Trials and Tribulations of Pancreatic Cancer Immunotherapy

Abstract

Immunotherapy has revolutionized cancer treatment in the last decade, and strategies to re-activate cytotoxic immunity are now standard of care in several malignancies. Despite rapid advances in immunotherapy for most solid cancers, progress in immunotherapy against pancreatic ductal adenocarcinoma (PDAC) has been exceptionally difficult. This is true for several approaches, most notably immune checkpoint inhibitors (ICIs) and GM-CSF cell-based vaccines (GVAX). Though many immunotherapies have been explored in clinical trials, few have shown significant therapeutic efficacy. Further, many have shown high rates of serious adverse effects and dose-limiting toxicities, and to date, immunotherapy regimens have not been successfully implemented in PDAC. Here, we provide a comprehensive summary of the key clinical trials exploring immunotherapy in PDAC, followed by a brief discussion of emerging molecular mechanisms that may explain the relative failure of immunotherapy in pancreas cancer thus far.

1 -. INTRODUCTION

Cancer immunotherapy has shown remarkable efficacy in the treatment of several solid tumors. This rapid progress is primarily due to the advent of immune checkpoint inhibitors (ICIs), which have yielded significant antitumor activity in most cancers [1–7]. These strategies consist of neutralizing antibodies against negative regulators of immune function, such as cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), programmed cell death protein 1 (PD-1), and PD-1 ligand 1 (PD-L1), impeding the ability of tumor cells to escape immune surveillance [8]. However, despite significant advances in immunotherapy in nearly all cancer types, progress for immunotherapy in pancreatic ductal adenocarcinoma (PDAC) has been difficult, particularly for ICIs. Though several clinical trials have evaluated ICIs or other immunotherapies in PDAC, the results have been mostly disappointing, particularly compared to successes in other cancers. In this review, we discuss past and present efforts to advance immunotherapy in the treatment of PDAC, as well as emerging evidence that may explain the relative failure of immunotherapy in PDAC and the means through which these obstacles can potentially be overcome.

2 -. SINGLE/DUAL IMMUNE CHECKPOINT INHIBITORS

Based on its success in melanoma, the anti-CTLA-4 antibody Ipilimumab was the first ICI to be evaluated in PDAC. In a 2010 Phase 2 trial, 27 patients with PDAC were enrolled, 20 with metastatic disease, and 7 with localized disease. Patients were administered a 3 mg/kg/dose of Ipilimumab every 3 weeks in 4 dose cycles for a maximum of 2 cycles. Though 3 patients experienced severe, grade ≥3 immune-mediated adverse effects (colitis, encephalitis, and hypophysitis), indicating that checkpoint inhibition was sufficiently robust to elicit autoimmune-mediated side effects [9]. Yet, no objective responses were observed, and the median overall survival on treatment was 5 months [9]. In 2012, as part of a pan-cancer trial, 14 patients with advanced PDAC were treated with escalating 0.3 to 10 mg/kg doses of the anti-PD-L1 antibody BMS-936559 (MDX-1105) every 14 days in 6-week cycles for up to 16 cycles. However, as observed with single-agent Ipilimumab, no objective responses were observed in any patient [10]. Pan-cancer studies evaluating the anti-PD-L1 antibody MPDL3280A (maximum dose 20 mg/kg every 3 weeks) and the anti-PD-1 antibody Pembrolizumab (maximum dose 10 mg/kg every 2 weeks) have also included patients with advanced PDAC and, despite marked efficacy in other cancer types, these agents failed to produce any objective responses in PDAC [11, 12].

The anti-PD-L1 antibody Durvalumab (1500 mg every 4 weeks for up to 12 months) has also been evaluated in a Phase 2 trial for PDAC, which showed an objective response rate of 0% in 32 patients who had previously received either fluorouracil–based or gemcitabine-based treatment [13]. This study also evaluated the combination of the Durvalumab in combination with the CTLA-4 inhibiting antibody Tremelimumab. In addition to the 32 patients receiving Durvalumab alone, 32 patients received Durvalumab (1500 mg every 4 weeks) combined with Tremelimumab (75 mg every 4 weeks) for 4 cycles followed by Durvalumab for up to 12 months or until the onset of progressive disease or unacceptable drug toxicity. The objective response rate of this combination was 3.1% compared to 0% for patients receiving Durvalumab alone. Grade 3 or higher adverse effects were observed in 7/32 (22%) patients receiving Durvalumab and Tremelimumab, compared to 2/32 (6%) patients receiving Durvalumab monotherapy. However, due to the low response rate in both arms, patients were not enrolled in the planned study expansion [13].

While these and other trials strongly suggest that single agents or combined ICIs do not offer clinical benefit to PDAC patients, there is a notable exception regarding tumors deficient in DNA mismatch repair (dMMR) with high microsatellite instability (MSI-H). Mismatch repair deficiency has been shown to strongly predict response to anti-PD-1 therapy in various solid tumor types [14]. The Phase 2 KEYNOTE-158 cancer study evaluated single-agent Pembrolizumab in 233 patients with advanced non-colorectal MSI-H/dMMR cancer, all of whom had been previously treated. Of the 22 PDAC patients enrolled, the authors observed an overall response rate of 18.2%, with a median overall survival of 4 months, median progression-free survival of 2.1 months, and a median duration of response of 13.4 months [15]. This is consistent with previous data suggesting that MSI positive PDACs have increased tumor-infiltrating lymphocytes than MSI negative tumors [16], and a smaller study demonstrating that 4/7 PDAC patients with Lynch Syndrome derived clinical benefit from ICI-based regimens [17].

Combined, these data strongly suggest that single-agent ICIs offered benefit to MSI-H/dMMR patients and led to the tissue-agnostic FDA approval of Pembrolizumab for MSI-H/dMMR solid tumors, including PDAC, in 2017. It is important to note that the MSI-H/dMMR phenotype is relatively uncommon in PDAC. While it has been suggested that up to 22% (24/109) of surgically resected PDAC tumors can be categorized as MSI-H based on immunohistochemistry [18], other studies report that as few as 0.8% (7/833) of PDAC tumors are deficient in mismatch repair, particularly in the metastatic setting [17]. Additionally, the 18.2% ORR seen in MSI-H PDAC in the single arm KEYNOTE-158 trial was substantially lower than the response rates seen in MSI-H cholangiocarcinoma (40.9%), and small intestine (42.1%), gastric (45.8%), and endometrial (57.1%) cancers (15). It is important to note that though 18.2% of patients showed a radiographic response by Response Evaluation Criteria in Solid Tumors (RECIST) v1.1 criteria, this did not correspond to meaningful changes in clinical outcomes when compared to other trials using standard of care chemotherapy. Hence, the true rate of MSI-H/dMMR and its predictive value for ICIs warrants continued exploration in the treatment of PDAC. Pivotal trials examining single-agent ICIs in PDAC are summarized in Table 1.

Table 1.

