Emerging avenues in immunotherapy for the management of malignant pleural mesothelioma

Background The role of immunotherapy in cancer is now well-established, and therapeutic options such as checkpoint inhibitors are increasingly being approved in many cancers such as non-small cell lung cancer (NSCLC). Malignant pleural mesothelioma (MPM) is a rare orphan disease associated with prior exposure to asbestos, with a dismal prognosis. Evidence from clinical trials of checkpoint inhibitors in this rare disease, suggest that such therapies may play a role as a treatment option for a proportion of patients with this cancer. Main text While the majority of studies currently focus on the established checkpoint inhibitors (CTLA4 and PD1/PDL1), there are many other potential checkpoints that could also be targeted. In this review I provide a synopsis of current clinical trials of immunotherapies in MPM, explore potential candidate new avenues that may become future targets for immunotherapy and discuss aspects of immunotherapy that may affect the clinical outcomes of such therapies in this cancer. Conclusions The current situation regarding checkpoint inhibitors in the management of MPM whilst encouraging, despite impressive durable responses, immune checkpoint inhibitors do not provide a long-term benefit to the majority of patients with cancer. Additional studies are therefore required to further delineate and improve our understanding of both checkpoint inhibitors and the immune system in MPM. Moreover, many new potential checkpoints have yet to be studied for their therapeutic potential in MPM. All these plus the existing checkpoint inhibitors will require the development of new biomarkers for patient stratification, response and also for predicting or monitoring the emergence of resistance to these agents in MPM patients. Other potential therapeutic avenues such CAR-T therapy or treatments like oncolytic viruses or agents that target the interferon pathway designed to recruit more immune cells to the tumor also hold great promise in this hard to treat cancer.


Background
MPM is an aggressive inflammatory cancer associated with exposure to asbestos. Despite having been banned in the western world, current data from the US has shown that the rate of MPM in males has remained constant from 1994, while the rate of MPM in females has remained unchanged for decades [1]. Indeed, while the use of asbestos has declined in industrialized nations, asbestos is still being exported to developing nations [2,3]. Moreover, environmental exposure is still widespread due to (a) previous industrial use; (b) its difficulty to remove; (c) natural deposits are being disturbed by human activities; and (d) housing proximity to these natural deposits [1,[4][5][6][7].
The economic burden for MPM is significant both at the level of total cost for hospital care [8,9], and economic burden [9,10].

Open Access
*Correspondence: sgray@stjames.ie 1 Thoracic Oncology Research Group, Central Pathology Laboratory, CPL 30, TCDSJ Cancer Institute, St James's Hospital, Dublin D08 RX0X, Ireland Full list of author information is available at the end of the article Clinically, if untreated, MPM has a median survival time of 6 months, and most patients die within 24 months of diagnosis. The current standard of care (SOC) is a combination of pemetrexed/raltitrexed and cisplatin chemotherapy) [11] is non-curative and results in a response rate of ∼ 40% [12], and there is no standard second line therapy once treatment fails. Recently, the addition of an anti-vascular endothelial growth factor (anti-VEGF) therapy (Bevacizumab) has been shown to enhance OS when given in the first line setting [13]. And whilst this therapeutic combination is now the new standard of care in France [14], it has not yet been approved by the FDA, issues with cost and lack of reimbursement prevent it from being added to the SOC in many countries, and other anti-angiogenic combinations have not been successful [15].
The power of the human immune system to prevent cancer (often described as immune-surveillance) was first mooted by Ehrlich in 1909 [16,17]. One of the mechanisms used by cancer cells to evade immune surveillance involves a series of surface regulatory markers (called checkpoint molecules), and has led to the development of checkpoint inhibitors for cancer therapy, an area of active investigation in MPM. Other prominent new treatment options emerging in MPM (and other cancers) involve cancer immunotherapy, a situation where the patient's own immune system (antibodies, cells, cytokines, etc.) is exploited to eliminate tumor cells [17,18]. In the following review we examine some of the current clinical studies of immunotherapies in mesothelioma, explore some of the issues potentially linked to lack of objective responses, and discuss alternative immunotherapy targets which may translate into mesothelioma clinical trials moving forwards.

