Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The importance of antigen presenting cell (APC)-derived, cognate second signal pathways in the activation of T lymphocytes is well established (1). While a number of such pathways have been described, one of the best studied is the interaction of CD28/CTLA-4 with the B-7 homologues, B-7.1 (CD80) and B-7.2 (CD86) (2). Engagement of this co-stimulatory pathway in the context of T cell receptor (TCR)-mediated signaling optimizes interleukin (IL)-2 production from human CD4+ T cells, while disruption of this pathway promotes T cell anergy (3, 4). Although the B-7 homologues bind CTLA-4 with 20-fold higher avidity than CD28, CTLA-4 is present at only 3% of the surface density of CD-28 (5). Evidence for the differential roles of these second signal molecules in the regulation of T cell activation states is evolving (6). Thus, selective intervention in T cell responses through manipulations of these interactions is of potential therapeutic interest.

Phenotypic specificity of T lymphocyte responses has been shown in both mice and humans (9, 10). Type 1 (Th1) responses, characterized by the production of IL-2 and interferon  (IFN), govern delayed-type hypersensitivity responses, while type 2 (Th2) responses promote IgE-mediated humoral responses through IL-4 (11). The potential for selective induction of Th1 or Th2 responses through specific engagement of the B-7 homologues has recently been described in mice (12). In this model, engagement of murine CD28 through B7.1 promoted a Th1 response while engagement through B7.2 promoted a Th2 response. Although the molecular basis for this differential signaling effect is unclear, CD28 has been shown to exhibit different signaling characteristics depending on the method of engagement (15). Confirmation of this selectivity in humans has been difficult to establish (16).

In the present study, we report the effects of selective blockade of CD80, CD86, or both on the antigen-driven proliferation and cytokine responses of both human peripheral blood mononuclear cells (PBMCs) and nontransformed, antigen-specific human Th1 and Th2 clones.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Preparation of PBMCs

Preparation of PBMCs was performed as previously described (17). Briefly, 50 ml of heparinized whole blood were diluted 1:1 with serum-free RPMI 1640 supplemented with 1% penicillin/streptomycin. This mixture was overlaid onto 10 ml Ficoll-Paque in 50-ml centrifuge tubes and centrifuged at 800 × g at room temperature for 30 min. PBMCs were collected from the interface and washed twice in the serum-free medium. The cells were then resuspended in RPMI 1640 with 5% human AB serum (complete medium) and either irradiated for use as antigen presenting cells (APCs) in clonal experiments (see below) or used directly in mixed cell cultures. Platelet contamination of these preparations was < 1%; viability by trypan blue exclusion was uniformly greater than or equal to 99%.

Derivation of Antigen-specific T Lymphocyte Clones

The antigen-specific T cell clones used in these experiments were derived as previously described (18, 19). Briefly, PBMCs from an atopic subject with epicutaneous skin test reactivity to short ragweed (RW) (Ambrosia artemisiifolia) and short ragweed-specific IgE = 1761 ng/ml of serum by RAST were cultured in the presence of short ragweed antigen (10 µg/ml). Successive biweekly re-stimulations of the antigen-specific T cells with the major short ragweed antigen, Amb a 1, in the presence of autologous, irradiated PBMCs as a source of APCs was conducted for two cycles. The resulting Amb a 1-specific, CD4+ T cell line was then cloned and subcloned using the limiting dilution technique. Cytokine profiles were determined by both reverse transcription polymerase chain reaction (RT-PCR) and ELISA. Based on these data, phenotypic assignment to Th0, Th1, or Th2 was made. Of 30 well-characterized clones obtained by this procedure, two Th1 and two Th2 clones were selected for further analysis, matched on the basis of their similar degrees of proliferation to Amb a 1 antigen. Antigen-driven proliferative responses of the clones were specific for Amb a 1 (data not shown).