Select clinical trials examining the single agent efficacy of immune checkpoint inhibitors (ICIs) in pancreatic cancer

ICI	Phase	Number of Patients	Prior Lines of Therapy	Response Rate	Median PFS	Median OS	Ref
Ipilimumab	II	27	Any	0%	NR	5.0	[9]
MDX-1105	I	14	≥1	0%	NR	NR	[10]
MPDL3280A	Ib	5	Any	0%	NR	NR	[11]
Durvalumab	II	32	1	0%	NR	NR	[13]
Durvalumab+Tremelimumab	II	32	1	3.1%	NR	NR	[13]
Pembrolizumab	II	22 (MSI-H/dMMR)	≥1	18.2%	2.1	4.0	[15]

Abbreviations: Microsatellite Instability High/Deficient in DNA mismatch repair (MSI-H/dMMR); NR (Not Reported)

3 -. COMBINATION STRATEGIES

Despite the clinical benefit observed in MSI-H/dMMR patients, overall single-agent or combined ICIs have shown little efficacy in PDAC. This has led to several trials evaluating ICIs in combination with chemotherapy, radiation, other immunotherapies, and targeted therapies. Overall, these trials have shown mixed results, with key findings summarized below.

3.1 -. ICIs & CHEMOTHERAPY

While chemotherapy has long been considered immunosuppressive, there is emerging data that cytotoxic chemotherapy may potentiate the effects of ICIs through various mechanisms [19]. Most notably, in addition to their well-documented effects on nucleotide synthesis, DNA polymerization, and translation, chemotherapeutic agents alter various additional cell functions, including the processing and presentation of self-peptide [19–21]. Accordingly, the combination of chemotherapy and immunotherapy has shown efficacy in several solid cancer types [19]. Several such combinations have been explored in PDAC (Table 2), particularly in light of evidence suggesting that neoadjuvant chemotherapy may have an extensive impact on the pancreatic tumor microenvironment (TME) and increase the intra-tumoral accumulation of CD4+ and CD8+ T-cells in PDAC patients [22]. One of the seminal trials investigating the intersection between chemotherapy and immunotherapy in PDAC evaluated the combination of gemcitabine (1000 mg/m² given on days 1, 8, and 15 of 28-day cycles) with the anti-CTLA-4 antibody Tremelimumab (dose escalation 6, 10, or 15 mg/kg) on day 1 of each 84-day cycle for a maximum of 4 cycles. This approach showed an overall response rate of 10.5%, with 2/19 patients achieving a partial response, and a median overall survival of 7.4 months. The combination of gemcitabine and Tremelimumab had a relatively low rate of grade 3/4 toxicity, most commonly asthenia (11.8%) and nausea (8.8%), with one patient developing a serious drug-related complication in the form of diarrhea with dehydration [23].

Table 2.

Select clinical trials examining the efficacy of immune checkpoint inhibitors (ICIs) combined with cytotoxic chemotherapy in pancreatic cancer

Chemotherapy	ICI	Phase	Number of Patients	Prior Lines of Therapy	Response Rate	Median PFS	Median OS	Ref
Gemcitabine	Tremelimumab	I	19	No Anti-CTLA-4 or Radiation	10.5%	NR	7.4	[23]
Gemcitabine	Ipilimumab	Ib	16	No Gemcitabine	12.5%	2.5	8.5	[24]
Gemcitabine	Ipilimumab	Ib	21	Any	14%	2.78	6.9	[25]
Gemcitabine nab-Paclitaxel	Pembrolizumab	Ib	11	Any	18%	NR	8.0	[28]
nab-Paclitaxel	Nivolumab	I	9	≥1	22.2%	NR	NR	[26]
nab-Paclitaxel±Gemcitabine	Nivolumab	I	6	0	50%	NR	NR	[26]
Gemcitabine nab-Paclitaxel	Nivolumab	I	50	0	18%	5.5	9.9	[27]
Gemcitabine nab-Paclitaxel	Pembrolizumab	II	11	0	27.3%	9.1	15.0	[29]
Gemcitabine nab-Paclitaxel Cisplatin	Pembrolizumab	Ib/II	25	0	70.8%	NR	16.5	[30]
Gemcitabine nab-Paclitaxel	Durvalumab Tremelimumab	II	11	0	73%	7.9	NR	[31]

Abbreviations: NR (Not Reported)

Two Phase 1b studies also explored the similar combination gemcitabine and the anti-CTLA-4 antibody Ipilimumab. The first study included 16 patients with PDAC, and was a single-institution study with dose escalation performed in a 3+3 fashion to determine the maximum tolerated dose. Patients received 750 mg/m² gemcitabine with 3 mg/kg Ipilimumab, 1000 mg/m² gemcitabine with 3 mg/kg Ipilimumab, or 1000 mg/m² gemcitabine with 6 mg/kg Ipilimumab. Therapy consisted of 12 weeks of induction, and gemcitabine was administered on weeks 1-7 and 9-11, and Ipilimumab on weeks 1, 4, 7, and 10. Based on the incidence of dose-limiting toxicities (19%, 25%, and 38%, respectively), this established a maximum tolerated dose of 1000 mg/m² gemcitabine and 3 mg/kg Ipilimumab. Overall, the disease control rate was 43%, with 2/16 (12.5%) patients showing a partial response, 5/16 (31.25%) showing stable disease, and the remaining 9/16 (56.25%) experiencing progressive disease. Median progression-free survival was 2.5 months, and the median overall survival 8.5 months [24].

The second study evaluated gemcitabine and Ipilimumab in 21 PDAC patients. Gemcitabine was administered to a maximum dose of 1000 mg/m² and given once weekly for 7 weeks followed by 1 week off. Subsequent gemcitabine treatments were administered in cycles of 3 weeks on and 1 week off. Ipilimumab was given at a maximum dose of 3 mg/kg on treatment weeks 1, 4, 7, and 10, and every 12 weeks thereafter. Thirteen patients were enrolled during dose escalation, and the remaining 8 treated at the maximum tolerated dose. The median overall survival was 6.9 months, progression-free survival 2.8 months, and the overall response rate was 14% (3/21). Of the three responding patients, one showed a complete response, with a median response duration of 11 months. This study was associated with an increased rate of grade 3/4 toxicities compared to gemcitabine and Tremelimumab, with 48% of patients developing anemia, 48% developing leukopenia, and 43% developing neutropenia [25].

Several clinical trials have also evaluated the combination of chemotherapy with PD-L1/PD-1 inhibiting antibodies. For instance, a Phase 1 trial explored the combination of the anti-PD-L1 antibody Nivolumab with chemotherapy. This study consisted of two arms, one of which comprised 11 patients who had received one prior chemotherapy regimen, and another consisting of 6 treatment-naïve patients. Previously treated patients were administered Nivolumab (3 mg/kg on days 1 and 15 of a 28-day cycle) and nab-paclitaxel (125 mg/m2 on days 1, 8, and 15), and treatment-naïve patients were treated similarly with the addition of gemcitabine (1000 mg/m² on days 1, 8, and 15). Of the previously treated 11 patients, 9 were evaluable. Of this group, 2/9 (22.2%) had a partial response, 4/9 (44.4%) had stable disease, and 3/9 (33.3%) had progressive disease. For the 6 treatment-naïve patients, 3 (50%) had a partial response and 3 (50%) had stable disease. Grade 3/4 adverse effects were observed in 2/11 (18%) of previously treated patients and included pulmonary embolism, neutropenia, and anemia. For the treatment-naïve patients, 2/6 (33.3%) developed grade 3/4 anemia. At the time of publication, only one dose-limiting toxicity was observed in the form of non-immune hepatitis in a treatment-naïve patient, most likely due to gemcitabine. Once resolved, this patient continued receiving Nivolumab and nab-paclitaxel without the addition of gemcitabine [26].

