Looking Beyond Checkpoint Inhibitor Monotherapy: Uncovering New Frontiers for Pancreatic Cancer Immunotherapy

Pancreatic ductal adenocarcinoma (PDAC) represents a formidable challenge in oncology, poised to become the second leading cause of cancer-related deaths by 2030 (1, 2). The bleak prognosis associated with PDAC largely stems from its resistance to the majority of available treatment options. While surgical resection of the tumor can offer long-term benefits, it remains feasible for only a small subset of patients due to the limited resect ability of pancreatic tumors, accounting for only 10% to 20% of cases (3). Multiagent chemotherapy regimens, such as FOLFIRINOX (folinic acid, fluorouracil, irinotecan, oxaliplatin), have demonstrated modest improvements in patient survival, albeit with short-lived partial remissions (4).

Despite the transformative impact of immunotherapies in various solid tumor types, PDAC remains largely refractory, particularly to immune checkpoint inhibitor therapies. Clinical trials investigating the efficacy of immune checkpoint monotherapies in PDAC have consistently failed to elicit objective responses (5, 6). This failure is attributed to PDAC’s inherently poor immunogenic nature coupled with its immunologically “cold” tumor microenvironment (TME). PDAC TME is characterized by a significant infiltration of myeloid cells but lacks CD8⁺ T cells (Teffector) and shows low expression of activation markers like granzyme B (GZMB) and IFNγ (7). These features indicate a deficiency or impairment in adaptive T-cell immunity, rendering PDAC unresponsive to checkpoint blockade therapies. Specifically, the absence of effector lymphocyte activity in the PDAC TME hinders the effectiveness of checkpoint inhibitor therapy.

However, despite these challenges, studies investigating immunotherapy in PDAC have provided invaluable insights for future approaches. They have prompted further investigation into the underlying drivers of immunotherapy failure in PDAC. By considering the unique interplay between the tumor, immune system, and genetic landscape in PDAC, a new wave of immunotherapy strategies is emerging. These strategies not only serve as standalone therapies but also sensitize tumors to immune checkpoint blockade (ICB), offering great promise for improved patient outcomes. As our comprehension of the PDAC TME and its immune evasion mechanisms continues to evolve, it unveils a plethora of promising targets. These targets hold the potential to facilitate immune-mediated control over this formidable disease.

In this review, we explore several therapeutic strategies that capitalize on our understanding of the complex tumor-immune landscape of PDAC. Additionally, we delve into recent discoveries regarding mediators of therapeutic resistance, with a particular focus on ICB resistance. Through these discussions, we aim to provide insights that can inform the development of more effective treatments for patients with PDAC.

As our grasp of the intricacies within the PDAC TME deepens, a range of emerging immunotherapy strategies has surfaced. These strategies stand as potential breakthroughs, either as standalone treatments or in combination with conventional therapies that previously showed limited effectiveness against PDAC (1, 5–7). These promising immunotherapies for PDAC encompass various distinct therapeutic approaches:

(i) Activation and Establishment of Immunological Memory: This approach aims to activate the immune system against PDAC and establish a lasting memory to prevent tumor recurrence. Strategies may include immune checkpoint inhibitors, cancer vaccines, and adoptive T-cell therapy. (ii) Stroma Disruption/Modulation: PDAC is characterized by a dense stroma that creates a barrier to effective immune response. Disrupting or modulating the stroma can enhance immune cell infiltration and improve treatment efficacy. Therapies in this category may include stroma-targeting agents, such as hedgehog pathway inhibitors, or strategies to remodel the extracellular matrix (ECM). (iii) Targeted Therapies: Targeted therapies aim to inhibit specific molecular pathways implicated in PDAC growth and progression. Small molecule inhibitors may target key signaling pathways involved in tumor proliferation, survival, and angiogenesis, among others. (iv) Adaptive T-cell Therapy: This approach involves harnessing the power of T cells to target PDAC cells specifically. Adoptive T-cell therapy, such as chimeric antigen receptor T-cell (CAR-T) therapy, can be engineered to recognize and eliminate tumor cells. (v) Targeting Myeloid Cell Population: Additionally, targeting the myeloid population, which plays a critical role in immune suppression within the TME, may enhance antitumor immune responses.