Immunotherapy in MPM in the historical setting
Historically, immunotherapy in mesothelioma is not new, and studies involving this cancer have been attempted for over 25 years [19]. Examples of early trials in this arena predominantly used Interleukin-2 (IL-2) and Tumor Necrosis Factor alpha (TNF-α), were ineffective and suffered particularly from a lack of scalability and logistical issues [19,20]. Some encouraging clinical responses were observed for patients with good performance status [21] while more recent studies in animal models suggest that direct injection of IL-2 plus an agonist anti-CD40 antibody induces regression of large mesothelioma tumors through a mechanism involving natural killer (NK) cells driven acquisition and/or maintenance of systemic immunity and long-term effector/memory anti-mesothelioma responses [22].
Some of the earliest trials involved the infusion of interferon (IFN) gamma to treat malignant pleural effusions [23,24], oftentimes with complete responses in Stage I patients [24]. Follow up studies using intrapleural infusion of interferon-gamma in a larger cohort of (n-89) patients observed a 20% overall response rate with most responses in early stage disease especially if the tumor was confined to the parietal or diaphragmatic pleura [25]. Whilst these and other studies of interferon therapy combined with chemotherapy regimens suggest that this strategy could be useful [26,27] with median survival rates of approximately 8-12 months, other studies found significant toxicities [28]. Later studies using intrapleurally infused autologous human activated macrophages combined with interferon gamma found limited antitumor activity [29], while a study involving debulking surgery coupled with interferon based immunotherapy also demonstrated limited overall survival benefit [30], suggesting that interferon therapy has limited clinical benefit in MPM.
Another potential immunotherapy target for MPM involves Granulocyte macrophage colony-stimulating factor or GM-CSF, used as an immune-stimulatory adjuvant to elicit antitumor immunity [31,32]. Initial studies in MPM involved infusions of GM-CSF [33][34][35] but few or no responses were observed [34,35], and high toxicity [33] or a poor OS (median survival of 7 months) were the outcomes. A small clinical trial (n = 22 patients) was conducted involving a vaccination strategy comprising autologous mesothelioma tumor cell lysate combined with GM-CSF was conducted. The trial was found to be safe, and induced tumor specific immunity in 32% of patients, but saw only stable disease ad no tumor objective responses [36]. More recently, tumor derived GM-CSF was shown to actually promote immunosuppression in mesothelioma suggesting that actually targeting this molecule may be more effective in augmenting immunotherapy in MPM [37].

Checkpoint inhibitor immunotherapy within the neo-adjuvant setting
Although not SOC, there is compelling evidence that a select subgroup of mesothelioma patients benefit from a surgery-based multimodal approach, particularly if they have an epithelioid histological subtype, lowervolume disease, and/or minimal to no nodal involvement [38]. In MPM microscopic complete resection is considered to not be achievable, and patients who have surgically resectable disease often undergo an aggressive multi-modality therapy for which the optimal combination therapy has not yet been identified [39]. Various taskforces have been set to explore the various options, and some proposed consensus reports have recently been published [40][41][42]. In this regard neoadjuvant immunotherapy prior to surgery has been mooted as an advantageous prospect in the management of solid tumors as they enhance T-cell activation the moment antigen is encountered, and encouraging findings from early-phase clinical trials in various cancers support this notion [43][44][45]. A series of Phase I/ II clinical trials involving neo-adjuvant immunotherapy prior to surgical resection have been initiated in MPM (Table 1) but as these trials are still running the results are not yet mature.

Checkpoint inhibitor immunotherapy in the surgical setting
Aside from the neo-adjuvant setting, there is one ongoing multicenter, randomized, controlled, open-label Phase II study which is designed to assess the efficacy of standard chemotherapy combined with nivolumab in the context of multimodal management of early-stage MPM (Table 1) [50]

Checkpoint inhibitor immunotherapy within the front-line (first-line) setting
A number of studies have been completed or are ongoing which aim to examine the potential utility of checkpoint inhibitor immunotherapy in the front-line setting. Several of these are ongoing (Table 2), but interim and published results from some of these trials as discussed by us and others [47,48,51,52] suggest that checkpoint inhibitors will be important new agents in the frontline setting for the management of MPM. For example, interim results of PrE0505 (Table 2) demonstrate a median OS of 20.4 months, and a 1-year OS rate of 70.4% [52] with new trials such as DREAM3R (Table 3) initiated on the basis of these interim results.
Most recently, analysis of the Checkmate-743 trial ( Table 2) has resulted in the FDA approval of a combination therapy of Nivolumab/Ipilimumab as a first line treatment for unresectable MPM [53,54]. The median OS with this treatment was consistent between patients with epithelioid histology (18·7 months) and non-epithelioid histology (18·1 months) [53]. The OS benefit observed in the non-epithelioid subgroup for the checkpoint inhibitor combination versus standard chemotherapy is notable (18.7 months vs 8.8 months), but can be attributed to the established inferior effect of chemotherapy in the non-epithelioid subtype [53].
The approval of Nivolumab/Ipilimimab by the FDA as a front-line therapy for the treatment of MPM is greatly encouraging [54], and the results of the various ongoing trials will help improve the utility of checkpoint inhibitors in the front-line setting moving forwards.

Checkpoint inhibitor immunotherapy within the salvage setting
A large number of studies are also currently investigating the potential use of checkpoint inhibitors within the second or third-line (salvage therapy) setting, and are discussed in more detail in the following sections.

