Immune responses and immunotherapeutic interventions in malignant pleural mesothelioma

Cytotoxic CD8⁺ T-cell infiltration

The correlation of the presence of tumor-infiltrating CD8⁺ cells with MPM patient survival was demonstrated by immunohistochemistry in extrapleural pneumonectomy specimens. Yamada et al. [4], in a multivariate analysis examining 27 patients undergoing extrapleural pneumonectomy as initial treatment, demonstrated CD8⁺ lymphocyte infiltration as an independent prognostic factor. In addition, Yamada et al. observed positive MHC I expression in all MPM specimens, possibly accounting for the effective anti-tumor activity of CD8⁺ cells. The protective association of tumor-infiltrating CD8⁺ T cells has also been demonstrated in patients who have received induction chemotherapy. In an immunohistochemical evaluation of 32 patients who underwent induction chemotherapy followed by extrapleural pneumonectomy, a high density of tumor-infiltrating CD8⁺ cells was found to be an independent predictor of prolonged survival (3-year overall survival: 83% vs. 28% for high vs. low infiltration, respectively) [3]. Furthermore, this study observed that high levels of tumor-infiltrating CD8⁺ T cells were correlated with reduced frequency of mediastinal lymph node metastasis. Interestingly, the authors observed higher CD8⁺ lymphocyte infiltration in patients receiving cisplatin/pemetrexed (currently the most efficacious MPM chemotherapy regimen) versus cisplatin/vinorelbine. These results suggest that strategies that promote tumor-infiltrating CD8⁺ T cells may be beneficial to MPM patients.

Regulatory T cells (T_regs) and mast cells

Clinical studies examining the presence of CD4⁺CD25⁺FoxP3⁺ regulatory T cells (T_regs) in patients with MPM and their prognostic significance are quite scarce [5]. Hegmans et al. confirmed the presence of significant numbers of T_regs on immunohistochemistry; however, no correlation with survival was possible due to the small number of tumor samples examined. Additionally, the dual and seemingly contradictory role of tumor-associated mast cells has been detailed in breast, lung, and ovarian cancer [22–26], but a paucity of data exist on mast cell function in MPM and their role is unknown. Studies elucidating the natural history of T_reg and mast cell infiltration upon MPM progression are required.

Peripheral immune cell relative ratio as a predictive biomarker in MPM

Peripheral blood neutrophil-to-lymphocyte ratio (NLR), a marker of systemic inflammation shown to be prognostic in solid malignancies [27, 28], was also found to be an independent predictor of survival on multivariate analysis in the recent study examining 173 MPM patients undergoing systemic therapy [29]. One-year survival rate for NLR < 5 and ≥5 was 60% versus 26%, with median overall survivals of 16.7 versus 6.6 months, respectively. Furthermore, subgroup analysis revealed a 3-month survival advantage in patients whose NLR normalized post-treatment. The clinical utility of NLR in mesothelioma, however, remains to be seen, as the promise of this simple, low-cost, peripheral blood test needs to be reproduced by other investigators.

Although MPM was initially thought to be a poorly immunogenic tumor due to a paucity of reported cases of spontaneous remissions [30], clinical investigation into MPM patient samples demonstrates that increased intratumoral CD8⁺ lymphocyte infiltration is beneficial in MPM patients’ prognosis.

T_regs depletion

CD25, a marker of activation on T lymphocytes, is constitutively expressed by T_regs [31]; the presence of CD4⁺CD25⁺ T_regs has been implied in the inhibition of anti-tumor immunity and has also been associated with worse prognosis in several cancers [32–35]. Depletion of T_regs using intratumoral or intraperitoneally injected anti-CD25 antibody in mice prior to intraperitoneal inoculation with mesothelioma tumor resulted in the restriction of tumor growth and significantly increased median survival from 19 to 33 days [5, 36, 37]. Flow cytometric analysis of tumor cell suspensions showed that T_reg depletion resulted in a large accumulation of intratumoral interferon-γ (IFN-γ)-producing CD4⁺ and CD8⁺ cytotoxic T cells, which the investigators proposed to be responsible for the observed anti-tumor response [36]. CD4⁺ cells appeared to be important mediators that lure CD8⁺ cells into the tumor microenvironment, as depletion of CD4⁺ cells significantly reduced intratumoral CD8⁺ cell infiltration by approximately 2.5-fold [36].