Each of these strategies offers unique avenues for combating PDAC, addressing different aspects of the TME and immune escape mechanisms. By targeting these specific pathways, these immunotherapeutic approaches hold promise for improving outcomes in patients with PDAC. Detailed elucidation of the platforms underlying each strategy is detailed in Fig. 1. Additional immunotherapy strategies are encompassed in Table 1.

Immunological memory is a hallmark of immunity and characterized by a rapid and robust activation in the presence of antigens that have been encountered previously by the host. Given the potential for durable, precise antitumor immunity, establishment and activation of immunological memory is being explored to combat PDAC progression. Advances of strategy: Due to the low mutational burden of PDAC, the usage of OX40 agonists and delivery of exogenous antigen combined with established therapies offer a potential strategy in forming robust immunological memory. OX40 is a secondary costimulatory molecule belonging to the tumor necrosis factor superfamily that promotes T-cell proliferation, survival, and memory (28). Ma and colleagues (29) investigated the therapeutic efficacy of programmed death receptor 1 (PD-1) blockade in combination with an OX40 agonist. Utilizing an orthotopic KPC-luc PDAC tumor model, the treatment resulted in a durable tumor rejection that was sustained for over 100 days following completion of treatment. Additionally, when treated animals were rechallenged with KPC-luc cells, no discernable tumor was established, suggesting rapid recall of tumor-specific memory T cells (29). Establishment of immunological memory was confirmed through identification of enriched intratumoral CD4⁺ CD127⁺ effector memory T-cell populations, and the depletion of CD4⁺ T cells resulted uncontrolled tumor growth following rechallenge due to abrogating CD4⁺ T-cell memory. CD4⁺ memory T cells also played an integral role in a bacterial based immunotherapy using Listeria monocytogenes (29). To overcome the poor immunogenicity and suppressive TME of PDAC, Listeria was engineered to deliver immunogenic tetanus toxoid (TT) directly to the tumor (30). Listeria-based tumor immunotherapies provide the benefit of effective delivery of antigen to professional antigen presenting cells, to initiate antigen specific T-cell immunity (30, 31). Using the Panc-02 orthotopic model, animals receiving Listeria-TT in combination with prior TT immunization had reduced tumor weight and metastasis than animals that received Listeria-TT treatment without immunization. TT-specific memory was further confirmed in that CD4⁺ and CD8⁺ T cells derived from Listeria-TT–treated KPC mice were robustly activated in the presence of TT, as measured by IFNγ production, compared to control animals. Listeria-TT treatment also demonstrated enhanced efficacy when combined with gemcitabine chemotherapy, effectively reducing primary tumor burden and metastasis by 80% and 87%, respectively. Thus, it can be inferred that pre-immunization with TT followed by combination treatment of Listeria-TT and gemcitabine would elicit even greater antitumor effects due to the established memory against TT. Because many individuals receive TT in the form of childhood immunizations, reactivation of TT-specific memory T cells with this treatment greatly adds to its clinical impact and translatability. Limitations of strategy: T-cell memory–based immunotherapy is often hindered by T-cell exhaustion and dysfunction associated with the immunosuppressive TME. T cells exposed to inhibitory cytokines such as IL10 or TGFβ produced by regulatory T cells (Treg) or tumor-associated macrophages (TAM) can upregulate expression of inhibitory receptors PD-1, cytotoxic T-lymphocyte antigen 4 (CTLA-4), and TIM-3 on T cells, resulting in failure to establish immunological memory (32, 33).