Proliferation Assays

Proliferation assays were performed as previously described (17, 19). Briefly, either 2 × 10⁵ PBMCs/well or 2 × 10⁴ clonal T cells/well with 1.5 × 10⁵ APCs/well were cultured in the absence and presence of antigen in 96-well plates. Culture conditions were designated by the presence or absence of saturating concentrations (as determined by flow cytometry, data not shown) of anti-CD80 and/or anti-CD86 monoclonal antibodies (Ancell Corp., Bayport, MN and Immunotech Inc., Westbrook, ME). The actual concentrations used were 10 µg/ml and 20 µg/ml respectively for the two different CD80 monoclonal antibodies and 5 µg/ml for the CD86 monoclonal antibody; these values are in close agreement with data from the suppliers. No exogenous cytokine was used in these assays. Ragweed was used at 10 µg/ml; tetanus toxoid (TT) was used at 0.1 lF/ml. The cells were preincubated with the antibodies for 30-60 min prior to the addition of antigen. All culture conditions were performed in triplicate and incubated for 72 h. The cultures were then pulsed with 1 µCi/well of [³H]thymidine for an additional 20 h, harvested onto glass fiber filter strips in a PHD multichannel harvester (both from Cambridge Technologies Inc., Watertown, MA), and counted in a beta counter (LS 5000 TD; Beckman Instruments Inc., Fullerton, CA).

Gene Expression Assays

Cytokine gene expression was determined as previously described (20, 21). Briefly, either 5 × 10⁶ PBMCs/condition or 2 × 10⁵ clonal T cells/condition with 3 × 10⁶ APCs/ condition were cultured in the absence and presence of antigen in slanted 14-ml polypropylene tubes. Culture conditions were designated by the presence or absence of saturating concentrations of anti-CD80 and/or anti-CD86 monoclonal antibodies, as determined by flow cytometry. Again, no exogenous cytokine was used. Ragweed was used at 10 µg/ml; tetanus toxoid was used at 0.1 lF/ml. The cells were preincubated with the antibodies for 30-60 min prior to the addition of antigen and further cultured for 12 h. Cultured cells were then pelleted, washed, and subjected to RNA isolation by the RNAzol^B method (Tel-Test, Inc., Friendswood, TX) according to the manufacturer's instructions. Diethylpyrocarbonate-treated water was used in the final resuspension. Strict RNase-free conditions were maintained at all times. RNA was stored at 70°C until further study. Normalization of RNA to approximately 100 ng/µl was achieved with a combination of spectrophotometry, ethidium bromide-stained gel electrophoresis, and RT-PCR for a constitutive marker gene ( actin at subsaturating cycle number), as previously described. A_260/280 values > 1.7 were uniformly obtained. Semi-quantitative RT-PCR was performed with 5 mM magnesium and oligo dT priming, using standard reagents (Perkin-Elmer Cetus, Norwalk, CT) and cytokine-specific primer pairs designed in our laboratory and made at the DNA Core Facility of the Johns Hopkins University (Table 1). The specificity of RT-PCR fragments has been confirmed previously by Southern analysis and specific restriction fragment analysis. All PCR products were visualized by ethidium bromide-stained gel electrophoresis, photographed, and quantitated on a Molecular Dynamics densitometer (Sunnyvale, CA) running ImageQuant software according to manufacturer's instructions.

Cytokine Protein Secretion Assays

Cytokine protein secretion was assessed by ELISA (Biosource, Int., Camarillo, CA) according to the manufacturer's instructions. Sample content was quantitated using the WHO standards provided by the company. Briefly, duplicate cultures to those used in the clonal cytokine gene expression experiments were constructed and incubated for 12 h. Supernatants from these cultures were harvested, and cellular debris was removed by centrifugation. The supernatants were stored at 20°C until assayed. Dilutions of samples, when necessary, were performed in culture medium. All standards and samples were run in duplicate. Most samples were run in two different dilutions and compared for internal consistency.

Statistical Analysis

Mean and standard error values, as well as t test comparisons, were derived using StatView (BrainPower, Inc., Calabasas, CA) on a Macintosh PowerBook 145B computer. Probability values are paired, two-tailed. Percent control values for each condition were calculated relative to stimulated, antibody-free mean values, corrected in every case for background with media alone.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Proliferative Responses of PBMCs