The same investigators more recently reported data from an extended Phase 1 study, in which 50 treatment-naïve patients (including 6 from their previous report) were treated with Nivolumab with gemcitabine and nab-paclitaxel as described above. In this larger cohort, they observed one complete response (2%), 8 partial responses (16%), and 23 patients with stable disease (46%) for a disease control rate of 64%. Median progression-free survival was 5.5 months, and the median overall survival 9.9 months. Despite the improved efficacy, 48 patients (96%) had at least 1 grade 3/4 treatment-related adverse event, most frequently anemia (36%), neutropenia (36%), gastrointestinal events (24%), hepatotoxicity (22%), peripheral neuropathy (16%), thrombocytopenia (12%), and colitis (12%). One patient had a grade 5 adverse event in the form of respiratory failure, and 7 patients (14%) discontinued treatment due to toxicity [27].

Similarly, a recent Phase 1 study evaluated Pembrolizumab in combination with chemotherapy in several advanced cancer types. The 11 PDAC patients included in the study were administered gemcitabine (1000 mg/m²) and nab-paclitaxel (125 mg/m²) on days 1 and 8 every 21 days, with Pembrolizumab (2 mg/kg) every 21 days. Of these patients, 2/11 (18%) showed a partial response, with a median overall survival of 8.0 months. Two patients developed dose-limiting toxicities in the first cycle of treatment in the form of grade 3 thrombocytopenia, both of whom had been previously treated with systemic chemotherapy [28].

A single-center Phase 2 study subsequently evaluated the combination of Pembrolizumab, gemcitabine, and nab-paclitaxel in patients with metastatic PDAC, using the dose regimens described previously. Of the 11 evaluable, treatment-naïve patients in this study, the authors observed a disease control rate of 100%, with 3/11 (27.3%) patients showing a partial response and the remaining 8/11 (72.7%) showing stable disease. The overall survival for the treatment-naïve group was 15.0 months and progression-free survival 9.1 months. However, the primary endpoint of a >15% complete response rate was not met. Grade 3 adverse events occurred in 53% of patients in this arm, though 100% of patients experienced at least one treatment-emergent adverse event. Overall, the authors concluded that this regimen can be safely administered to treatment-naïve PDAC patients, and has slightly improved efficacy compared to the combination regimen of gemcitabine and nab-paclitaxel alone [29].

A Phase 1b/2trial has also evaluated the addition of cisplatin (maximum tolerated dose 25 mg/m²) to Pembrolizumab, gemcitabine, and nab-paclitaxel in metastatic PDAC. Of the 25 patients enrolled, 24 were evaluable. With this combination, 2/24 (8.3%) had complete responses, 15/24 (62.5%) had partial responses, 4/24 (16.7%) had stable disease, and the remaining 3/24 (12.5%) patients experienced disease progression. At the time of publication, the authors note a median overall survival of 16.5 months. Grade 3/4 adverse events were common, with 76% of patients developing thrombocytopenia, 32% developing anemia, 24% contracting an infection, and 16% had diarrhea. Treatment-related death occurred in 3 patients, with one patient each developing infection, cardiac arrest, or stroke [30].

Finally, a recent Phase 2 study examined the combination of gemcitabine, nab-paclitaxel (both administered as described above), Durvalumab (1500 mg on day 1 in 28 day cycles) and Tremelimumab (75 mg on day 1 for the first 4 cycles only) in the first-line treatment of metastatic PDAC. While final results from the 190 enrolled patients are not published at this time, the results of 11 patients are in press. In this initial phase, the disease control rate was 100%, and 8/11 (73%) patients had a partial response. The most frequently observed grade 3 or higher adverse events included hypoalbuminemia (45%), abnormal serum lipase (45%), anemia (36%), fatigue (27%), abnormal white blood cell counts (27%), and hyponatremia (27%), with colitis in one patient (9.1%). While the median overall survival had not been reached at the time of publication, the 6-month survival rate in this study was 80% [31].

3.2 -. ICIs & RADIOTHERAPY

The combination of radiotherapy (RT) and ICIs has shown tremendous efficacy in treating various solid cancers [32]. Radiation delivered to a local site has also been shown to cause regression of distant metastatic sites outside of the radiation field, a phenomenon known as the abscopal effect [33–36]. This is presumably due to RT-induced DNA damage in the form of base substitutions, frameshift mutations, and small deletions that manifest as de novo tumor antigens [37, 38]. However, PDAC tumors are not generally radiosensitive, with some studies even suggesting that the addition of RT offers no clinical benefit when added to standard of care chemotherapy [39]. However, given the paucity of antigen presentation in PDAC [40], RT has been suggested as a potential means to enhance tumor cell antigenicity and sensitize PDAC to ICIs. This approach has shown efficacy in preclinical models [41] and is now being explored in PDAC patients. There are several ongoing clinical trials evaluating RT in combination with immunotherapy, as described in Table 3.

Table 3.

Ongoing clinical trials examining the efficacy of immune checkpoint inhibitors (ICIs) combined with radiation therapy (RT)

Clinical Trial Number	Phase	Combination Therapy	Notes
NCT02305186	I/II	Pembrolizumab, RT	Borderline resectable PDAC
NCT03161379	II	GVAX, Cyclophosphamide, Nivolumab, RT
NCT01595321		GVAX, Cyclophosphamide, FOLFIRINOX, RT	Surgically resected PDAC
NCT03104439	II	Nivolumab, Ipilimumab, RT
NCT02648282	II	GVAX, Cyclophosphamide, Pembrolizumab, RT	Locally advanced PDAC
NCT02311361	I/II	Tremelimumab and/or Durvalumab, RT	Unresectable PDAC
NCT04327986	I/II	M7824, M9241, RT	Advanced PDAC

Though relatively few studies have been fully completed, early Phase 2 results for NCT03104439 were published in 2019. Twenty-five previously treated patients with metastatic PDAC received Ipilimumab (1 mg/kg every 6 weeks), Nivolumab (240 mg every 2 weeks), and 3 fractions of 8 Gy of RT at cycle 2. For the 22 evaluable patients, the authors reported an overall response rate of 13.6%, with 1 patient having a complete response and 2 having a partial response. Three out of 22 patients had stable disease, for a disease control rate of 27.3%. Median progression-free survival was 2.53 months, and median overall survival was not yet reported. Eight out of 22 (36.4%) patients experienced grade 3/4 adverse events, most frequently in the form of hyperglycemia, mucositis, and hepatitis. One patient developed an unspecified grade 5 adverse event [42]. However, it should be noted that 7/22 patients did not receive radiation due to disease progression. Hence, the efficacy of RT combined with immunotherapy in PDAC is not currently established, though its potential for therapeutic benefit may become clear pending the results of these and other ongoing clinical trials.

4 -. CD40 AGONISTS

In addition to therapeutic inhibition of negative immune checkpoints, such as CTLA-4, PD-L1, and PD-1, select clinical trials have also explored agonists of positive co-stimulatory molecules, most notably CD40. CD40 is a TNF family receptor expressed on a variety of leukocytes. The association of CD40 with its ligands, particularly CD40L (also known as CD154), is paramount to T-cell priming and adaptive immune responses [43, 44]. As tumor-infiltrating T-cells are often highly dysfunctional and/or exhausted [45], several clinical trials have explored CD40 agonists as a means of reinvigorating immune surveillance in PDAC.