Single agent checkpoint inhibitor studies
Several clinical trials of checkpoint inhibitors as single agents have been completed in the salvage setting and are summarized in Table 3. In particular, the MERIT trial, a Phase II multi-center, open-label, uncontrolled, trial of patients within the second-line setting, observed an OS of 17.3 months which resulted in Nivolumab being approved by the Japanese Ministry of Health, Labor and Welfare for salvage therapy in MPM [65]. Despite this, several other single-agent trials of checkpoint inhibitors such as DETERMINE, PROMIS-MESO, JAVELIN or Nivo-Mes for example have had mixed results (Table 3) and these data have been reviewed extensively by us and others [19,47,66]. Another ongoing single institute phase II trial (IRB14-1381-NCT02399371) has reported interim results suggesting that checkpoint inhibitors show robust activity in the salvage setting, with a median OS of 11.5 months [64] (Table 3). Most recently, interim results from the Phase III trial CONFIRM trial (Table 3) have been presented [67]. In the interim results presented superiority for nivolumab over and above placebo was observed for OS with a hazard ratio of 0.72 (p = 0.018). The same was true for PFS with a hazard ratio of 0.61 (p < 0.001). Interestingly PD-L1 expression had no bearing on OS, whereas an epithelioid histology was found to have a significant survival advantage with a 12 month OS (40 vs. 26.7 months) with a hazard ratio of 0.71 (p = 0.021) [67]. Despite this, data from the Dutch expanded access program, suggest that in a real-world setting patients with recurrent malignant pleural mesothelioma, nivolumab did not provide the same benefits as observed in clinical trials with worse ORR and a median OS of only 6.7 months [68,69].

Combination checkpoint inhibitor studies in MPM
Several studies have now combined checkpoint inhibitors in the salvage setting summarized in Table 4.
The INITIATE trial (Table 4) which had an estimated OS of approximately 12.7 months (Table 4) [72][73][74] along with the NivoMes trial (Table 3) were recently re-examined to complete a comprehensive immune cell profiling of samples [74], and the results demonstrated that the     combination therapy induced a profound increase in the proliferation and activation of effector memory T cells which was not observed in the monotherapy, suggesting a clear benefit for the combination therapy, and therefore this observation warrants a larger Phase III trials of this combination therapy in the salvage setting.
Other trials of combination therapies in the combination setting have not shown as good responses. For example NCT03075527 a single institute trial examining a Durvalumab/ Tremelimumab combination was prematurely terminated as it did not meet its primary endpoint of ORR at interim analysis [75] ( Table 4). The results of these trials continue to support the further development of checkpoint inhibitors as both single agents or as combination therapies in MPM.

Are there other checkpoint inhibitor therapy options?
However, checkpoint inhibitors and anti-tumor immunity are not restricted to just the three candidates (CTLA-4, PD-1, PD-L1) currently described in the previous sections. Many other potential immunotherapy targets have been identified as shown in Table 5, and some potentially actionable candidates are discussed in the following sections.
LAG-3: Lymphocyte activation gene-3 (LAG-3, also known as CD223) is a checkpoint inhibitor, where it acts as an inhibitory co-receptor, playing pivotal roles in autoimmunity, tumor immunity, and anti-infection immunity [79]. A number of agents targeting this receptor are in active clinical development [79]. LAG-3 has been proposed as a candidate checkpoint inhibitor target in MPM [80], and expression of LAG-3 has been identified on immune cell infiltrates isolated from patients with MPM [81]. Most recently a study found that whilst the immune phenotype of pleural fluid cells had no prognostic significance, the presence of PD-1 + /LAG-3 + / TIM-3 + CD4 + tumor-infiltrating lymphocytes in pleural biopsy samples correlated with worse overall survival [82]. Intriguingly, analysis of The Cancer Genome Atlas (TCGA) dataset indicates that high mRNA expression of LAG-3 is associated with better OS (Table 5), and a recent analysis suggests that high LAG-3 mRNA expression is a common feature in mesothelioma at both the mRNA and protein level [83,84].
In a recent report Marcq et al., have shown that in pre-clinical models of MPM, a combination of an anti-PD-1/anti-LAG-3 results in delayed tumor growth and survival benefit [85]. Interestingly, a bispecific antibody Tebotelimab (in development by Macrogenics) targets both LAG-3 and PD-1 and is currently in a Phase I dose escalation study (NCT03219268). Preliminary data from this study suggests it has an acceptable safety profile with encouraging early evidence of anti-tumor activity, with one confirmed partial response in a mesothelioma patient [86].
VISTA: V-type immunoglobulin domain-containing suppressor of T cell activation (VISTA) (also known as VSIR or B7H5) is an immune-checkpoint gene which was first reported as having strong expression in epithelioid MPM, above and beyond that seen in other solid cancers, with obvious implications for the immune response to MPM and for its immunotherapy [87]. A subsequent study has confirmed that VISTA expression is higher in epithelioid subtype [88,89]. High expression of VISTA in epithelioid cancers is associated with a better OS, in both the TCGA dataset (Table 1) and in a separate analysis in the French MESOBANK samples [90]. CA-170 is a small molecule inhibitor of VISTA in development by Curis, and NCT04475523 is an open-label, multicenter doseescalation study of CA-170, assessing the safety and tolerability of this agent in patients with relapsed/refractory solid tumors. This trial had a cohort of (n = 12) MPM patients, and the results for this cohort were recently presented, which effectively showed that while CA-170 was well tolerated and showed favorable clinical pharmacokinetics, no partial or complete responses were reported in MPM [91].
B7-H3: B7H3 (also known as CD276) is another candidate checkpoint, whose expression has been observed in mesothelioma [92]. In 2018, it was reported that expression of B7H3 was positive in 41 of 44 mesothelioma samples tested, and of these 39/44 highly expressed B7H3 [93]. The histological subtype of the mesothelioma specimens examined was not provided. A separate study has confirmed that almost all MPM patients across all histological subtypes were positive for B7-H3 (epithelioid − 90.9%; non-epithelioid − 88.9%) [94]. In this analysis albeit of a small number of patients (n = 31), it was found that the expression level of B7-H3 was significantly higher than that of PD-L1 in the epithelioid type, whereas in non-epithelioid samples, there was no significant difference in the expression levels of PD-L1 and B7-H3 [94]. Analysis of the TCGA dataset demonstrates that high expression of B7H3 mRNA is associated with a worse OS (Table 1).
Several compounds targeting B7-H3 are under active development by companies such as Daichii-Sankyo (DS-7300-a humanized antibody drug (topoisomerase inhibitor) conjugate) or Macrogenics (Enoblituzumab/ MGA271-monoclonal antibody; MGC018-a humanized monoclonal antibody (DNA alkylating agent) conjugate). All are currently in Phase I/II clinical trials. Interim data from the MGC018 trial (NCT03729596) has been presented which indicate that this antibody drug conjugate (ADC) has a manageable safety profile with early evidence of clinical activity [95].

