CD137 stimulation and p38 MAPK inhibition improve reactivity in an in vitro model of glioblastoma immunotherapy

Abstract

Dendritic cell vaccination has become an interesting option for cancer immunotherapy. Tumor-lysate-pulsed dendritic cells (DC) can prime naïve T cells and induce the regression of established tumors including gliomas as shown in various animal models. Despite hopeful results even in clinical studies, the outcome for many patients is still unsatisfying. In the present study, we tested the combination of tumor-lysate-pulsed dendritic cells (TPDC) with a monoclonal antibody against CD137, a monoclonal antibody against CD25 (daclizumab) and a specific p38 mitogen-activated protein kinase (p38 MAPK) inhibitor (SB203580) for improving immunostimulation in an in vitro model of immunotherapy for human gliomas. We observed a higher secretion of interferon gamma by TPDC-primed peripheral blood mononuclear cells (PBMC) that were incubated with an antibody against CD137 or the p38 MAPK inhibitor. In addition, we observed higher specific lysis of tumor cells after incubation of PBMC with the p38 MAPK inhibitor or the anti-CD137 antibody. In contrast, incubation of TPDC-primed PBMC with the anti-CD25 antibody did enhance neither interferon gamma secretion nor cellular cytotoxicity. Cell depletion experiments demonstrated that the immune reaction induced by TPDC is strongly dependent on CD4-positive and CD8-positive cells. Incubation of DC during maturation and antigen loading with the anti-CD137 antibody did not enhance cytotoxicity and interferon gamma secretion in comparison with application of the anti-CD137 antibody during priming. In conclusion, our data suggest that p38 MAPK inhibition and anti-CD137 antibodies can enhance the immune response against glioblastoma cells.

Electronic supplementary material

The online version of this article (doi:10.1007/s00262-013-1484-9) contains supplementary material, which is available to authorized users.

Introduction

Dendritic cell vaccination (DCV) represents an interesting option for treatment for high-grade gliomas and has already been used in various clinical trials [1]. Although promising results with prolonged event-free and overall survival after DCV have been reported for certain patient subgroups with glioblastoma multiforme (GBM), the general prognosis for patients with GBM still remains very poor [2]. Further improvements in DCV strategies are obviously needed. Thus, we investigated in the present study whether the combination of DCV with immunomodulatory signals can improve the overall immune response against glioma cells. To this end, we combined DCV with an anti-CD137 antibody, a p38 mitogen-activated protein kinase (MAPK) inhibitor and an anti-CD25 antibody (Fig. 1).

Fig. 1.

Overview of the used immunomodulatory elements. Tumor-lysate-pulsed dendritic cells were used for priming of PBMC with the aim of stimulating cytotoxic T cells (CTL) that can kill tumor cells. The simultaneous activation of helper T cells can increase the activity of these cells, whereas regulatory T cells might inhibit the activity of effector cells. Antibodies directed at CD137 can increase survival of effector T cells. Antibodies against CD25 can inhibit CD25-positive regulatory cells. The p38 MAPK inhibitor can inhibit regulatory T cells. In addition, this agent might positively affect the TH1 polarization activity of dendritic cells and inhibit apoptosis of CTL

The co-stimulatory receptor CD137 (a type I transmembrane protein) is a member of the tumor necrosis factor receptor superfamily (TNFRSF), a group of molecules that play key roles in the control of survival of immune cells during immune reactions [3]. CD137, also known as ILA (induced by lymphocyte activation), was first detected on activated human and murine T cells [4]. In addition to activated T cells, also dendritic cells, NK cells and mast cells express CD137 [5–8]. The natural ligand of CD137 is CD137L (4-1BBL), a type II transmembrane protein [9, 10]. CD137L was detected on activated antigen-presenting cells (APC), e.g., on dendritic cells [10–12]. Stimulation of CD137 with its natural ligand or a monoclonal antibody (mAB) leads to an activation of nuclear factor kappa B (NFκB) and stress-activated protein kinase/c-jun N-terminal kinase (SAPK/JNK) pathways and finally to the regulation of gene transcription [13–15]. Stimulation of CD137 on T cells leads to cell proliferation and cytokine production [16–18]. Another effect of stimulation of CD137 on activated T cells is the protection from activation-induced cell death (AICD) [19].