The initial trial exploring a CD40 agonist in PDAC was based on safety data from a multi-cancer Phase 1 trial reported in 2007 [46]. Twenty-one PDAC patients were administered the CD40 agonist monoclonal antibody CP-870,893 (0.1-0.2 mg/kg on day 3 of each 8-day cycle) in combination with gemcitabine (1000 mg/m² on days 1, 8, and 15 of each 28-day cycle). In this study, 4/21 (19%) patients had a partial response, 11 (52.4%) patients had stable disease, and 4 (19%) patients had progressive disease. Two patients were not evaluable; one due to a grade 4 cerebrovascular accident after one dose of CP-870,893, and the other due to clinical deterioration from disease progression. Median progression-free survival was 5.6 months, and the median overall survival 7.4 months. Overall, treatment was well tolerated, with the most common treatment-related adverse event being mild-to-moderate cytokine release syndrome, which was observed in 20/21 patients (95.2%), followed by fatigue (19/21 or 90.4%) and nausea (18/21 or 85.7%). However, in addition to the aforementioned grade 4 cerebrovascular accident, one patient also developed a grade 4 pulmonary embolism 5 weeks into therapy [47]. Final results of this study were published two years later, reporting an adjusted progression free survival of 5.2 months, and an overall survival of 8.4 months, with the same overall response rate of 19% [48].

A more recent Phase 1b study explored the more potent CD40 agonist APX005M in combination with gemcitabine and nab-paclitaxel with or without Nivolumab as first-line therapy in metastatic PDAC. The 30 patients in this study received gemcitabine and nab-paclitaxel per the standard dosing described above with APX005M (0.1 or 0.3 mg/kg), with or without Nivolumab (dose not described). Of the 24 evaluable patients, 13 (54%) experienced an adverse event leading to discontinuation, and 10 (42%) experienced a treatment-related serious adverse event. The authors noted 2 dose-limiting toxicities in the form of grade 3 and grade 4 febrile neutropenia. Four subjects died in this study, 2 due to progressive disease and 2 due to septic shock in the setting of neutropenia. Fourteen of the 24 patients (58%) showed a partial response, and 8 (33%) had stable disease. Though survival data was not yet reported, the 0.3 mg/kg APX005M dose was selected for the Phase 2 portion of the study in which the primary endpoint is one-year overall survival [49].

5 -. ICIs & TGFβ Inhibition

Though the reasons that PDAC tumors respond poorly to immunotherapy are multifaceted, recent evidence suggests that the highly immunosuppressive TME is a significant barrier to the efficacy of ICIs. In addition to the desmoplastic tumor stroma exerting mechanical stress on the intra-tumoral vasculature and lymphatics, the pancreatic TME is abundant in soluble immune suppressants. One such example is Transforming Growth Factor β (TGFβ), a potent and pleiotropic cytokine with established roles in benign and malignant cell processes [50, 51]. TGFβ plays a central role in both physiologic and pathologic immune tolerance. In the absence of TGFβ signals, T-cells rapidly acquire high levels of FasL and Granzyme B. Similarly, T-cells with constitutively active TGFβ signals remain refractory to full activation [51]. Accordingly, the combined inhibition of TGFβ and PD-L1/PD-1 shows early efficacy in several cancers [52]. This includes metastatic urothelial cancer [53], esophageal squamous cell carcinoma [54], and colon cancer [55], with emerging preclinical evidence in breast cancer [56].

Importantly, TGFβ has established roles in both fibrosis and immune evasion in PDAC, and the combination of TGFβ pathway inhibition and anti-PD-1 has shown significant preclinical efficacy in transgenic models of PDAC [57, 58]. Early trials exploring the combined inhibition of TGFβ and PD-L1/PD-1 in PDAC are ongoing. A Phase 1 trial in several solid cancers has examined the effects of M7824, a bifunctional fusion protein composed of a monoclonal antibody against PD-L1 fused to a TGFβ ligand trap. In the 3+3 dose-escalation component of this study, patients received M7824 at 1, 3, 10, or 20 mg/kg once every 2 weeks, with an additional cohort receiving an initial 0.3 mg/kg dose to evaluate pharmacokinetics, followed by 10 mg/kg. Of the 19 heavily pretreated cancer patients enrolled in the study, 4 patients (21%) developed grade 3 or greater adverse events in the form of skin infection secondary to localized bullous pemphigoid, increased lipase levels without pancreatitis, colitis with associated anemia, and gastroparesis with hypokalemia. Efficacy was seen across all treatment groups and the maximum tolerated dose was not reached. Of the 5 PDAC patients included in the study, only one patient with locally advanced MSI-H/dMMR PDAC showed a partial response at a dose of 3 mg/kg. This patient had a durable response that persisted until disease progression after 10.5 months [59].

Another Phase 1b study explored the combination of the Type 1 TGFβ receptor inhibitor Galunisertib (LY2157299) with Durvalumab in metastatic PDAC. Patients were given Durvalumab (1500mg every 28 days) with Galunisertib in 28-day cycles (14 days on and 14 days off). Galunisertib was tested at 4 doses during the dose-escalation portion of the study, including 50 mg once per day, 50 mg twice per day, 80 mg twice per day, or 150 mg twice per day. This study included 42 patients, the majority of which were heavily pretreated. Galunisertib was well tolerated at the highest dose of 150 mg twice per day, which was selected as the recommended Phase 2 dose. For the 32 patients receiving this dose, the authors observed one partial response (3%) and 7 patients with stable disease (22%) for a disease control rate of 25%. Median progression-free survival was 1.9 months, though the median overall survival was not yet reported. This combination had a relatively low rate of grade 3/4 adverse events, with 2 patients developing AST and GGT elevations, and one patient each developing ALT and alkaline phosphatase elevations or neutropenia [60].

Recent preclinical evidence suggests that the limited efficacy of concomitant TGFβ and PD-L1/PD-1 inhibition may be due to the limited presentation of tumor antigen, which was remedied by the addition of gemcitabine [61]. In an ongoing Phase 2 trial (N=156), Galunisertib (150 mg twice per day) was added to single-agent gemcitabine (standard 1000 mg/m² dosing), which extended overall survival from 7.2 to 10.9 months [62]. This is supported by other recent trials, also showing a significant survival benefit when Galunisertib is added to gemcitabine [63, 64]. Given the evidence supporting chemo-immunotherapy in PDAC, the addition of Galunisertib to such regimens warrants consideration in future clinical trials.

6 -. ICIs & OTHER TARGETED THERAPIES

Beyond the combination strategies described above, several clinical trials have explored the combination of ICIs and additional targeted therapies. For instance, 34 PDAC patients were included in the recent Phase 1/2 ECHO-203 study that combined the indoleamine 2,3-dioxygenase 1 (IDO1) inhibitor Epacadostat and Durvalumab. These patients were part of the Phase 1 component and received Epacadostat (25, 50, 75, 100, or 300 mg twice per day) and Durvalumab (3 or 10 mg/kg every 2 weeks). This approach was well tolerated, with grade ≥3 fatigue or rash observed in 3 patients each (8.8%), and 5 patients (14.7%) withdrawing from therapy due to treatment-related adverse effects in the form of pneumonitis, grade 2 diarrhea, grade 2 subarachnoid hemorrhage, grade 2 peripheral edema, and grade 3 dyspnea. The median duration of disease control was 156 days, and no objective responses were observed. Based on these data, the authors did not elect to conduct a Phase 2 expansion for PDAC [65].

