Costimulation was originally defined and characterized during primary T cell activation. The signaling events that regulate subsequent antigen encounters by T cells are less well defined. In this study we examined the role of CD30 in T cell activation and defined factors that regulate expression of CD30 on T cells. We demonstrate that CD30 expression is restricted to activated T cells and regulated by CD28 signal transduction. In contrast to CD28-expressing TCR Tg cells, CD28-deficient TCR Tg cells did not express CD30 when cultured with peptide and APCs. However, rIL-4 reconstituted CD30 expression on CD28-deficient TCR Tg cells. Blockade of CD28 interactions or depletion of IL-4 inhibited the induction of CD30, suggesting that both CD28 and IL-4 play important roles in the induction of CD30 expression on wild-type cells. However, CD28 signaling did not up-regulate CD30 expression solely through its ability to augment IL-4 production because IL-4-deficient T cells stimulated with anti-CD3 and anti-CD28 expressed CD30. Induction of CD30 in the absence of IL-4 was not due to the IL-4-related cytokine IL-13. CD30, when expressed on an activated T cell, can act as a signal transducing receptor that enhances the proliferation of T cells responding to CD3 crosslinking. Collectively, the data suggest that T cell expression of CD30 is dependent on the presence of CD28 costimulatory signals or exogenous IL-4 during primary T cell activation. Once expressed on the cell surface, CD30 can serve as a positive regulator of mature T cell function.
The activation requirements of naive T cells have been well studied. In most cases signaling through the TCR is insufficient to effectively sustain a primary immune response. In addition to TCR signaling, a second costimulatory signal is also required. A large body of work has indicated that CD28 is the predominant costimulatory molecule that functions at early stages of naive T cell activation (1, 2, 3, 4). However, the potential role of additional molecules serving as costimulatory receptors in later stages of T cell immune responses is not as well defined.
The TNF receptor (TNFR)3 family is a rapidly expanding group of related proteins which includes TNFR I and II, Fas, 4-1BB, OX40, CD27, CD40, and CD30 (5, 6). The members of the TNFR family are likely candidates as potential regulators of late stage T cell activation. For example, CD40 is a potent costimulatory receptor for B cell proliferation, differentiation, and survival (7, 8). Mechanistically, CD40 has been shown to prevent apoptosis through the induction of Bcl-xL (9, 10) and to promote cell-cycle progression by regulating cdk4 and cdk6 expression in a B cell line (9). The importance of CD40 in B cell physiology suggests that there may be a related family member(s) that could regulate T cell costimulation.
TNFR family members share significant homology in the extracellular domains but limited homology in their intracellular domains (6). However, investigation of the binding properties of the cytoplasmic domains to intracellular proteins has demonstrated that there are also important similarities in their intracellular domains. For example, the cytoplasmic domain of CD30 has recently been shown to have TNFR-associated factor (TRAF) binding domains with significant similarity to the TRAF binding domains of CD40 (11). In both receptors these domains have been shown to be necessary and sufficient for NF-κB induction (12, 13, 14). In addition, CD30 and CD40 can augment cytokine production by cell lines derived from Reed-Sternberg cells of Hodgkin’s disease (15). These observations suggested that CD30 may serve as a positive regulator of T cell function.
CD30 has been reported to augment proliferation in human T cells during a primary stimulation and in some cell lines derived from Hodgkin’s lymphoma, whereas in other cell lines of the same origin no effect has been observed (16, 17). CD30 signaling can also enhance cytokine production by a murine CTL line (18). In contrast, CD30-deficient mice have impaired negative selection of thymocytes (19). In addition, CD30 signaling can augment TCR-dependent apoptosis of a T cell hybridoma (20), and down-regulate CD28 expression and inhibit cytotoxicity of a large granular lymphoma cell line (21). Together these reports indicate that CD30 in some instances can promote cell survival and in others induce apoptosis. A similar dichotomy of function has been reported for other members of the TNFR family, such as TNFR II, and can perhaps be explained by the cellular context in which a signal is delivered (22, 23, 24).
