Pancreatic Cancer and Immunotherapy: A Clinical Overview

Simple Summary Pancreatic ductal adenocarcinoma (PDAC) is an aggressive cancer with a dismal prognosis. While immunotherapy has been deemed a breakthrough treatment for various subtypes of cancer, its efficacy in PDAC is limited. This review discusses a wide range of immunotherapies, providing a general introduction to their working mechanism as well as current evidence on their clinical efficacy and immune eliciting abilities in PDAC. Utilizing combination (immuno)therapies to generate synergistic anti-tumor effects may provide the key to successful PDAC treatment. Abstract Pancreatic ductal adenocarcinoma (PDAC) is an aggressive disease with high mortality. The vast majority of patients present with unresectable, advanced stage disease, for whom standard of care chemo(radio)therapy may improve survival by several months. Immunotherapy has led to a fundamental shift in the treatment of several advanced cancers. However, its efficacy in PDAC in terms of clinical benefit is limited, possibly owing to the immunosuppressive, inaccessible tumor microenvironment. Still, various immunotherapies have demonstrated the capacity to initiate local and systemic immune responses, suggesting an immune potentiating effect. In this review, we address PDAC’s immunosuppressive tumor microenvironment and immune evasion methods and discuss a wide range of immunotherapies, including immunomodulators (i.e., immune checkpoint inhibitors, immune stimulatory agonists, cytokines and adjuvants), oncolytic viruses, adoptive cell therapies (i.e., T cells and natural killer cells) and cancer vaccines. We provide a general introduction to their working mechanism as well as evidence of their clinical efficacy and immune potentiating abilities in PDAC. The key to successful implementation of immunotherapy in this disease may rely on exploitation of synergistic effects between treatment combinations. Accordingly, future treatment approaches should aim to incorporate diverse and novel immunotherapeutic strategies coupled with cytotoxic drugs and/or local ablative treatment, targeting a wide array of tumor-induced immune escape mechanisms.


Introduction
Pancreatic ductal adenocarcinoma (PDAC) is an aggressive disease, with a dismal five-year overall survival (OS) rate of 6% [1]. Symptoms are often nonspecific with a late clinical onset, allowing the tumor to progress freely and silently. As a result, over 80% of patients present with unresectable locally advanced (LAPC) or metastatic PDAC (mPDAC) [2,3]. In these patients, survival can be moderately improved with palliativeintent systemic chemotherapy consisting of gemcitabine/nab-paclitaxel or FOLFIRINOX (folinic acid, 5-fluorouracil, irinotecan and oxaliplatin) [4,5].
Immunotherapy has led to a paradigm shift in the treatment of a subset of various solid cancers. However, in PDAC the currently available immunotherapies have only demonstrated marginal efficacy in terms of survival [6,7]. The treatment insensitivity of PDAC to immunotherapies can be attributed to the low mutational burden as well as the densely packed, inaccessible tumor microenvironment (TME) with fibrotic, hypoxic and immunosuppressive features, rendering the tumor immunologically 'cold' (Figure 1) [8].
Nonetheless, systemic and local immune responses have been reported in PDAC patients after treatment with immunotherapy, implying they can instigate an anti-tumor effect [9], albeit not potent enough. In this regard, combination with other immunotherapies, chemotherapy and/or local ablation may work synergistically. Certain cytotoxic drugs and ablative therapies have demonstrated the ability to sensitize the TME to immunotherapeutic agents by inducing immunogenic cell death, disrupting immune evasive mechanisms and reducing immune suppression [10,11]. Here, we provide an overview of currently available immunotherapies and their anti-tumor effects in PDAC in terms of immune activation and survival, with a focus on clinical research from the past decade. Additionally, we explore the potential of multimodal treatment strategies and illustrate how synergistic combinations may provide the key to successful PDAC therapy.

Mechanisms of Immune Evasion
The pancreatic TME is considered extremely immunosuppressive and generally lacking in immune infiltration ( Figure 1) [12]. Poor infiltration by T cells may in part be attributed to the low mutational burden of PDAC and consequent lack of neoantigens, which serve as immune targeting molecules and have the ability to elicit high-affinity T cell responses [13]. Further immune suppression and limited anti-tumor T cell infiltration is achieved through various mechanisms that enable immune evasion and immune exclusion from the TME. The TME comprises extracellular matrix (ECM), cancerous cells and non-cancerous cells including (mainly) pro-tumorigenic stromal, immune and endothelial cells. The cancerous cells release pro-tumorigenic chemokines, such as C-X-C motif chemokine ligand (CXCL)12, and pro-tumorigenic cytokines, including interleukin (IL)-6, IL-8, IL-10, transforming growth factor-beta (TGF-β), macrophage colony-stimulating factor (M-CSF) and vascular endothelial growth factor (VEGF) [14,15]. Abundant release of these molecules sways the immune balance from effective immune surveillance to immune tolerance by activating regulatory T cells (Tregs), tumor-associated macrophages (TAMs), T-helper2 (Th2) cells and myeloid-derived suppressor cells (MDSCs). This change in immune status contributes to tumor cell proliferation, migration and angiogenesis [16][17][18]. These immune cell subsets block anti-tumor activity of natural killer (NK) cells and effector CD4 + and CD8 + T cells. Additionally, maturation and survival of dendritic cells (DCs) is also impeded by these pro-tumorigenic cytokines [19,20]. DCs are powerful antigen presenting cells (APCs) essential in establishing a potent anti-tumor T cell response. Higher DC levels in blood and tumor tissue of PDAC patients have been correlated with improved survival [21][22][23]. Further immune evasion by the tumor is established through downregulation of the antigen presentation machinery, including major histocompatibility complex (MHC) class I [24], upregulation of inhibitory immune checkpoint ligands and increased apoptotic resistance through augmented expression of apoptotic regulatory proteins, signal transducer and activator of transcription (STAT)3 and B-cell lymphoma (BCL)-2 [19,25]. In addition to immune evasion, PDAC is able to physically exclude anti-tumor immune cells from its TME through activation of cancer-associated fibroblasts (CAFs) [12]. These CAFs promote fibrosis through collagen deposition and rearrangements of the ECM, leading to a desmoplastic reaction. The desmoplastic stroma, comprising up to 50-80% of tumor volume, creates a physical barrier preventing adequate vascularization, hindering infiltration of anti-tumor immune cells and resisting systemic treatments [26,27]. This barrier is present in primary as well as metastatic pancreatic tumors [28]. Collectively, these mechanisms contribute to an overall immune-inhibitory PDAC TME and explain the limited response to both chemo-and immunotherapy [29].