The thick, fibrotic stroma surrounding PDAC tumors provides a significant barrier to therapeutic for both chemotherapy and immunotherapy alike in that it precludes access to the tumor (34, 35). The PDAC comprises a complex interplay of various components, including ECM proteins, tumor vasculature, fibroblasts, and immune cells. In PDAC, ECM proteins play a multifaceted role, not only providing structural support but also influencing intratumoral pressure and acting as a barrier to effective drug delivery (36, 37). Despite efforts in clinical trials to leverage this understanding of the PDAC ECM, improvements in patient survival have remained elusive. Preclinical models have underscored the significance of tumor-associated vasculature in PDAC pathogenesis. However, clinical trials targeting PDAC vasculature have yet to demonstrate prolonged patient survival. Strategies aimed at pruning PDAC vasculature, known as normalization, hold promise for enhancing chemotherapy delivery and fostering antitumor immune responses. Moreover, molecular subtypes of pancreatic cancer–associated fibroblasts (CAF) have been delineated, notably including inflammatory CAFs and myofibroblastic CAFs. These distinct subtypes may exhibit either protumoral or antitumoral properties, contributing to the dynamic TME in PDAC. Understanding the roles of these CAF subtypes may offer insights into potential therapeutic interventions. Advances of strategy: Attenuated Salmonella typhimurium engineered to secrete cytolysin A (SL-ClyA), a bacterial pore-forming toxin, has been positioned as a potential PDAC stroma-targeting therapy. Treatment of AsPC-1 tumors with SL-ClyA bacteria altered the composition of PDAC stroma, characterized by reduced expression of neural/glial antigen 2, PDGFRβ, and CD31, which constitute markers for CAFs, pericytes, and lymphatic endothelial cells, respectively. Accordingly, this reduction in stromal components permitted an increase in immune cell filtration within the tumor, namely through macrophage, CD4⁺ T cell, and CD8⁺ T-cell populations. Unsurprisingly, these changes in the composition of the stroma and TME following SL-ClyA treatment ultimately resulted in a significant reduction in tumor growth (38). As this study highlights, stromal disruption also renders PDAC significantly more permissible to immune cell infiltration and a target for eradication by the immune system. CAR-T has also been utilized to disrupt PDAC stroma and reverse tumor immunosuppression. Specifically, CAR-T were developed to target fibroblast activation protein (FAP), a protease highly expressed on stromal cells. Mesothelin (MSLN) positive 4662 PDAC tumors treated with FAP-CAR-T experienced effective depletion of FAP+ CAFs followed by reduced collagen remodeling and fibronectin within the stroma (39). Consequently, FAP-CAR-T treatment was accompanied by an increase in CD3⁺ T cells, effectively altering the composition of the TME. This change in the immune landscape rendered PDAC tumor sensitive to anti–PD-1 blockade as well as MSLN targeting CAR-T, as measured by reduced tumor growth and improved survival (39). Limitations of strategy: Indiscriminate depletion of PDAC stroma may indeed hinder tumor outcomes. Stroma residing antigen-presenting CAFs express MHC class II molecules, and have displayed T-cell activation capacity in a co-culture setting (40). Although the role of antigen-presenting CAFs needs to be further elucidated, the findings thus far suggest complete depletion of the stroma and CAF populations may impede tumor rejection.

Utilizing targeted therapies offers the advantage of potent treatment with minimal toxicities, owing to their precise targeting capabilities. In the realm PDAC, these therapies target both established immunological pathways and protein targets with emerging roles in tumor immunotherapy. Advances of strategy: Dipeptidyl peptidases (DPP), initially recognized for their involvement in metabolic hormone regulation, have recently emerged as promising targets for anticancer interventions. DPP inhibitors, such as BXCL701, have demonstrated notable antitumor effects in melanoma and colorectal cancer models. In murine tumor models, BXCL701, acting as a DPP4 and DPP8/9 inhibitor, exhibited significant therapeutic efficacy by suppressing tumor growth (41). Analysis of cytokine profiles in BXCL701-treated mice revealed a marked upregulation of Th1 response-associated cytokines, suggesting its immunomodulatory potential. Moreover, BXCL701 sensitized tumors to anti–PD-1 blockade, indicating its ability to enhance antitumor immunity, particularly through CD8⁺ T-cell– and NK cell–dependent mechanisms (41). Similarly, SUMOylation inhibitor TAK-981 has shown promise in bolstering antitumor immunity. In KPC3 tumors, TAK-981 treatment elicited enhanced effector properties in CD8⁺ T cells and NK cells, as evidenced by increased expression of cytotoxic molecules GZMB and perforin (42). However, the efficacy of TAK-981 was compromised when combined with blockade of CD8⁺ T-cell and IFNAR1 (42). Additionally, small molecule immunomodulators have demonstrated synergistic effects with chemotherapy. Axl, a receptor tyrosine kinase implicated in cell proliferation and migration, has garnered attention as a therapeutic target. BGB324, an Axl tyrosine kinase inhibitor, exhibited anticancer effects in breast and lung tumor models. When administered alongside gemcitabine, BGB324 significantly prolonged survival in mouse models of PDAC (43). This combinatorial approach was associated with suppression of TBK1/NF-κB signaling and reduced populations of myeloid-derived suppressor cells (MDSC) and TAMs, further underscoring the potential of small molecule immunomodulators to enhance therapeutic outcomes in PDAC (43). Limitations of strategy:A major consideration of using targeted therapies to treat PDAC is the development of resistance to these treatments over time. Because of the heterogeneous nature of PDAC, some cell populations may be sensitive to these treatments and are effectively eradicated. Consequently, the remaining treatment-resistant cells become now dominant in the tumor, establishing a tumor that is increasingly difficult to treat.

