecipitation and analyzed by SDS-PAGE according to published methods [18]. Autoradiograms were analyzed on a laser densitometer (model 100 A, Molecular Dynamics, Sunnyvale, CA), using PDQuest analysis software (Protein Database Inc. , Huntington Station, NY). 2.4 Monoclonal antibodies and purification Hybridomas that produced monoclonal antibodies were obtained from the following sources: HO-13-4 an IgM mouse anti-Thy-1.2 antibody (ATCC TIB 99) [19], 3.155 an IgM rat anti-Lyt-2 antibody (ATCC TIB211) [20, 211, and R17.217 1.3 an IgGza rat anti-mouse transferrin receptor antibody (ATCC TIB 219) 1221 were obtained from the American Type Culture Collection (Rockville, MD). 2B6, an IgM rat anti-L3T4b antibody, was obtained from Dr. Ethan Shevach (Bethesda, MD) [23]. GK1.5, a rat IgGzb anti-L3T4a antibody and 2.43, a rat IgG2b anti-Lyt-2.2 antibody were obtained from Dr. David Flyer (Hershey, PA) and were originally developed in the laboratory of Dr. Frank Fitch (Chicago, IL) [20, 211. T24/40.7, a rat monoclonal antibody to Thy-1 [21], was also obtained from Dr. Flyer. These clones were maintained in high glu-
2.5 Depletion of T lymphocyte populations .
To deplete SpF of T lymphocytes or T lymphocyte subsets, SpF were incubated twice with mixtures of monoclonal antibody and rabbit complement (Pel Freez, Rogers, AK, or Cedarlane low-tox-M, Accurate Chemical and Scientific Co., Westbury, NY). In individual experiments, each incubation time was either 1 or 2 h as indicated. Flow cytometry was used to assess the extent of cell depletion in some experiments. After treatment with monoclonal antibody and complement, SpF were incubated overnight. SpF were then disrupted using a spleen homogenizer (E-C Cellector, E-C Apparatus Corporation, St. Petersburg, FL) and lymphocytes were isolated by centrifugation at 2000 x g for 1.5 min on a Ficoll-Hypaque gradient (Histopaque 1083, Sigma). For staining, 106 lymphocytes were incubated with IgG monoclonal antibodies that had been purified by passage over protein G columns (Pierce) and conjugated with biotin-LC hydrazide (Pierce) according to the manufacturer's instructions. This was followed by incubation with streptavidin-fluorescein conjugate (Pierce). All dilutions were made in 0.02 M phosphatebuffered saline containing 20 YO fetal calf serum (GibcoBRL), 3 '30bovine serum albumin and 200 pg/ml goat IgG (Sigma). Cells were analyzed on an Epics V Flow Cytometer/Cell Sorter (Coulter Electronics, Hialeah, FL) in the College of Medicine Flow Cytometry Facility.To determine the extent of depletion of CD8+ cells,T lymphocytes were enriched prior to FACS analysis by passage of the FicollHypaque purified lymphocytes over a mouse T lymphocyte column (Biotex, Edmonton, Alberta, Canada). This was
Eur. J. Immunol. 1992. 22: 3103-3109
T cells and control of C3 gene expression
necessary because of the low percentage of CD8+ T cells in the splenic population.
3105
3.1 Role of T cells
SDS-PAGE analysis of nascent radiolabeled C3 that was immunoprecipitated from the supernatants and lysates of SpF cultures at different times after treatment with the immune or nonimmune serum. To determine if Tregulatory cells are required for antigenic suppression of C3, the effect of reducing the total T cell population prior to inducing suppression was investigated.
We have studied antigenic suppression of C3 and the immunologic control of C3 gene expression by taking advantage of an in vitro system that preserves the in vivo microenvironment of C3-producing macrophages [181. Uniform splenic fragments, containing approximately 10 % macrophages and all of the requisite cell types necessary for establishing the regulatory network, were culturied several days in the presence of anti-C3 antiserum or nonimmune serum. At the end of the treatment period, C3 bimynthesis gradually returned i n the antibody-treated cultures but total combined intracellular and extracellular levels were reduced compared to controls. C3 synthesis was assessed by
Fig. 1depicts an experiment in which SpF were treated with different dilutions of ascites-derived mouse IgM monoclonal antibody HO-13-4 that is specific for theThy-1.2 antigen. IgM monoclonal antibodies were chosen because they have been demonstrated to be the most cytotoxic in vitro [26]. SpF were treated with anti-Thy-1.2 and rabbit complement for 1 h at 37 "C, washed, and then treated with either 10 YO nonimmune rabbit serum or 10 % rabbit anti-mouse C3 for 4 days. The fragments were then washed three times and cultured thereafter in media containing 2 % agammaglobulinemic horse serum. For the data depicted in fig. 1A and B, SpF were maintained throughout the experiment in the
3 Results
pro C3
pro C3
C3 a
C3 a
c3 0
c3 0
Anti-C3 Ab
0
+
Anti-T cell Ab
0
0
+
t
1:200 1 : l O O
Anti-C3 Ab
0
+
Anti-T cell Ab
0
0
(B)
(D)
pro C3
pro C 3
C3a
c3 (1
c30
c30
Anti-C3 Ab Anti-T cell
Ab
0
t
0
0
+
+
+
1:200 1 : l O O 1;50
0
+
Anti-T cell Ab 0
0
AntkC3 Ab
+
t
1:200 1 : l O O
+
+
t
112001:lOO 150
Figure I . The effect of T cell depletion and CTon C3 suppression. SpF from DBA/1 female mice were pretreated for 1 h at 37 "C with an IgM anti-Thy-1.2 at a final dilution of 1: 200, 1:100, or 1: 50 plus rabbit complement at a 1:10 dilution.The immunofluoresce titer of the monoclonal antibody by indirect staining was 1 :800. After pretreatment and washing, the SpF were treated 4 days with either 10 % nonimmune rabbit serum (0) or 10 YOrabbit anti-mouse C3 immune serum (+),which contained 3.4 mg/ml of specific anti-C3 antibody (undiluted).The SpF were radiolabeled with 100 KCi of [?j]methionine 4 days after treatment with rabbit serum (A and C) or 8 days after treatment with rabbit serum (B and D).Throughout the experiment half the cultures were maintained without added CT (A and B) and half were maintained in the presence of CT (C and D). Rabbit anti-mouse C3 and rabbit complement used in these experiments were prepared in our laboratory. This experiment is representative of five similar experiments.