Based on strong preclinical evidence supporting the benefits of FAK inhibition [66], a Phase 1 dose-escalation trial explored the combination of the FAK inhibitor Defactinib (200-400 mg twice per day) with Pembrolizumab (200 mg) and gemcitabine (500-1000 mg/m²). This study included 17 patients with advanced, treatment-refractory cancer, 6 of which had PDAC. No dose-limiting toxicities were observed, and this combination had a favorable safety profile with common adverse effects including fatigue (35%), nausea (29%), myalgia (29%), vomiting (24%), anorexia (24%), pruritus (24%), and fever (18%). Of the 13 evaluable patients, 7/13 (54%) showed stable disease, though no partial or complete responses were observed. The expansion cohort of only PDAC patients is ongoing [67].

As the desmoplastic tumor stroma is emerging as a barrier to the efficacy of ICIs, another ongoing study is evaluating the combination of PEGylated recombinant human hyaluronidase (PEGPH20) and Pembrolizumab. This study aims to treat patients who have failed on ≥2 lines of therapies and administer 3 μg/kg PEGPH20 once per week, and Pembrolizumab (200 mg) every 3 weeks [68]. Though survival data is not available, it is important to note that several groups have urged caution when targeting the tumor stroma due to its complex composition and the possibility that certain stromal cells may exert beneficial effects that restrain PDAC progression [69]. PEGPH20 acts on a specific feature of PDAC stroma, its richness in hyaluronic acid, which leads to increased interstitial pressure that compresses the tumor vasculature [70]. However, PEGPH20 can have serious side effects, including a relatively high incidence of thromboembolic events. Moreover, when the addition of PEGPH20 to FOLFIRINOX was recently evaluated in a Phase 1b/2 trial, median survival was reduced from 14.4 months in the FOLFIRINOX group to 7.7 months in the combination cohort [71]. PEGPH20 was also evaluated with gemcitabine and nab-paclitaxel in the HALO-301 study, though it failed to significantly improve outcomes and future development of PEGPH20 has been halted [72].

More encouraging are early results from the COMBAT Phase 2a trial exploring the combination of the C-X-C Motif Chemokine Receptor 4 (CXCR4) inhibitor BL-8040, Pembrolizumab, and cytotoxic chemotherapy consisting of liposomal Irinotecan and 5-fluorouracil (5-FU) administered with leucovorin (LV) in 22 patients. These patients received a 5-day course of BL-8040 monotherapy followed by liposomal Irinotecan/5-FU/LV every 2 weeks, Pembrolizumab every 3 weeks, and BL-8040 twice per week. Of the 15 evaluable patients, 2 (13.3%) were discontinued due to adverse effects, 4 (26.6%) showed a partial response, and 8 (53.3%) had stable disease [73].

7 -. OTHER IMMUNOTHERAPY APPROACHES

Though the majority of clinical trials exploring PDAC immunotherapy involve ICIs, several other treatments have also been attempted with varying degrees of success and/or toxicity.

7.1 -. GVAX & THERAPEUTIC VACCINES

Several studies have also explored the efficacy of a variety of therapeutic vaccines in PDAC, often in combination with ICIs, RT, or chemotherapy. Among the best-studied therapeutic vaccines in PDAC is GVAX. A 2001 Phase 1 trial was the first to explore GVAX in PDAC therapy. This study used irradiated and non-dividing, allogeneic pancreatic cancer cells (PCCs) secreting granulocyte-macrophage colony-stimulating factor (GM-CSF) that were administered as a therapeutic vaccine to 14 patients with Stage I, II, or III PDAC (73). Eight weeks after surgery, patients received an injection of between 1 x 10⁷ and 50 x 10⁷ vaccine cells. Subsequently, 12/14 patients received a 6-month course of adjuvant chemotherapy/RT. One month after completing adjuvant treatment, the 6 patients still in clinical remission received ≤3 additional doses of vaccine cells at the original dose. No dose-limiting toxicities were observed, and 3 patients who received ≥10 x 10⁷ vaccine cells appeared to have improved progression-free survival, remaining disease-free for a minimum of 25 months after diagnosis [74]. Of the 3 patients who showed post-vaccination delayed-type hypersensitivity responses, the authors identified consistent induction of CD8+ T-cell responses to multiple HLA-A2-, HLA-A3-, and HLA-A24-restricted epitopes of mesothelin, a tumor antigen upregulated in most PDAC tumors [75].

In light of evidence supporting the use of cyclophosphamide to enhance anti-tumor immune responses by depleting regulatory T-cells, the authors next explored the combination of cyclophosphamide and GVAX in an open label study of 50 patients. Thus, 30 patients were administered 2.5 × 10⁸ cells of two GM-CSF-secreting PCC lines (CG8020 and CG2505) at 21-day intervals to a maximum of 6 doses, and 20 patients received the same regimen preceded by 50 mg/m² intravenous cyclophosphamide. Both arms showed minimal toxicity, with a median overall survival of 2.3 months for patients receiving GVAX alone and 4.3 months for patients receiving GVAX and cyclophosphamide. Similar to the previous study, the authors also identified CD8+ T-cell responses to HLA Class I restricted mesothelin epitopes, predominantly in patients treated with both GVAX and cyclophosphamide [76].

GVAX was also evaluated in a Phase 2 trial that enrolled 60 PDAC patients after surgical resection. Here, patients were treated with 5 x 10⁸ GM-CSF-secreting PCCs (Panc 6.3 and Panc 10.05) starting 8-10 weeks after surgery followed by a combination of RT and 5-FU-based chemotherapy. Upon completing chemoradiation, disease-free patients received 2 to 4 GVAX treatments, each 1 month apart, with a fifth and final booster 6 months after the fourth GVAX treatment. This approach was well tolerated, demonstrated a median disease-free survival of 17.3 months, and a median overall survival of 24.8 months. Once again, the authors found a significant post-immunotherapy induction of mesothelin-specific CD8+ T-cells, which was associated with improved disease-free survival [77].

A subsequent Phase 1 trial sought to exploit the mesothelin-specific CD8+ T-cell phenomenon by administering a live-attenuated strain of Listeria monocytogenes modified to express human mesothelin (CRS-207). This approach was well tolerated up to 1 x 10⁹ cfu, and listeriolysin O and mesothelin-specific T-cell responses were detected in 37% of the subjects who lived ≥15 months [78]. These approaches were combined in a larger Phase 2 trial of pretreated patients with metastatic PDAC randomized into one of 3 arms. Patients received CRS-207 alone, single-agent chemotherapy, or a combination of cyclophosphamide, GVAX, and CRS-207. Of the 213 patients included in the primary cohort, the median overall survival for CRS-207 alone, chemotherapy alone and triple therapy were 5.4 months, 4.6 months, and 3.7 months, respectively. This study did not meet its primary endpoint, and the authors concluded that the combination of cyclophosphamide, GVAX, and CRS-207 did not improve survival compared to chemotherapy [79].