Regulation of the T Cell Response
To establish an anti-tumor immune response, both a priming phase and effector phase are required. In the priming phase, APCs present the acquired tumor antigens to T cells, initiating clonal expansion of antigen-specific T cells. The effector phase covers the direct anti-tumor T cell activity. T cell activation in both phases requires T cell receptor (TCR) complex ligation with antigen-loaded MHC class I (expressed on all nucleated cells) or class II (expressed on APCs). This is known as signal 1. However, signal 1 by itself is ineffective at T cell activation. Co-stimulation through specialized receptors is necessary to fully establish an anti-tumor T cell response, referred to as signal 2. The most critical co-stimulation is provided by binding of cluster of differentiation (CD)80 or CD86 on APCs with CD28 on T cells. Other co-stimulatory receptors include OX40, CD137 (4-1BB), CD27 and inducible T cell co-stimulator (ICOS) on T cells and CD40 on APCs or tumor cells. Their respective ligands are OX40L, CD137L, CD70 and ICOSL on APCs or tumor cells and CD40L on T cells (Table 1). Co-inhibitory receptors, or immune checkpoints, on T cells play an important role in the restriction of inflammatory responses by providing negative feedback upon T cell activation, thus overriding these activation signals. Co-inhibitory receptors include programmed death (PD)-1, cytotoxic T-lymphocyte-associated protein (CTLA)-4, T cell immunoglobulin and mucin domain-containing protein (TIM)-3, T cell immunoglobulin and ITIM domain (TIGIT) and lymphocyte-activation gene (LAG)-3 (Table 1) [30]. Their respective ligands include, but are not limited to, PD-L1/2, CD80/86, galectin (Gal)-9, CD155/112 and MHC-I/II. Tumor cells are able to abuse these regulatory mechanisms, upregulating inhibitory checkpoint ligands on their cellular membrane or mobilizing immune-suppressive myeloid cells with high expression levels of these checkpoint ligands and downregulating activating immune receptors on T cells, thereby avoiding recognition by patrolling immune cells [31]. Table 1. An overview of co-inhibitory (red) and co-stimulatory (green) receptor pairs.

Antigen Presenting Cell or Tumor Cell (Ligand) T Cell (Receptor)
MHC class I or II (antigen loaded)

Antigen Presenting Cell or Tumor Cell (Receptor) T Cell (Ligand)
CD40 CD40L

Immunotherapy
Immunotherapies can be categorized into immunomodulators (i.e., immune checkpoint inhibitors, immune stimulatory agonists, bispecific antibodies (not discussed in this review), cytokines and adjuvants), oncolytic viruses, adoptive cell therapies (i.e., T cells and NK cells) and cancer vaccines, each with a unique mechanism of action. Below, various immunotherapies are addressed within the context of PDAC. Of note, only a selection of clinical trials listed in the tables will be discussed more in-depth in the text.  [32]. Moreover, signaling pathways can interfere with T cell activation through signaling interference downstream from the TCR complex and CD28 [33,34] and promote apoptosis [35,36]. The PD-1 pathway can be blocked by immune checkpoint inhibitors anti-PD-1 or anti-PD-L1. If PD-L1 is overexpressed, although uncommon in PDAC [37], it inversely correlates with CD8 + TIL frequency and clinical prognosis [38][39][40][41]. Despite encouraging results of anti-PD-1 and anti-PD-L1 therapy in patients with other solid malignancies [42,43], results in patients with PDAC are less impressive. To date, various clinical trials have described the effects of anti-PD-1/anti-PD-L1 in PDAC as monotherapy [44][45][46] or in combination with other systemic therapies [46][47][48][49][50][51][52] (Table 2). The most promising outcomes were reported in a cohort of mPDAC patients (n = 17), partially chemotherapynaïve (n = 11), who received anti-PD-1 concurrently with a gemcitabine/nab-paclitaxel regimen, resulting in a median progression-free survival (mPFS) of 9.1 months and median overall survival (mOS) of 15 months [48]. As reference, mOS outcomes of 8.5 and 11.1 months may be achieved with gemcitabine/nab-paclitaxel and FOLFIRINOX, respectively [4,5]. Recently, results from the phase 2a COMBAT trial suggested that chemotherapy concurrent with combined anti-PD-1 and anti-C-X-C motif chemokine receptor (CXCR)4 may augment chemotherapeutic effects after attaining an mOS of 7.8 months from the start of immunotherapy in pre-treated mPDAC patients [49]. Results from both trials are in line with the expanding recognition that cytotoxic drugs (i.e., chemotherapy) can enhance immunotherapeutic effects by stimulating immunogenic tumor cell death, reducing tumor-induced immune suppression and increasing effector T cell function and infiltration [53][54][55]. Definitive advice regarding combined treatment with chemotherapy (gemcitabine/paclitaxel or FOLFIRINOX) and anti-PD-1 may be provided by two large RCTs (ClinicalTrials.gov Identifier: NCT04674956; NCT03983057). In contrast, combining anti-PD-L1 with the Bruton's tyrosine kinase inhibitor ibrutinib, which targets various oncogenic driver pathways, in pre-treated mPDAC patients (n = 48) showed limited anti-tumor activity (mOS 4.2 months, mPFS 1.7 months) [52]. . This process can be reversed by using anti-PD-1 or anti-PD-L1 antibodies, allowing CD8 + T cell-induced killing of tumor cells. Of note, recent studies also point to decreases in suppressive Treg rates and decreased T cell priming by CD28 signaling interference upon CTLA-4 and PD-1 blockade, respectively.  Randomization: (1) − (2) + GVAX 30 Previously treated LAPC/mPDAC (1) 3.6 (t) (2) 5.7 (t)

Anti-TIM-3
Dysfunctional and (terminally) exhausted CD4 + and CD8 + T cells as well as tumoral DCs are characterized by overexpression of TIM-3 [69,70], which is likely related to tumor cell invasion, metastatic processes and recurrence [71]. TIM-3 has multiple ligands, including Gal-9, phosphatidylserine (PtdSer), carcinoembryonic antigen-related cell adhesion molecule 1 (CEACAM1) and high mobility group protein B1 (HMGB1), all of which are implicated in its immune inhibitory functions [70,[72][73][74]. Ligation of Gal-9, the first known and most studied TIM-3 ligand, with TIM-3 on T cells and NK cells leads to diminished activation and, hence, inhibited anti-tumor activity. Expression of Gal-9 is increased in PDAC tissue, both on tumor and immune cells, as well as in blood of PDAC patients when compared to healthy pancreas tissue and healthy individuals, respectively [75]. There are no (upcoming) data from (pre-)clinical research on the therapeutic effects of anti-TIM-3 in PDAC.