TIM3:
The T-cell inhibitory receptor Tim3 (T-cell immunoglobulin and mucin-domain containing-3) is a heavily investigated immune-checkpoint [96], and demonstrating significant pre-clinical activity [97,98]. Tim3 expression has been examined in MPM, and its expression is found on both tumor cells and immune cells [81,99], and double-positive PD-1 + /TIM-3 + CD8 + T cells are more commonly found in PD-L1-positive tumors [99]. Whilst expression of this receptor does not have any prognostic value (Table 1) in the MPM TCGA dataset, its expression suggests that it may be a potential new target in mesothelioma [85,100].
TIGIT: The role of inhibitory repressors (IRs) on tumor infiltrating lymphocytes (TILs) is generally associated with T-cell exhaustion. In such a situation, when exposed to chronic tumor antigens, T cells become dysfunctional/exhausted and upregulate various checkpoint inhibitory receptors (IRs) that limit their survival and function [101]. In a recent analysis of TILs isolated from patients with MPM [102], it was observed that the levels of TIGIT were significantly greater on TILs isolated from MPM compared with those isolated from tumor free lungs (TFLs), with high levels of TIGIT on ~ 60% of CD8 + T-cells [102]. Functionally, the expression of TIGIT was associated with TIL hypofunction [102], suggesting that an anti-TIGIT therapy may have potential for therapeutic use in mesothelioma [102], and a number of clinical trials and anti-TIGIT therapies are in progress [103]. BTN3A1/ BTN3A3: Butyrophilin subfamily 3 (BTN3) genes are emerging as checkpoints critical to the regulation of immune responses for specific γδ T cell (Vγ9Vδ2T) subsets which can exert anti-tumoral effects [104]. Two of these BTN3A1 (CD277) and BTN3A3 (BTF3) are upregulated in MPM and high expression is associated with MPM OS (Table 5). Vγ9Vδ2T cell infiltration into tumor tissues is associated with a positive prognosis across multiple cancers [105], which makes the BTN3A subfamily an interesting target for enhancing anti-tumor immunity. Several companies have developed Bold value represents significance at p < 0.05 a Assessed using oncomine analysis [76] of the Gordon MPM dataset (normal pleura versus malignant) [77] b Assessed using UALCAN [78] [106]. OX40/OX40L: These are members of the TNF receptor superfamily (TNFRSF), and are key co-stimulators of T cells during infection, and there has been an increasing interest in harnessing these receptors to augment tumor immunity. OX40 (TNFRSF4) and OX40L (TNFSF4) have been implicated in mesothelioma. In a recent study of an animal model of mesothelioma, tumor resident regulatory T-cells were shown to co-express high levels of CTLA-4 and OX40 on a large proportion of cells. Individually targeting OX40 generated an effective response against tumor development, and was found to be synergistic with anti-CTLA4 agents [107]. Whilst there appears to be little information as regards OX40L in mesothelioma, analysis suggests that OX40L is overexpressed in MPM and high expression is associated with poorer OS (Table 1). At present a Phase I clinical trial (NCT03894618) of SL-279252 (PD1-Fc-OX40L) is assessing the safety, tolerability, pharmacokinetics, antitumor activity and pharmacodynamic effects of this bifunctional fusion protein [108] in patients with advanced solid tumors or lymphomas. The trial is expected to complete in April, 2022.
Other candidate checkpoints which could be therapeutically targeted include PVR (CD155), CD47 (MER6, IAP), CD84 (SLAMF5), TNFRSF14 (HVEM), and various members of the nectins (Table 5). Clearly, as our knowledge of checkpoint inhibitor therapy improves, the wealth of candidate targets and agents currently under investigation coupled with emerging data from patients with MPM suggest that further investigations of combination immune checkpoint inhibitor therapy are warranted.