Beneficial observations of T_reg depletion prior to tumor implantation were reproduced by Anraku et al. [38]. Enhanced survival was dependent on tumor infiltration by CD8⁺ cells, as CD8-depleted mice were ineffective at mounting a robust anti-tumor response despite T_reg blockade. T_reg depletion by anti-CD25 antibody did not improve survival, however, if administered as a monotherapy in established MPM tumors. Following the observation of increased T_reg percentage in response to pemetrexed therapy, Anraku et al. hypothesized that T_reg blockade in concert with pemetrexed may prove effective. In fact, combination of pemetrexed chemotherapy and anti-CD25 antibody administration—given intraperitoneally 4 days after tumor injection—functioned synergistically, improving survival in mice with established MPM tumors. In addition to pemetrexed, gemcitabine also eliminated established MPM tumors with the addition of T_reg depletion via cyclophosphamide or anti-CD25 antibody [39]. Again, anti-tumor response was demonstrated to be reliant upon CD8⁺ T cells, with nearly unaffected tumor growth occurring in CD8⁺ depleted mice. Cyclophosphamide, a chemotherapeutic with established immunostimulatory properties including the ability to selectively deplete T_regs, was found to sensitize MPM tumor cells to TRAIL-mediated apoptosis, suggesting that anti-tumor immune responses provoked by cyclophosphamide depend on increased tumor cell susceptibility, rather than effector cell proliferation [40]. A more recent investigation observed retarded MPM tumor growth when either anti-CD25 antibody or cyclophosphamide was given following cisplatin cycles [41]. These data implied that T_reg depletion between cisplatin cycles may hinder MPM tumor cell repopulation, potentially improving outcomes. Although T_reg depletion combined with chemotherapeutic agents was beneficial in MPM, T_reg depletion by anti-CD25 did not improve the efficacy of intratumoral IL-2 immunotherapy in an investigation by Jackaman et al. [42].

Although human studies are scarce, the aforementioned investigations imply that the balance between cytotoxic T cells and T_regs within the MPM tumor microenvironment contributes to whether a tumor will be subject to (a) an immunosuppressive environment in which unrestrained tumor growth can occur or (b) an anti-tumor setting in which cytotoxic T cells can elicit an effective cytotoxic immune response.

Tumor-associated macrophage depletion

In various solid tumors, in the presence of appropriate cytokines or ligands, macrophages can polarize into two types: M1 and M2 tumor-associated macrophages (TAMs) [43]. M1 TAMs demonstrate immunostimulatory properties and confer enhanced tumor resistance and cytotoxicity, while M2 TAMs secrete immunosuppressive cytokines, promote angiogenesis, and support tumor progression [43], invasion, and metastasis and overall have been associated with poorer prognosis in lung, breast, prostate, and cervical cancer [44–50]. The role of TAMs in MPM is still largely unknown, however. Interaction with MPM cells appears to shift mature macrophages toward a M2 phenotype, characterized by poor antigen presentation and increased immunosuppressive activity [51]. Upon cocultivation with MPM cells, macrophages released a significant amount of prostaglandin E₂, an arachidonic acid metabolite with considerable anti-inflammatory properties. The presence and production of prostaglandin E₂ contribute to T_reg development, promoting an immunosuppressive tumor microenvironment [52, 53]. With this background, the depletion of pro-tumorigenic M2-TAMs by liposome-encapsulated clodronate (CLIP) in a peritoneal mesothelioma model was performed [54]. Mice with established tumors demonstrated a 50% reduction in both the number of tumors and tumor burden when treated with CLIP compared to control treatments, although CLIP treatment of mice bearing established tumors did not prolong survival. The results of this study imply that TAMs contribute to the development and progression of mesothelioma; however, additional investigations are needed to clarify the role of TAMs in MPM.