Anti-TIGIT
Activated CD4 + and CD8 + T cells, Tregs and NK cells express TIGIT. Ligation of TIGIT with its ligands CD155 and CD112, expressed on tumor and myeloid cells, promotes NK and T cell tolerance [76,77]. There is abundant expression of CD155 in tumor tissue (primarily tumor cells) of PDAC patients, with expression levels inversely correlated with TIL frequency and survival [78]. TIGIT competes with its opposing receptor CD226 for its ligands, the latter receptor enforcing anti-tumor responses through NK and T cell activation. Accordingly, it was shown that anti-TIGIT treatment selectively affects CD226 hi CD8 + T cells in a pre-clinical setting [79]. Modified FOLFIRINOX treatment in mPDAC patients increased the proportion of CD226 hi CD8 + T cells, implying that this chemotherapeutic combination may increase tumor sensitivity to anti-TIGIT treatment. A phase 1b/2 randomized trial is currently using anti-TIGIT combined with gemcitabine/nabpaclitaxel in mPDAC patients (ClinicalTrials.gov Identifier: NCT03193190).

Anti-LAG-3
LAG-3 potentiates inhibitory signals in T cells upon TCR/MHC interaction. In addition, it is known for its synergism with PD-1. While anti-LAG-3 has shown little effect as monotherapy, dual anti-LAG-3/anti-PD-1 treatment exerted potent anti-tumor effects in mice with melanoma and colon carcinoma [80]. Consequently, a bispecific antibody targeting LAG-3 and PD-1 was developed and tested in a murine pancreatic tumor model [81]. Initial results demonstrated that this antibody targeted CD4 + T cells, increasing their effector functions by releasing cytolytic granzyme B and interferon (IFN)-γ, followed by complete suppression of tumor growth. Clinical trials have yet to determine its clinical value and none are currently ongoing for PDAC.

MHC Class II Agonist
Eftilagimod alpha, or IMP321, is an MHC class II agonist that triggers maturation of APCs followed by activation of CD8 + T cells. To date, one phase 1 clinical trial has explored the safety profile of IMP321 combined with gemcitabine in treatment-naïve advanced PDAC patients (Table 3) [82]. A total of 18 patients were enrolled, of whom none reported serious adverse events. Overall, combination treatment with IMP321 and gemcitabine was well-tolerated. No significant differences were noted in pre-and post-treatment serum levels of monocytes (CD11b + CD14 + ), conventional DCs (CD11c + ) and T cell subsets (CD4 + or CD8 + ). From the lack of immunological response, the investigators concluded that higher doses of IMP321 may be required for future clinical trials. (1) 7 (t) (2) 4.4 (t) NR = not reported; i.t. = intratumoral; gem = gemcitabine; LAPC = locally advanced pancreatic cancer; mPDAC = metastatic pancreatic ductal adenocarcinoma; mPFS = median progression-free survival; mOS = median overall survival; t = from treatment/registration; d = from diagnosis; x = unknown whether from diagnosis or treatment/registration.

OX40 Agonist
OX40-OX40L ligation enhances effector functions, memory formation and survival of CD4 + and CD8 + T cells [87]. Pre-clinical research in a PDAC mouse model demonstrated that OX40 agonists significantly improved survival compared to anti-PD-1 monotherapy [88]. Moreover, when the OX40 agonists and anti-PD-1 were combined the proportion of Tregs and exhausted T cells decreased, whereas memory CD4 + and CD8 + T cell numbers increased. While OX40 agonists have demonstrated their potent immune-stimulating capacities in other advanced cancer patients [89], clinical data for PDAC is lacking. A phase 1b/2 trial is currently recruiting in which an OX40 agonist will be combined with a toll-like receptor (TLR)-9 ligand for, among other things, patients with mPDAC (ClinicalTrials.gov Identifier: NCT04387071).

CD137 (4-1BB) Agonist
CD137, or 4-1BB, is expressed on primed T and NK cells. CD137-CD137L ligation protects antigen-specific T cells from undergoing apoptosis and promotes their effector functions and differentiation into memory T cells [90]. Pre-clinical research, in which CD137 agonists were evaluated in murine PDAC models, provided promising results in terms of survival and immune stimulation [91][92][93]. Clinical reports on the use of CD137 agonists in PDAC patients are limited to a single phase 1 clinical trial, in which an adoptive T cell strategy was employed (discussed in "Adoptive Cell Therapies") [94].

ICOS Agonist
The ICOS pathway plays a dual role in cancer immunology [95,96]. Activation of this pathway has been shown to induce immunosuppression through CD4 + T cell subsets including Tregs. On the other hand, it can also unleash Th1 cell-mediated immunity and has demonstrated the ability to enhance the anti-tumor efficacy of GVAX [97,98]. Besides initial pre-clinical results in a PDAC model, in which combined co-stimulation with ICOS and CD137 has shown to increase the persistence of chimeric antigen receptor (CAR) T cells (CAR T cell therapy will be discussed in "Adoptive cell therapies") [92], no clinical trials have yet been published.

CD40 Agonist
CD40 is a co-stimulatory receptor found on APCs rather than on T cells. Its ligand, CD40L, is present primarily on activated CD4 + T cells [99]. CD40 ligation on DCs induces positive signaling that leads to their maturation and the release of IL-12, resulting in antitumor activity through T cell proliferation and differentiation towards a Th1 dominant state. Remarkably, one study in an in vivo PDAC model showed this efficacious antitumor activity to be directly mediated by CD40-expressing macrophages rather than T cells [83]. Since PDAC has a high macrophage content this triggered high interest in the clinical application of agonistic CD40 antibodies in PDAC. CD40 agonists, in combination with mesothelioma-lysate loaded DCs [100] or IL-15 [101], have demonstrated promising pre-clinical results in PDAC models. Beatty et al. [83,84] combined a CD40 agonist with gemcitabine in 22 chemotherapy-naïve advanced PDAC patients (Table 3). Results from this phase 1 trial revealed an mOS and mPFS of 8.5 and 5.2 months, respectively. While this treatment did not improve survival compared to gemcitabine/nab-paclitaxel, the immune system was activated, evidenced by the increase in inflammatory cytokines, a boost in B-cell expansion of co-stimulatory molecules and overall depletion of B cells. In their multi-cohort phase 1b trial, O'Hara et al. [85] utilized a similar regimen ± anti-PD-1 in synchronous and metachronous mPDAC patients (allowed prior chemoradiotherapy). They found encouraging initial mOS (12.7-20.1 months) and mPFS (10.8-12.5 months) outcomes, leading to the ongoing randomized phase 2 portion of the study.

CD27 Agonist/Anti-CD70
Activation of the co-stimulatory receptor CD27 promotes survival and differentiation of T cells towards effector and memory subtypes [102]. Hence, an agonistic CD27 antibody was tested in patients with (non-pancreatic) solid malignancies, resulting in biological and clinical activity [103]. No trials are currently ongoing that employ a CD27 agonist in PDAC patients. The CD27 ligand CD70 is generally only expressed on activated immune cells as it is regulated by the presence of antigens. However, aberrant CD70 expression has been noted in tumor tissue of 25% of PDAC patients [104]. Importantly, although the CD27-CD70 axis is known for its immune stimulatory effects, CD70-expressing (pancreatic) cancer cells may also instigate immune tolerance by increasing the frequency of activated Tregs [105]. Since normal tissues have negligible expression, CD70 may present as a therapeutic target.
Anti-CD70 antibodies have demonstrated potent anti-tumor activity in a murine PDAC model overexpressing CD70 [104] and good tolerability and preliminary anti-tumor activity in various CD70 positive advanced (non-pancreatic) solid cancers [106,107]. However, no clinical trials are currently recruiting CD70-overexpressing PDAC patients.