The p38 MAPK was first discovered in mice [20], followed by the identification of the human p38 MAPK [21]. Several p38 MAPK isoforms were described (reviewed in [22]); the one that is characterized best is isoform p38α, encoded by the MAPK14 gene. For CD4-positive cells, it is known that full activation of p38 MAPK needs ligation of the T-cell receptor (TCR) and co-stimulatory molecules such as CD137, CD28, or inducible T-cell co-stimulator (ICOS) (reviewed in [23]). p38 MAPK is necessary for production of interferon gamma and differentiation of CD4-positive cells into effector TH1 cells [24]. In CD8-positive cells, p38 MAPK activation can induce apoptosis [25]. It was shown that p38 MAPK is necessary for maturation of dendritic cells [26–29]. Furthermore, p38 MAPK has a negative effect on expression of MHC class II proteins [30]. Production of interleukin (IL) 10 after incubation of dendritic cells with Toll-like receptor (TLR) agonists is also dependent on p38 MAPK. Inhibition of p38 MAPK increased IL12 production by dendritic cells and amplified the therapeutic effects of a TLR-ligand-activated DC immunotherapy against tumors by suppression of regulatory T cells (Treg) [31]. It was shown that p38 MAPK inhibition can also increase the ratio of IL12–IL10 in human DC. Inhibition of p38 MAPK reduces expression of IL10 by human DC. Increased as well as decreased expression of IL12 has been described and might depend on the activation status of the human DC [32, 33].

CD25 (alpha chain) belongs together with CD122 (beta chain) and CD132 (gamma chain) to the trimeric high-affinity IL2 receptor [34]. It is expressed on activated T cells, on forkhead box P3 (FOXP3)-positive CD4+CD25+ Treg, and at lower levels on pre-B cells, thymocytes and NK cells [35]. CD25 is up-regulated after ligation of the TCR on CD8-positive T cells and on activated CD4-positive T cells. Binding of IL2–CD25 leads to proliferation of CD4-positive and CD8-positive cells (reviewed in [36]). Neutralization of IL2 leads to autoimmune diseases with evidences of a dysfunction or deficiency of Treg [35, 37]. Other experiments showed a long-term immunity against glioma in mice after blocking CD25 by daclizumab, a mAB against human CD25 [38].

Materials and methods

Cell culture

The HLA-A0201-positive human glioma cell lines U87MG [39] and T98G [40] and the human fibrosarcoma cell line HT1080 [41] were grown at 37 °C in a humidified atmosphere with 5 % CO₂ in Dulbecco’s modified Eagle medium (DMEM, PAA, Coelbe, Germany) with 10 % fetal calf serum (FCS, Biochrom, Berlin, Germany), 100 U/mL penicillin, and 100 μg/mL streptomycin (PAA) in cell culture flasks. Twice a week, cells were washed with phosphate-buffered saline (PBS) and detached with trypsin/EDTA solution (0.05 % trypsin, 0.02 % EDTA in PBS, PAA). After centrifugation at 350×g for 7 min, cells were resuspended in DMEM at a ratio of 1:10 in new cell culture flasks.

Generation and maturation of dendritic cells

Human dendritic cells were generated from monocytes from peripheral blood mononuclear cells (PBMC) of voluntary healthy HLA-A2-positive donors. PBMC and CD14-positive cells were isolated as previously described [42]. Isolated monocytes were then cultured with IL4, granulocyte–macrophage colony-stimulating factor (GM-CSF), IL1β, IL6, tumor necrosis factor alpha (TNFα), and prostaglandin E2 (PGE2) for 7 days in a humidified atmosphere with 5 % CO₂ at 37 °C. On day 6 of culture, tumor lysate from human glioma cell line U87MG (50–100 μg/mL) was added for 2 days. Tumor-lysate-loaded dendritic cells (TPDC) were harvested, washed with PBS, and suspended in DMEM with 1 % human AB serum (Lonza, Basel, Switzerland), 100 U/mL penicillin, and 100 μg/mL streptomycin. Purity of DC was assessed by flow cytometry. A representative result is shown in Supplementary Figure 1.