Figure 1 shows the modulation of antigen-driven proliferation of PBMCs by anti-CD80, anti-CD86, and the combination, each at saturating concentrations as determined by flow cytometry. Results are depicted for both Th2-promoting antigen (RW) and Th1-promoting antigen (TT) as the percent of positive control responses, subtracted for background. While anti-CD86 caused a significant downregulation of proliferation for both antigens (% control = 45 ± 4 and 39 ± 12, respectively, for RW and TT, P < 0.03), anti-CD80 was ineffective in both antigen-driven systems (P  0.4). No synergy was evident when anti-CD80 and anti-CD86 were used in combination (P = 0.28 and 0.91, respectively, for RW and TT). Finally, no differences were evident between the responses to the Th1-promoting antigen (TT) and the Th2-promoting antigen (RW). These data parallel those of Freeman and colleagues in anti-CD3-driven purified peripheral human T cells (22). Similar results were obtained in our system with two different anti-CD80 blocking antibodies; isotype-matched control antibodies (Cappel, Organon Teknika Corp., West Chester, PA) were ineffective in modulating proliferative responses to either antigen (data not shown). While both antigens were used at optimal concentrations throughout, titration of antigen concentrations had no effect on any of these data.

View larger version (44K): [in this window] [in a new window]   Figure 1.   Inhibition of antigen-driven proliferation of PBMCs by anti-CD86 but not anti-CD80. PBMCs were stimulated with either RW or TT, as indicated. Data are depicted as % of positive control (antibody-free culture) values, corrected in each case for background (antigen-free culture). Isotype-matched, azide-free control antibodies had no effect on proliferation. Positive control and background values (mean ± SEM), respectively: RW = 5,165 + 1,390 and 1,051 + 152 cpm; TT = 5,553 ± 1,178 and 1,255 ± 84 cpm. n = 4 individual experiments on 4 different subjects.

Cytokine Gene Expression in PBMCs

Figure 2a shows the  actin- and cytokine-specific RT-PCR amplification products from a representative study of ragweed-driven cytokine gene expression in PBMCs in the absence and presence of blocking antibodies to CD80, CD86, or both. Adequate normalization of RNA was confirmed by the equality of RT-PCR amplification products at sub-saturating cycle number for  actin gene expression (top row). Unstimulated PBMCs did not express message for proinflammatory cytokines (column 1). While blockade of CD86 consistently resulted in downregulation of mRNA for IL-4, IL-5, and IFN (column 4, P < 0.02), blockade of CD80 resulted in neither independent efficacy nor synergy with anti-CD86 (columns 3 and 5, respectively). Again, similar results were obtained with a second anti-CD80 blocking antibody, and isotype-matched control antibodies were ineffective in modulating proliferative responses to either antigen (data not shown). Representative data from the same cell donor in the tetanus toxoid-driven system are shown in Figure 3a. Again, while anti-CD86 significantly downregulated proinflammatory cytokine gene expression (P < 0.03), anti-CD80 blocking antibodies provided neither independent efficacy nor synergy with anti-CD86. Although results for both RW and TT are depicted only at the 12-h time point, assays at various other time points yielded similar results. A quantitative representation of these data for the entire group of study subjects is depicted in Figures 2b and 3b. These data are in agreement with the findings of Freeman and colleagues (22); in conjunction with the proliferation data in Figure 1, these results clearly suggest that the primary B7 signaling homologue in antigen-driven PBMCs is CD86.

View larger version (99K): [in this window] [in a new window]   View larger version (30K): [in this window] [in a new window]   Figure 2.   Inhibition of ragweed-driven cytokine gene expression from PBMCs by anti-CD86 but not anti-CD80. (a) Representative data are depicted for IL-4, IL-5, and IFN in rows 2, 3, and 4, respectively. Adequate normalization of RNA was confirmed by  actin gene expression at subsaturating cycle number (row 1). Culture conditions are noted at the top of each lane. The DNA size marker is shown in lane 6. (b) Densitometry data are depicted for all subjects for each RT-PCR target and culture condition. The numbers represent percentage optical density ± SEM of PCR amplification product bands normalized to stimulated, antibody-free controls, corrected for background. The variability of  actin gene expression was less than 3%. n = 4 individual experiments on 4 different subjects.