Adjuvants
In immunotherapy, adjuvants ( Figure 3) usually refer to products that are added to vaccines in order to modulate or increase an immune response against the antigens contained within them [118]. However, these adjuvant compounds may also be used as monotherapy or as a supplement to other types of cancer treatment. They target the priming phase or the effector phase but can also be utilized as immune modulators to condition the microenvironment of both tumors and their draining lymph nodes in order to support both phases. Adjuvants bind onto pattern recognition receptors (PRRs), including toll-like receptors (TLRs), stimulator of IFN genes (STING) and NOD-like receptors (NLRs), which are present on epithelial (cancer) cells and innate immune cells, including DCs, and initiate immune responses against pathogens or tumor cells in secondary lymphoid organs [119]. Moreover, in tumors they may trigger release of chemokines and cytokines that can attract T cells and drive their effector response. Adjuvants can be delivered systemically, locally through vaccination or by peri-or intratumoral injection. A considerable advantage of intratumoral adjuvant administration over systemic approaches is that lower dosages will suffice, limiting severe toxicities [120,121]. Moreover, peritumoral administration has demonstrated superior anti-tumor efficacy in terms of DC and tumor-specific CD8 + T cell activation and long-lasting tumor protection in mice when compared to intravenous or intradermal administration [122].  [123,124]. However, TLR expression has also been observed in pancreatic tumor cells and, likely context-dependent, has been linked to both pro-and anti-tumor effects [125]. High cytoplasmic expression of various TLRs in tumor cells positively correlated with a favorable prognosis in PDAC patients [126][127][128]. Conversely, TLR ligation in pancreatic cancer cells has also been linked to tumorigenesis through inflammatory responses that stimulate their anti-apoptotic properties and angiogenesis [129][130][131]. Pre-clinical evidence in PDAC models demonstrated encouraging results of TLR agonist monotherapy [132][133][134] or combined with chemotherapy [135], local tumor eradication (i.e., radiotherapy or ablation) and/or other immunotherapies [136][137][138]. Initial clinical results in PDAC patients are also encouraging. Dalgleish et al. [86] published data of a randomized phase 2 trial in which they employed intradermally injected IMM-101, a TLR-2/1 agonist, in combination with gemcitabine (n = 75) and compared clinical efficacy to gemcitabine (n = 35) only in 110 LAPC and mPDAC patients (Table 3). A significant (p = 0.01) improvement in survival was noted for mPDAC patients in the IMM-101 + gemcitabine group (mOS 7 months) compared to gemcitabine monotherapy (mOS 4.4 months). Two clinical trials are currently investigating an intratumoral TLR-9 agonist for mPDAC, both in combination with anti-PD-1 and a local form of tumor eradication (ClinicalTrials.gov Identifier: NCT04612530 (PANFIRE-III trial [139]); NCT04050085).

Stimulator of Interferon Genes (STING) Agonists
Cyclic GMP-AMP synthase (cGAS)-STING is an intracellular sensor pathway in innate immune cells that detects cytosolic double stranded DNA fragments [140]. Similar to  [123,124]. However, TLR expression has also been observed in pancreatic tumor cells and, likely context-dependent, has been linked to both pro-and anti-tumor effects [125]. High cytoplasmic expression of various TLRs in tumor cells positively correlated with a favorable prognosis in PDAC patients [126][127][128]. Conversely, TLR ligation in pancreatic cancer cells has also been linked to tumorigenesis through inflammatory responses that stimulate their anti-apoptotic properties and angiogenesis [129][130][131]. Pre-clinical evidence in PDAC models demonstrated encouraging results of TLR agonist monotherapy [132][133][134] or combined with chemotherapy [135], local tumor eradication (i.e., radiotherapy or ablation) and/or other immunotherapies [136][137][138]. Initial clinical results in PDAC patients are also encouraging. Dalgleish et al. [86] published data of a randomized phase 2 trial in which they employed intradermally injected IMM-101, a TLR-2/1 agonist, in combination with gemcitabine (n = 75) and compared clinical efficacy to gemcitabine (n = 35) only in 110 LAPC and mPDAC patients (Table 3). A significant (p = 0.01) improvement in survival was noted for mPDAC patients in the IMM-101 + gemcitabine group (mOS 7 months) compared to gemcitabine monotherapy (mOS 4.4 months). Two clinical trials are currently investigating an intratumoral TLR-9 agonist for mPDAC, both in combination with anti-PD-1 and a local form of tumor eradication (ClinicalTrials.gov Identifier: NCT04612530 (PANFIRE-III trial [139]); NCT04050085).

Stimulator of Interferon Genes (STING) Agonists
Cyclic GMP-AMP synthase (cGAS)-STING is an intracellular sensor pathway in innate immune cells that detects cytosolic double stranded DNA fragments [140]. Similar to TLR activation, stimulation of the cGAS-STING pathway results in the production of type 1 IFN and other pro-inflammatory cytokines, instigating activation of effector T cells and NK cells, and maturation of DCs (Figure 3). In pre-clinical PDAC mouse models, STING agonists have demonstrated potent anti-tumor efficacy by increasing cytotoxic T cell activity and decreasing Treg levels [141]. Moreover, STING activation augmented costimulatory receptor expression on DCs and converted immune-suppressive macrophages into immune-activating subtypes. Combined with other immunotherapies [138,142,143], or radiotherapy [144], STING agonists in PDAC murine models have demonstrated the ability to induce durable tumor regression and to improve survival. Nonetheless, two initial clinical trials (ClinicalTrials.gov Identifier: NCT03010176; NCT03172936) that employed respectively intratumoral and intravenous STING agonists in patients with solid cancers and lymphomas failed to demonstrate favorable results in terms of tumor regression [145]. The ongoing clinical trial ClinicalTrials.gov Identifier: NCT04144140 is administering intratumoral STING agonists in various tumor types.

NOD-Like Receptor (NLR) Agonists
NLRs can cooperate with TLRs and are able to detect pathogen-associated molecular patterns (PAMPs) and danger or stress signals, thereby regulating inflammatory and apoptotic responses [146]. However, the therapeutic potential of targeting these pathways in PDAC has not yet been explored.