Oncolytic therapy?
While early studies of interferon or GM-CSF based MPM therapy based on infusions proved disappointing, new therapeutic strategies which involve oncolytic virus mediated expression of these agents may have more clinical activity and benefit. [109][110][111]. In the most recent of these (ClinicalTrials.gov NCT01119664), 40 patients received two intra-pleural doses of a replication-defective adenoviral vector containing the human IFNα2b gene (Ad. IFN) concomitant with a 14-day course of celecoxib followed by chemotherapy (either first line with pemetrexed, or second-line with pemetrexed or gemcitabine. whilst patients in the first-line cohort had median OS of 12.5 months, in second-line settings the median OS was 21.5 months with 32% of patients alive after 2 years [109]. A new Phase III study-(INFINITE-NCT03710876) is currently recruiting for a trial involving intra-pleural administration of TR002 an adenovirus-delivered Interferon Alpha-2b (rAd-IFN) and examining its efficacy and safety in combination with celecoxib and gemcitabine in patients with mesothelioma.

Oncolytic adenovirus overexpression of IFN: Several Phase I trials involving intra-pleural infusion of adenoviral mediated interferon therapy have been attempted in recent years
Oncolytic measles virus overexpression of IFN: On a separate note, in 2015 defects within the interferon type-I response were found to render MPM cells sensitive to oncolytic measles virus [112], and a follow up study found that the defects in IFN-I responses that renders them sensitive to oncolytic activity induced by exposure to the measles virus were most frequently homozygous deletions of all the 14 IFN-I genes (IFN-α and IFN-β) [113]. These results suggest that the interferon pathway continues to be potentially important immunotherapy target in MPM.
Intriguingly, a recent report indicates that IFN-γ treatment of mesothelioma cells results in both the upregulation of membranous PD-L1 [114], which suggest that interferon therapy, could be combined with anti-PDL1 checkpoint inhibitors for the treatment of MPM.
Oncolytic adenovirus overexpression of GM-CSF: ONCOS-102 is an immune-priming GM-CSF coding oncolytic adenovirus in development by Targovax. The safety, immune and clinical results of an open-label Phase I/II clinical trial of ONCOS-102 in combination with pemetrexed/cisplatin (NCT02879669) for 1st and 2nd line unresectable MPM have just been reported, and indicate that the immune priming function of ONCOS-102 was both safe and had robust immune activation, with increased T-cell infiltration [115]. Moreover upregulation of PD-L1 was noted, which could potentially allow for future combinations with checkpoint inhibitors [115]. Currently, Targovax has been granted a European Patent for combining this oncolytic virus with checkpoint inhibitors (European Patent no 3293201) [116], and has further announced a collaboration with Merck to evaluate ONCOS-102 with Pembrolizumab in MPM [117]. The envisaged trial will be a randomized phase II of up to 100 patients comparing this investigational triple combination against Pembrolizumab and SOC, with multiple centers in both the USA and EU participating, and the aim will be to start enrolling patients into the trial within twelve months. Moving forwards it will be interesting to see the results of any clinical trials combining ONCOS-102 and checkpoint inhibitors.

CAR-T based approaches
Chimeric antigen receptors (CARs) therapy functions by coupling the Human leukocyte antigen (HLA)-independent binding of a cell surface target to the delivery of a tailored T-cell activating signal by recognizing and binding to specific tumor-associated antigens [118,119]. The potential to use CAR-T therapy in mesothelioma has been explored fairly extensively, and pre-clinical models using various targets including mesothelin (MSLN) [120,121], Fibroblast activation protein (FAP) [122], Met Proto-oncogene (cMET) [123], pan-ErbB [124] and others have been extensively tested [125,126].
Various clinical trials of CAR-T based approaches in MPM have been conducted and were recently summarized by us and others [47,118,119,127].
One factor which may currently limit the use of CAR-T strategies in solid tumors could be the issue of T-Cell exhaustion [128]. However, recent studies suggest that checkpoint inhibitors may be a mechanism for improving the potency of CAR-T cell therapies in this regard [129][130][131], and other approaches such as co-stimulation induction and cytokine based approaches may also have merit [128].
Dendritic cell (DC) therapy is a cell based vaccination approach used to initiate an anti-tumor immune response [127]. In mesothelioma initial approaches used autologous tumor lysate loaded DCs, and have showed excellent long lasting clinical responses with survival up to 66 months post treatment [132][133][134][135][136]. While greatly encouraging, the main disadvantage of this approach remains that it is time-consuming and may not often generate sufficient amounts of the required quality for DC therapy. Allogenic tumor lysates have the possibility to circumvent this drawback [137], and a Phase I clinical trial MesoCancerVa (NCT02395679) has recently completed. In this trial, no dose-limiting toxicities were established and radiographic responses were observed. The median PFS was 8.8 months and median OS was not reached at a median follow-up of 22.8 months [137]. In a follow up analysis of the peripheral blood T cell receptor β (TCRβ) chain repertoire of nine MPM patients before and 5 weeks after the start of dendritic cell (DC)-based immunotherapy, it was found that clinical responses to DC-mediated immunotherapy was dependent on both the pre-existing TCRβ repertoire of total CD3 + T cells and on therapy-induced changes, in particular expanding PD1 + CD8 + T cell clones, and therefore TCRβ repertoire profiling could potentially allow for the selection of MPM patients that might benefit from DC-based immunotherapy [138].
These promising results have led to the establishment of the Phase II/III DENIM trial (NCT03610360) which aims to recruit n = 230 patients to examine the OS in patients treated with DCs loaded with this allogeneic tumor cell lysate as maintenance treatment after chemotherapy [139]. This trial is estimated to complete in January, 2021, and the results will be eagerly awaited.