Dendritic cell vaccination

Dendritic cell (DC) vaccination exploits the ability of DC to present antigens and initiate a cascade of immunologic events leading to the destruction of cells expressing these specific tumor antigens [55, 56]. Hegmans et al. [57] showed that tumor lysate-pulsed DCs can prevent engraftment and halt mesothelioma progression by stimulating tumor-directed CD8⁺ cytotoxicity. The protective effect of DC immunotherapy appeared to be long-lived as tumor re-challenge after 3 months in this cohort did not result in subsequent mesothelioma outgrowth. Given therapeutically, in mice with established peritoneal mesothelioma, DC treatment was associated with decreased velocity of mesothelioma outgrowth and prolonged survival. Furthermore, adoptively transferred CD8⁺ T cells from mice treated with tumor lysate-pulsed DCs retained tumor-specific cytotoxicity when injected into mesothelioma-bearing mice, prolonging survival in mice when given either 1 week before or 2 days after tumor inoculation. Veltman et al. [58] demonstrated that combination treatment with DC immunotherapy and cyclophosphamide potentiated therapy, significantly improving mice survival compared with cyclophosphamide or DC immunotherapy monotherapy. Timing and dose of cyclophosphamide administration have previously been demonstrated to be vital to whether the drug displays immunosuppressive or immunostimulatory effects [59].

Tumor antigen-targeted immunotherapy

Although tumor lysates contain a repertoire of MPM-associated antigens, it is possible that critical MPM antigens are neither shed nor highly expressed. Therefore, studies have sought to identify MPM antigens as specific targets for immunotherapies. Two such candidate cancer antigens are mesothelin and Wilms’ tumor-1 (WT-1) protein. These proteins demonstrate tumor-specific overexpression and the ability to elicit both active and humoral immune responses [60–64]. Mesothelin, a cell-surface cancer-associated antigen with uncertain function, is overexpressed in the vast majority of epithelioid MPM [65]. While it is highly expressed in MPM, normal mesothelin expression is restricted to the mesothelial lining of the peritoneum, pleura, and pericardium. In addition, soluble mesothelin-related protein (SMRP) may be detected in the serum of patients with MPM and may prove a useful biomarker to follow treatment response [66, 67]. Studies in a murine model of mesothelin-expressing epidermoid carcinoma suggest that targeting mesothelin-expressing cells using SS1P, a recombinant immunotoxin consisting of the variable fragment (Fv) of an anti-mesothelin antibody fused to a Pseudomonas exotoxin, can cause tumor regression with enhanced anti-tumor efficacy observed when combined with chemo- or radiotherapy [68, 69]. Similarly, MORAb-009, a chimeric mouse-human monoclonal IgG_1κ antibody with high affinity for human mesothelin, was found to prevent the adhesion of mesothelin-expressing MPM cells, conferring antibody-dependent cellular cytotoxicity upon cells exhibiting mesothelin positivity [70].

Another approach targeting tumor-associated antigens is the chimeric antigen receptor (CAR). CARs are comprised of a tumor antigen-binding domain linked to an intracellular signaling domain able to activate T cells [71]. CARs unite high-affinity antigen recognition with T-cell signal transduction, two functions that fundamentally arise from separate complexes, the TCR αβ-heterodimer and the CD3 complex. The extracellular, antigen-binding domain is usually derived from a murine or human monoclonal antibody, but other receptors have also been utilized. It follows that antigen recognition by CARs is not MHC-restricted, as is the case for TCR-mediated antigen recognition. Antigen binding by CARs activates the signaling cascade that induces antigen-specific cytolysis, effector cytokine secretion, and T-cell proliferation. A recent investigation utilizing mesothelin-targeted CARs to target mesothelin-expressing MPM (derived from the malignant pleural effusion of a mesothelioma patient) showed that these genetically engineered T cells were able to proliferate, express effector cytokines, and eliminate tumor cells both in vitro and in vivo in a flank MPM model [72]. Engineered T lymphocytes demonstrated anti-tumor efficacy at effector-to-target ratios of approximately 1:10 and 1:40 in vitro and in vivo, respectively. In vivo, both intratumoral administration and intravenous administration were found to be efficacious, with intravenous administration being associated with delayed treatment response, however. In an attempt to improve T-cell trafficking into MPM tumors, T cells bearing mesothelin-targeted CARs were transduced to express the CCL2 chemokine receptor, as CCL2 was found to be highly expressed by MPM tumors [73]. In mice with established flank MPM tumors, mesothelin CAR-transduced T cells bearing CCL2 receptors infiltrated tumors significantly greater than T cells transduced with mesothelin CARs only, improving anti-tumor efficacy.