Oncolytic Viruses
Oncolytic viral therapy utilizes the destructive capacity of a virus, specifically targeted at tumor cells whilst normal healthy cells remain unaffected (Figure 4). Activated oncogenic pathways and a defective IFN response (normally limiting viral replication), both specific to tumor cells, allow for selective replication of the virus, leading to production of abundant novel viral particles and eventual tumor cell lysis. The newly produced oncolytic viruses, virus-derived PAMPs, DAMPs and tumor antigens are released into the TME, infecting other tumor cells and activating DCs followed by T cell priming in the draining lymph nodes (i.e., in vivo vaccination). Subsequently, effector T cells are attracted to the tumor site via a gradient of chemokines such as CXCL9 and CXCL10 [147,148]. A major benefit of oncolytic viruses is their genetically modifiable genome, allowing for incorporation of immune modulator transgenes ("armed oncolytic viruses"), which are also released by infected tumor cells. Such immunological arming may be aimed at decreasing immune suppression in the TME and/or increasing immune activation.
In murine PDAC models, oncolytic viruses have demonstrated the capacity to reduce tumor burden and prolong survival by downregulating TAMs and increasing infiltration and function of Th1 CD4 + cells and CD8 + T cells [149,150]. A few clinical trials have utilized oncolytic viruses in PDAC (Table 4) [151][152][153][154][155]. Noonan et al. [154] reported data of their RCT, in which patients received intravenous pelareorep (an oncolytic reovirus) combined with carboplatin/paclitaxel (n = 36) or carboplatin/paclitaxel alone (n = 37). The primary outcome, mPFS, did not significantly differ between the treatment groups (4.9 vs. 5.2 months, p = 0.6). However, addition of pelareorep did result in demonstrable immune responses with increased systemic levels of pro-inflammatory cytokines, Th1 CD4 + and CD8 + T cells, yet also enhanced Treg numbers. Using the same oncovirus in combination with anti-PD-1 and chemotherapy, Mahalingam et al. [50] showed the generation of new In murine PDAC models, oncolytic viruses have demonstrated the capacity to reduce tumor burden and prolong survival by downregulating TAMs and increasing infiltration and function of Th1 CD4 + cells and CD8 + T cells [149,150]. A few clinical trials have utilized oncolytic viruses in PDAC (Table 4) [151][152][153][154][155]. Noonan et al. [154] reported data of their RCT, in which patients received intravenous pelareorep (an oncolytic reovirus) combined with carboplatin/paclitaxel (n = 36) or carboplatin/paclitaxel alone (n = 37). The primary outcome, mPFS, did not significantly differ between the treatment groups (4.9 vs. 5.2 months, p = 0.6). However, addition of pelareorep did result in demonstrable immune responses with increased systemic levels of pro-inflammatory cytokines, Th1 CD4 + and CD8 + T cells, yet also enhanced Treg numbers. Using the same oncovirus in combination with anti-PD-1 and chemotherapy, Mahalingam et al. [50] showed the generation of new T cell clones, recirculating systemically and transcriptional evidence of systemic immune activation, both associated with clinical benefit. This suggests the priming and mobilization of new T cell clones, providing systemic protection and possibly setting the stage for subsequent successful immune checkpoint blockade. Hirooka et al. [155] utilized intratumoral HF-10, a natural oncolytic HSV-1 virus, in combination with erlotinib (anti-EGFR antibody) and gemcitabine in 10 LAPC patients, achieving an mOS of 15.5 months. Multiple clinical trials incorporating an oncolytic virus in PDAC patients are currently recruiting (ClinicalTrials.gov Identifier: NCT02705196; NCT04637698; NCT03252808). Several . Oncolytic viral therapy. Oncolytic viruses may be equipped with a transgene (e.g., immune checkpoint inhibitor (ICI) or cytokine), creating an armed oncolytic virus. Oncolytic viruses can be administered intravenously or intratumorally and will infect both healthy cells and tumor cells. In the former, viruses are cleared, whereas in the latter, due to activation of oncogenic pathways and a defective interferon (IFN) response, oncolytic viruses thrive, leading to production of novel viral particles to a point where the cell lyses due to viral overload. Newly created oncolytic viruses, pathogen-associated molecular patterns (PAMPs), damage-associated molecular patterns (DAMPs) and tumor antigens are released into the microenvironment. New waves of released oncolytic viruses can then infect other tumor cells. Dendritic cells (DCs) take up, process and present the released antigens and upon activation by the released DAMPs and PAMPs via their pattern recognition receptors (PRRs), will transport them to the draining lymph nodes where T cell priming takes place. Primed effector T cells (Th1 CD4 + , CD8 + ) subsequently provide systemic immune surveillance and will migrate towards the original tumor site (but also to distant, uninjected metastatic sites) via a chemokine gradient, often provided by activated DCs.

Adoptive Cell Therapies
Adoptive cell therapies utilize autologous or allogeneic immune effector cells, such as T cells or NK cells, to eradicate cancer ( Figure 5). After harvesting, either from the patient's tumor or from blood, the immune effector cells are selected and their functionality may be improved in vitro-e.g., through genetic modification to express a chimeric antigen receptor (CAR) that targets a specific protein or a TCR to specifically recognize a peptide/MHC complex-followed by re-administration back into the patient. The application of CAR-engineered cells in solid tumors is still in an early developmental stage and requires optimization to achieve long-term persistence post-transfer and sufficient infiltration into the tumor fields.

Adoptive Cell Therapies
Adoptive cell therapies utilize autologous or allogeneic immune effector cells, such as T cells or NK cells, to eradicate cancer ( Figure 5). After harvesting, either from the patient's tumor or from blood, the immune effector cells are selected and their functionality may be improved in vitro-e.g., through genetic modification to express a chimeric antigen receptor (CAR) that targets a specific protein or a TCR to specifically recognize a peptide/MHC complex-followed by re-administration back into the patient. The application of CAR-engineered cells in solid tumors is still in an early developmental stage and requires optimization to achieve long-term persistence post-transfer and sufficient infiltration into the tumor fields.  Similarly, autologous tumor tissue can be collected to isolate tumor-infiltrating lymphocytes (TILs), generally T cells. These isolated cells are cultured in vitro and can be selected for expansion based on their anti-tumoral capacity or enhanced with T cell receptors (TCRs) or chimeric antigen receptors (CARs). Following in vitro manipulation and stimulation and, in case of systemic administration, subsequent to a lymphodepleting chemo(radio)therapy regimen, these cells are injected intratumorally (i.t.) or intravenously (i.v.) to (systemically) detect and kill cancer cells.
A pre-clinical study in a murine PDAC model has yielded promising data after treatment with mesothelin-directed CAR T cells combined with armed oncolytic viruses expressing IL-2 and TNF-α [150]. Clinically (Table 5), CAR T cell therapy has not yet been widely explored for PDAC, but initial results, using mesothelin-directed (± CD137 and CD3ζ) CAR T cells, showed limited activity [94,157]. Aoki et al. [158] utilized Vγ9Vδ2 T cells, a subset of T cells that secrete Th1-related cytokines and have the capacity to exert potent cytotoxicity against tumor cells through MHC-unrestricted recognition of phosphoantigens. After surgical resection of the primary pancreatic tumor, patients (n = 40) received gemcitabine as adjuvant therapy, with (n = 23) or without (n = 17) Vγ9Vδ2 T cell infusion. This supplemental therapy was unable to prolong mPFS or mOS, but patients with >15% peripheral Vγ9Vδ2 T cells after two injections had a more favorable outcome. Conversely, Kumai et al. [159] recently demonstrated that αβ T cell therapy ± chemotherapy resulted in an mOS of 11.3 months from start of immunotherapy and 18.7 months from diagnosis, indicating a survival benefit compared to previous reports in a similar patient population receiving chemotherapy only (6.8-11.1 months from diagnosis) [4,5]. In addition, increased levels of CD8 + T cells were found in peripheral blood samples. Several clinical trials are currently ongoing to further explore CAR T cell potential in PDAC (ClinicalTrials.gov Identifier: NCT04037241; NCT03638193; NCT03323944) and TIL therapy (ClinicalTrials.gov Identifier: NCT03935893; NCT04426669). A trial employing NK cells combined with ablation will be discussed in 'Immunotherapy + IRE'. Several NK cell therapy trials are recruiting (ClinicalTrials.gov Identifier: NCT03634501; NCT03093688) or have been completed without publication (ClinicalTrials.gov Identifier: NCT03008304).