Outstanding issues and other therapeutic considerations
Clearly immunotherapy will in the future play important roles in the management of this cancer. As we continue to develop our understanding and knowledge of these exciting therapeutic options and avenues of approach, additional issues and possibilities arise summarized in Table 6, and are discussed in more detail in the following sections. What is the best way to monitor response to checkpoint inhibitors?
8 Is the cost prohibitive for the use of checkpoint inhibitors in the second-line/salvage setting?

Tumor-treating fields and checkpoint inhibitors
Developed by Novocure, the NovoTTF ™ -100L System is a device which uses alternating electric fields at specific frequencies and intensities (called Tumor Treating Fields or TTF) to selectively disrupt mitosis in cancerous cells [140]. This technology has received FDA approval for use in MPM [141], though concerns exist as to whether potential inherent biases and lack of sufficient controls can allow for a true interpretation of the therapeutic value of this system in MPM [141][142][143][144]. Novocure has recently initiated Phase III clinical trials (e.g. NCT02973789) of its platform in combination with immune checkpoint inhibitors in NSCLC [145]. If these clinical trials show efficacy it will interesting to see if similar clinical trials of the NovoTTF ™ -100L System combined with checkpoint inhibitors will be conducted with MPM moving forwards.

Patient stratification: Is there a role for tumor mutational burden in predicting response to immunotherapy?
Tumor mutational burden (TMB) is emerging as a strong predictor for identifying cohorts of patients who may respond to checkpoint inhibitor based therapy [146]. Theoretically, a higher TMB should therefore increase the likelihood for tumor neo-antigen production and as such the probability for immune recognition and tumor cell killing [147]. Even though MPM is considered to have a low TMB [48,148,149], TMB has been assessed is some available studies of checkpoint inhibitors.
Keynote-028-Expanding on a more detailed analysis of the entire trial cohort (n = 475) it was found that T-cell-inflamed gene expression profiles (GEP), PD-L1 expression and/or tumor mutational burden was associated with a higher likelihood of response to therapy. Within this analysis of the n = 25 mesothelioma patients n = 19 had GEP; n = 12 had PD-L1 positivity and n = 9 had TMB data available [61]. However, no subgroup analysis was available for the mesothelioma cohort alone.
In an analysis of the Keynote-158 study with a prespecified cutpoint of at least 10 mutations per megabase as TMB-high, 9/84 MPM patients who were assessed as being TMB-low had an ORR [63], although in terms of PFS and OS TMB-high status with Pembrolizumab treatment was not significant for the overall population [150]. In a single case study, an MPM patient who derived a prolonged response to a checkpoint inhibitor (45 months to 52 cycles of Pembrolizumab) was also assessed for TMB. The baseline biopsy was found to have 0.92 somatic mutations per megabase, while the relapse biopsy had 0.26 [151]. The issue of TMB therefore remains to be resolved for its potential utility in predicting or stratifying MPM patients to checkpoint inhibitor based immunotherapy.

Is there any utility for the use of PD-L1 expression as a biomarker to predict response?
The role of PD-L1 expression as a biomarker to predict outcome in MPM is still an ongoing issue that has yet to be resolved. If one considers the results of Checkmate-743, PD-L1 status does not predict for OS benefit as similar survival was seen in the subgroups with less than 1% vs 1% or greater PD-L1 status [53]. Similar results have been observed in other clinical trials [58,59,71]. The results of the MERIT trial found that differences in OS and PFS favored positive PDL-1 expression (although not-significant) [55]. What is emerging from these studies is that expression of PD-L1 is associated with higher ORR [60,69,70,72,152], and in analysis of the Dutch expanded access program, long survival for patients with partial responses suggested a clinical benefit that is correlated with ORR [69]. Moreover, expression of PD-L1 and non-epithelioid histology is associated with higher ORR [55,69] One feature that emerged from Checkmate-743 was that patients who had tumor PD-L1 expression of less than 1% had better survival with chemotherapy which suggests that absence of PD-L1 might be indicative for chemotherapy based regimens. Support for this comes from a recent analysis of the immune microenvironment in MPM which identified that chemotherapy treated patients deriving the best OS were PD-L1 negative and had a higher percentage of stromal CD8 + lymphocytes [153,154]. Likewise, the Dutch nivolumab EAP study also found that patients no PD-L1 expression had very poor responses to Nivolumab with significantly worse ORR and mOS [68,69].
Other interesting developments as regards PD-L1 expression as a candidate biomarker are emerging from the CONFIRM trial which found that PD-L1 expression had no bearing on OS [67].
As such PD-L1 remains a contentious biomarker in this sphere, and a significant number of patients exist who whilst being PD-L1 negative, demonstrate ORRs to checkpoint inhibitors. The challenge will be to identify new markers or ways to identify such patients, perhaps using transcriptomic or other approaches [53,[154][155][156][157].