Myeloid cell populations dominate the immune landscape of PDAC. Myeloid populations such as MDSCs, TAMs, and in certain contexts neutrophils, all contribute to the immunosuppressive or “cold” nature of the PDAC TME and diminish the potential for developing T-cell immunity. These MDSCs manifest in two primary subsets: polymorphonuclear (PMN)-MDSCs, akin to granulocytes/neutrophils, and mononuclear-MDSCs (M-MDSC), resembling monocytes. In mice, MDSCs are discerned by CD11b⁺ Gr-1⁺, with Ly-6C and Ly-6G utilized for population stratification. In humans, PMN-MDSCs are typically identified by CD11b⁺ CD14⁻ CD15⁺ or CD11b⁺ CD14⁻ CD66b⁺, while M-MDSCs express CD11b⁺ CD14⁺ HLA-DR-/lo CD15⁻ (44). Functionally, MDSCs exert immune suppression predominantly by depleting L-arginine via high levels of Arginase 1 (ARG1) expression, hindering T-cell function. PMN-MDSCs constitute the predominant MDSC subset within blood and tumors, while M-MDSCs, albeit fewer in number, often exhibit heightened immunosuppressive activity. Both subsets produce ARG1, yet possess distinct mechanisms: PMN-MDSCs generate reactive oxygen species, while M-MDSCs produce nitric oxide, facilitated by tumor-derived prostaglandin E2 (44). Advances of strategy: Targeting MDSCs emerges as a promising therapeutic avenue, exemplified by preclinical studies utilizing zoledronic acid to curtail MDSC recruitment, yielding delayed tumor progression and enhanced T-cell infiltration. Inhibition of CXCR2, a key MDSC recruitment mediator, and neutralization of granulocyte-macrophage colony-stimulating factor (GM-CSF), pivotal in MDSC attraction, further highlight potential strategies to alleviate MDSC-mediated immune suppression, fostering improved outcomes in PDAC treatment.

Activation of the stimulator of interferon genes (STING) pathway is a key mediator in inflammation and the production of type I interferon (45). Thus, investigating its role as an immunostimulator in cancer is being extensively studied. STING agonist IACS-8803 sensitized mice bearing primary and metastatic mT4-LS tumors to a combination of anti–PD-1 and anti–CTLA-4 blockade, leading to complete clearance of all legions (46). As determined by high-parameter flow cytometry, the combination therapy resulted in diminished TGFβ producing PMN-MDSC and TAM populations, while expanding the CD8⁺ T-cell compartment, indicating proinflammatory remodeling of the tumor through STING pathway activation (45). The functions of neutrophils in cancer are largely context dependent, thus defining their role in cancer to be much more ambiguous. In PDAC, neutrophils have been implicated in stromal maintenance and immune escape of malignant cells (46). Recently, PDAC cells have been identified to activate non receptor tyrosine kinase FES on neutrophils, which is a protooncogene. Lorlatinib is a potent tyrosine kinase inhibitor that can target FES. Thus, lorlatinib treatment of KPC mT4 orthotopic tumors diminished PDAC growth and reduced CD11b⁺ F4/80- Ly6c^mid Ly6G⁺ neutrophil populations within the tumor (47). Further confirming the tumorigenic role of neutrophils, depletion through Ly6G depletion resulted in comparable tumor growth reduction to lorlatinib. Anti-human MARCO (anti-hMARCO) antibody targeting triggered the repolarization of TAMs and activated the inflammasome machinery, resulting in IL18 production. This in turn enhanced T-cell and NK-cell functions. The targeting of MARCO thus remodels the TME and represents a rational approach to make immunotherapy more efficient in patients with PDAC (48). Limitations of strategy: The PDAC TME encompasses a number of distinct myeloid phenotypes (44), however, many treatment modalities being investigated target only a single myeloid population. Ultimately, this may not be sufficient to overcome tumor immunosuppression and establish durable antitumor immunity.