3106
Eur. J. Immunol. 1992. 22: 3103-3109
M. B. Goldman, M. A . Knovich and J. N. Goldman
absence of cholera toxin. For the data depicted in Fig. 1C and D, SpF were maintained throughout the experiment in media containing M CT. Fig. 1A and C depict nascent C3 synthesis 4 days after treatment with anti-C3 or nonimmune rabbit serum. Fig. 1B and D depict nascent C3 synthesis 8 days after treatment with antLC3 or nonimmune rabbit serum. Four days after treatment with nonimmune rabbit serum, secreted C3 was readily immunoprecipitated from culture supernatants and detected as C3 a and C3 0 chains as well as precursor molecules on reduced SDS-PAGE gels. In contrast, C3 was barely detectable when SpF supernatants from cultures treated with anti-C3 antibody were analyzed. Pretreatment of SpF with antiThy-1.2 and complement partially or completely abrogated suppression in a dose-related manner. Pretreatment with a 1:200 dilution of antibody had no effect while pretreatment with 1:100 and 1:50 dilutions showed increasing inhibition of suppression. This was most apparent 8 days after antibody treatment (Fig. 1). 3.2 Effect of altered cAMP levels
We had previously demonstrated that in vitro suppression of guinea pig C4 could be augmented by elevating intracelMar levels of cAMP by adding lo-* M CT to all culture media [13].Therefore, half of the SpFcultures were treated with cholera toxin (Fig. 1C and D) and compared to SpF that were cultured simultaneously in the absence of CT (Fig. I A and B). The ability of CT to augment C3 suppression can be seen by comparing lane 2 in Fig. 1B and D. In the absence of CT, SpF treated with anti-C3 antibody still produce less C3 than SpF treated with nonimmune serum, i.e. lane 2 compared to lane 1 of Fig. 1B. In contrast, however, in the presence of CT, SpF treated with anti-C3 antibody produce markedly reduced amounts of C3 even 8days after treatment with antLC3 antibody, i.e. lane 2 of Fig. D.Therefore, moreT cells had to be depleted to achieve the same effect in the presence of CT. Experiments in progress with an SV40-immortalized macrophage cell line demonstrated that suppression was augmented whenT lymphocytes alone were pretreated to elevate their cAMP levels (M. B. Goldman, M. A. Knovich, and J. N. Goldman, manuscript in preparation). When CT was present, a higher proportion of T lymphocytes needed to be depleted to block suppression since the lymphocytes that survived were better effector cells. Thus, artificially elevating intracellular levels of cAMP with CT can partially substitute for the effects mediated by T cells in this reaction. It remains to be proven specifically which of the variety of effects attributed to CT is important for immune modulation of C3. However, since our earlier published work demonstrated that suppression was enhanced by elevating intracellular cAMP by several methods [13], this is almost certainly the mechanism of enhancement in this closely related system. 3.3 Role of T cell subsets
Having established that maintenance of suppression in vitro requires the presence of T lymphocytes, we next addressed the question of whichT cell subsets participate in the active maintenance of suppression. In the next series of experiments, SpF were pretreated with rabbit complement and
monoclonal antibodies to CD4, CD8 or an irrelevant control antibody, anti-mouse transferrin receptor. Both experimental and control cultures received the same pretreatment. This ensured that similar cell populations would be compared, thereby minimizing modulation of C3 levels by cytokines produced by T cells present in one culture but not the other. SpF were then treated for 7 days with either goat anti-mouse C3 or nonimmune goat serum. C3 biosynthesis was compared in cultures treated with the two different sera. Fig. 2 depicts the effects of reduction of individual subpopulations of T lymphocytes prior to treatment with anti-C3 antibody on the ability of antLC3 antibody to induce the regulatory network that maintains suppression. Those SpF that were pretreated with an irrelevant antibody and complement followed by nonimmune goat serum synthesized C3 2 days after removal of the nonimmune serum as detected by radiolabeled precursor C3 and C3 a and chains on SDS-PAGE (Fig. 2, lane 1). Suppression of C3 synthesis was achieved (Fig. 2, lane 2) in those SpF treated with anti-C3 antibody. This demonstrates that pretreatment with an irrelevant antibody and complement did not interfere with the induction of suppression. In marked contrast, if SpF were pretreated with either anti-CD4 antibody (lanes 3 and 4) or anti-CD8 antibody (lanes 5 and 6), subsequent treatment with anti-C3 antibody failed to induce suppression (lanes 4 and 6). Table 1 summarizes three similar experiments. In experiments 1 and 2, SpF were pretreated twice with 1:50 dilutions of monoclonal antibody and a 1:10 dilution of complement each time for 1 h at 37 C. Experiment 1 was carried out in the absence of cholera toxin. Experiment 2 was carried out 1
2
3
4
5
6
proC3
c3a
c3 P
Anti-CS Ab Anti-subset
Ab
0 + o + o + - - irr
aCD4
aCD8
Figure 2. Pretreatment of SpF with antibodies to T cell subsets interferes with the ability to induce C3 suppression by treatment with anti-C3 antibody. SpF obtained from a C57BLl6 female mouse were treated twice for 1h at 37°C each time with a 1:50 dilution of the appropriate antibody (used as an ascites fluid) and a 1:10 dilution of rabbit complement. Irrelevant antibody was R17 217 1.3 reactive with the mouse transferrin receptor. AntiCD4 was B6 anti L3T4 and anti-CD8 was 3.155 anti-Lyt-2. SpF were then washed and treated 7 days either with 10 YO nonimmune normal goat serum ( 0 , lanes 1, 3 and 5 ) or 10 YO goat anti-mouse C3 (+, lanes 2, 4, and 6) which contained 1.8 mglml of specific antLC3 antibody (undiluted). Two days after removal of anti-C3 antibody, SpF were radiolabeled with 100 pCi methionine, and nascent C3 was immunoprecipitated and analyzed on 7.5 % SDS-PAGE gels run under reducing conditions.
T cells and control of C3 gene expression
Eur. J. Immunol. 1992. 22: 3103-3109
in the presence of M CT and is the experiment that is depicted in Fig. 2. Experiment 3 was carried out in a manner analogous to the other two but pretreatment was with monoclonal antibodies that had been purified and whose concentration was increased fivefold over previous experiments. The incubation time during pretreatment was doubled and cultures were gently rocked throughout the pretreatment period. These changes were to optimize conditions for cell depletion based on flow cytometry studies. Either 2 or 4 days after treatment with anti-C3 antibody, C3 suppression was determined from analysis of autoradiograms by laser densitometry by comparing nascent C3 synthesis by SpF treated with nonimmune serum to C3 synthesis by SpF treated with immune serum. The percent suppression was calculated as [l-(A antibody treated/A nonimmune serum treated)] x 100. At the time point shown for each of the three experiments summarized inTable 1,those cultures pretreated with irrelevant antibody and complement achieved at least 75 YOsuppression of C3 biosynthesis after treatment with antLC3 antibody. However, pretreatment with either anti-CD4 or anti-CD8 antibody and complement blocked the ability of the anti-C3 antibody to induce C3 suppression. C3 suppression in the range of only 8 YO-28 YO was observed in those cultures depleted of CD8+ cells compared to suppression. C3 suppression in the range of 75%-80% in the cultures pretreated with irrelevant antibody and complement. In two out of three experiments, depletion of CD4+ cells prevented C3 suppression. In one experiment, only 16 YO C3 suppression was achieved and in the other, C3 synthesis was actually stimulated, i.e.-13 Yo suppression. In the third experiment, suppression decreased to 57 Yo from 75 YO,an insignificant change. To carry out these experiments, we assumed that the IgM monoclonal antibodies would diffuse into the SpF and effectively lyse T lymphocytes in the presence of complement. To determine the actual extent of depletion of the T cell subsets. SpF were incubated with IgM monoclonal antibodies and complement according to our usual procedure, and then cultured overnight.The following day spleen cells were dispersed by teasing apart the SpE Lymphocytes obtained after Ficoll-Hypaque centrifugation were then analyzed by FACS using biotinylated IgG monoclonal antibodies. Because only a third of murine spleen cells are lymphocytes and only a third of these lymphocytes are CDW [21], any determination of the extent of depletion in this cell population required partial purification of T lymphocytes before analysis.These analyses demonstrated that following the procedures used in these experiments, all of Table 1. Effect of depletion of T cell subsets on C3 suppresion
Exp. 1
om
Pretreatment Ab Irrelevant Ab Anti-CD4 Anti-CD8
Exp.2
+cr
Exp.3
+cr
Days after anti-C3 2 2 4
75%=) 80% -13%b) 16% 27% 28%
a) Percent suppression of C3. b) Negative numbers represent stimulation of synthesis.
75% 57% 8%
3107
the detectable CD8+ Tcells and more than half of the CD4+T cells were eliminated. More complete depletion of CD4+ T cells was not attempted since the initial experiments demonstrated that a partial reduction in this population was sufficient to significantly diminish suppression. Therefore, we concluded that reduction of either CD4+ or CD8+ T cell subsets blocks the induction of C3 suppression that occurs when C3-producing macrophages are treated with anti-C3 antibody.