Other studies have combined GVAX with ICIs, including a seminal Phase 1b trial that evaluated Ipilimumab alone (10mg/kg induction doses administered every 3 weeks for a total of 4 doses followed by maintenance dosing every 12 weeks) or in combination with GVAX in 30 previously treated patients with advanced PDAC (79). In this study, GVAX consisted of 2 irradiated PCC lines (Panc 6.03 and Panc 10.05) that expressed GM-CSF, and 2.5 × 10⁸ of each were combined into a single vaccine and administered as intradermal injections at weeks 1, 4, 7, and 10. In both arms of the study the authors reported an objective response rate of 0%, with 2/15 (13.3%) patients in each arm having stable disease. However, despite the similar disease control rates, the GVAX/Ipilimumab arm had a median overall survival of 5.7 months compared to 3.6 months with Ipilimumab monotherapy. Consistent with previous reports, in patients with overall survival of >4.3 months, the authors report an increase in mesothelin-specific CD8+ T-cells. Similar to other trials, 20% of patients in each arm developed Grade 3/4 adverse events [80].

While GVAX may have limited efficacy as monotherapy or in combination with cyclophosphamide, it may warrant continued exploration in combination with ICIs and other treatment modalities. Several such trials are currently underway, including the Phase 2 trial NCT03161379 evaluating GVAX with cyclophosphamide, RT, and Nivolumab; NCT01595321 evaluating GM-CSF cell-based vaccines (GVAX), cyclophosphamide, RT, and FOLFIRINOX in surgically resected PDAC; and NCT02648282 evaluating GVAX with cyclophosphamide, RT, and Pembrolizumab in locally advanced PDAC.

Beyond GVAX, several other PDAC vaccines have been explored in clinical trials. These include a Phase 1/2 trial of a telomerase peptide vaccine (GV1001) in 38 patients with non-resectable PDAC [81]. Though this study showed immune responses in 25/38 patients (65.8%), the large international Phase 3 TeloVac trial of 1062 PDAC patients showed that the addition of GV1001 gemcitabine and capecitabine (standard dosing) failed to improve overall survival [82]. KRAS peptide vaccines have also been explored [83–86], as have gastrin-based vaccines [87], survivin-targeting vaccines [88, 89], HSP-peptide complex-based vaccines [90], viral CEA and MUC1 vaccines [91, 92], dendritic cell-based vaccines [93, 94], and the murine α-1,3-galactosyltransferase-expressing live, attenuated cell vaccine Algenpantucel-L [95]. While these have shown mixed results, several have shown early promise and may warrant continued exploration in larger trials [96].

Despite the marginal results observed with GVAX and other therapeutic vaccines, mesothelin remains a target of interest. For example, a tri-specific recombinant protein construct known as HPN536 is under development for mesothelin-expressing tumors, including PDAC. HPN536 binds to mesothelin, the CD3ε subunit of the T-cell receptor, and human serum albumin, which prolongs its half-life [97]. The subsequent T-cell activation and cytotoxicity against mesothelin-expressing tumor cells has demonstrated promising results in preclinical human cancer cell lines and animal models. The phase 1, first-in-human clinical trial of HPN536 in PDAC and other mesothelin-expressing tumors is currently ongoing [98].

7.2 -. CAR T-CELL THERAPY

Chimeric antigen receptor (CAR) T-cell therapy is emerging in the clinical management of several hematologic malignancies. As this approach shows early promise in solid tumors, select ongoing trials include patients with PDAC [99]. For example, NCT02850536 and NCT02349724 both involve using CEA-directed CAR T-cells to treat solid cancers, including PDAC. Other studies such as NCT00924287, NCT02713984, and NCT00889954 are evaluating CAR T-cells against HER2-Neu, some of which include patients with PDAC. Though these and other trials are ongoing and efficacy/toxicity data are not available, such approaches may offer benefit to PDAC patients, particularly given the apparent success of CAR T-cell therapies directed against mutant KRAS in colon cancer [100].

8 -. BARRIERS TO EFFICACY

In light of the above clinical trials, it is more important than ever to further delineate the mechanisms that contribute to the low clinical efficacy of immunotherapy in PDAC and advance our understanding of the pathophysiological processes that promote the suppression of cancer-directed immune mechanisms (Figure 1). Indeed, recent findings have offered at least a partial explanation as to why PDAC tumors typically respond poorly to conventional immunotherapy regimens and have led to the consideration of potential new strategies for overcoming resistance and improving immune therapy strategies. Select barriers to the efficacy of ICIs, both established and emerging, are briefly described below.

Figure 1. Barriers to the efficacy of immunotherapy within the pancreatic tumor microenvironment.

The pancreatic tumor microenvironment (TME) harbors several immunosuppressive cell types and cytokines that impede the therapeutic efficacy of various immunotherapy regimens. In addition to cytokines such as TGFβ and various CXCR4 ligands that are produced by a different cell types, the stroma comprises a diverse population of cancer-associated fibroblasts (CAFs) and pancreatic stellate cells (PSCs), which in addition to producing a variety of extracellular matrix (ECM) proteins, have emerging roles in regulating immune tolerance. The TME also contains several highly immunosuppressive leukocyte subsets, including regulatory T-cells (T-regs), myeloid-derived suppressor cells (MDSCs), tumor-associated macrophages (TAMs), and several species of bacteria, all of which cooperate to suppress the anti-cancer immune program.

8.1 -. LIMITED TUMOR ANTIGEN

Classically, PDAC is poorly immunogenic with a low mutational burden and limited presentation of tumor antigen [40], often displaying loss of HLA Class I expression [101]. This combination of events is a plausible reason for the relative failure of single-agent ICIs in PDAC patients, while explaining their efficacy in patients with MSI-H disease. As discussed, mismatch repair deficiency strongly predicts for response to anti-PD-1 in several solid tumors. The presumptive increase in mutational burden increases abnormal peptides and enhances antigen presentation, sensitizing patients to ICIs [14]. In preclinical models of PDAC, strategies to promote antigen presentation, such as irreversible electroporation, have helped overcome the innate resistance of PDAC to ICIs [102]. A similar approach is now under evaluation in the recently announced PANFIRE-3 trial (NCT04612530), which seeks to combine irreversible electroporation with Nivolumab and intratumoral administration of a Toll-like Receptor 9 Ligand (CpG) as a means of in vivo vaccination for metastatic PDAC. In addition to irreversible electroporation, cytotoxic chemotherapy has also been shown to increase MHC Class I expression, and cooperate with combination immunotherapy in vivo [61]. Therefore, further exploration of these and other strategies to enhance the presentation of self-peptide in PDAC is warranted and may augment responses to select immunotherapies.

8.2 -. AUTOPHAGY

In addition to a paucity of tumor antigen, the direct loss of HLA Class I is a known mechanism of immune escape in cancer, and a well-established mediator of resistance to ICI-based therapy [103–105]. While often caused by genetic events such as loss of heterozygosity or inactivating mutation, these events are rare in PDAC [106]. However, HLA Class I is frequently lost in pancreatic cancer, likely contributing to the innate resistance of PDAC to ICIs [107]. A recent study determined that MHC Class I molecules are selectively targeted for lysosomal degradation by an autophagy-dependent mechanism, involving the ubiquitin-binding receptor Neighbor of BRCA1 gene 1 (NBR1). The authors noted scant MHC Class I expression on the surface epithelium of tumor cells, and determined that MHC Class I molecules were concentrated in autophagosomes/lysosomes. Importantly, they demonstrated that genetic ablation or therapeutic inhibition of autophagy via chloroquine restored MHC Class I expression, and potentiated responses to ICIs in vivo [108]. Hence, repurposing agents such as chloroquine to mitigate the contribution of autophagy to the degradation of MHC Class I molecules may serve as a novel and clinically useful strategy to improve responses to ICI-based therapy. This approach warrants further study, particularly given the importance of autophagy in PDAC pathobiology [109].