An alternative approach to stimulating tumor-specific T-cell activity in patients is to adoptively transfer tumor-reactive T cells. This approach bypasses the need for in vivo T-cell priming and allows for assessment of downstream mechanisms that may regulate T-cell infiltration and effector activity within tumors (49). The association between T-cell infiltration and improved clinical outcomes in PDAC suggests a potential avenue for therapeutic intervention through tumor-infiltrating lymphocyte (TIL) therapy. In PDAC tumors, there exists a moderate infiltration of CD4⁺ and CD8⁺ T cells, primarily localized within stromal areas surrounding the tumor core (50). While preclinical investigations have demonstrated the feasibility of isolating and expanding TILs ex vivo from PDAC tumor specimens, the lack of preexisting T-cell immunity in the majority of patients with PDAC poses a significant challenge to the broad applicability of TIL therapy in this patient population (51).

Recent advances in antigen identification, T-cell biology, gene therapy, and gene editing have spurred the development of strategies aimed at redirecting T-cell antigen specificity and enhancing T-cell function (52). These advancements have yielded unprecedented clinical outcomes in patients with hematologic malignancies treated with CAR-T therapy and T-cell receptor T-cell (TCR-T) therapy. However, the translation of adoptive T-cell therapy (ATCT) to solid tumors, including PDAC, has been hindered by issues such as poor homing, proliferation, and survival of transferred cells. To overcome these challenges, various strategies have been proposed, including combinatorial treatment with immune-modulating agents and the expression of transgenes to enhance T-cell homing, penetration, and persistence. Additionally, the advent of CRISPR/Cas9 technology has facilitated the development of clinically translatable gene-editing strategies aimed at enhancing the antitumor activity of adoptively transferred T cells.

However, antigen selection remains a significant limitation in CAR-T and TCR-T strategies for PDAC (50). Most studies to date have targeted tumor-associated antigens (TAA), which may exhibit variable or heterogeneous expression on tumor cells, thereby increasing the risk of on-target, off-tumor toxicity. Advances in strategy: Antigenic targets under investigation for CAR-T therapy in PDAC include MSLN, prostate stem cell antigen, carcinoembryonic antigen (CEA), MUC1, and human epidermal growth factor receptor 2 (HER2), among others (50). Notably, serious adverse events have been reported in patients treated with T-cell products engineered to target HER2, limiting the evaluation of these antigens as targets for CAR-T therapy in PDAC. Conversely, CAR-T directed against MSLN has demonstrated efficacy in preclinical mouse models and safety in early-phase clinical studies in patients with chemotherapy-refractory metastatic PDAC, warranting its continued clinical evaluation.

In addition to TAAs, TCRs targeting neoantigen epitopes offer a promising alternative approach. These TCRs are not subject to thymic negative selection, allowing for the identification of high-avidity TCRs with enhanced antitumor activity and reduced toxicity (53). The high prevalence and conserved mutational profile of mutant KRAS (mKRAS) in PDAC provide a unique opportunity to develop neoantigen-targeted TCR therapy with broad generalizability. Mutant KRAS–specific TCRs have been identified using TILs and peripheral blood of patients with mKRAS epithelial cancers and induced in vaccinated human leukocyte antigen–transgenic mice (54). A case report highlights the therapeutic potential of targeting mKRAS neoantigens using ATCT, and ongoing clinical trials further investigate this strategy.