4 Discussion Our laboratory has been studying immunologic control of gene expression in a family of evolutionarily related complement components, C5 [ 5 , 61, C4 [12-171 and C3 [18]. Our initial studies were analogous to previous experiments with immunoglobulin allotype suppression [3, 41. In these experiments F1hybrid mice of C5 sufficient (C5') and C5 deficient (C5-) genotypes, that were treated at birth with antLC5 antibody, exhibited reduced or absent levels of serum C5. Suppression has also been accomplished in vitro by treatment of guinea pig or murine tissue macrophages with antibody directed against the fourth component of guinea pig complement or the third component of mouse complement. In each system there was specific suppression of the complement component recognized by the inducing antibody. Nonspecific stimulation of other proteins synthesized by the target cell was often noted. For example, serum levels of C5 (most of which is derived from hepatocyte synthesis) were suppressed but serum levels of C6 were elevated when newborn mice were treated with anti-C5 antibody [5]. C4 secreted by macrophages in cultured spleen fragments could be suppressed but other macrophage products such as C2 and @-glucuronidase were concurrently stimulated [12, 141. Evidence was obtained that when treatment was terminated, residual antibody was not present in quantities sufficient to give the false appearance of suppression due to formation of immune complexes [5, 6, 14, 161. Suppression, therefore, cannot be explained simply be neutralization of the suppressed component because (a) the effectiveness of an antibody in inducing suppression of a component is unrelated to its ability to bind to or to neutralize that component, (b) the suppressed component disappears intracellularly as well as extracellularly, (c) synthesis of new molecules is suppressed, (d) the quantity of antibody required to suppress is less than that amount that would be required to complex all of the components that could be produced by an untreated animal or a cell culture system in that time period and (e) suppressed animals may exhibit suppression for a brief period, break suppression, and then re-establish suppression without additional administration of antibody. Data from each of the models suggested that immunologic suppression of a normally expressed complement component required cellular components or factors in addition to the inducing antibody. The following suggest that the initial exposure to antibody induces a regulatory network that actively down-regulates biosynthesis. Some mice that broke suppression and exhibited normal C5 serum levels would re-establish C5 suppression months after the initial exposure to antibody [5]. When newborn (C5-C5+)F1
3108
M. B. Goldman, M. A. Knovich and J. N . Goldman
hybrid mice were treated with either (CS-CS-)CS hyperimmune lymphoid cells or (CS-C5-) nonimmune lymphoid cells plus antibody [ 6 ] ,suppression was more effective than with antibody alone [5].That is, suppression was induced in a higher proportion of littermates, suppression was of much longer average duration (in excess of the 9 months observation period), and suppression could be achieved in F1 hybrid strain combinations that were unaffected by antibody treatment alone. Irradiation (900 rad) of donor cells prior to transfer abrogated suppression. Long-lasting suppression of C4 biosynthesis by guinea pig peritoneal macrophages [17] and C3 biosynthesis by an SV40-immortalized macrophage cell line (unpublished observations, M. B. Goldman, M. A . Knovich, and J. N. Goldman) only occurred in the presence of added lymphoid cells. Guinea pig SpF suppressed for C4 or their supernatants could transfer suppression to untreated SpF [16]. Supression was mediated by at least one genetically nonrestricted soluble factor that was produced by SpF treated with anti-C4 antibody from normal guinea pigs but not by antibodytreated SpF from animals with a genetic deficiency of C4. The gel filtration, ion exchange and solubility properties of this soluble factor were different from those of guinea pig immunoglobulins [l6]. Based on our previous results and because of the demonstrated role of T lymphocytes in suppression of immunoglobulin allotypes [7-101 and idiotypes 1111, we undertook experiments to determine if T lymphocytes actively maintain C3 suppression. Pretreatment of SpF with anti-Thy-1.2 antibody and complement diminished suppression. The degree of interference increased with increasing doses of the antibody used for depletion of T lymphocytes (Fig. 1). When suppression was augmented by the presence of cholera toxin in the culture media, the effect of pretreatment with anti-Thy-1.2 was reduced and did not become apparent until later in the time course of the experiment. Therefore, although the presence of cholera toxin augments suppression, it cannot completely overcome the effects of T cell depletion. For each of our models of immune regulation, data have been obtained to demonstrate a need for a cellular component in addition to the inducing antibody. Our previous experiments with mouse CS and guinea pig C4 had demonstrated a role for lymphocytes but had not identified the specific cell populations or subpopulations required for suppression. We now show with depletion of T cell subsets as well as the whole T cell population, that both CD4+CD8- and CD8+CD4T cells are required for suppression (Fig. 2 and Table 1). Since the initial exposure to anti-C3 antibody during the first week in culture must induce the cellular network that is necessary for regulation, the requirement for both T cell subsets is not unexpected. CD4+ cells are known to act as inducers for CD8+ suppressor precursors and effector cells in other systems [27], including allotype suppression [28].At the level of protein synthesis, antigen-specific suppression of complement components has many parallels to antigen-specific suppression of immunoglobulin allotypes in mice. The following similarities exist: The duration of suppression induced by antibody in the F1 hybrids is highly dependent on the parental strain combination. A variable numbcr of suppressed animals are obtained within experimental groups. Animals are seen with fluctuating levels of suppression. Suppression can be mediated by lymphoid cells obtained from a donor that has been
Eur. J. Immunol. 1992. 22: 3103-3109 hyperimmunized. The administration of lymphoid cells at birth may lead to a higher percentage of suppressed animals and a longer duration of suppression. Benaroch and Bordenave have recently demonstrated in vivo the effect of T cell subsets in mouse allotype suppression [7-101. Their results with an immunoglobulin regulatory network [8] are strikingly similar to our results with the complement regulatory network [6] except that unsensitized cells were effective in inducing suppression but not to the extent of sensitized cells [7]. Allotype suppression could not be induced by the cell donor pool if (a) CD4+ or CD8+ cells were depleted in vitro prior to transfer or if (b) CD4+ or CD8+ cells were depleted in vivo subsequent to transfer into newborn mice that were F1 hybrids at an immunoglobulin allotype locus [lo]. Similarly purified CD4+ or CD8+ cells alone had no effect while their mixture was highly effective when administered to newborns. In suppressed animals, elimination of CD8+ but not CD4+ T lymphocytes immediately abrogated suppression 191. This implies that once suppression is achieved, only CD8+ cells are essential for its active maintenance and this is not consistent with a role for clonal deletion. Bartnes and Hannestad [28] have achieved suppression of the Igh-lb mouse immunoglobulin allotype both in vivo and in vitro with cloned CD4+ cells of the Thl subset. However, while in vivo suppression by adoptive transfer of these cloned CD4+ cells was long lasting 1281, it was of much shorter duration than the Igh-lb suppression that resulted from the transfer of both CD4+ and CD8+ cells [lo]. The parallels of the two networks at the level of molecular modulation of gene expression are unknown. In the murine model we have determined that the changes in total C3 mRNA that occur after antibody treatment are insufficient to account for the complete disappearance of C3 protein [181.Therefore, immunologic control of C3 gene expression through this regulatory network is likely t o involve predominantly interference with translation with only minor control of processing or transcription. At the time of preparation of this manuscript, we are unaware of any published report that has determined if the regulatory networks that are responsible for suppression of allotype expression lead to interference with translation as has already been demonstrated in the complement regulatory networks [181. Both regulatory networks are dependent on the presence of CD4+ and CD8+ lymphocytes for induction. Other similarities may exist in cellular interactions that actively maintain suppression or posttranscriptional control of gene expression. The authors wish to thank Drs. David Flyer and Betsy OhlssonWilhelm for their helpful discussionc and advice on the use offlow cytometry in these studies and Drs. David Flyer and Ethan Shevach for supplying relevant hybridomas. The authors thank Ms. Joanne Hutton for assistance in preparation of the manuscript. Received March 3, 1992; in final revised form August 11, 1992.