No pre-clinical investigations are available regarding the targeting of WT-1 in MPM.

Cytokine therapy

Cytokine/chemokine blockades

A dual strategy aimed at bolstering the anti-tumor immune response while dampening the immunosuppressive forces within the tumor environment may be an effective way to permit unabated anti-tumor immunity. Studies by Fridlender et al. [87] suggest that blocking CCL2 and CCL12—chemoattractants that attract T_regs and dampen CD8⁺ T-cell effector functions—augments anti-tumor immune responses by inhibiting T_regs trafficking and immunosuppression within the tumor microenvironment. Blocking CCL2/12 alone with intraperitoneally administered monoclonal antibodies had limited efficacy, leading to significantly slowed tumor growth, but not regression. When combined with mesothelin-vaccine immunotherapy, however, CCL2/12 blockade resulted in a significant decrease in both tumor burden (~50% cure rate) and intratumoral T_regs, as well as an increase in the number and activity of intratumoral CD8⁺ T cells. Importantly, blockade of CCL2/12, in concert with vaccine immunotherapy, induced ~twofold higher expression of the pro-inflammatory molecules CXCL10, ICAM, IL-10, IL-12, IFN-γ, and TNF-α.

A similar approach has used TGF-β blockade to potentiate immunotherapy [88]. TGF-β is produced by MPM at high levels and is believed to play a pivotal role in generating an immunosuppressive tumor microenvironment via promoting T_reg-mediated immunosuppression, impeding tumor infiltration by T cells, and inhibiting anti-tumor T-cell functions [89–91]. TGF-β blockade given for 7 days prior to intratumoral injection of adenovirus expressing IFN-β caused significant reduction in the size of established MPM flank tumors; monotherapy with TGF-β blockade or intratumoral IFN-β gene transfer caused only minimal reduction in tumor burden. Compared with untreated controls, mice receiving TGF-β blockade demonstrated an increased intratumoral CD8⁺ T-cell infiltration and activation and increased expression of pro-inflammatory cytokines, chemokines, and cell adhesion molecules.

TGF-β has also been shown to polarize macrophages toward a pro-tumorigenic M2 phenotype [92] and theorized to polarize neutrophils toward a pro-tumorigenic state N2 phenotype (as opposed to the anti-tumor, pro-inflammatory N1 phenotype); as such, TGF-β blockade’s effect on tumor-associated neutrophil (TAN) phenotype was examined in a recent study [93]. TANs derived from flank MPM tumors in mice treated with TGF-β blockade demonstrated an elevated mRNA expression of pro-inflammatory molecules (TNF-α, CCL3, and ICAM-1), as compared to untreated control MPM tumors. Interestingly, arginase levels were decreased 2- to 5-fold in TANs isolated from mice treated with TGF-β blockade, compared to untreated controls; the immunosuppressive functions of arginase have been shown previously [94]. The aforementioned data led the authors to hypothesize that TGF-β blockade resulted in the polarization of TANs toward the anti-tumor, N1 phenotype. Depletion of TANs in untreated mice (that is, removal of the immunosuppressive, N2 TANs) led to ~twofold increase in CD8⁺ T-cell activation as shown by flow cytometry, as well as a small reduction in tumor growth. Conversely, depletion of pro-inflammatory, N1 TANs in mice treated with TGF-β blockade significantly blunted T-cell activation, allowing for increased tumor growth. The results of this study imply that following TGF-β blockade, TANs that exhibit an anti-tumor phenotype (N1 TANs) are recruited into the MPM tumor microenvironment, leading to elaboration of pro-inflammatory cytokines with subsequent activation of intratumoral CD8⁺ T cells. Although the MPM tumor-neutrophil relationship has not been extensively studied, it warrants further investigation as the pro-inflammatory environment setup by N1 TANs may be an important mechanism for tumor regression.