8.3 -. IMMUNOSUPPRESSIVE TUMOR MICRENVIRONMENT

A key barrier to PDAC immunotherapy is its highly immunosuppressive TME. To understand the role of the TME in PDAC immune evasion, it is important to review the complex interactions between the various components of the PDAC TME. By the time a patient with PDAC presents clinically, the tumor is highly desmoplastic due to the proliferation of several different types of cancer-associated fibroblasts (CAFs) and the deposition of collagen, fibronectin, and hyaluronic acid in the extracellular matrix (ECM), resulting in a marked increase in interstitial pressures and ECM stiffness [110]. Since only 5% to 10% of the tumor mass consists of PCCs in advanced PDAC, these physical alterations constitute a physical and mechanical barrier that impedes drug delivery.

More importantly, PDACs are infiltrated with varying amounts of myeloid-derived suppressor cells (MDSCs), immune cells such as CD4+ T-cells, B-cells, tumor-associated macrophages (TAMs), neutrophils and other inflammatory cell types [111, 112]. MDSCs, which derive from immature myeloid cells, consume cysteine and downregulate L-selectin expression, thereby impeding T-cell activation due to the essential requirement of T-cells for cysteine and their dependence on L-selectin for efficient trafficking to lymph nodes. MDSCs can be abundant in PDAC due to its richness in MDSC chemoattractant molecules such as vascular endothelial growth factor (VEGF), interleukin 6 (IL-6), tumor necrosis factor-alpha, and GM-CSF [111, 113]. MDSCs also act to promote immune evasion by multiple additional mechanisms such as increasing the population of immunosuppressive M2 TAMs and FoxP3+ T regulatory cells (T-regs), upregulating their surface expression of PD-L1 which binds to PD-1 on the surface of T-cells to induce a state of T-cell anergy, and interfering with the beneficial actions of natural killer cells [114].

Recent data suggests the contribution of these leukocyte cell types to the immune landscape of PDAC tumors is extremely complex, which must be considered when translating strategies against these cell types to the clinic. For instance, despite the clinical association between increased T-regs and poor prognosis in PDAC [115], depletion of T-regs has recently been shown to accelerate tumor development in vivo [116]. Macrophages also seem to have contradictory roles in PDAC tumors. While select macrophage subpopulations may contribute to T-cell exclusion [117], other macrophage subtypes may restrain tumor development [118]. Hence, further dissecting the contributions of these and the many different cell types within the pancreatic tumor microenvironment may provide the opportunity for new therapeutic approaches.

PDAC is hypovascular and its blood vessels are compressed by the high interstitial pressures that combine with the abundance of collagen and hyaluronic acid to promote a TME that is dense, highly hypoxic, and acidified. Consequently, PDAC cells rely on alternative strategies to obtain their nutrients. These include autophagy and micropinocytosis, and altered metabolic cascades that lead to the generation of nutrients utilized by the PDAC cells, such as alanine and lipids derived from the surrounding activated pancreatic stellate cells (PSCs) that have transformed into CAFs [119].

It is now appreciated that are several different types of CAFs in PDAC and that their actions are very complex and often deleterious. CAFs originate from activated PSCs, but also derive from activated resident fibroblasts, from the recruitment of bone marrow-derived mesenchymal stem cells, and potentially from other cell types [120, 121]. CAFs nurture the PCCs and produce immunosuppressive cytokines such as TGFβ or IL-6 [120, 121]. IL-6-producing CAFs are also known as inflammatory CAFs [120, 121]. Moreover, the immunosuppressive pathways in any PDAC are context-dependent. They are modulated by the genetic makeup of a particular tumor [120–123], underscoring the complexity of the pathways that must be activated or suppressed in order to efficiently promote cancer-directed immune mechanisms. Nonetheless, recent advances in our understanding of these pathways and technologies that allow for precision medicine approaches suggest that the field is ready to devise strategies for effective immunotherapies in PDAC. For example, a very recent study demonstrated that targeting IL-20 can suppress PD-L1 expression and prolong survival in pre-clinical models of PDAC [124]. Similarly, the combination of antibody-mediated neutralization of IL-1 β and anti-PD1 was also effective in mice, enhancing the recruitment and effector function of cytotoxic T-lymphocytes [125], as was combined inhibition of IL-6 and PD-1 [126].

8.4 -. MICROBIOME

The microbiome is a newly emerging barrier to the efficacy of immunotherapy in several cancers [127–129]. Before recent findings, the pancreas had long been considered a sterile organ. However, a recent study identified an abundance of select bacterial species in both human and murine pancreatic cancer tissues [130]. Interestingly, the presence of Pseudoxanthomonas, Streptomyces, Saccharopolyspora, and Bacillus clausii predicts for long-term survival in PDAC patients [131], and the ablation of intratumoral bacteria enhanced the efficacy of immunotherapy in vivo. Following these observations, others have suggested that several commensal intestinal microbes may also be involved in PDAC.

The bacterial species that colonize the pancreas have a significant overlap with those of the gastrointestinal tract [131]. This is corroborated by evidence suggesting that bacteria from the gastrointestinal tract can access the pancreas via both the biliary tract and bloodstream [130, 132, 133]. More compelling is that the ablation of the gut microbiome appears to significantly reduce tumor burden in mouse models of PDAC, though this phenomenon was not observed in mice lacking functional T and B-cells [134]. Additionally, human-into-mice fecal microbiota transplantation modulates the tumor microbiome in vivo, altering tumor growth and local immune responses [131].

8.5 –. EPIGENETICS & EPITRANSCRIPTOMICS

Beyond the more established mechanisms described above, there are several still emerging and largely underappreciated mechanisms that may also contribute to immune escape and the corresponding failure of ICIs in pancreatic cancer. For instance, PDAC is associated with extensive dysregulation of the epigenome, which can occur due to a combination of genetic, environmental, and metabolic cues [135, 136]. These alterations, most notably regarding DNA methylation, have been demonstrated to reprogram local immune function and contribute to immune evasion in several cancer types [137]. For example, a recent study has demonstrated that sustained cell proliferation results in the eventual demethylation of late-replicating partial methylation domains, where several genes associated with immunomodulatory pathways are concentrated [137]. The authors concluded that global methylation loss was associated with immune evasion signatures, which was independent of tumor mutational burden and aneuploidy [137]. Pancreatic cancer has its own unique methylation signatures that may similarly contribute to immune escape and ICI resistance [138]. For instance, a recent study found that H3K4 trimethylation (H3K4me3) is enriched in the CD274 (PD-L1) promoter in PDAC cells. The authors determined that this phenomenon is mediated by the histone methyltransferase MLL1, and that MLL1 inhibition cooperates with anti-PD-L1 or anti-PD-1 in vivo [139]. Hence, MLL1 may be an attractive therapeutic target to improve the efficacy of ICIs in PDAC through reprogramming of the epigenome.