Tumor Vaccines
Tumor vaccines clinically pursued in the treatment of PDAC mainly include wholetumor-cell, peptide-based and peptide-pulsed DC vaccines ( Figure 6). They may contain one or several tumor-associated antigens (TAAs) or tumor-specific antigens (TSAs). TAAs are proteins overexpressed or aberrantly expressed in tumor cells but are also expressed to some extent on non-cancerous tissues, hence making them non-specific. TAA vaccines include, but are not limited to, those containing wilms tumor (WT)-1, kinesin family member (KIF)20A, VEGFR1/2, survivin, mucin (MUC)-1, mesothelin and human telomerase reverse transcriptase (hTERT) ( Table 6) [160]. Mutated TSAs, or neoantigens, are expressed exclusively on tumor cells. Neoantigen vaccines are highly specific, immunogenic and avoid self-antigen-induced T cell tolerance [161]. However, their working mechanism relies on the mutational burden of a tumor, which is limited in PDAC [37]. In other words, these vaccines may not be able to kill (subgroups of) tumor cells lacking expression of that neoantigen [13]. Neoantigen targets, an example of which are kirsten rat sarcoma (KRAS) mutations, may be identified through mutational analyses. Vaccines can be ad-ministered intratumorally, intranodally, intradermally, intravenously, subcutaneously and intramuscularly [162]. intratumorally, intranodally, intradermally, intravenously, subcutaneously and intramuscularly [162]. Figure 6. Cancer vaccines and in vivo vaccination following chemotherapy, ablation or oncolytic viral therapy. Cancer vaccines: Peptide-based vaccines can be produced on the basis of a patient's mutational profile (i.e., KRAS (or other) neoantigens) or using (a combination of) generic tumor-associated antigens (TAAs). Whole-tumor-cell vaccines can be generated with autologous tumor lysate or using (a combination of) allogeneic PDAC cell lines. Peptide vaccines and whole-tumor-cell vaccines may be supplemented with cytokines, such as granulocyte macrophage colony-stimulating factor (GM-CSF), or adjuvants, which attracts and activates dendritic cells (DCs). For DC vaccines, autologous or allogeneic DCs are generated in vitro from patient or donor derived precursors in blood (e.g., monocytes), followed by antigen pulsing. Cancer vaccines may be administered intratumorally, intravenously, intranodally or subcutaneously, intradermally or intramuscularly. In vivo vaccination: Chemotherapy, oncolytic viruses and local ablative treatments (e.g., stereotactic ablative body radiotherapy (SABR) and irreversible electroporation (IRE)) have the capacity to initiate immunogenic cell death, thereby releasing damage-associated molecular patterns (DAMPs) and antigens, followed by alleviation of tumor-induced immune suppression and tumor-specific T cell priming. The immune response may be further leveraged are generated in vitro from patient or donor derived precursors in blood (e.g., monocytes), followed by antigen pulsing. Cancer vaccines may be administered intratumorally, intravenously, intranodally or subcutaneously, intradermally or intramuscularly. In vivo vaccination: Chemotherapy, oncolytic viruses and local ablative treatments (e.g., stereotactic ablative body radiotherapy (SABR) and irreversible electroporation (IRE)) have the capacity to initiate immunogenic cell death, thereby releasing damage-associated molecular patterns (DAMPs) and antigens, followed by alleviation of tumor-induced immune suppression and tumor-specific T cell priming. The immune response may be further leveraged by employing other immunotherapies such as immune checkpoint inhibitors (ICI), adoptive cell therapies (ACT) or adjuvants. Systemic immune response: With the exception of antigen-pulsed DC vaccines, all other forms of vaccination require in vivo antigen uptake of DCs. Activated DCs transport the antigens to the draining lymph nodes and establish antigenspecific priming and expansion of T cells. Primed effector anti-tumor T cells (Th1 CD4 + and CD8 + cytolytic T cells) roam the system in search of tumor cells and upon discovery may destroy these cancerous cells.

GVAX
Pancreas GVAX contains irradiated allogeneic pancreatic cancer cells, unable to grow, administered intradermally and comprising a plethora of PDAC antigens. In addition, the vaccine has been virally transduced with GM-CSF, an immune stimulatory cytokine with the ability to attract and activate conventional DCs (cDCs) [193], thus facilitating cross-priming of antigens contained within the constituent allogeneic cell lines. Thus far, results have demonstrated that GVAX, with or without concurrent treatments, is safe and well-tolerated and can induce antigen-specific T cell responses [57,[108][109][110][111][112]. However, it has yet to show superior efficacy compared to current standard chemotherapy. In their phase 2b RCT of 169 mPDAC patients, Le et al. [110] showed no improved survival for patients receiving GVAX plus cyclophosphamide (Cy, aiming to eliminate Tregs) plus CRS-207 (i.e., a Listeria Monocytogenes-based bacterial vaccine expressing mesothelin) or CRS-207 only, compared to chemotherapy (mOS 3.7 months, 5.4 months and 4.6 months, respectively, p > 0.05). Additionally, Tsujikawa et al. [111] showed no benefit of supplemental anti-PD-1 when added to a GVAX regimen (GVAX + Cy + CRS-207) in 93 mPDAC patients, with mOS of 5.9 months versus 6.2 months. However, patients whose TME showed an increase in CD8 + T cells and a decrease in TAMs and MDSCs, had a better OS. The same group also published an RCT in which the value of GVAX + anti-CTLA-4 was explored in a maintenance setting where mPDAC patients had already received 4-6 doses of FOLFIRINOX. Patients (n = 82) were randomized to receive the experimental treatment (GVAX + anti-CTLA-4, n = 40) or continue standard FOLFIRINOX (n = 42). While dual immunotherapy did promote memory T cell differentiation and increased the M1 macrophage population, it proved significantly inferior in terms of survival compared to FOLFIRINOX (mOS 9.4 months vs. 14.7 months, ORR 6% vs. 14%, p = 0.019).   (1) 7.9 (t) (2) 6.9 (t) (3) 8.4 (t)

KIF20A
KIF20A belongs to the superfamily of motor proteins which play an essential role in the trafficking of molecules during pancreatic cancer growth [197] and is highly expressed in >90% of PDAC patients [198]. Initial results have been published of KIF20A vaccines for PDAC [172,173,176,177]. Asahara et al. [172] utilized a subcutaneously injected KIF20A-66 protein vaccine in gemcitabine refractory, unresectable or recurrent metastatic PDAC patients (n = 29), resulting in a peptide-specific CD8 + effector T cell response and an mOS of 4.2 months, compared to 2.2 months after best supportive care (p = 0.047).