Is there a role for epigenetic priming in the use of checkpoint inhibitor therapy for MPM?
Epigenetic priming is emerging as a mechanism to potentially prime solid tumors for enhanced targeting of immune checkpoint inhibitors via the induction or upregulation of PD-L1. It is now well established that epigenetic targeting agents such as decitabine (a DNA methyltransferase inhibitor) can induce or upregulate PD-L1 expression [158,159]. In this regard a clinical trial (NCT03233724) designed primarily for NSCLC (but includes MPM) ( Table 7), aims to assess if epigenetic targeting with Decitabine can prime solid tumors for enhanced targeting of immune checkpoint inhibitors (in this instance Pembrolizumab) [160].
Other epigenetic targeting agents such as histone deacetylase inhibitors (HDACi) are also well established as candidate agents with the capacity to induce PD-L1 in cancer cells [161][162][163][164]. However, in MPM cell lines HDACi by themselves had modest effects on PD-L1, but when combined with decitabine, higher induction of this checkpoint inhibitor were observed [165].

Can a patients BAP1 status inform therapy decisions?
Given the potential sensitivity of BAP1 mutated MPM to Enhancer of Zeste 2 (EZH2) inhibitors [166], is there an opportunity to combine EZH2 inhibitors with checkpoint inhibitors? In a non-mesothelioma setting, a patient with SMARCB1-deleted, metastatic, poorly differentiated chordoma was treated with Tazemetostat (EZH2 inhibitor), and had a significant increase in intratumoral and stromal infiltration by immune cells expressing checkpoint regulators PD-1 and LAG-3 [167]. In this regard, preliminary data from the EZH-203 (NCT02860286) trial of Tazemetostat in MPM had a 12 week DCR of 47% (n = 35), with mostly stable disease with no complete responses and only 2 partial responses [168]. Given the observation that Tazemetostat results in enhanced infiltration of immune cells it may be possible to conceive of a strategy which could include Tazemetostat/anti-PD1 in BAP1 mutant patients. In this regard, a recent study has shown that while macrophages can be directly cytotoxic for mesothelioma cells, inhibition of EZH2 reduced that activity because it induced PD-1 overexpression. A combination of PD-1 blockade and EZH2 inhibition restores macrophage cytotoxicity [169]; and suggests that combination therapy with EZH2 inhibitors plus checkpoint inhibitors may have potential for clinical efficacy in MPM.
For those patients with wild-type BAP1, there may be a possibility to combine gemcitabine with immunotherapy. Initial pre-clinical studies suggest that it did not change the expression of PD-L1 on human mesothelioma cell lines in vitro [170]. Additional evidence now suggests that wild-type (WT) BAP1 positivity may be a factor in the sensitivity of MPM to gemcitabine [171,172]. Moreover, a recent study using PET demonstrated that gemcitabine based therapy in a murine colon cancer model strongly induced PD-L1 Expression [173]. Furthermore, a synergistic effect for gemcitabine combined with anti-PD1 was observed in pre-clinical models of mesothelioma, and similar responses were seen in two patients who were resistant to gemcitabine or anti-PD-1 (Pembrolizumab) Not yet recruiting as monotherapy, but who achieved observable clinical responses following combination therapy [170], and has led to the ORIGIN trial (Table 7), which will examine if a Gemcitabine/Atezolizumab combination can overcome resistance in either advanced NSCLC or Mesothelioma patients progressing under immune-checkpoint inhibitors or gemcitabine. Intriguingly, gemcitabine has been shown by us to act as a DNA methyltransferase inhibitor, reactivating silenced genes in mesothelioma cells [174], and as such the observed responses to gemcitabine on PD-L1 expression changes may reflect an epigenetic priming event, although functional studies will be required to delineate this.