View larger version (70K): [in this window] [in a new window]   View larger version (29K): [in this window] [in a new window]   Figure 3.   Inhibition of tetanus toxoid-driven cytokine gene expression from PBMCs by anti-CD86 but not anti-CD80. (a) Representative data are depicted for IL-4, IL-5, and IFN in rows 2, 3, and 4, respectively. Adequate normalization of RNA was confirmed by actin gene expression at subsaturating cycle number (row 1). Culture conditions are noted at the top of each lane. The DNA size marker is shown in lane 6. (b) Densitometry data are depicted for all subjects for each RT-PCR target and culture condition. The numbers represent percentage optical density ± SEM of PCR amplification product bands normalized to stimulated, antibody-free controls, corrected for background. The variability of  actin gene expression was less than 3%. n = 4 individual experiments on 4 different subjects.

Proliferative Responses of Th1 and Th2 Clones

Figure 4 depicts the effects of saturating concentrations of anti-CD80, anti-CD86, or both on the ragweed-driven proliferation of Amb a 1-specific Th1 and Th2 clones. Results are depicted as the percent of positive control responses, subtracted for background. Neither antibody, alone or in combination, was effective in downregulating the antigen-driven proliferative response (P < 0.2 for each culture condition in Th1 and Th2 clones). These results were obtained consistently from multiple T cell clones, each assayed on more than one occasion; similar results were obtained from a Th0 clone (data not shown).

View larger version (37K): [in this window] [in a new window]   Figure 4.   Lack of efficacy of anti-CD80, anti-CD86, or the combination in downregulating proliferative responses of Th1 and Th2 clones. Clones were stimulated with autologous, irradiated APCs and antigen (RW). Data are depicted as % of positive control (antibody-free culture) values, corrected in each case for background (antigen-free culture). All P values for antibody-treated cultures were 0.2 or greater, compared with the positive control. Positive control and background values (mean ± SEM), respectively: Th1 = 41,677 ± 16,091 and 699 ± 62 cpm; Th2 = 44,991 ± 7,922 and 623 ± 10 cpm. n = 4 individual experiments on two clones of each phenotype.

Cytokine Gene Expression in Th1 and Th2 Clones

Figure 5a depicts the  actin- and cytokine-specific RT-PCR amplification products from a representative study of ragweed-driven cytokine gene expression in the same Amb a 1-specific Th1 and Th2 clones in the absence and presence of blocking antibodies to CD80, CD86, or both. Data from the Th2 clone is shown in the top three rows, while data from the Th1 clone is shown in the fourth and fifth rows. Adequate normalization of RNA was confirmed by the equality of RT-PCR amplification products at sub-saturating cycle number for  actin gene expression (first and fourth rows, respectively, for Th2 and Th1). Sub-saturating conditions for all cytokine-specific RT-PCR amplifications were maintained throughout these experiments. Neither antibody, alone or in combination, was effective in downregulating the antigen-driven gene expression of proinflammatory cytokines (rows 2 and 3 for IL-4 and IL-5, respectively, in a Th2 clone; row 5 for IFN in a Th1 clone; P > 0.6). These results were obtained consistently from multiple T cell clones, each assayed on more than one occasion; similar results were obtained from a Th0 clone (data not shown). A quantitative representation of these data for the entire group of clones studied is depicted in Figure 5b.

View larger version (66K): [in this window] [in a new window]   View larger version (51K): [in this window] [in a new window]   Figure 5.   Lack of efficacy of anti-CD80, anti-CD86, or the combination in downregulating cytokine gene expression by Th1 and Th2 clones. (a) Representative data are depicted for IL-4 and IL-5 in Th2 clones (rows 2, 3) and IFN in Th1 clones (row 5). Clones were stimulated with autologous, irradiated APCs and antigen (RW). Adequate normalization of RNA was confirmed by  actin gene expression at subsaturating cycle number (rows 1 and 4, respectively, for Th2 and Th1). Culture conditions are noted at the top of each lane. The DNA size marker is shown in lane 6. (b) Densitometry data are depicted for all subjects for each RT-PCR target and culture condition. The numbers represent percentage optical density ± SEM of PCR amplification product bands normalized to stimulated, antibody-free controls. The variability of  actin gene expression was less than 3%. n = 4 individual experiments on two clones of each phenotype.