Immuno-adjuvant therapies

An alternative approach to immunotherapy aims to enhance the intrinsic activity of immune system. In a flank model of MPM, Currie et al. [95] intratumorally injected Toll-like receptor (TLR) ligands to activate DCs and induce T-cell-dependent cytotoxicity in a type I interferon (IFN)-dependent manner. Intratumorally administered TLR7 agonist induced systemic, CD8⁺ T-cell and type-I IFN anti-tumor responses that prevented primary MPM tumor growth but not the growth of distant, untreated tumor foci [96]. The anti-tumor effects were obviated following CD8⁺ cell depletion as well as in nude mice. The relative paucity of intratumoral CD4⁺ T cells, along with the lack of observed systemic CD4⁺ T-cell activation in mice with response, implied that CD4⁺ T-cell assistance may be required for a robust systemic anti-tumor response. Indeed, distant untreated tumor foci were targeted when an agonistic anti-CD40 antibody (mimicking CD4⁺ T-cell assistance) was administered systemically in combination with local TLR7 agonist. The authors hypothesized that systemic anti-CD40 antibody administration, a surrogate for CD4⁺ T helper cells, increased tumor cell sensitivity to cytotoxic T cells or increased the quality of the CD8⁺ T-cell response.

Cytotoxic T lymphocyte-associated antigen 4 (CTLA-4) is an inhibitory molecule found on T cells, hypothesized to regulate T-cell activation [97]. Anti-CTLA-4 antibodies (ipilimumab and tremelimumab) that block interaction between CTLA-4 and its associated ligands (B7.1 and B7.2) have been observed to promote anti-tumor immune responses in pre-clinical investigations of various solid tumors [98–100]. Recent clinical successes in treating melanoma with ipilimumab (Yervoy) have led to its FDA approval [101, 102]. While the use of CTLA-4 has been investigated in other non-melanoma solid tumors, there have yet to be any notable pre-clinical investigations of its use in MPM models. Interestingly, however, a phase II clinical trial utilizing tremelimumab as second-line therapy for unresectable MPM is in progress in a single Italian institution [Trial MESOT-TREM-2008].

Although the vast majority of these studies were performed in either flank or peritoneal MPM models, orthotopic MPM models have been previously well described [103–105] and should be utilized in future studies to more accurately mimic the pleural tumor microenvironment.

Cytokine-based therapy

As a result of the aforementioned promising pre-clinical results, a recent phase I clinical trial assessing intrapleural interferon-β gene transfer in patients with MPM and malignant pleural effusions was performed [106]. Successful interferon-β gene transfer was detected in seven of ten patients. Variable anti-tumor immune responses were observed in seven out of ten MPM patients, as demonstrated by the activation of natural killer cells and cytotoxic T cells, as well as the generation of antibodies to MPM-associated antigens. An update of this cohort, following repeated intrapleural administration of adenoviral encoded interferon-β, published more recently observed that nine out of ten MPM patients had demonstrated humoral immune responses to MPM-associated antigens following treatment [107]. While the majority succumb to progressive disease, three out of ten MPM patients were alive at 18, 39, and 42 months following treatment, demonstrating either mixed response or stable disease on follow-up PET imaging. Another study by Boutin et al. [108] suggested that intrapleural administration of recombinant interferon-γ may be beneficial to MPM patients with early (stage I) disease; eight patients demonstrated significant histologic response and nine patients yielded partial responses for an overall response rate of 20%.