It has been reported that CD30 expression on human and murine T cells is activation dependent (18, 25, 26). However, little is known about the exact stimuli that regulate CD30 expression. Here we report that CD30 expression on activated T cells requires costimulation via CD28 or the addition of IL-4 as an exogenous cytokine. CD28 induces CD30 expression through both IL-4-independent and IL-4-dependent mechanisms. In support of the hypothesis that CD30 may function as a positive regulator of T cell function, we also report that CD30 provides a costimulatory signal for T cell proliferation during a secondary stimulation.
Materials and Methods
BALB/cJ mice and BALB/cJ IL-4-deficient mice were purchased from The Jackson Laboratories (Bar Harbor, ME). The CD28-deficient mice were generated as previously described (27). The CD28 null allele has subsequently been bred from the C57BL/6 background onto the BALB/cJ background. TCR transgenic mice specific for OVA peptide fragment 323–339 (DO11.10) and I-Ad restricted were initially provided as a generous gift from D. Loh (Washington Uuniversty, St. Louis, MO). Wild-type and CD28-deficient BALB/cJ mice carrying the DO11.10 transgene were generated by appropriate breeding. All mice were bred and maintained in a specific pathogen-free facility at the University of Chicago (Chicago, IL).
For all CD30 expression experiments, bulk lymph node cells were isolated and single cell suspensions were cultured in complete medium as previously described (28). The anti-CD3 stimulations were conducted in 48-well plates coated with a 1 μg/ml solution of anti-CD3 in 50 mM Tris (pH 9.4) with 2.5 × 105 cells/well. For the OVA TCR Tg experiments, 4 × 105 cells/well were cultured in 48-well plates with 100 μM OVA peptide 323–339. As indicated in the text and figure legends, additional reagents were added to the above conditions as follows: rIL-2 (100 U/ml, Boehringer Mannheim, Indianapolis, IN), rIL-4 (10,000 U/ml, R&D Systems, Minneapolis, MN), rIL-13 (50 ng/ml, R&D Systems), anti-CD28 (PV1, 10 μg/ml, kindly provided by C. June, Naval Medical Research Institute, Bethesda, MD), anti-IL-4 (11B11, 10 μg/ml, a gift from R. Coffman, DNAX Corp., South San Francisco, CA), anti-IL-13 (50 μg/ml, R&D Systems), and CTLA4-Ig (100 μg/ml, G. Gray, Genetics Institute, Cambridge, MA). This dose of CTLA4-Ig was chosen because it gave maximal inhibition of a one way mixed lymphocyte reaction. Anti-IL-13 was used at a dose that was two times greater than the dose necessary to achieve one-half maximal inhibition of 10 ng/ml of recombinant murine IL-13 as assayed using the factor-dependent cell line, TF-1.
Bulk lymph node cells from the indicated animals were harvested on day 5 of primary culture and stained with either a phycoerythrin (PE)-conjugated-hamster (anti-TNP; 2,4,6-trinitrophenol) Ig control, anti-CD30-PE, or anti-CD28-PE (37.51). In some instances two color staining was performed, using the described PE-conjugated Abs in combination with either FITC-coupled rat IgG2a isotype control or anti-Thy-1.2-FITC. All of the Abs used for the described FACS experiments were purchased from PharMingen (San Diego, CA). In all experiments 10,000 live events were acquired and positive staining with specific Ab was analyzed relative to the isotype control for the same culture condition using a FACScalibur flow cytometer and Cellquest software (Becton Dickinson, Mountain View, CA).
Bulk lymph node cells from BALB/c mice were cultured in 24-well plates coated with a 1 μg/ml anti-CD3 50 mM Tris (pH 9.4) solution at 1 × 106 cells/well. Where indicated, 10 μg/ml anti-CD28 was also added to the culture supernatant. After 5 days, cells were harvested from the plate, washed two times, and placed in fresh medium for 24 h in a 37°C incubator. For restimulation, cells were washed two more times, and 5 × 104 cells/well were stimulated in 96-well flat bottom microtiter plates with the indicated concentration immobilized anti-CD3, in wells coated with a 20 μg/ml anti-CD30 in a 50 mM Tris (pH 9.4) solution, as described by Bowen et al. (18), or 10 μg/ml soluble anti-CD28, or no additional stimulus. CD30 expression was determined at both the time of harvest and restimulation. Each well was pulsed with 1 μCi tritiated thymidine ([3H]TdR, ICN Biochemicals, Costa Mesa, CA) for the final 8 h of a 48-h culture. All plates were harvested using a Tomtec Mach II cell harvester and counted on a Betaplate liquid scintillation counter (Wallac Inc., Gaithersburg, MD). All results are expressed as the mean ± SD of triplicate cultures.