Cytokine Protein Secretion from Th1 and Th2 Clones

Figure 6 shows the levels of secreted IL-4 and IFN from Th2 and Th1 clones, respectively, in the absence and presence of blocking antibodies to CD80, CD86, or both. Variability of  3% for individual values from a clone was confirmed by both duplicate culture experiments as well as replicate ELISA assays at different dilutions of a culture supernatant (data not shown). Once again, neither antibody, alone or in combination, was effective in downregulating the antigen-driven secretion of proinflammatory cytokines (P > 0.44 and 0.16, respectively, for IL-4 and IFN). These results were obtained consistently from multiple T cell clones, each assayed on more than one occasion; similar results were obtained from a Th0 clone (data not shown). While results are depicted only at the 12-h time point, cytokine protein secretion is not detected prior to 6 h and has reached plateau by 24 h; thus, the 12-h time point provided maximal discrimination. Supernatants from PBMC cultures produced no detectable T cell cytokines, probably due to the low number of antigen-specific responder cells in the periphery (data not shown).

View larger version (39K): [in this window] [in a new window]   Figure 6.   Lack of efficacy of anti-CD80, anti-CD86, or the combination in downregulating cytokine protein production by Th1 and Th2 clones. Clones were stimulated with autologous, irradiated APCs and antigen (RW). Data are depicted as % of positive control (antibody-free culture) values (mean ± SEM). All P values for antibody-treated cultures were 0.16 or greater, compared with the positive control. Controls: IL-4 = 1,730 ± 411 pg/2 × 105 cells; IFN = 938 ± 50 pg/2 × 105 cells. n = 4 individual experiments on two clones of each phenotype.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

In the present study, we have addressed the hypothesis that selective activation of human T cell subsets may be achieved by selective engagement of CD28/CTLA4 by specific B7 homologues. We observed the relative contributions of CD80 and CD86 to T cell activation parameters, using multiple lots of functionally characterized blocking antibodies to the B7 homologues (23). While anti-CD86 significantly suppressed activation of either RW- or TT-driven PBMCs, the same anti-CD86 antibody became ineffective in modulating the antigen-driven responses of either Th1 or Th2 clones; anti-CD80 was consistently ineffective in both PBMCs and the T cell clones. These data represent one of the few studies of B7 homologue function in humans, and the only study, to our knowledge, to include data using nontransformed, antigen-specific human Th1 and Th2 clones and autologous antigen-presenting cells.

The prior investigations on which this study was predicated deserve further discussion in light of the present data; a number of methodological differences may account for the disparate results. First, the hallmark studies showing signaling specificity for T cell subsets associated with specific B7 isomers have all been performed in mice (12- 14); differences between mice and humans are not without precedent in comparative studies of immunobiology. Interestingly, if one accepts the hypothesis that diabetes in non-obese diabetic (NOD) mice and experimental allergic encephalomyelitis (EAE) in SJL/J mice are both Th1- mediated diseases, the data of Kuchroo and colleagues and Lenschow and associates appear to support opposite hypotheses (12, 13). Treating mice with the same anti-B7.1 monoclonal antibody, Kuchroo and colleagues showed a reduction in incidence and severity of murine EAE while Lenschow and colleagues showed an acceleration of disease progression and enhanced susceptibility of NOD mice to diabetes. Both groups found an opposite effect with anti-B7.2 treatment in their model systems, and both groups showed corraborative histologic data. Thus, the issue of signaling specificity through the B7 homologues remains controversial even in the murine system. Second, while the study of peripheral blood CD4+ cells by Freeman and colleagues yielded data similar to ours in PBMCs, their system used either antibody-coated plastic plates or one-way MLRs (mixed lymphocyte reactions) for cellular activation (22); we feel that this system, although more convenient, is less physiologic than antigen-driven stimulation with autologous APCs. Moreover, in a similar study of peripheral blood CD4+ cells, Levine and associates were unable to show differences in co-stimulatory efficacy between B7.1 and B7.2 (16). While both of these studies focused on the responses of peripheral lymphocytes, our data are the first to involve human, antigen-specific Th1 and Th2 clones. Finally, while other investigators have studied the ability of the B7 homologues to provide co-stimulatory signals in systems using B7-transfected NIH 3T3 or CHO cells standardized for the level of surface expression, our data look at the effects of removal of these signals in antigen-specific T cell/autologous APC interactions (16, 22). Thus, our results may more closely reflect the intact, native human immune response. While our data do not dispute the ability of the B7 homologues to provide co-stimulation in an antigen-driven system, they argue against an essential role for these molecules, a hypothesis that has previously been suggested (26).