VEGFR
Under physiological conditions, VEGF is essential for angiogenesis during embryonic growth, normal growth and tissue repair. However, in cancer this process has become deregulated due to abnormal expression of VEGF, resulting in anomalous blood vessel structures, thereby promoting tumor growth, invasion and spread, while hindering proper lymphocyte homing and extravasation [199]. VEGF and its receptors, VEGFR1/2, are co-expressed in 77-93% of PDAC patients and overexpression is linked to worse clinical prognosis [200,201]. A few articles have presented clinical data of VEGFR vaccines in PDAC [174][175][176][177]. Yamaue et al. [175] reported results of a phase 2/3 RCT in which 153 chemoradiation naïve LAPC and mPDAC patients were included and allocated either a subcutaneously injected VEGFR2 peptide vaccine combined with gemcitabine (n = 100) or a placebo with gemcitabine (n = 53). Results revealed a comparable mOS between the two treatment groups (8.4 and 8.5 months, p = 0.9, respectively). Similar results were presented by Miyazawa et al. [174]

Survivin
Survivin is physiologically expressed during embryonic and fetal development and plays a critical role in cell cycle control and apoptosis. In PDAC it is re-expressed in about 80% of patients [202], with elevated expression associated with worse prognosis and treatment resistance [203]. Shima et al. [178,179] utilized a survivin 2B-vaccine (SVN-2B) in a phase 2 clinical trial with randomization design in which 83 advanced PDAC patients were included, all pre-treated with at least one line of chemotherapy. Patients were allocated to receive SVN-2B + IFN-α (n = 30), SVN-2B only (n = 34) or a placebo (n = 19). While survivin-specific CD8 + T cells were increased, mOS outcomes were similar between the three arms (3.4 months, 3.2 months and 3.6 months, respectively). 4.7.6. MUC-1 MUC-1 is expressed at intermediate levels by pancreatic tissue, orchestrating cell signaling and differentiation functions. However, atypical expression is seen in >60% of PDAC patients and correlates with tumor size and dysplasia, suggesting a pivotal role in tumor progression [204]. Rong et al. [180] employed a MUC-1-pulsed DC vaccine in 7 advanced PDAC patients previously treated with chemotherapy and surgery, all of whom had aberrant MUC-1 expression. ELISPOT assays showed an increase in IFN-γ and granzyme B secreting peripheral blood mononuclear cells (PBMCs) in 29% of patients, but a clinical response was not observed (ORR 0%, DCR 0%).

Mesothelin
Mesothelin is minimally expressed in several healthy tissues and is almost always (~100%) highly expressed in PDAC at its invading edges, but its exact functions remain unknown [205]. It is often co-expressed with CA-125 (MUC-16). Binding of CA-125 with secreted mesothelin enhances tumor cell motility and invasion [206]. In chemotherapy refractory PDAC patients, CRS-207 (the mesothelin-expressing bacterial vaccine) increased mesothelin-specific CD8 + T cells in 60% of patients and resulted in an mOS of 7 months [181].

hTERT
Telomeric ends of DNA strings become progressively shortened with cellular replication, eventually leading to senescence and death. Pancreatic cancer cells are able to reverse this process by re-activating telomerase enzymes, which can elongate shortened telomeres, hereby inducing replicative immortality [207]. Middleton et al. [115] published a phase 3 RCT, including LAPC and mPDAC patients (n = 1082) of whom a subset was treated with an intradermal telomerase peptide vaccine (GV1001). Patients were randomized to receive chemotherapy (n = 358), sequential chemo-immunotherapy (GV1001/GM-CSF) (n = 350) or concurrent chemo-immunotherapy (GV1001/GM-CSF) (n = 354). Adding GV001 did not improve survival; the mOS outcomes were 7.9 months, 6.9 months and 8.4 months, respectively (p = 0.11). For both chemo-immunotherapy groups an immunological response was measured; in the sequential group 12% of patients were DTH positive and 31% demonstrated GV1001-specific T cell proliferation, whereas in the concurrent group 20% of patients were DTH+ and 15% demonstrated GV1001-specific T cell proliferation.

Neoantigen: Mutant KRAS
The KRAS protein is part of the RAS/Mitogen-Activated Protein Kinase (MAPK) signaling pathway, involved in cell proliferation and differentiation. In more than 90% of PDAC patients the KRAS proto-oncogene is mutated, its expression being associated with overall worse prognosis and treatment insensitivity [208]. In clinical trials, a mutant KRAS vaccine has been utilized through an Epstein Barr Virus-transformed lymphoblastoid cell line (CLC) [183] or combined with adjuvant GM-CSF [113,114] in PDAC patients with a confirmed KRAS mutation. Mutant KRAS vaccines combined with gemcitabine as adjuvant therapy have demonstrated encouraging initial results in resected PDAC patients. Palmer et al. [114] published data from a phase 1/2 clinical trial, in which 32 PDAC patients with resected primary tumors received an intradermal injection of a seven-peptide vaccine covering most known mutations of KRAS (TG01), co-administered with recombinant GM-CSF (TG01/GM-CSF), and combined with gemcitabine. Their survival outcomes (mOS 33.1-34.3 months; mPFS 13.9-19.5 months) compared favorably with current literature reporting survival after resection followed by adjuvant gemcitabine (mOS 17-27 months) [209][210][211][212]. In addition, patients presented with enhanced levels of immune activation; 62-95% of patients were skin-tested and showed a positive DTH reaction, 74-92% of patients had peptide-specific T cells.

Neoantigen: Other
Neoantigen vaccines are fairly novel but have already seen some preliminary clinical success in multiple solid tumors, even poorly immunogenic glioblastoma, in terms of safety, immunogenicity and feasibility [213][214][215][216]. A pre-clinical study using a murine PDAC model demonstrated that a combined neoantigen vaccine and STING adjuvant regimen resulted in transient tumor regression [142]. Moreover, adding anti-PD-1 and an OX40 agonist enhanced these results. Clinically, there is only a proof-of-antigen discovery feasibility study done in three resected PDAC patients [184]. A multitude of clinical trials are currently recruiting PDAC patients for neoantigen vaccine treatments (e.g., ClinicalTrials.gov Identifier: NCT03558945, NCT03953235, NCT03956056).