CD28 signaling is required for maximal induction of CD30 expression on T cells
Previously, Bowen et al. (18) reported that splenic T cells stimulated with anti-CD3 express CD30 at peak levels 4 to 6 days following activation. To obtain a population of CD30-expressing T cells for our studies, bulk lymph node T cells from wild-type mice were stimulated in wells coated with a 1 μg/ml anti-CD3 solution, and CD30 expression was analyzed. Although in some instances a small population of T cells expressing CD30 was observed on day 5, CD30 was not detectable on the majority of T cells using anti-CD3 stimulus alone (Fig. 1,B, upper panel). In fact, when T cells were cultured in wells coated with as much as a 10 μg/ml anti-CD3 solution there was no observable induction of CD30 expression (data not shown). This suggested that CD30 was either expressed at levels below the detection limit of flow cytometry or that an additional signal was required for induction of CD30 expression on T cells. Therefore we tested whether costimulation through CD28 could induce CD30 expression. Stimulation of bulk lymph node cells with anti-CD3 in the presence of anti-CD28 resulted in a uniform population of CD30 expressing Thy-1.2+ cells after 5 days in culture (Fig. 1,C, upper panel). Anti-CD30 stained T cells were first detected at 3 days following stimulation, and maximal CD30 expression was reproducibly observed on day 5 (data not shown and Fig. 1 C, upper panel).
Activated cells were also analyzed for CD28 expression. CD28 surface expression was up-regulated after stimulation with anti-CD3 alone (Fig. 1,B, lower panel). However after treatment with anti-CD3 and anti-CD28, expression of CD28 was dramatically reduced (Fig. 1 C, lower panel). This observation was not due to competition between the anti-CD28 used for stimulation and the CD28 Ab used for flow cytometry, because incubation of either a CD28-positive T cell hybridoma or lymph node cells at 4°C with the CD28 Ab used for stimulation did not block binding of the Ab used for FACS analysis (data not shown). Thus, in vitro activation of T cells with anti-CD3 and anti-CD28 results in expression of high levels of CD30 and low levels of CD28. This result suggests that under the culture conditions described here, chronic ligation of CD28 leads to receptor down modulation from the T cell surface. It is possible that CD30 serves to regulate the late function and/or differentiation of this population of T cells and may potentially substitute for CD28 function in this context.
CD30 can function as a costimulatory receptor during secondary T cell stimulation
To address the effects of CD30 signaling in T cells stimulated with anti-CD3 and anti-CD28, we examined the effects of CD30 crosslinking during a secondary stimulation with anti-CD3. Bulk lymph node cells were stimulated with anti-CD3 and anti-CD28 for 5 days. Cells were next washed to remove the residual anti-CD3 and anti-CD28 and cultured in medium alone for 24 h. The cells were then restimulated in the presence of anti-CD3 alone or anti-CD3 and anti-CD30 or anti-CD3 and anti-CD28. In these experiments, the restimulated Thy-1.2+ population of cells was comprised of ∼90% CD4+ of which 75% displayed a memory phenotype as defined by low Mel-14 mean fluorescence (data not shown). Figure 2, A and B, show that ligation of CD30 augments proliferation of T cells in the presence of a suboptimal dose of anti-CD3. Anti-CD30 also had a small effect in the absence of additional anti-CD3; this effect may be due to synergy with residual anti-CD3 from the primary culture. The costimulatory effect of anti-CD30 is comparable to that of anti-CD28 (Fig. 2 B). These data indicate that CD30 can deliver a costimulatory signal to previously activated T cells.