Stimulation of PBMC

HLA-A2-positive PBMC (1 × 10⁶/mL) from voluntary healthy donors were stimulated for 5 or 7 days in 6-well plates with TPDC (2.5 × 10⁵/mL) of the same donors in 5 mL DMEM with 1 % human AB serum, 100 U/mL penicillin, and 100 μg/mL streptomycin). Depletion of single-cell populations was performed as described before [42]. When monoclonal antibodies (mAB) against CD137 (clone 26G6, a kind gift from R. Mittler, Atlanta, USA) were used, 6-well plates were coated with mAB solution (10 μg/mL in PBS) overnight at 4 °C before starting the co-incubation of PBMC and TPDC. This antibody increases the proliferation of T cells that have been stimulated with suboptimal concentrations of αCD3 (data not shown), suggesting that it has an intrinsic stimulating activity. The p38 MAPK inhibitor SB203580 (Biaffin, Kassel, Germany) was dissolved in dimethyl sulfoxide (DMSO) and was used at a concentration of 1 μM. SB203580 was added either at the beginning of the co-incubation of TPDC with PBMC or together with the tumor lysate during pulsing of dendritic cells. As control we added dimethyl sulfoxide (DMSO, Sigma, Taufkirchen, Germany). Antibodies against CD25 (Daclizumab, Roche, Mannheim, Germany) were used at 5 μg/mL and added at the beginning of the co-incubation of PBMC and TPDC. For analyses cells were harvested and centrifuged at 300×g. Supernatants were stored at −80 °C for further usage; cells were washed with PBS and analyzed.

Fluorescence-activated cell scanning (FACS) analysis

FACS analysis was performed essentially as described before [43]. About 0.5–1 × 10⁶ cells were stained with fluorochrome-labeled antibodies (αCD3, αCD4, αCD8, αCD25, αCD56, αCD137L, αFOXP3, and appropriate isotype controls, Becton–Dickinson, Heidelberg, Germany). After incubation for 30 min in the dark at 4 °C, cells were washed twice with PBS and resuspended in 0.5 mL PBS. For detection of FOXP3, cells were permeabilized and stained for 30 min at 4 °C. FACS analysis was performed using a FACSscan or FACSCalibur (Becton–Dickinson). Representative plots are presented in Supplementary Figures 1–3.

ELISPOT analysis

Interferon gamma ELISPOT was used to quantify tumor-cell-specific interferon-gamma-releasing cells. On days 5 and 7 of the PBMC/TPDC co-culture experiments, cells were harvested, washed, and resuspended in cell culture medium (DMEM with 1 % human AB serum, 100 U/mL penicillin, and 100 μg/mL streptomycin) at a concentration of 4 × 10⁵ cells/mL. Tumor cells from cell lines U87MG, T98G, and HT1080 were harvested, washed with PBS, and resuspended in cell culture medium (1 × 10⁵ cells/mL). ELISPOT analyses were performed with an interferon gamma ELISPOT Set (Becton–Dickinson) according to the manufacturer’s instructions. Spots were counted with software KS ELISPOT (Zeiss, Jena, Germany).

Cytotoxicity assay

Cytotoxicity exerted by PBMC after priming with TPDC was assessed by lactate dehydrogenase (LDH) release essentially as described [44]. On days 5 and 7 of PBMC/TPDC co-culture, cells were harvested, washed, and resuspended in cell culture medium (DMEM with 1 % human AB serum, 100 U/mL penicillin, and 100 μg/mL streptomycin) at a concentration of 5 × 10⁶ cells/mL. Tumor cells from cell lines U87MG, T98G, and HT1080 were harvested, washed with PBS, and resuspended in cell culture medium (2 × 10⁵ cells/mL). Tumor cells and primed PBMC were co-incubated in triplicates in 96-well plates for 16 h. Thereafter, LDH release from the cells was analyzed by using a LDH assay kit (Roche) according to the manufacturer’s instructions.

Cytometric bead array

Cell culture supernatants were harvested on days 5 and 7, centrifuged at 300×g, and stored for further usage at −80 °C. Cytokines were measured with a cytometric bead array (CBA) using the Human TH1/TH2 Cytokine Kit (Becton–Dickinson) according to the manufacturer’s instructions.

Results

Effects of anti-CD137 stimulation on PBMC priming with tumor-lysate-pulsed DC

Priming of PBMC with DC that had been loaded with tumor lysate from human glioma cell line U87MG led to a significant increase in U87MG-reactive cells as detected by IFNγ ELISPOT analysis (Fig. 2a). The highest number of IFNγ spots was detected after re-stimulation with U87MG tumor cells, which were used for priming of the PBMC.

Fig. 2.