5 References 1 Gause, A.,Yoshida, N., Kappen, C. and Rajewsky, K., Eur. J. Immunol. 1987. 17: 981. 2 Lalor, P. A . , Stall, A. M., Adams, S. and Herzenberg, L. A , , Eur. J. Immunol. 1989. 19: 501. 3 Mage, R. G., Transplant. Rev. 1975. 27: 84.
Eur. J. Immunol. 1992. 22: 3103-3109
4 Jacobson, E. B. and Herzenberg, L. A., J. Exp. Med. 1972. 135: 1151. 5 Goldman. J. N. and Goldman, M. B., J. Immunol. 1978. 120: 400. 6 Goldman, M. B., French. D. L. and Goldman, J. N., J. Immunol. 1979. 122: 246. 7 Benaroch, €? and Bordenave, G., Eur. J. Immunol. 1987. 17: 167. 8 Benaroch. €? andBordenave, G., Eur. J. Immunol. 1988.18: 51. 9 Benaroch, P , Georgatsou, E. and Bordenave, G., J. Exp. Med. 1988. 168: 891. 10 Benaroch. l? and Bordenave, G., J. Immunol. 1989. 142: 1. 11 Weiler, I., Weiler, E., Sprenger, R . and Cosenza, H., Eur. J. Immunol. 1977. 7: 591. 12 Goldman, M. B., O’Rourke, K. S. and Goldman, J. N., Cell. Immunol. 1982. 70: 118. 13 Goldman, M. B., Becker, D. S., O’Rourke, K. S. and Goldman. J. N., J. Immunol. 1985. 135: 2702. 14 Goldman, M. B., O’Rourke, K. S., Becker, D. S. and Goldman, J. N.. J. Immunol. 1985. 134: 3298. 15 Goldman, J. N., O’Rourke, K. S.. McMannis, J. D. and Goldman, M. B., Complement 1988. 5: 13. 16 Goldman, J. N., O’Rourke, K. S. and Goldman, M. B., Cell. Immunol. 1985. 96: 26.
T cells and control of C3 gene cxpression
3109
17 McMannis, J. D., Goldman, M. B. and Goldman, J. N., Cell. Immunol. 1987. 106: 22. 18 Goldman, M. B., Knovich, M. A. and Goldman, J. N., J. Immunol. 1991. 147: 4248. 19 Marshak-Rothstein, A , , Fink, €!, Gridley, T., Raulet, D. H., Bevan, M. J. and Gefter, M., J. Immunol. 1979. 122: 2491. 20 Sarmiento, M., Glasebrook, A. L. and Fitch, F. W., J. Immunol. 1980. 125: 2665. 21 Dialynas, D. l?,Wild, D. B., Marrack, €?,Pierres, A., Kappler, J. and Fitch, F. W., Immunol. Rev. 1983. 74: 29. 22 Lesley, J., Hyman, R., Schulte, R. and Troter, J., Cell. Immunol. 1984. 83: 14. 23 Logdberg, L. ,Wassmer, P. and Shevach, E. M., Cell. Immunol. 1985. 94: 299. 24 Harlow, E. and Lane, D. (Eds.) Antibodies: a Laboratory Manual. Cold Spring Harbor Laboratory, New York 1988, p. 274. 25 Garcia-Gonzalez, M., Bettinger, S., Ott, S., Oliver, l?, Kadouche, J. and Philippe, I?, J. Immunol. Methods 1988. 111: 17. 26 Cobbold, S. P., Jayasuriya, A., Nash, A., Prospero. T. D. and Waldmann, H., Nature 1984. 312: 548. 27 Sercarz, E. and Krzych, U., Immunol. Today 1991. 12: 111. 28 Bartnes, K. and Hannestad, K., Eur. J. Immunol. 1991. 21: 2365.
Eur. J. Immunol. 1992. 22: 3103-3109
Margaret B. GoldmanAo, Mary Ann KnovichA and John N. Go1dman.O Department of MedicineA and Department of Microbiology and Immunology", The Pennsylvania State University College of Medicine, Hershey
3103
T lymphocytes mediate immunologic control of C3 gene expression* Immunologic control of C3 gene expression by tissue macrophages can be accomplished by treatment of spleen fragments with anti-C3 antibody. We now demonstrate that suppression of C3 requires participation of T lymphocytes of both the CD4+ and CD8+ phenotypes. Pretreatment of splenic tissue with anti-Thy-1.2 monoclonal antibody blocks the ability of the antLC3 antibody to induce C3 suppression. Reduction in either the CD4+ or CD8+ subpopulations of T lymphocytes also abrogates C3 suppression demonstrating that both T cell subsets are required in addition to the inducing antibody. Artificially elevating intracellular levels of CAMP with cholera toxin can partially substitute for the effects mediated by T cells in this reaction. Therefore, normal expression of the C3 gene can be suppressed by a regulatory network that requires the presence of a specific inducing antibody and Tlymphocytes of both the CD4+ and CD8+ subsets.This regulatory network has many similarities to regulatory networks that have been well documented in suppression of specific murine immunoglobulin allotypes.
1 Introduction Administration of antibody directed against individual proteins to newborn animals or to cells in tissue culture may result in suppression of synthesis and secretion of those proteins. This has been accomplished with immunoglobulins [I-41 and with individual complement components [S,61. Several laboratories have demonstrated that the suppression of production of immunoglobulin allotypes [7-101 or idiotypes [111 can be the result of regulatory networks that actively block the normal expression of those proteins. Our laboratory has previously demonstrated that in vivo treatment of newborn mice with anti-C5 antibody can suppress CS production in hepatocytes for months [5, 61. I n vitro treatment of guinea pig spleen cells with antiLC4 antibody can block C4 biosynthesis by macrophages [12-171. Similar regulation of mouse C3 can be accomplished by in vitro treatment of mouse spleen cells with antiX3 antibody (181. Although C4 and C3 biosynthesis gradually resumes, the total amount produced (intracellular plus extracellular) never returns to normal.
C3 protein synthesis and C3 mRNA levels in macrophages were compared during the suppression phase and the recovery phase for spleen cell populations treated with anti-C3 antibody or nonimmune serum. These studies demonstrated that the rate determining steps of immunologic control of C3 gene expression were posttranscriptional
[HI. Immediately after antibody treatment, C3 mRNA levels were either unchanged or were elevated compared to controls even though C3 protein was not detectable as metabolically labeled molecules or as antigen by sensitive immunoassays for C3 protein. Therefore, posttranscriptional mechanisms must account for the block in C3 gene expression. Two days after antibody treatment C3 mRNA levels were reduced SO % . Although this reproducible change in C3 mRNA was demonstrated, it was too small to account for the total disappearance of C3 protein. During the recovery phase, C3 mRNA increased as C3 biosynthesis resumed. The overall conclusion of these studies is that immunologic control of C3 gene expression resulted mainly from impaired translation. With suppression of each of the the three complement components, data were obtained that demonstrated a requirement for lymphocytes in addition to the inducing antibody. There are also many other parallels between in vivo suppression of mouse C5 and in vivo suppression of mouse immunoglobulin allotypes where suppression is actively induced and maintained with T lymphocytes. We, therefore, undertook the present studies to verify that T lymphocytes are important for immunologic control of C3 gene expression and to determine whether either CD4+ or CD8+ lymphocytes or both are required for the regulatory network.