Treatment with rIL-4 can substitute for CD28 in the up-regulation of CD30 expression on CD28-deficient TCR Tg cells
To confirm the importance of CD28 in the induction of CD30 expression, wild-type and CD28-deficient OVA TCR Tg lymph node cells were treated with 100 μM OVA peptide 323–339 for 5 days and analyzed for CD30 expression. Whereas bulk lymph node cells from wild-type TCR Tg mice expressed high levels of CD30 on T cells when stimulated with peptide alone, CD28-deficient T cells bearing the same TCR transgene failed to express CD30 (Fig. 3,A, upper panel and lower panel, respectively). Comparable levels of CD30 were induced on wild-type T cells in cultures containing 10-fold less peptide, suggesting that the failure to induce CD30 expression on CD28-deficient T cells was not due to a suboptimal Ag dose. These data indicated that under this set of culture conditions no other surface receptor was able to substitute for CD28. CD28 has been shown to mediate costimulatory effects both directly through induction of intracellular signaling and indirectly through the augmentation of cytokine production (29, 30, 31, 32). Therefore, we tested whether cytokines could induce CD30 expression on CD28-deficient T cells. Because CD28 is known to increase the expression of IL-2 (32, 33), the ability of rIL-2 to induce CD30 expression was tested in the absence of a CD28 signal. Addition of rIL-2 to peptide stimulated bulk lymph node cells from CD28-deficient TCR Tg mice did not augment CD30 expression (Fig. 3,B, lower panel). In vitro IL-4 production is also impaired in CD28-deficient mice (31). Therefore, we also tested whether rIL-4 could function to induce CD30 expression in the absence of CD28. Addition of rIL-4 was able to induce CD30 on CD28-deficient TCR Tg lymph node cells stimulated with peptide (Fig. 3,C, lower panel). It also appeared that rIL-4 could further enhance CD30 levels on wild-type TCR Tg cells over those levels observed with peptide alone or peptide and rIL-2 (Fig. 3 C, upper panel).
Blockade of CD28 signaling or of IL-4 prevents CD30 expression in wild-type TCR Tg+ cells
To further test the roles of CD28 ligation and IL-4 in the induction of CD30, wild-type TCR Tg lymph node cells were incubated with peptide alone, peptide and CTLA4-Ig, or peptide and anti-IL-4. As observed earlier, CD30 expression on wild-type T cells was up-regulated when incubated with peptide alone (Fig. 4,A). However when bulk lymph node cells were incubated with peptide in the presence of either CTLA4-Ig or anti-IL-4, induction of CD30 expression was significantly suppressed (Fig. 4, B and C, respectively). These data are indicative of a requirement for both CD28 and IL-4 signal transduction to up-regulate CD30 expression on T cells activated by peptide in the presence of APCs. These data also suggest that the ability of CD28 to induce CD30 expression is mediated, at least in part, through augmentation of IL-4 expression.
CD28 stimulation can function to up-regulate CD30 in IL-4-deficient mice
If CD28 functions to up-regulate CD30 solely via indirect effects on the IL-4 pathway, one prediction is that IL-4-deficient mice would fail to express CD30 when stimulated with anti-CD3 and anti-CD28. However, when bulk lymph node cells from IL-4-deficient mice were stimulated with anti-CD3 and anti-CD28 for 5 days, CD30 expression was induced in the majority of Thy-1.2+ cells (Fig. 5,B, upper panel). This finding suggested that CD28 could exert effects on CD30 expression independent of IL-4. The results in Figure 5,B were mirrored when IL-4 was depleted from the anti-CD3 and anti-CD28 treated wild-type bulk lymph node cultures (Fig. 5,C, lower panel). That is, the addition of anti-IL-4 decreased but did not eliminate the expression of CD30 on activated T cells provided with maximal anti-CD28 costimulation. Although the expression of CD30 in the presence of anti-IL-4 appeared bimodal, the populations that displayed differential CD30 expression did not partition by cell surface memory markers nor T cell subclass as defined by CD4, CD8, Mel-14, and CD44 (data not shown). The lack of an effect of anti-IL-4 in the IL-4-deficient mice assures the specificity of this reagent (Fig. 5 C, upper panel). This dose of anti-IL-4 was sufficient to neutralize the effects of 10,000 U/ml rIL-4; therefore the result obtained with the wild-type mice is most likely not due to a partial block of IL-4 produced in the culture (data not shown). Therefore, it appears that the addition of anti-IL-4 to wild-type cultures reduces CD30 expression to levels comparable to that observed on IL-4-deficient T cells cultured with anti-CD3 and anti-CD28 alone. Taken together, these data suggest that CD28 induces CD30 expression through both IL-4-dependent and IL-4-independent mechanisms.