Characterization of the response elicited by tumor-lysate-pulsed DC. a HLA-A2-positive PBMC were cultured with U87MG tumor-lysate-pulsed DC (DC/TL), immature DC (imDC), or without DC (w/o DC). After 5 days of incubation in a humidified atmosphere, cells were re-stimulated with tumor cells from lines U87MG, T98G, and HT1080 and interferon gamma ELISPOT analysis was performed. b Before culture of PBMC (HLA-A2-positive) with TPDC, depletion of CD4-positive, CD8-positive, or CD56-positive cells was performed. Depleted and non-depleted PBMC were cultured for 5 and 7 days in a humidified atmosphere. During interferon gamma ELISPOT analyses, PBMC were re-stimulated with U87MG tumor cells. The differences between non-depleted and CD4-depleted cells as well as between non-depleted and CD8-depleted cells are significant (p < 0.05). c Interferon gamma ELISPOT analyses of PBMC (HLA-A2-positive) at three different time points. DC/TL: PBMC were cultured with TPDC; DC/TL (αCD137): PBMC were cultured with TPDC that matured on coated plates with anti-CD137 antibodies; DC/TL + αCD137: PBMC were cultured with TPDC on plates coated with anti-CD137 antibodies; DC/TL + αCD137 (d4): PBMC were cultured with TPDC. After 4 days, cells were transferred to anti-CD137-antibody-coated plates. During interferon gamma ELISPOT analyses, PBMC were re-stimulated with U87MG tumor cells. d Analysis of the cytotoxicity of PBMC (HLA-A2-positive) after stimulation with U87MG tumor-lysate-pulsed DC. PBMC were stimulated as described in c. Cytotoxic activity of stimulated PBMC was assessed by LDH release assay. If not explicitly mentioned, differences between groups are not statistically significant

Depletion of CD4-positive or CD8-positive cells reduced the number of interferon gamma spots significantly (p < 0.05). Depletion of CD56-positive cells only displayed a minor effect (Fig. 2b). Priming of PBMC in the presence of an anti-CD137 antibody increased the number of U87MG-reactive cells (Fig. 2c). No increase in the number of U87MG-reactive cells was observed when the antibodies were present during pulsing of DC with tumor lysate. Antibodies that were applied after 4 days of co-culture of DC and PBMC did not increase the number of U87MG-reactive cells on day 5 of the co-culture. However, on day 7 the number of spots was significantly (p < 0.05) higher than in cultures without αCD137 or in cultures where the αCD137 antibody was applied during maturation of DC. In addition, we tested the cytotoxicity of primed PBMC. Similar to the results of the ELISPOT assays, incubation of dendritic cells with αCD137 during maturation and loading with tumor lysate did not enhance cytotoxicity of the PBMC (Fig. 2d). An increased cytotoxicity was measured in co-cultures where αCD137 antibodies had been present from the beginning of the incubation of TPDC and PBMC. Cultures where PBMC were re-seeded after 4 days into new wells that had been coated with αCD137 also showed an increased cytotoxicity on days 5 and 7. Taken together, our data indicated that the strongest immune response against U87MG cells was induced in cultures where incubation with αCD137 started at the beginning of the co-culture of DC and PBMC. In the following experiments, we therefore applied αCD137 immediately on the first day of co-incubation of DC and PBMC.

FACS analyses on days 5 and 7 of the co-culture did not show major differences in T-cell subsets (CD3+/CD4+; CD3+/CD8+), NK cells (CD3−/CD56+) and NKT cells (CD3+/CD56+) after incubation of PBMC with TPDC in the presence or absence of αCD137 (Fig. 3a). We observed a slightly higher percentage of regulatory T cells (CD3+/CD4+/CD25+/FOXP3+) in the cultures with αCD137 (10.95 %) compared with cultures without αCD137 (7.85 %), but these differences were not statistically significant. On both days, we counted the highest number of spots after challenging the different stimulated PBMC with cells of the tumor cell line U87MG (Fig. 3b). We found higher spot numbers in the assays where αCD137 was used. The number of interferon gamma spots after challenging with the other two tumor cell lines (T98G, HT1080) was also higher after priming in the presence of αCD137. The highest cytotoxicity of stimulated PBMC was shown in LDH release assays against U87MG cells with αCD137 (Fig. 3c). Supernatants were harvested and cytokine analyses were done by CBA (Fig. 3d). On both days, supernatants from cultures with αCD137 showed a higher amount of interferon gamma compared with medium control. In addition, analyses of TNFα showed higher levels on both days in supernatants from cultures with αCD137.