2 Materials and methods 2.1 Animals [I 104671 *' This work was supported by National Institutes of Health Grant
ROI A1 25775 and a grant from the Eleanor Naylor Dana Charitablc Trust. Correspondence: Margaret B. Goldman, Division of lnfectious Diseases and Epidemiology, The Milton S . Hershey Mcdical Center, SO0 University Drive, Hershey, PA 17033, USA Abbreviations:
SpF: Splenic fragments CT: Cholera toxin
0 VCH Vcrlagsgesellschaft mbH. D-6940 Weinheim, 1992
DBA/l, CS7BL/6, BALB/c, BALB/cBy female mice were obtained from The Jackson Laboratory (Bar Harbor, ME). Har: athymic nude mice were obtained from Harlan Sprague Dawley, Inc. (Madison, WI). All animals were housed in a facility that is approved by the American Association for the Accreditation of Laboratory Care and were maintained according to standards outlined in Guide to Care and Use ojlahoratoryAnimals as established by the National Institutes of Health (Bethesda, MD). OO14-29SO/92/1212-3103$3.SO + .25/O
3104
Eur. J. Immunol. 1992. 22: 3103-3109
M. B. Goldman, M. A. Knovich and J. N. Goldman
cose DMEM (Gibco-BRL) supplemented with 25 mM Hepes (Sigma, St. Louis, MO), L-asparagine (36 mg/liter), L-arginine (116 mg/liter), folic acid (6 mg/liter), 1% Lglutamine, 100 U/ml penicillin G (Squibbmarsham, Princeton, NJ), 100 pg/ml streptomycin sulfate (Pfizer Laboratories, New York, NY) and 10 YO heat-inactivated horse or fetal calf serum (Gibco-BRL). Fetal calf serum was used with HO-13-4, 2B6, GK1.5, and 2.43. For 3.155 and T24 the final serum concentration was increased to 20 '70.Horse serum (10 YO)was used for R17.217.1.3.2-ME was added to a final concentration of 0.05 mM for HO-13-4, 2B6, and R17.217.1.3. MEM nonessential amino acids (Gibco-BRL) were added to a final concentration of 0.001 mM for HO-13-4 and 2B6. Media for 2B6 was also supplemented with 1 mM sodium pyruvate. Ascites fluids were prepared according to published procedures [24] by injection of athymic nude mice with either R17.217.1.3, 3.155, or 2B6 cells and by injection of BALB/cBy mice with HO-13-4 cells. IgG monoclonal antibodies were purified from tissue culture supernatants by passage over an ImmunoPure Plus Immobilized Protein G column (Pierce) following the manufacturer's instructions. IgM monoclonal antibodies were purified from ascites fluids by euglobulin precipitation [25]. The specificity of the reagents prepared in our own laboratory was verified by FACS analysis of thymic lymphocytes. The immunofluorescent titer of each monoclonal antibody was determined against mouse thymocytes by indirect staining with a second antibody to IgM that was conjugated to FITC. Each antibody had the following titers: HO-13-4 1:800, 2B6 1:100, and 3.155 1:100. The irrelevant antibody, R17.217 did not stain mouse thymocytes.
2.2 Antisera Rabbit anti-mouse C3 antibody was prepared in our laboratory as previously described [ 181. Goat anti-mouse C3 antibody was purchased from Organon Teknika Cappel (Durham, NC). The specificity of the commercial antibody was identical to that of our own rabbit antibody [lS]. Specific anti-C3 antibodies were purified from the antisera by affinity chromatography and their protein concentration determined in a bicinchoninic acid assay (Pierce, Rockford, IL).The undiluted goat antiserum contained 1.8 mg/ml of specific antLC3 antibody. The undiluted rabbit serum contained 3.4 mg/ml of specific antibody.
2.3 Cell culture Using a Mcllwain tissue chopper (Sybron Brinkmann, Westbury, NY), mouse spleens were sectioned into fragments (1 mm x 1 mm x 1 spleen thickness) and placed in tissue culture with four fragments per well. Visual assessment was used to ensure that each culture well had an equal mass of spleen tissue. Spleen fragments (SpF) were maintained according to published procedures that are standard for our laboratory [18]. Briefly, experimental cultures were treated several days with heat-inactivated (56" for 30 min) anti-C3 antisera at a 1:10 dilution and control cultures were treated with a 1:10 dilution of heat-inactivated nonimmune serum. When the hyperimmune and normal sera were derived from rabbits, treatment lasted 4 days. When the hyperimmune and normal serum were derived from goats, the treatment lasted 7 days. After antibody treatment, fragments were washed three times and maintained thereafter in media containing 2 YO agammaglobulinemic horse serum (Gibco-BRL, Grand Island, NY). At various times after removal of anti-C3 antibody, SpF were placed for 4-18 h in 0.5 ml DMEM without L-methionine (GibcoBRL) to which was added 100 pCi of [35S]L-methionine (New England Nuclear Research Products, Boston, MA) and 2 % agammaglobulinemic horse serum. Individual radiolabeled proteins were isolated by immunoprecipitation and analyzed by SDS-PAGE according to published methods [18]. Autoradiograms were analyzed on a laser densitometer (model 100 A, Molecular Dynamics, Sunnyvale, CA), using PDQuest analysis software (Protein Database Inc. , Huntington Station, NY). 2.4 Monoclonal antibodies and purification Hybridomas that produced monoclonal antibodies were obtained from the following sources: HO-13-4 an IgM mouse anti-Thy-1.2 antibody (ATCC TIB 99) [19], 3.155 an IgM rat anti-Lyt-2 antibody (ATCC TIB211) [20, 211, and R17.217 1.3 an IgGza rat anti-mouse transferrin receptor antibody (ATCC TIB 219) 1221 were obtained from the American Type Culture Collection (Rockville, MD). 2B6, an IgM rat anti-L3T4b antibody, was obtained from Dr. Ethan Shevach (Bethesda, MD) [23]. GK1.5, a rat IgGzb anti-L3T4a antibody and 2.43, a rat IgG2b anti-Lyt-2.2 antibody were obtained from Dr. David Flyer (Hershey, PA) and were originally developed in the laboratory of Dr. Frank Fitch (Chicago, IL) [20, 211. T24/40.7, a rat monoclonal antibody to Thy-1 [21], was also obtained from Dr. Flyer. These clones were maintained in high glu-
2.5 Depletion of T lymphocyte populations .
To deplete SpF of T lymphocytes or T lymphocyte subsets, SpF were incubated twice with mixtures of monoclonal antibody and rabbit complement (Pel Freez, Rogers, AK, or Cedarlane low-tox-M, Accurate Chemical and Scientific Co., Westbury, NY). In individual experiments, each incubation time was either 1 or 2 h as indicated. Flow cytometry was used to assess the extent of cell depletion in some experiments. After treatment with monoclonal antibody and complement, SpF were incubated overnight. SpF were then disrupted using a spleen homogenizer (E-C Cellector, E-C Apparatus Corporation, St. Petersburg, FL) and lymphocytes were isolated by centrifugation at 2000 x g for 1.5 min on a Ficoll-Hypaque gradient (Histopaque 1083, Sigma). For staining, 106 lymphocytes were incubated with IgG monoclonal antibodies that had been purified by passage over protein G columns (Pierce) and conjugated with biotin-LC hydrazide (Pierce) according to the manufacturer's instructions. This was followed by incubation with streptavidin-fluorescein conjugate (Pierce). All dilutions were made in 0.02 M phosphatebuffered saline containing 20 YO fetal calf serum (GibcoBRL), 3 '30bovine serum albumin and 200 pg/ml goat IgG (Sigma). Cells were analyzed on an Epics V Flow Cytometer/Cell Sorter (Coulter Electronics, Hialeah, FL) in the College of Medicine Flow Cytometry Facility.To determine the extent of depletion of CD8+ cells,T lymphocytes were enriched prior to FACS analysis by passage of the FicollHypaque purified lymphocytes over a mouse T lymphocyte column (Biotex, Edmonton, Alberta, Canada). This was
Eur. J. Immunol. 1992. 22: 3103-3109
T cells and control of C3 gene expression
necessary because of the low percentage of CD8+ T cells in the splenic population.
3105
3.1 Role of T cells
SDS-PAGE analysis of nascent radiolabeled C3 that was immunoprecipitated from the supernatants and lysates of SpF cultures at different times after treatment with the immune or nonimmune serum. To determine if Tregulatory cells are required for antigenic suppression of C3, the effect of reducing the total T cell population prior to inducing suppression was investigated.