Beyond methylation, other epigenetic mechanisms may also facilitate immune evasion in PDAC, including those pertaining to post-translational histone modification [140]. This long standing hypothesis has resulted in new avenues of research exploring the intersection between epigenetic and immune checkpoint inhibitors [141]. Select histone-binding proteins have been suggested as potential drug targets in PDAC [142, 143], namely Bromodomain and Extra-Terminal motif (BET) proteins [144–147]. BET proteins such as Bromodomain-containing protein 2 (BRD2), BRD3, and BRD4 recognize acetylated lysine residues via their bromodomains, thereby acting as epigenetic scanners [148–150]. While classically associated with transcriptional activation of oncogenes such as c-MYC [145], BET proteins have been shown to regulate PD-L1 expression in a variety of cell types, including several within the pancreatic TME [151, 152]. Accordingly, though BET inhibitors such as JQ-1 have shown only modest preclinical efficacy in murine PDAC [144, 153], BET inhibitors seem to cooperate with ICIs and synergistically inhibit tumor growth in vivo [154], offering another potential strategy to improve therapeutic responses to immunotherapy in pancreatic cancer patients.

PDAC is also associated with myriad alterations to the epitranscriptome, which may also be linked to immune evasion and clinical resistance to ICIs. For instance, the reversible addition of N6-methyladenosine (m6A) to mRNA has important consequences regarding tumor cell proliferation and differentiation [155]. This modification is mediated by m6A methyltransferases, and recognized by several “reader” proteins that subsequently regulate several aspects of RNA metabolism including translation, splicing, export, degradation and microRNA processing [155]. In PDAC, an m6A-related mRNA signature is associated with poor clinical outcomes, and seems to correspond to alterations in select tumor infiltrating immune cell populations [156]. While the specific m6A methyltransferases responsible for this phenomenon are poorly described, one recent study has examined the role of methyltransferase-like 3 (METTL3) in PDAC [157]. The authors first determined that cells with loss of METTL3 are more sensitive to cell death induced by either chemotherapy or radiation [157]. They then found that the loss of METTL3 corresponds to several transcriptomic alterations, namely the upregulation of genes primarily associated with immune or interferon reactions [157]. Hence, the action of METTL3 and other m6A methyltransferases may serve as an underappreciated barrier to the efficacy of immunotherapy in PDAC.

Several additional epitranscriptomic mechanisms of immune evasion are also emerging. In a recent study using melanoma cells, the authors found that loss of the RNA-editing enzyme Double-stranded RNA-specific Adenosine Deaminase 1 (ADAR1) reduces A-to-I editing of interferon-inducible RNAs, leading to double-stranded RNA ligand sensing by Protein Kinase R (PKR) and Melanoma Differentiation-associated Protein 5 (MDA5). The authors state that this culminates in reduced tumor growth and inflammation, and helps overcome resistance of PD-1 inhibition caused by diminished antigen presentation by tumor cells [158]. Though this mechanism has yet to fully be explored in PDAC, there is mounting evidence that ADAR1 has important roles in PDAC pathobiology including regulating cell death and inflammation, as well as resistance to BET inhibition [159–161]. Hence, this warrants continued exploration in PDAC, particularly in the setting of ICIs.

8.6 -. HYPOXIA

As mentioned, PDAC is associated with a dense, desmoplastic stroma that exerts mechanical forces to compress the intratumoral vasculature [162, 163]. Consequently, PDAC tumors are generally hypoxic, particularly when compared to other tumor types. While we have briefly discussed the effects of the hypoxic TME on mediating epithelial-stromal cell interactions and how this feeds into local immune reprogramming, hypoxia also directly contributes to several other aspects of immune dysfunction in PDAC, including the evasion of cell-mediated cytotoxicity and resistance to immunotherapy [164]. In response to hypoxic conditions, PDAC cells produce increased levels of VEGF in a Hypoxia-Inducible Factor 1α (HIF1α) dependent manner [165]. Similarly, low oxygen levels lead to enhanced VEGF biosynthesis by PSCs [166]. While this combined increase in intratumoral VEGF is best known for enhancing pro-fibrogenic and pro-angiogenic responses, VEGF is also an important immunomodulator. Specifically, VEGF can inhibit the function of T-cells, enhance the recruitment of several immunosuppressive cell types, as well as hinder the differentiation/activation of dendritic cells [167]. VEGF has similarly been shown to modulate immune checkpoint expression on CD8+ T-cells, thereby contributing to a state of functional exhaustion [168]. In murine models of PDAC, increased concentrations of VEGF are associated with the accumulation of MDSCs and local immune suppression. However, the addition of phosphodiesterase-5 (PDE₅) inhibitor sildenafil reduced intratumoral VEGF levels, presumably due to improved profusion through the tumor vasculature and increased oxygen availability. Sildenafil also reduced intratumoral MDSCs and enhanced anti-tumor immune responses [169]. Thus, PDE₅ inhibitors such as sildenafil may warrant consideration as an adjuvant to immunotherapy regimens in PDAC, though this has yet to be explored.

Additional strategies to relax the tumor vasculature and improve profusion are also showing preclinical efficacy in PDAC. One such study explored the use of the Angiotensin Converting Enzyme (ACE) inhibitor losartan, which effectively reduced stromal collagen and hyaluronan production, and was associated with decreased expression of several pro-fibrotic signals including TGFβ1 [162]. Though the authors did not explore alterations in immune function, should ACE inhibitors lead to similar immunological changes as those observed with sildenafil, this may offer yet another potential strategy to target tumor-associated hypoxia in hopes of improving therapeutic responses to ICIs.

In addition to strategies to reverse hypoxia, other approaches targeting hypoxia-induced signals are now showing early promise in animal models, particularly those directly targeting VEGF. Though the combination of VEGF inhibition and ICIs are better explored in other cancer histologies [170, 171], a recent study evaluated antibody-mediated inhibition of both VEGF and PD-1 in murine models of pancreatic neuroendocrine tumors. This approach was highly effective in promoting a cytotoxic immune response, which was mediated by the enhanced formation of high endothelial venules [172]. Though such combinations have also yet to be explored in PDAC, as VEGF is itself clinically actionable, anti-VEGF agents such as bevacizumab may provide a benefit to PDAC patients when combined with select ICIs.

9 -. SUMMARY

There is currently no effective treatment for PDAC. Given the lack of early detection or chemoprevention strategies [173], most patients will present with advanced disease. Nearly all are administered palliative chemotherapy, though the majority will eventually progress on treatment. As immunotherapy has been highly effective in other difficult-to-treat cancers, investigators have been eager to explore such approaches in the treatment of PDAC. Though progress for immunotherapy in PDAC has been difficult, select combination strategies are beginning to show early promise in clinical trials. While these are mostly still under investigation, it is clear that there remain several barriers to efficacy that have yet to be overcome. Therefore, the success of future trials depends on an improved understanding of the complex and dynamic tumor microenvironment in hopes of identifying new strategies to reverse the innate resistance of PDAC to immunotherapy.

Highlights.

• Pancreatic ductal adenocarcinoma (PDAC) is a leading cause of cancer related death, with an overall five-year survival rate of less than 10%.

• Despite rapid progress in the treatment of many solid cancers, immunotherapy has yet to show clear efficacy in PDAC.

• Several clinical trials have explored immunotherapy in the treatment of PDAC. However, the results of these trials have been mostly disappointing, particularly regarding immune checkpoint inhibitors.

• Here, we provide a comprehensive summary of the key clinical trials exploring immunotherapy in PDAC, and a brief discussion of emerging mechanisms that may explain the innate resistance of PDAC tumors to immunotherapy.