Fig. 3.

Treatment with anti-CD137 antibodies increases the response against glioma cells of TPDC-stimulated PBMC. a PBMC (HLA-A2-positive) were stimulated with TPDC in the presence or absence of αCD137 antibodies. After 5 and 7 days of co-culture, flow cytometry with antibodies directed at the indicated antigens was performed. b After 5 and 7 days of co-culture, cells were re-stimulated with tumor cells from lines U87MG, T98G, and HT1080 and interferon gamma ELISPOT analysis was performed. c After 5 and 7 days of co-culture, cytotoxic activity of primed PBMC against cell lines U87MG, T98G, and HT1080 was assessed by LDH release assay. d After 5 and 7 days of co-culture, supernatants were harvested. Interferon gamma and tumor necrosis factor were analyzed by cytometric bead array. Statistically significant differences are indicated

Absence of immune modulation of DCV with monoclonal antibodies directed at CD25

In addition to αCD137, we analyzed whether we could modify DCV by using a monoclonal antibody (daclizumab) against the α chain of the IL2 receptor (CD25). Data from the literature suggested a positive influence on the response against glioma cells by blocking CD25 and reducing the inhibitory function of regulatory T cells. The monoclonal antibody against CD25 (5 μg/mL) was present from the beginning of the incubation of TPDC and PBMC. Analyses were done after 5 and 7 days (Fig. 4).

Fig. 4.

Anti-CD25 treatment did not change the response of TPDC-stimulated PBMC against glioma cells. a PBMC (HLA-A2-positive) were stimulated with TPDC in the presence or absence of αCD25 (daclizumab) antibodies. After 5 and 7 days of co-culture, flow cytometry with antibodies directed at the indicated antigens was performed. b After 5 and 7 days of incubation, cells were re-stimulated with tumor cells from lines U87MG, T98G, and HT1080 and interferon gamma ELISPOT analysis was performed. c After 5 and 7 days of co-culture, cytotoxic activity of primed PBMC against cell lines U87MG, T98G, and HT1080 was assessed by LDH release assay. d After 5 and 7 days of co-culture, supernatants were harvested. Interferon gamma and tumor necrosis factor were analyzed by cytometric bead array. Statistically significant differences are indicated

On day 5 of co-culture, we found lower percentages of all T and NK cell subsets in the cultures with αCD25 antibodies compared with control cultures (Fig. 4a). Regulatory T cells (CD3+/CD4+/CD25+/FOXP3+) were undetectable after co-incubation of PBMC and TLDC in the presence of αCD25 antibodies. On day 7, we found a higher percentage of CD3+/CD8+ cells in co-cultures with αCD25 antibodies. Again, no regulatory T cells in the cultures with αCD25 could be detected. Surprisingly, we found a significantly higher number of U87MG-reactive cells in the control cultures compared with the cultures with αCD25 antibodies (Fig. 4b). Similar results were found in LDH release assays (Fig. 4c). Compared with medium control, the cytotoxicity was significantly lower after incubation with αCD25 antibodies. Culture supernatants were harvested and cytokine analyses were done by CBA. On both days, the cultures with the antibody against CD25 showed a lower amount of interferon gamma compared with medium control. Analyses of TNFα showed similar levels on both days in the cultures with αCD25 and without antibodies.

Immune modulation of DCV with the p38 MAPK inhibitor SB203580

MAPK inhibitors have been suggested to improve cancer immunotherapy by augmenting IL12 secretion by DC and inhibition of regulatory T cells. Therefore, we tested the p38 MAPK inhibitor SB203580 in our model. Analyses of CD3+/CD4+, CD3+/CD8+ and CD56+/CD3− cells on day 5 of incubation showed no clear differences between MAPK-inhibited and control cells (Fig. 5a). Only a slight decrease in regulatory T cells and increased numbers of NKT cells were observed. After 7 days of incubation, these differences were even more pronounced. After 5 days of incubation of PBMC with TPDC, the highest number of U87MG-reactive cells was found in the cultures with the p38MAPK inhibitor (Fig. 5b). The results after 7 days of incubation were similar. The highest number of spots was found for the cultures with p38 MAPK inhibitor. However, specificity of the immune reaction by using the inhibitor could not be enhanced. The analyses of the LDH release assay supported the results of the interferon gamma ELISPOT analyses (Fig. 5c). On both days, the highest cytotoxicity against U87MG cells was observed for the cultures with p38 MAPK inhibitor. The LDH release assay showed no clear differences in the specificity of the immune reaction in the presence or absence of the p38MAPK inhibitor. The analyses of the cytokines showed a higher amount of IFNγ and TNFα in the supernatants with p38MAPK inhibitor compared with vehicle control or medium control.