A number of possible confounding issues with this study have been addressed without presenting the data formally in this manuscript. First, the possibility that irradiation of PBMCs might alter B7 expression was considered; however flow cytometry revealed no alterations in either CD80 or CD86 surface expression. The possibility remains that irradiation may cause an alteration in B7 function rather than expression, but currently available techniques do not allow this hypothesis to be tested. Second, the possibility that different sub-populations of APCs could differentially regulate B7 signaling was considered. However, we have performed identical experiments using purified (> 99%) autologous monocytes as APCs for the T cell clones; these experiments yielded identical results (data not shown). We have chosen to present the data using unfractionated PBMCs since this more closely approximates a physiologically relevant environment and allows for more direct comparisons between PBMC and clone experiments in this model. Third, although our data suggest that alterations in the activity of B7 homologues may be a function of differentiation or memory, we have been unable to pursue this hypothesis in detail. While the T cell clones used in these studies are CD3+CD4+CD45RO+ by flow cytometry, we have found that the majority of the CD3+CD4+ cells in PBMCs are CD45RA+. Unfortunately, available techniques preclude accurate assessment of the relative numbers of CD45RO+ and CD45RA+ antigen-specific responder cells in the PBMC cultures. However, since TT is a recall antigen for these subjects, the differences are probably not simply a function of naive versus memory responses. Of interest in this regard are the data of Van de Velde and colleagues showing more stringent requirements for B7 co-stimulation in anti-CD3-driven CD45RA+ T cells compared with CD45RO+ T cells, and the data of Freeman and associates showing that only B7-2 costimulates CD4+CD45RA+ T cells to produce IL-4 (22, 27). Finally, the possibility that plastic-adherent (activated) monocytes, and not activated T cells, might have produced some or all of the measured IFN was considered. However, blocking culture plates with 10% serum had no effect on IFN production in these experiments, while this blocking step does prevent activation and gene expression of IL-1, TNF, and IFN from elutriator- purified monocytes. Thus, the contribution of APCs to the IFN signal in these studies was likely negligible.

We have continued to take a number of precautions in our methodology to ensure the reliability of the results. First, we have optimized the time interval for incubation of the proliferation assays to maximize the signal/noise ratio. Second, we have repeated the analysis of phenotypic profiles of the T cell clones used in these experiments over a 9-mo period; the phenotypes have remained constant, precluding the effects of cellular differentiation events on these data. Third, replicate experiments using the same clones over a 3-mo period have yielded nearly identical results. Fourth, we have continued to use a complex, multi-step normalization process with the RT-PCR assay to ensure valid, quantitative comparisons between culture conditions; a 12-h culture interval for gene expression assays precludes the effects of cellular proliferation on these data. Fifth, we have studied in detail the kinetics of cytokine gene expression and protein production in this system, and use optimized incubation times for these assays. Sixth, the possibility of cellular senescence as an explanation of the findings is negated by a > 99% cell viability by trypan blue exclusion at the beginning and end of the culture period. Anergy is also an unlikely explanation, since cells cultured with the blocking antibodies are readily restimulated with antigen and APCs. Finally, the meticulous use of subsaturating cycle numbers in the RT-PCR assays assures the validity of the comparative data. The close correlation of gene expression with cytokine protein secretion in the clonal study supports this contention.

In conclusion, evidence is presented that, in PBMCs, anti-CD86 downregulates proliferation and cytokine gene expression induced by either Th1- or Th2-promoting antigens while anti-CD80 is ineffective. Furthermore, in antigen-specific Th1 and Th2 clones, neither anti-CD80 nor anti-CD86 is effective in downregulating proliferation, cytokine gene expression, or cytokine protein secretion. Thus, signaling through the B7 homologues shows no specificity for Th1 or Th2 cells in humans, and modulation of B7 signaling has little effect on differentiated, antigen-driven human T cells. CD86 is the predominant B7 signaling molecule in human PBMCs, and the function of CD86 may be restricted to more naive responder cells.