IL-13 does not mimic the effect of IL-4 in the induction of CD30
One possible mechanism of IL-4-independent induction of CD30 could be through an IL-13-mediated signal(s). IL-4 and IL-13 are cytokines that display significant sequence similarity (34). In addition, they have been reported to share common receptor subunits (35, 36) and can induce many of the same biologic responses in B cells including the up-regulation of both CD23 and MHC class II expression (37). Therefore, we sought to determine whether IL-13 shared the ability of IL-4 to activate CD30 expression in T cells. Figure 6 demonstrates that T cells treated with anti-CD3 and rIL-4 expressed CD30. However, T cells stimulated with anti-CD3 alone or with anti-CD3 and rIL-13 failed to express CD30 (Fig. 6). IL-13 was originally defined as a cytokine that is induced by CD28 (34). Therefore we also tested whether the depletion of IL-13 in the context of anti-CD3 and anti-CD28 stimulation would attenuate CD30 expression. Addition of anti-IL-13 had no effect on the induction of CD30 expression by CD28. However, expression of CD30 by a population of T cells was blocked when IL-4 was depleted from anti-CD3 and anti-CD28 stimulated cultures (Figs. 5 and 6). These data suggest that although IL-4 and IL-13 functionally overlap in many biologic systems, in the case of CD30 induction, the two cytokines are distinct.
CD30 expression on T cells requires prior activation, suggesting that CD30 can function only following primary T cell activation. The present study demonstrates that TCR engagement alone is not sufficient to induce CD30 expression on activated T cells. Activated T cells require either CD28 costimulation or exogenous IL-4 for the induction of CD30 expression. The ability of CD28 costimulation to induce CD30 expression is due, at least in part, to augmentation of IL-4 production, because treatment of OVA TCR Tg T cells with peptide and anti-IL-4 prevented the induction of CD30 comparable to that of blockade of CD28/B7 interactions. However, we also found that a high percentage of T cells in IL-4-deficient mice express CD30 after treatment with anti-CD3 and anti-CD28. This finding suggests that CD28 can regulate CD30 expression by IL-4-dependent and -independent mechanisms. IL-2, IL-7, and IL-12 were unable to substitute for a CD28 signal in wild-type or CD28-deficient T cells (Figs. 3 and 6, and data not shown). In addition, the CD28 effects that are independent of IL-4 are not due to the IL-4-related cytokine, IL-13 (Fig. 6). Although IL-13 can mimic many of the effects of IL-4, IL-13 failed to mimic the ability of IL-4 to promote CD30 expression on activated T cells.
While our manuscript was in preparation, Nakamura et al. (38) reported a study of cytokine regulation of CD30 expression. They concluded using CD4+ TCR Tg mice that the presence of IL-4 was essential for the induction of surface CD30 expression; however, the significance of costimulation was not addressed in their study. In addition, in contrast to our report, they found that in cultures stimulated by the addition of peptide alone T cells failed to express CD30. This could be attributed to at least three apparent differences: genetic background of mice, use of purified CD4+ T cells with irradiated APCs, and/or use of different TCR transgenes. In our experiments utilizing TCR Tg mice we determined that blockade of the CD28/B7 interaction(s) or depletion of IL-4 during peptide stimulation of wild-type mice prevented CD30 expression. This finding may be due to the reported requirement for CD28 costimulation for the acquisition of IL-4 responsiveness by both Th1 and Th2 clones (39, 40). CD30 expression has been shown to be associated with Th2 clones (41, 42); this observation is most likely explained by the use of IL-4 in the induction and maintenance of Th2 clones, and the fact that Th2 clones produce IL-4. Since we demonstrated that IL-4 in trans augments CD30 expression on CD28-deficient T cells, T cells from CD28-deficient mice may be able to induce CD30 expression and receive subsequent costimulatory signals by CD30 as a result of IL-4 produced by accessory cells in vivo. Nakamura et al. (38) also observed that Th0 cells which produce both IL-4 and IFN-γ expressed CD30 and concluded through additional studies that CD30 expression correlated with the ability of T cells to respond to IL-4. We have observed that wild-type lymph node cells stimulated with anti-CD3 and anti-CD28 for 5 days retain a Th0 phenotype and produce IFN-γ and IL-4 upon restimulation (data not shown).