Fig. 5.

p38MAPK inhibition antibodies increase the response of TPDC-stimulated PBMC against glioma cells. a PBMC (HLA-A2-positive) were stimulated with TPDC in the presence or absence of a p38MAPK inhibitor (p38I). After 5 and 7 days of co-culture, flow cytometry with antibodies directed at the indicated antigens was performed. b PBMC (HLA-A2-positive) were cultured with U87MG tumor-lysate-pulsed DC in the presence or absence of p38I. DMSO was used as vehicle control. After 5 and 7 days of incubation, cells were re-stimulated with tumor cells from lines U87MG, T98G, and HT1080 and interferon gamma ELISPOT analysis was performed. c After 5 and 7 days of co-culture, cytotoxic activity of primed PBMC against cell lines U87MG, T98G, and HT1080 was assessed by LDH release assay. d After 5 and 7 days of co-culture, supernatants were harvested. Interferon gamma and tumor necrosis factor were analyzed by cytometric bead array. Statistically significant differences are indicated

Combination of immune modulation with αCD137 antibodies and p38 MAPK inhibitor

The enhanced anti-tumor activity of PBMC that had been primed in the presence of mAB against CD137 and the p38 MAPK inhibitor SB203580 prompted us to test whether the combination of both reagents have additive or synergistic effects. The FACS analyses after 5 days of incubation showed similar percentages of CD3+/CD4+, CD3+/CD8+ and CD56+/CD3− cells. The presence of the p38 MAPK inhibitor slightly decreased the number of CD3+/CD56+ cells (Fig. 6a). The analysis of the Treg showed a significant (p < 0.04) higher percentage in the cultures with the αCD137 antibody alone compared with the cultures with the p38 MAPK inhibitor. After 7 days of incubation, the percentage of NK cells was increased in the presence of p38 MAPK inhibitor. Again the fraction of Treg was clearly decreased in the cultures in the presence of p38 MAPK inhibitor. In ELISPOT analyses, we did not find significant differences after 5 and 7 days of incubation between the three conditions (Fig. 6b; Supplementary Figure 4). The combination of both agents could not enhance the number of spots in the interferon gamma ELISPOT compared with both assays with a single additive. The comparison of the cytotoxicity between the three different conditions showed on both time points a significantly (day 5: p < 0.05; day 7: p < 0.02) higher LDH release in the culture with the p38 MAPK inhibitor compared with the culture with the monoclonal αCD137 antibody (Fig. 6c; Supplementary Figure 5). The combination of both additives could not enhance the anti-tumor activity compared with the use of both additives alone. Finally, we also analyzed the cytokines IFNγ and TNFα in the supernatants of the different stimulated PBMC. The analysis of IFNγ showed the highest amount in the assay with the p38 MAPK inhibitor without the additional stimulation by the antibody against CD137 (Fig. 6d). The highest TNFα amount was found with the p38 MAPK inhibitor alone. We found no synergistic or additive effects of the combination of the p38MAPKI and the αCD137 antibody compared with the use of both agents as single additives.

Fig. 6.

Combination of αCD137 antibodies and p38MAPK inhibition. a HLA-A2-positive PBMC were incubated together with U87MG tumor-lysate-pulsed DC on plates coated with αCD137 antibodies (αCD137), in the presence of p38MAPK inhibitor (p38I) or in the presence of both reagents (p38I + αCD137). Analyses were performed as described before. After 5 and 7 days of co-culture, flow cytometry with antibodies directed at the indicated antigens was performed. b After 5 and 7 days of incubation, cells were re-stimulated with tumor cells from lines U87MG, T98G, and HT1080 and interferon gamma ELISPOT analysis was performed. c After 5 and 7 days of co-culture, cytotoxic activity of primed PBMC against cell lines U87MG, T98G, and HT1080 was assessed by LDH release assay. d After 5 and 7 days of co-culture, supernatants were harvested. Interferon gamma and tumor necrosis factor were analyzed by cytometric bead array. Statistically significant differences are indicated

Discussion

DC vaccination has become an interesting field in the treatment for malignant diseases; especially for the treatment of malignant gliomas, results of in vitro studies [45, 46], experiments with rodents [47, 48], and even phase I/II clinical trials or case reports showed remarkable effects [49–60]. Priming of T cells needs tumor antigen presentation by MHC class I and class II molecules. Loading of DC with tumor antigens, e.g., by pulsing with tumor cell lysate is necessary to establish a cytotoxic response against tumor cells [45, 61, 62].