Immunotherapy and Local Ablation
Focal ablation techniques have, in addition to their tumor destructive capacity, demonstrated immune potentiating abilities by inducing immunogenic cell death and temporarily lifting immune-suppressive conditions [11,217,218]. In effect, this form of tumor destruction may prompt in vivo vaccination against patient-specific pancreatic tumor cells ( Figure 6). Combining ablation with immunotherapy might act synergistically to elicit an anti-tumor immune response. Of the available ablation techniques, thus far only irreversible electroporation (IRE) and stereotactic ablative body radiotherapy (SABR) have been combined with immunotherapy in PDAC patients (Table 7). (1) 3.3-9 (t) (2) 2.1-4.2 (t) NR = not reported; IT = immunotherapy; IRE = irreversible electroporation; SABR = stereotactic ablative body radiotherapy; perc = percutaneous; chemo = chemotherapy; LAPC = locally advanced pancreatic cancer; mPDAC = metastatic pancreatic ductal adenocarcinoma; mPFS = median progression-free survival; mOS = median overall survival; t = from treatment/registration; d = from diagnosis; x = unknown whether from diagnosis or treatment/registration.

Immunotherapy + IRE
IRE has demonstrated the ability to temporarily lift immune-suppressive barriers and to elicit an anti-tumor immune response in PDAC patients [224,225]. Consequently, the feasibility of combined IRE with immunotherapy, coined as 'electroimmunotherapy' [226], was examined in pre-clinical immunocompetent PDAC murine models, resulting in significant tumor regression, increased CD8 + T cell infiltration and improved survival [137,[227][228][229]. In a clinical setting, percutaneous IRE has been combined with allogeneic NK cell therapy [219]. Compared to IRE alone, IRE + NK cells resulted in a modestly improved mOS in patients with LAPC (12.2 vs. 13.6 months, p = 0.033, n = 35) and mPDAC (9.1 vs. 10.2 months, p = 0.037, n = 32), as well as improved mPFS in LAPC (9.1 months vs. 7.9 months, p = 0.043). Furthermore, patients receiving NK cell therapy demonstrated significantly increased systemic levels of total and CD8 + T cells, NK cells and pro-inflammatory cytokines. Recently, an RCT was published in which IRE + allogeneic Vγ9Vδ2 T cells (n = 30) was compared to IRE alone (n = 32) in 62 LAPC patients, of whom 49 had previously received chemotherapy [220]. Clinical outcomes for IRE + Vγ9Vδ2 T cells were superior compared to IRE alone (mOS 14.5 months vs. 11 months, p = 0.01; mPFS 11 months vs. 8 months, p = 0.03). Interestingly, patients receiving multiple T cell infusions survived significantly longer compared to patients receiving a single infusion (mOS 17 months vs. 13.5 months, p > 0.05). O'Neill et al. [221] provided initial results of a phase 1b trial in which a combination of surgical IRE and systemic anti-PD-1 was administered to ten LAPC patients. All patients were stable after first line chemotherapy (8 FOLFIRINOX, 2 gemcitabine/nab-paclitaxel). The mOS was 18 months and mPFS was 6.3 months, thus providing a solid base for their subsequent, currently ongoing phase 2 clinical trial. Finally, the PANFIRE-III trial (NCT04612530) is currently evaluating safety and efficacy of IRE + systemic anti-PD-1 ± an intratumoral TLR-9 agonist in mPDAC patients with at least stable disease after pre-treatment with a minimum of eight FOLFIRINOX cycles [139].

Immunotherapy + SABR
Pre-clinical treatment strategies in murine PDAC models combining SABR and intratumoral anti-CD40 or intratumoral IL-12 microspheres presented promising results. IFN-γ production increased and CD8 + T cells were activated, followed by significant tumor regression and formation of long-term immunologic memory [230,231]. Clinically, Xie et al. [223] examined combination treatment with SABR and immune checkpoint inhibitors in 59 pre-treated mPDAC patients. Interestingly, SABR + anti-PD-L1 (mOS of 3.3-9 months) resulted in superior survival compared to SABR + anti-PD-L1 + anti-CTLA-4 (mOS 2.1-4.2 months). An increase in CD3 + and CD8 + T cells was observed independent of clinical response. While this study was limited due to an absent control arm, SABR + anti-PD-L1 may confer a minor survival benefit in previously treated mPDAC patients. In another clinical trial, a multimodal treatment strategy including chemotherapy, SABR, nelfinavir (anti-retroviral drug, i.e., a radiosensitizer), an anti-CA-125 antibody ± pancreatic resection in LAPC patients (n = 11) resulted in an mOS and mPFS of 13 and 8.6 months, respectively [222]. Of the five patients tested, 40% developed CA-125 specific CD8 + T cells. Overall, this treatment approach was safe and feasible, but did not result in a major clinical benefit. Multiple trials are ongoing that combine SABR with various immunotherapies in PDAC (e.g., ClinicalTrials.gov Identifier: NCT03563248; NCT04390399; NCT04327986; NCT04098432).

Conclusions
Results from clinical trials incorporating immunotherapy strategies have been disappointing in PDAC. The limited efficacy, although of multifactorial causation, can largely be attributed to the highly immunosuppressive TME. In order to overcome this major obstacle, current immunotherapies such as oncolytic viruses, cancer vaccines and adoptive (CAR) T or NK cell therapies should be improved and optimized. In addition, novel immunotherapy targets identified in a pre-clinical context for which limited or no clinical data in PDAC exists should be further explored to determine their potential. These include novel immune checkpoint inhibitors, stimulatory agonists and adjuvants. Equally, the complex interplay of immunosuppressive mechanisms likely requires rational multimodal treatment approaches, targeting several immune evasion mechanisms to establish a durable anti-tumor response. In order to achieve this response, antigen presentation should be improved, anti-tumor immune cells activated, checkpoint-induced inhibition removed and tumor-site infiltration of cytotoxic immune cells increased. To make such optimized treatment choices, careful dissection of immune evasion mechanisms at play in the human PDAC TME and its draining lymph nodes is required. The motive for employing combination therapies is reinforced by the demonstrable potential of various immunotherapies to elicit a tumor-specific immune response. Moreover, therapies with the ability to provoke immunogenic cell death (e.g., chemotherapy and ablative treatment) may work synergistically with immunotherapy. For these strategies, optimal timing, dosage and choice of combination therapies may emerge as the primary challenge. Ultimately, the aim is to develop a (personalized) multimodal treatment strategy which will lift immunosuppressive barriers, effectively kill all tumor cells and create long-lasting protective immune memory.