A phase I-II clinical trial utilizing intrapleural administration of IL-2 in 23 stage I-IIA patients detailed some anti-tumor efficacy and an acceptable side-effect profile, with high intrapleural IL-2 levels observed with low corresponding serum levels [109]. Pleural effusion-derived mononuclear cells displayed significantly higher lymphokine-activated killer (LAK) activity than peripherally obtained mononuclear cells. No correlation between IL-2 dose and LAK activity or clinical response was observed. Additional phase II clinical trials assessing intrapleural administration of IL-2 in MPM have been carried out reporting marginal histopathologic responses and modest improvements in median survival [110–113]. Unfortunately, immunologic parameters following treatment were not detailed in these investigations. The results from clinical trials assessing intrapleural IL-2 efficacy in MPM are conflicting and, although some increase in overall survival appears to be achieved in these studies, the potential of IL-2 to significantly impede disease progression has not been convincingly demonstrated. While phase II intrapleural IL-2 has demonstrated some effectiveness in treating MPM, other routes of administration (intravenous and/or subcutaneous) as first-line treatment have failed to do so [114, 115]. In a group of 60 MPM patients who received pre-operative intrapleural IL-2 treatment, a significantly higher number of T_regs was observed in surgical specimens, as compared to 33 patients who did not receive IL-2 [116]. Within the IL-2-treated group, high T_regs, defined as greater than the median on immunohistochemical assessment, was correlated with both decreased patient survival and increased tumor recurrence on subgroup analysis. Furthermore, pre-operative, intrapleural IL-2 therapy in MPM was associated with a high level of tryptase-positive mast cell infiltration, which correlated with both increased overall survival and time to progression [117]. When taken together, the density of tryptase-positive mast cell and T_reg infiltration was able to more accurately predict prognosis than either parameter alone; a high tryptase-positive mast cell and low T_reg tumor infiltrate were found to possess the best overall patient survival. The results from these studies must be interpreted with caution however, since IL-2 is known to induce a wide range of immunomodulatory effects, several of which may contribute to the aforementioned survival benefit.

Active immunization

Based on pre-clinical studies demonstrating the efficacy of DC-based immunotherapy [57], Hegmans et al. [118] conducted a small, phase I clinical trial where DCs were pulsed with autologous tumor lysate, matured ex vivo, and then administered intradermally and intravenously to MPM patients following cytoreductive surgery and cisplatin-based chemotherapy. Although DCs immunotherapy was well tolerated and appeared to elicit an immunologic response, no survival benefit was noted. Interpreting this data is challenging, however, due to small sample size (n = 10) and inclusion of a highly selected patient population; only patients with the epithelioid subtype of MPM and who had either a partial response or stable disease following multimodality therapy were included.

To specifically target MPM tumors, cancer antigens, such as mesothelin and WT-1, that are highly expressed and thought to be physiologically relevant for the tumor may be targeted. Recently, the NCI cancer panel identified WT-1 as the leading tumor-associated antigen to be targeted [119]. Studies assessing WT-1 as a target for immunotherapy have been carried out in leukemias and have shown clinical efficacy [62, 63, 120] and are the basis for expanding these therapies to MPM. Since MPM expresses WT-1 at high levels [121, 122] and provokes an immune response [63], it is an attractive target for MPM immunomodulatory therapy. A phase I peptide vaccination clinical trial from Memorial Sloan-Kettering Cancer Center for unresectable or relapsed MPM patients is ongoing. In these patients, inoculation with WT-1 peptide elicited WT-1-specific CD4 and CD8 T-cell responses, with minimal toxicity [123].

Immunotoxin therapy

Based on promising pre-clinical results targeting mesothelin with the immunotoxins SS1P and MORAb-009, clinical trials were initiated [65, 124–126]. Phase I clinical trials in patients with advanced MPM, pancreatic, and ovarian cancers, utilizing both bolus and continuous intravenous administration of SS1P while well tolerated, have shown only modest clinical responses [124, 125]. Similarly, a phase I MORAb-009 dose-escalation trial in MPM, pancreatic, and ovarian cancer patients failed to demonstrate any complete or partial responses [126]. Pre-clinical studies, however, have demonstrated significant anti-tumor efficacy using combination of MORAb-009 and chemotherapy treatment [70]; as such, a multicenter phase II clinical trial utilizing cisplatin/pemetrexed with MORAb-009 in MPM patients is currently ongoing [Clinicaltrials.gov Identifier-NCT00738582].

Despite early pre-clinical results utilizing genetically engineered T cells to target MPM-associated tumor antigens, the efficacy of adoptive T-cell therapy has yet to be evaluated in MPM patients.