In our in vitro model of human glioma immunotherapy, we tried to reduce allogeneic activation by using CD14-positive cells for establishing dendritic cells and PBMC for incubation with DC from the same HLA-A2-positive donor. In ELISPOT analyses, we re-stimulated primed PBMC with the glioma cell lines U87MG and T98G (both cell lines are HLA-A2 positive [63]) and the non-glioma cell line HT1080 (fibrosarcoma). We observed the highest number of interferon gamma spots after re-stimulation with U87MG tumor cells, suggesting that at least some U87MG-specific antigens were recognized by the primed PBMC. The number of interferon-gamma-secreting cells was lower after re-stimulation with tumor cells of glioma cell line T98G. The lowest number was measured after re-stimulation with cells from the non-glioma cell line HT1080. Part of the observed immune response was probably glioma specific; shared antigens of U87MG and T98G could be responsible for the higher number of interferon gamma spots after re-stimulation with glioma cell lines U87MG and T98G compared with re-stimulation with non-glioma cell line HT1080. However, with the lack of an autologous system, we cannot exclude the possibility that these antigens are related to (minor or major) histocompatibility antigens.

In order to modulate the immune response against glioma cells, we used a monoclonal antibody against the co-stimulatory molecule CD137 (4-1BB). The co-stimulatory receptor CD137 is primarily expressed on activated T cells and NK cells [64–66]. Ligation by cell surface 4-1BBL (CD137L) or specific antibodies against CD137 provides a co-stimulatory signal to both CD4-positive and CD8-positive T cells [18, 67] and promotes T cell survival [19, 68–70]. We observed only low expression of CD137L on the used dendritic cells (Supplementary Figure 3), suggesting that this co-stimulatory pathway cannot be activated by the natural ligand in our system. For full activation of T cells, the sole engagement of the TCR is not sufficient. An additional co-stimulatory signal is required for full activation of T cells. The best investigated pair of co-stimulatory molecules is CD28 and its ligand B7 [71–73]. It was shown that treatment of tumor-bearing mice with mAB against CD137 was effective in enhancing tumor immune response. Even in combination with DCV, co-stimulation by CD137 could augment the immune response [74–76]. In our experiments, we observed an enhanced tumor immune response when the antibody against CD137 was used. However, our data indicate that the specificity of the immune response could not be enhanced by application of the antibodies against CD137. Depletion of CD4-positive or CD8-positive cells strongly inhibited the reaction. Obviously, the reaction depends on both cell populations. This is in agreement with the model that we present in Fig. 1. Only class I-restricted immunocompetent cells can recognize class II-negative tumor cells. If the stimulation of these cells requires help from TH1 cells, it is expected that the depletion of one of the two populations will result in a strong suppression of the reaction.

It was reported that immature DC express CD137 and that this expression was enhanced by addition of lipopolysaccharide. A direct growth inhibition of tumor cells through application of antibodies directed at CD137 was observed [76]. In our experiments, we could not find expression of CD137 before and after pulsing and maturation of DC (data not shown). Consequently, the maturation of DC in the presence of antibodies directed at CD137 did not enhance the immune response against U87MG cells.

We observed that after depletion of CD4-positive or CD8-positive cells the number of interferon-gamma-secreting cells in our assays was very low. An immune response mediated by TPDC is dependent on both CD4-positive and CD8-positive cells. The antibody against CD137 did not influence these result in the interferon gamma ELISPOT. All in all, the number of interferon gamma spots was higher when the antibody against CD137 was admitted to the assays compared with assays without antibody. After depletion of CD4-positive or CD8-positive PBMC, the number of spots was not different between assays with or without αCD137 antibodies. In our experiments, the use of αCD137 antibodies had no positive effect in assays with CD4-depleted or CD8-depleted PBMC. Only in the assay with non-depleted PBMC, the antibody against CD137 could augment the number of interferon gamma spots.