We have studied antigenic suppression of C3 and the immunologic control of C3 gene expression by taking advantage of an in vitro system that preserves the in vivo microenvironment of C3-producing macrophages [181. Uniform splenic fragments, containing approximately 10 % macrophages and all of the requisite cell types necessary for establishing the regulatory network, were culturied several days in the presence of anti-C3 antiserum or nonimmune serum. At the end of the treatment period, C3 bimynthesis gradually returned i n the antibody-treated cultures but total combined intracellular and extracellular levels were reduced compared to controls. C3 synthesis was assessed by
Fig. 1depicts an experiment in which SpF were treated with different dilutions of ascites-derived mouse IgM monoclonal antibody HO-13-4 that is specific for theThy-1.2 antigen. IgM monoclonal antibodies were chosen because they have been demonstrated to be the most cytotoxic in vitro [26]. SpF were treated with anti-Thy-1.2 and rabbit complement for 1 h at 37 "C, washed, and then treated with either 10 YO nonimmune rabbit serum or 10 % rabbit anti-mouse C3 for 4 days. The fragments were then washed three times and cultured thereafter in media containing 2 % agammaglobulinemic horse serum. For the data depicted in fig. 1A and B, SpF were maintained throughout the experiment in the
3 Results
pro C3
pro C3
C3 a
C3 a
c3 0
c3 0
Anti-C3 Ab
0
+
Anti-T cell Ab
0
0
+
t
1:200 1 : l O O
Anti-C3 Ab
0
+
Anti-T cell Ab
0
0
(B)
(D)
pro C3
pro C 3
C3a
c3 (1
c30
c30
Anti-C3 Ab Anti-T cell
Ab
0
t
0
0
+
+
+
1:200 1 : l O O 1;50
0
+
Anti-T cell Ab 0
0
AntkC3 Ab
+
t
1:200 1 : l O O
+
+
t
112001:lOO 150
Figure I . The effect of T cell depletion and CTon C3 suppression. SpF from DBA/1 female mice were pretreated for 1 h at 37 "C with an IgM anti-Thy-1.2 at a final dilution of 1: 200, 1:100, or 1: 50 plus rabbit complement at a 1:10 dilution.The immunofluoresce titer of the monoclonal antibody by indirect staining was 1 :800. After pretreatment and washing, the SpF were treated 4 days with either 10 % nonimmune rabbit serum (0) or 10 YOrabbit anti-mouse C3 immune serum (+),which contained 3.4 mg/ml of specific anti-C3 antibody (undiluted).The SpF were radiolabeled with 100 KCi of [?j]methionine 4 days after treatment with rabbit serum (A and C) or 8 days after treatment with rabbit serum (B and D).Throughout the experiment half the cultures were maintained without added CT (A and B) and half were maintained in the presence of CT (C and D). Rabbit anti-mouse C3 and rabbit complement used in these experiments were prepared in our laboratory. This experiment is representative of five similar experiments.
3106
Eur. J. Immunol. 1992. 22: 3103-3109
M. B. Goldman, M. A . Knovich and J. N. Goldman
absence of cholera toxin. For the data depicted in Fig. 1C and D, SpF were maintained throughout the experiment in media containing M CT. Fig. 1A and C depict nascent C3 synthesis 4 days after treatment with anti-C3 or nonimmune rabbit serum. Fig. 1B and D depict nascent C3 synthesis 8 days after treatment with antLC3 or nonimmune rabbit serum. Four days after treatment with nonimmune rabbit serum, secreted C3 was readily immunoprecipitated from culture supernatants and detected as C3 a and C3 0 chains as well as precursor molecules on reduced SDS-PAGE gels. In contrast, C3 was barely detectable when SpF supernatants from cultures treated with anti-C3 antibody were analyzed. Pretreatment of SpF with antiThy-1.2 and complement partially or completely abrogated suppression in a dose-related manner. Pretreatment with a 1:200 dilution of antibody had no effect while pretreatment with 1:100 and 1:50 dilutions showed increasing inhibition of suppression. This was most apparent 8 days after antibody treatment (Fig. 1). 3.2 Effect of altered cAMP levels
We had previously demonstrated that in vitro suppression of guinea pig C4 could be augmented by elevating intracelMar levels of cAMP by adding lo-* M CT to all culture media [13].Therefore, half of the SpFcultures were treated with cholera toxin (Fig. 1C and D) and compared to SpF that were cultured simultaneously in the absence of CT (Fig. I A and B). The ability of CT to augment C3 suppression can be seen by comparing lane 2 in Fig. 1B and D. In the absence of CT, SpF treated with anti-C3 antibody still produce less C3 than SpF treated with nonimmune serum, i.e. lane 2 compared to lane 1 of Fig. 1B. In contrast, however, in the presence of CT, SpF treated with anti-C3 antibody produce markedly reduced amounts of C3 even 8days after treatment with antLC3 antibody, i.e. lane 2 of Fig. D.Therefore, moreT cells had to be depleted to achieve the same effect in the presence of CT. Experiments in progress with an SV40-immortalized macrophage cell line demonstrated that suppression was augmented whenT lymphocytes alone were pretreated to elevate their cAMP levels (M. B. Goldman, M. A. Knovich, and J. N. Goldman, manuscript in preparation). When CT was present, a higher proportion of T lymphocytes needed to be depleted to block suppression since the lymphocytes that survived were better effector cells. Thus, artificially elevating intracellular levels of cAMP with CT can partially substitute for the effects mediated by T cells in this reaction. It remains to be proven specifically which of the variety of effects attributed to CT is important for immune modulation of C3. However, since our earlier published work demonstrated that suppression was enhanced by elevating intracellular cAMP by several methods [13], this is almost certainly the mechanism of enhancement in this closely related system. 3.3 Role of T cell subsets
Having established that maintenance of suppression in vitro requires the presence of T lymphocytes, we next addressed the question of whichT cell subsets participate in the active maintenance of suppression. In the next series of experiments, SpF were pretreated with rabbit complement and
monoclonal antibodies to CD4, CD8 or an irrelevant control antibody, anti-mouse transferrin receptor. Both experimental and control cultures received the same pretreatment. This ensured that similar cell populations would be compared, thereby minimizing modulation of C3 levels by cytokines produced by T cells present in one culture but not the other. SpF were then treated for 7 days with either goat anti-mouse C3 or nonimmune goat serum. C3 biosynthesis was compared in cultures treated with the two different sera. Fig. 2 depicts the effects of reduction of individual subpopulations of T lymphocytes prior to treatment with anti-C3 antibody on the ability of antLC3 antibody to induce the regulatory network that maintains suppression. Those SpF that were pretreated with an irrelevant antibody and complement followed by nonimmune goat serum synthesized C3 2 days after removal of the nonimmune serum as detected by radiolabeled precursor C3 and C3 a and chains on SDS-PAGE (Fig. 2, lane 1). Suppression of C3 synthesis was achieved (Fig. 2, lane 2) in those SpF treated with anti-C3 antibody. This demonstrates that pretreatment with an irrelevant antibody and complement did not interfere with the induction of suppression. In marked contrast, if SpF were pretreated with either anti-CD4 antibody (lanes 3 and 4) or anti-CD8 antibody (lanes 5 and 6), subsequent treatment with anti-C3 antibody failed to induce suppression (lanes 4 and 6). Table 1 summarizes three similar experiments. In experiments 1 and 2, SpF were pretreated twice with 1:50 dilutions of monoclonal antibody and a 1:10 dilution of complement each time for 1 h at 37 C. Experiment 1 was carried out in the absence of cholera toxin. Experiment 2 was carried out 1
2
3
4
5
6
proC3
c3a
c3 P
Anti-CS Ab Anti-subset
Ab
0 + o + o + - - irr
aCD4
aCD8
Figure 2. Pretreatment of SpF with antibodies to T cell subsets interferes with the ability to induce C3 suppression by treatment with anti-C3 antibody. SpF obtained from a C57BLl6 female mouse were treated twice for 1h at 37°C each time with a 1:50 dilution of the appropriate antibody (used as an ascites fluid) and a 1:10 dilution of rabbit complement. Irrelevant antibody was R17 217 1.3 reactive with the mouse transferrin receptor. AntiCD4 was B6 anti L3T4 and anti-CD8 was 3.155 anti-Lyt-2. SpF were then washed and treated 7 days either with 10 YO nonimmune normal goat serum ( 0 , lanes 1, 3 and 5 ) or 10 YO goat anti-mouse C3 (+, lanes 2, 4, and 6) which contained 1.8 mglml of specific antLC3 antibody (undiluted). Two days after removal of anti-C3 antibody, SpF were radiolabeled with 100 pCi methionine, and nascent C3 was immunoprecipitated and analyzed on 7.5 % SDS-PAGE gels run under reducing conditions.