Can combined targeting of TLRs and checkpoint inhibitors improve responses to immunotherapy?
Toll-like receptors (TLRs) are expressed on many innate immune system cells and play a role in maturation of dendritic cells and priming of cytotoxic T lymphocytes [175]. A subset of TLRs has been shown to stimulate antitumor responses, and agonists to these receptors are being investigated in clinical trials [175,176]. Several studies have linked TLR3, TLR7 and TLR9 as potentially targetable in MPM [177][178][179], which could conceivably be trialed in combination with checkpoint inhibitors [180]. It is interesting to note that NCT02668770 is a clinical trial of ipilimumab and MGN1703 (a TLR Agonist) currently running in patients with advanced solid malignancies. Whether any mesothelioma patients are in this trial is unknown.
What is the best way to monitor immunotherapy response?
Hyper-progression, an accelerated growth or progression of a cancer after treatment is initiated, has been observed for a subset of patients undergoing checkpoint inhibitor therapy [181], and can emerge either during therapy, or can emerge post-therapy [182,183]. This further complicated by the issue of pseudo-progression where patients obtain an objective response following an initial progression with immunotherapy [184] The estimated occurrence of hyper-progression is estimated at 4 to 29%, while that of pseudo-progression ranges from 0 to 15% [184]. Whilst there is little evidence that hyperprogression occurs during treatments of mesothelioma, two patients have reported as showing pseudo-progression under treatment with Pembrolizumab, within the first 15-30 weeks of therapy followed by responses [185]. Additionally in the DREAM trial, two patients (4%) were also observed to undergo pseudo-progression in response to treatment with Durvalumab [51]. In a recent editorial on this topic key issues remain such as: why it occurs; is it simply a lead-time bias phenomenon; does it have a strong biological basis such as clonal selection; can we identify and predict those in whom it will occur; and if be stopped by additional therapies [186]. As more and more clinical trials of immunotherapies complete in mesothelioma, vigilance will be required to assess if hyper-progression does occur in MPM while undergoing treatment with immunotherapy.
Some efforts have been made to differentiate pseudoprogression from progression and hyper-progression, such as radiological responses DNA [184]. As PET/CT imaging has been used for the prediction of survival in response to Pembrolizumab in mesothelioma [187], and may be useful to incorporate into immunotherapy based regimens for the treatment of mesothelioma. Other methods that have been explored in other cancer types have involved analyses circulating-tumor DNA or cell-free DNA to assess response to immunotherapy, but larger prospective cohort studies will be required to confirm their potential use [184]. Pathologic scoring of responses to immunotherapy has also been explored [188,189], but may have limited utility in distinguishing between pseudo-, hyper-and progression in MPM. Clearly, new methods or modalities to monitor immunotherapy response will be required moving forwards.

Is the cost prohibitive for the use of checkpoint inhibitors in the second-line/salvage setting?
The combined cost for Ipilimumab/Nivolumab in the USA has been estimated approximately $153,800 for four cycles, while that of Nivolumab alone would be of the order of $87,000 [190]. One of the most commonly used chemotherapies in the treatment of MPM in the secondline or salvage setting is vinorelbine [191], which has been estimated to cost $515 for 24 weeks [190]. Given that the recent Dutch EAP program for Nivolumab in pre-treated MPM patients demonstrates a median OS of 6.7 months [68,69], whilst most trials of vinorelbine in the same setting have a median OS of approximately 9-11 months [191,192], the question arises if the cost of checkpoint inhibitors in the second line setting will limit use.

Conclusions
The following sections have described the current stateof-play as regards immunotherapy in MPM. A significant number of studies are investigating checkpoint inhibitors as both monotherapy or in combination therapy in both the front-line and salvage settings. Treatment combinations designed to recruit more immune cells to the tumor such as oncolytic viruses or those that target the interferon pathway hold promise. CAR-T therapy is emerging as a new avenue of approach for immunotherapy in MPM.
Despite impressive durable responses, immune checkpoint inhibitors do not provide a long-term benefit to the majority of patients with cancer [193]. The data arising from immune checkpoint inhibitor studies in MPM has resulted in one FDA approval for a combination checkpoint inhibitor for the first-line treatment of unresectable MPM. The only other approval is for second-line therapy in the salvage setting and is restricted to Japan. Overall, this would suggest that these agents will shortly become part of the frontline treatment options for MPM in the coming years. Given the data from Checkpoint-743 it would seem that nivolumab/ipilimumab should be used in the first line setting, however, cost reimbursement may limit their uptake [194]. The issue of whether or not to give it to all comers irrespective of histology and PD-L1 status has however yet to be resolved given the data that suggests PD-L1 negative tumors have better responses to chemotherapy, and that patients with the sarcomatoid histology may be better candidates for checkpoint inhibitors [53,153,195]. Indeed it may be that PD-L1 negative non-sarcomatoid patients should initially be treated with a chemotherapy regimen and then proceed to a checkpoint inhibitor in the salvage setting upon progression, whilst PD-L1 positive patients should be offered first-line nivolumab/ipilimumab. Overall, it would appear that additional studies will be required to further delineate these issues, and improve our understanding of the immune system as a therapeutic target in MPM. Moreover, many new potential checkpoints have yet to be studied for their therapeutic potential in MPM. All these plus the existing checkpoint inhibitors will require the development of new biomarkers for patient stratification, response and also for predicting or monitoring the emergence of resistance to these agents in MPM patients.