In our experiments, we used another potential immune-modulating substance, the specific p38 MAPK inhibitor SB203580. Former investigations showed an abrogation of the function of regulatory T cells through the specific p38 MAPK inhibitor [77]. Furthermore, p38 MAPK is related to signal transduction pathways inducing the expression of IL10 in T cells after T cell receptor/CD28 activation [78]. In another work, it was shown that inhibition of p38 suppresses Toll-like receptor ligand-induced IL10, enhances IL12 production by DC, and enhances their ability to induce TH1 over regulatory T cells [31]. In our work, we observed a considerably increased response against glioma cells in assays where the p38 MAPK inhibitor was added when incubation of TPDC and PBMC started. The highest amount of interferon gamma and TNFα was measured in assays with p38 MAPK inhibitor followed by assays with αCD137 antibodies. Analyses of CD3/CD4/CD25/FOXP3-positive cells showed a lower percentage of these regulatory T cells in assays with p38 MAPK inhibitor compared with medium control. In assays with αCD137 antibodies, the percentage of these cells was even higher then in medium control. After priming of T cells with TPDC, p38 MAPK could be activated and promote cytokine production (IL12, interferon gamma) by antigen-presenting cells and CD4-positive cells [79]. On the other side, the shift from regulatory T cells to TH1 cells [31, 78] can induce an increased number of interferon-gamma-secreting cells and a higher amount of interferon gamma and TNFα in assays with p38 MAPK inhibitor.

After positive results concerning the augmentation of the immune response with the specific p38 MAPK inhibitor and the monoclonal antibody against CD137, we tried to combine both reagents for potentially additive or synergistic effects. These effects were not observed. Probably p38 MAPK inhibition can inhibit the potential enhancing effects of an additional co-stimulation by CD137. After stimulation of the co-stimulatory receptor CD137with its natural ligand or antibodies, nuclear factor κB (NFκB) is activated via tumor necrosis factor receptor-associated factors (TRAF). Another pathway after TRAF activation is via apoptosis signal-regulating kinase-1 (ASK-1) to SAPK/JNK or p38 MAPK [13]. When using a specific inhibitor of the p38 MAPK, it is conceivable that parts of activating signals through CD137 are inhibited. On the other side, enhanced activation of T cells through CD137 induced a higher percentage of CD3/CD4/CD25/FOXP3-positive regulatory T cells that probably can inhibit the immune response. After application of the specific inhibitor of the p38 MAPK, a reduced number of these cells were described [31, 78]. Whether a combination of both p38 MAPK inhibition and antibodies against CD137 can augment a glioma-specific immune response in vivo requires further investigation. First results from vaccination studies using glioblastoma lysate-pulsed DC showed encouraging results and antibodies against CD137 as well as p38 MAPK inhibitors have been used in several clinical studies (reviewed in [80–82]). Therefore, the inclusion of these reagents in DC vaccination studies seems to be realizable.

In contrast to αCD137 antibodies, antibodies directed at CD25 did not enhance the immune reaction against glioma cells in our system. The up-regulation of high-affinity IL2 receptors on activated T cells is an important factor for IL2-driven expansion of these cells. It seems possible that αCD25 antibodies not only deplete regulatory T cells but also inhibit activated T effector cells. The decreased secretion of IFNγ in the presence of these antibodies (Fig. 4) is in agreement with such inhibition of effector T cells. In our experiments, we applied the antibodies simultaneously with DC because the aim of our study was the evaluation of different treatment elements in the same setting. Most investigators apply depletion of regulatory T cells before vaccination. However, the simultaneous application of anti-CD25 antibodies with vaccination can enhance immune responses even in vivo [83]. Whether the depletion of Tregs before co-culture with DC or alternative methods for Treg depletion can improve the immune reaction against glioma cells requires further investigations.

Today, there are some concerns about the usage of PEG2 in DC maturation cocktails and alternatives such as in vivo maturation with imiquimod are highly interesting especially for clinical studies. However, PEG2-containing protocols have been used in most studies, and it has been shown that PEG2 is an important component of the maturation cocktails in vitro [84–87]. Whether substitution of PEG2 can further increase the efficacy of DC vaccination protocols for glioma patients has to be investigated.

Electronic supplementary material

Below is the link to the electronic supplementary material.