T cells and control of C3 gene expression
Eur. J. Immunol. 1992. 22: 3103-3109
in the presence of M CT and is the experiment that is depicted in Fig. 2. Experiment 3 was carried out in a manner analogous to the other two but pretreatment was with monoclonal antibodies that had been purified and whose concentration was increased fivefold over previous experiments. The incubation time during pretreatment was doubled and cultures were gently rocked throughout the pretreatment period. These changes were to optimize conditions for cell depletion based on flow cytometry studies. Either 2 or 4 days after treatment with anti-C3 antibody, C3 suppression was determined from analysis of autoradiograms by laser densitometry by comparing nascent C3 synthesis by SpF treated with nonimmune serum to C3 synthesis by SpF treated with immune serum. The percent suppression was calculated as [l-(A antibody treated/A nonimmune serum treated)] x 100. At the time point shown for each of the three experiments summarized inTable 1,those cultures pretreated with irrelevant antibody and complement achieved at least 75 YOsuppression of C3 biosynthesis after treatment with antLC3 antibody. However, pretreatment with either anti-CD4 or anti-CD8 antibody and complement blocked the ability of the anti-C3 antibody to induce C3 suppression. C3 suppression in the range of only 8 YO-28 YO was observed in those cultures depleted of CD8+ cells compared to suppression. C3 suppression in the range of 75%-80% in the cultures pretreated with irrelevant antibody and complement. In two out of three experiments, depletion of CD4+ cells prevented C3 suppression. In one experiment, only 16 YO C3 suppression was achieved and in the other, C3 synthesis was actually stimulated, i.e.-13 Yo suppression. In the third experiment, suppression decreased to 57 Yo from 75 YO,an insignificant change. To carry out these experiments, we assumed that the IgM monoclonal antibodies would diffuse into the SpF and effectively lyse T lymphocytes in the presence of complement. To determine the actual extent of depletion of the T cell subsets. SpF were incubated with IgM monoclonal antibodies and complement according to our usual procedure, and then cultured overnight.The following day spleen cells were dispersed by teasing apart the SpE Lymphocytes obtained after Ficoll-Hypaque centrifugation were then analyzed by FACS using biotinylated IgG monoclonal antibodies. Because only a third of murine spleen cells are lymphocytes and only a third of these lymphocytes are CDW [21], any determination of the extent of depletion in this cell population required partial purification of T lymphocytes before analysis.These analyses demonstrated that following the procedures used in these experiments, all of Table 1. Effect of depletion of T cell subsets on C3 suppresion
Exp. 1
om
Pretreatment Ab Irrelevant Ab Anti-CD4 Anti-CD8
Exp.2
+cr
Exp.3
+cr
Days after anti-C3 2 2 4
75%=) 80% -13%b) 16% 27% 28%
a) Percent suppression of C3. b) Negative numbers represent stimulation of synthesis.
75% 57% 8%
3107
the detectable CD8+ Tcells and more than half of the CD4+T cells were eliminated. More complete depletion of CD4+ T cells was not attempted since the initial experiments demonstrated that a partial reduction in this population was sufficient to significantly diminish suppression. Therefore, we concluded that reduction of either CD4+ or CD8+ T cell subsets blocks the induction of C3 suppression that occurs when C3-producing macrophages are treated with anti-C3 antibody.
4 Discussion Our laboratory has been studying immunologic control of gene expression in a family of evolutionarily related complement components, C5 [ 5 , 61, C4 [12-171 and C3 [18]. Our initial studies were analogous to previous experiments with immunoglobulin allotype suppression [3, 41. In these experiments F1hybrid mice of C5 sufficient (C5') and C5 deficient (C5-) genotypes, that were treated at birth with antLC5 antibody, exhibited reduced or absent levels of serum C5. Suppression has also been accomplished in vitro by treatment of guinea pig or murine tissue macrophages with antibody directed against the fourth component of guinea pig complement or the third component of mouse complement. In each system there was specific suppression of the complement component recognized by the inducing antibody. Nonspecific stimulation of other proteins synthesized by the target cell was often noted. For example, serum levels of C5 (most of which is derived from hepatocyte synthesis) were suppressed but serum levels of C6 were elevated when newborn mice were treated with anti-C5 antibody [5]. C4 secreted by macrophages in cultured spleen fragments could be suppressed but other macrophage products such as C2 and @-glucuronidase were concurrently stimulated [12, 141. Evidence was obtained that when treatment was terminated, residual antibody was not present in quantities sufficient to give the false appearance of suppression due to formation of immune complexes [5, 6, 14, 161. Suppression, therefore, cannot be explained simply be neutralization of the suppressed component because (a) the effectiveness of an antibody in inducing suppression of a component is unrelated to its ability to bind to or to neutralize that component, (b) the suppressed component disappears intracellularly as well as extracellularly, (c) synthesis of new molecules is suppressed, (d) the quantity of antibody required to suppress is less than that amount that would be required to complex all of the components that could be produced by an untreated animal or a cell culture system in that time period and (e) suppressed animals may exhibit suppression for a brief period, break suppression, and then re-establish suppression without additional administration of antibody. Data from each of the models suggested that immunologic suppression of a normally expressed complement component required cellular components or factors in addition to the inducing antibody. The following suggest that the initial exposure to antibody induces a regulatory network that actively down-regulates biosynthesis. Some mice that broke suppression and exhibited normal C5 serum levels would re-establish C5 suppression months after the initial exposure to antibody [5]. When newborn (C5-C5+)F1
3108
M. B. Goldman, M. A. Knovich and J. N . Goldman
hybrid mice were treated with either (CS-CS-)CS hyperimmune lymphoid cells or (CS-C5-) nonimmune lymphoid cells plus antibody [ 6 ] ,suppression was more effective than with antibody alone [5].That is, suppression was induced in a higher proportion of littermates, suppression was of much longer average duration (in excess of the 9 months observation period), and suppression could be achieved in F1 hybrid strain combinations that were unaffected by antibody treatment alone. Irradiation (900 rad) of donor cells prior to transfer abrogated suppression. Long-lasting suppression of C4 biosynthesis by guinea pig peritoneal macrophages [17] and C3 biosynthesis by an SV40-immortalized macrophage cell line (unpublished observations, M. B. Goldman, M. A . Knovich, and J. N. Goldman) only occurred in the presence of added lymphoid cells. Guinea pig SpF suppressed for C4 or their supernatants could transfer suppression to untreated SpF [16]. Supression was mediated by at least one genetically nonrestricted soluble factor that was produced by SpF treated with anti-C4 antibody from normal guinea pigs but not by antibodytreated SpF from animals with a genetic deficiency of C4. The gel filtration, ion exchange and solubility properties of this soluble factor were different from those of guinea pig immunoglobulins [l6]. Based on our previous results and because of the demonstrated