The fact that IL-4-deficient cells express CD30 in response to anti-CD3 and anti-CD28 stimulation precludes IL-4 as the sole regulator of CD30 expression in all instances of T cell activation. It remains possible that responsiveness to IL-4 is indicative of a late stage differentiation event that is independent of Th cell phenotype. This event could potentially be marked by the expression of CD30. Our findings are consistent with the previously reported role of CD28 in general T cell differentiation (29, 31) and also suggest that CD30 expression may correlate with the transition from a naive to a committed or differentiated cell.
We have also shown that CD30 can function as a costimulatory receptor in a T cell secondary proliferative response. Recently another member of the TNFR family, 4-1BB, was reported to be expressed at peak levels 2 to 3 days following T cell activation (43), and was determined to augment proliferation of T cells (43, 44). CD30 expression peaks later than 4-1BB, therefore, it will be of interest to analyze whether cross-talk between the different TNFR family members occurs or whether their functions are temporally distinct. Another possible functional distinction could be differential effects of each receptor on CD4+ vs CD8+ T cells. In our culture system, 95% of the cells proliferating in response to anti-CD3 and anti-CD30 crosslinking are CD4+ cells. In contrast, Shuford et al. (45) recently reported that 4-1BB predominantly costimulates CD8+ T cells.
Our lab and others have previously reported that CD30 signal transduction induces NF-κB activity (12, 13, 46, 47), which is consistent with the positive effect of CD30 on proliferation. This finding is also in accord with the published report concerning CD30-deficient mice wherein negative selection of thymocytes is impaired, as marked by hypercellularity of the thymus and impaired deletion using the H-Y Tg model system (19). As signals which increase the avidity of thymocyte interaction with the thymic epithelium can augment negative selection (48, 49), it is likely that a signal which is perceived as costimulatory during peripheral activation such as CD30 could lead to deletion in the thymus. This exemplifies a persistent question in the biology of immune cell function. How do the same external stimuli result in divergent effects? For example, anti-CD3 which provides a mitogenic signal in primary T cell activation (4) can act as a death stimulus in immature T cells (50). Part of the answer may result from synergistic interaction between signaling pathways. This is seen in B cells that undergo apoptosis when signaled via either surface Ig or CD40 alone (7, 8, 9, 10, 51, 52). But simultaneous stimulation through both receptors potently triggers survival and proliferation (7, 8, 9, 10). How these distinct outcomes are regulated will presumably be answered when precise molecular signaling pathways are elucidated and quantitated for each condition. Nevertheless, the data presented in this report demonstrate the potential of CD30, expression of which is dependent on early costimulatory signals, to act as a positive regulator of activated T cell function.
We thank R. Gedrich, M. Alegre, and R. Arch for thoughtful comments on this work. We also acknowledge C. Rudin, E. Masteller, and A. Minn for critical review of the manuscript. In addition we are grateful to G. Sprull and C. Sampson for expert technical assistance, and M. Davis for outstanding animal husbandry.
M.C.G. is funded by the National Institutes of Health/MCB training grant. This work was also supported by a grant from the National Institutes of Health (AI35294)
Abbreviations used in this paper: TNFR, TNF receptor; TRAF, TNFR-associated factor; PE, phycoerythrin.