role of T lymphocytes in suppression of immunoglobulin allotypes [7-101 and idiotypes 1111, we undertook experiments to determine if T lymphocytes actively maintain C3 suppression. Pretreatment of SpF with anti-Thy-1.2 antibody and complement diminished suppression. The degree of interference increased with increasing doses of the antibody used for depletion of T lymphocytes (Fig. 1). When suppression was augmented by the presence of cholera toxin in the culture media, the effect of pretreatment with anti-Thy-1.2 was reduced and did not become apparent until later in the time course of the experiment. Therefore, although the presence of cholera toxin augments suppression, it cannot completely overcome the effects of T cell depletion. For each of our models of immune regulation, data have been obtained to demonstrate a need for a cellular component in addition to the inducing antibody. Our previous experiments with mouse CS and guinea pig C4 had demonstrated a role for lymphocytes but had not identified the specific cell populations or subpopulations required for suppression. We now show with depletion of T cell subsets as well as the whole T cell population, that both CD4+CD8- and CD8+CD4T cells are required for suppression (Fig. 2 and Table 1). Since the initial exposure to anti-C3 antibody during the first week in culture must induce the cellular network that is necessary for regulation, the requirement for both T cell subsets is not unexpected. CD4+ cells are known to act as inducers for CD8+ suppressor precursors and effector cells in other systems [27], including allotype suppression [28].At the level of protein synthesis, antigen-specific suppression of complement components has many parallels to antigen-specific suppression of immunoglobulin allotypes in mice. The following similarities exist: The duration of suppression induced by antibody in the F1 hybrids is highly dependent on the parental strain combination. A variable numbcr of suppressed animals are obtained within experimental groups. Animals are seen with fluctuating levels of suppression. Suppression can be mediated by lymphoid cells obtained from a donor that has been
Eur. J. Immunol. 1992. 22: 3103-3109 hyperimmunized. The administration of lymphoid cells at birth may lead to a higher percentage of suppressed animals and a longer duration of suppression. Benaroch and Bordenave have recently demonstrated in vivo the effect of T cell subsets in mouse allotype suppression [7-101. Their results with an immunoglobulin regulatory network [8] are strikingly similar to our results with the complement regulatory network [6] except that unsensitized cells were effective in inducing suppression but not to the extent of sensitized cells [7]. Allotype suppression could not be induced by the cell donor pool if (a) CD4+ or CD8+ cells were depleted in vitro prior to transfer or if (b) CD4+ or CD8+ cells were depleted in vivo subsequent to transfer into newborn mice that were F1 hybrids at an immunoglobulin allotype locus [lo]. Similarly purified CD4+ or CD8+ cells alone had no effect while their mixture was highly effective when administered to newborns. In suppressed animals, elimination of CD8+ but not CD4+ T lymphocytes immediately abrogated suppression 191. This implies that once suppression is achieved, only CD8+ cells are essential for its active maintenance and this is not consistent with a role for clonal deletion. Bartnes and Hannestad [28] have achieved suppression of the Igh-lb mouse immunoglobulin allotype both in vivo and in vitro with cloned CD4+ cells of the Thl subset. However, while in vivo suppression by adoptive transfer of these cloned CD4+ cells was long lasting 1281, it was of much shorter duration than the Igh-lb suppression that resulted from the transfer of both CD4+ and CD8+ cells [lo]. The parallels of the two networks at the level of molecular modulation of gene expression are unknown. In the murine model we have determined that the changes in total C3 mRNA that occur after antibody treatment are insufficient to account for the complete disappearance of C3 protein [181.Therefore, immunologic control of C3 gene expression through this regulatory network is likely t o involve predominantly interference with translation with only minor control of processing or transcription. At the time of preparation of this manuscript, we are unaware of any published report that has determined if the regulatory networks that are responsible for suppression of allotype expression lead to interference with translation as has already been demonstrated in the complement regulatory networks [181. Both regulatory networks are dependent on the presence of CD4+ and CD8+ lymphocytes for induction. Other similarities may exist in cellular interactions that actively maintain suppression or posttranscriptional control of gene expression. The authors wish to thank Drs. David Flyer and Betsy OhlssonWilhelm for their helpful discussionc and advice on the use offlow cytometry in these studies and Drs. David Flyer and Ethan Shevach for supplying relevant hybridomas. The authors thank Ms. Joanne Hutton for assistance in preparation of the manuscript. Received March 3, 1992; in final revised form August 11, 1992.
5 References 1 Gause, A.,Yoshida, N., Kappen, C. and Rajewsky, K., Eur. J. Immunol. 1987. 17: 981. 2 Lalor, P. A . , Stall, A. M., Adams, S. and Herzenberg, L. A , , Eur. J. Immunol. 1989. 19: 501. 3 Mage, R. G., Transplant. Rev. 1975. 27: 84.
Eur. J. Immunol. 1992. 22: 3103-3109
4 Jacobson, E. B. and Herzenberg, L. A., J. Exp. Med. 1972. 135: 1151. 5 Goldman. J. N. and Goldman, M. B., J. Immunol. 1978. 120: 400. 6 Goldman, M. B., French. D. L. and Goldman, J. N., J. Immunol. 1979. 122: 246. 7 Benaroch, €? and Bordenave, G., Eur. J. Immunol. 1987. 17: 167. 8 Benaroch. €? andBordenave, G., Eur. J. Immunol. 1988.18: 51. 9 Benaroch, P , Georgatsou, E. and Bordenave, G., J. Exp. Med. 1988. 168: 891. 10 Benaroch. l? and Bordenave, G., J. Immunol. 1989. 142: 1. 11 Weiler, I., Weiler, E., Sprenger, R . and Cosenza, H., Eur. J. Immunol. 1977. 7: 591. 12 Goldman, M. B., O’Rourke, K. S. and Goldman, J. N., Cell. Immunol. 1982. 70: 118. 13 Goldman, M. B., Becker, D. S., O’Rourke, K. S. and Goldman. J. N., J. Immunol. 1985. 135: 2702. 14 Goldman, M. B., O’Rourke, K. S., Becker, D. S. and Goldman, J. N.. J. Immunol. 1985. 134: 3298. 15 Goldman, J. N., O’Rourke, K. S.. McMannis, J. D. and Goldman, M. B., Complement 1988. 5: 13. 16 Goldman, J. N., O’Rourke, K. S. and Goldman, M. B., Cell. Immunol. 1985. 96: 26.
T cells and control of C3 gene cxpression
3109
17 McMannis, J. D., Goldman, M. B. and Goldman, J. N., Cell. Immunol. 1987. 106: 22. 18 Goldman, M. B., Knovich, M. A. and Goldman, J. N., J. Immunol. 1991. 147: 4248. 19 Marshak-Rothstein, A , , Fink, €!, Gridley, T., Raulet, D. H., Bevan, M. J. and Gefter, M., J. Immunol. 1979. 122: 2491. 20 Sarmiento, M., Glasebrook, A. L. and Fitch, F. W., J. Immunol. 1980. 125: 2665. 21 Dialynas, D. l?,Wild, D. B., Marrack, €?,Pierres, A., Kappler, J. and Fitch, F. W., Immunol. Rev. 1983. 74: 29. 22 Lesley, J., Hyman, R., Schulte, R. and Troter, J., Cell. Immunol. 1984. 83: 14. 23 Logdberg, L. ,Wassmer, P. and Shevach, E. M., Cell. Immunol. 1985. 94: 299. 24 Harlow, E. and Lane, D. (Eds.) Antibodies: a Laboratory Manual. Cold Spring Harbor Laboratory, New York 1988, p. 274. 25 Garcia-Gonzalez, M., Bettinger, S., Ott, S., Oliver, l?, Kadouche, J. and Philippe, I?, J. Immunol. Methods 1988. 111: 17. 26 Cobbold, S. P., Jayasuriya, A., Nash, A., Prospero. T. D. and Waldmann, H., Nature 1984. 312: 548. 27 Sercarz, E. and Krzych, U., Immunol. Today 1991. 12: 111. 28 Bartnes, K. and Hannestad, K., Eur. J. Immunol. 1991. 21: 2365.