Moreover, CAR-T therapy faces challenges due to the selective expression of cell surface antigens in pancreatic cancer. CEA and MSLN are highly expressed in pancreatic cancer (55, 56). Interestingly, dual-receptor CAR-T targeting both antigens have demonstrated precise tumor targeting and reduction of tumor burden in mouse models of pancreatic cancer. Additionally, hYP218 CAR-T, engineered to target MSLN at proximal membrane surface sites, have shown promising results in inducing durable antitumor immunity in murine models.

Despite these promising advances, the immunosuppressive TME remains a formidable obstacle to CAR-T therapy efficacy in pancreatic cancer. However, inducible IL18 CAR-T have shown effectiveness in inducing regression of advanced pancreatic cancer in mouse models (57). Furthermore, IL8 receptor–modified CARs, in combination with CD70, have enhanced CAR-T efficacy in pancreatic cancer therapy (58). Combinatorial production of IL7 and chemokine (C-C motif) ligand 19 (CCL19) in 7 × 19 CAR-T has demonstrated superior antitumor activity against pancreatic cancer compared to conventional CAR-T therapy (59). IL7/CCL19 anti-MSLN CAR-T have been effective in eliminating malignant tumors in situ in mouse models (59). Targeting trophoblast surface antigen 2 (Trop2) using CAR-T has also emerged as a promising approach for pancreatic cancer treatment (60). Similarly, CARs targeting chimeric PD-1 have shown increased survival in pancreatic cancer–bearing mice. Third-generation ICOSBBz CAR-T have been found to inhibit pancreatic cancer progression while promoting CAR-T survival in vivo (59). Additionally, microbial molecules have been identified to regulate immune cell motility, with valeric acid and butyric acid enhancing the antitumor activity of CD8⁺ CAR-T. Second-generation KD2—natural killer group 2D (NKG2D) —CAR-T have exhibited stronger antitumor activity compared to first-generation NKG2D-CAR-T in pancreatic cancer xenograft models (61). Given the challenges of CAR-T therapy, CAR NK-cell therapy has been recently investigated as an alternative for PDAC treatment. CAR NK cells possess the advantage of not requiring autologous immune cells thus are suitable allogeneic therapies. Further, CAR NK cells can exert cytotoxicity through CAR-mediated and CAR-independent mechanisms (62). CAR NK cells engineered to express CXCR2 demonstrated robust chemoattractive ability to Capan-2 tumors, as well as tumoricidal properties in patient-derived PDAC organoids (63). Targeting PDAC-associated tumor antigens with CAR NK cells is also being explored. MSLN-targeting CAR NK cells combined with Cyclic GMP-AMP (cGAMP) inhibited tumor progression in a subcutaneous model of PDAC. In a co-culture setting, this treatment directly induced apoptosis of AsPC-1 and Capan-2 cells (64). Although these studies are promising, more extensive in vivo studies are needed to establish the long-term efficacy of CAR NK-cell therapies in PDAC. Limitations of strategy: A constant challenge with adoptive cell therapies, notably CAR-based therapies, is loss of expression of the target antigen by the tumor over time (65). Given the poor immunogenicity of PDAC tumors, identifying alternative tumor-specific antigens as targets for CAR therapies is scarce. In addition, the stroma surrounding PDAC tumors also functions as physical barrier to cell therapies.

Therapeutic cancer vaccines typically comprise tumor antigens in the form of whole tumor cell lysate, tumor-derived peptides, or complete proteins enriched in cancer cells. These antigens are often combined immunological adjuvants allowing for successful uptake and activation of dendritic cells (DC), initiating tumor-specific immunity (66). Cancer vaccines following this paradigm have had limited success for the treatment of PDAC. GVAX is a cancer vaccine containing two irradiated allogeneic PDAC cell lines engineered to secrete GM-CSF, a DC maturation cytokine (67). A phase I clinical trial demonstrated promising results of GVAX combined with anti–CTLA-4 blockade in PDAC, however in a larger scale phase II study, GVAX provided no improvements to overall survival (68). The modest success of cancer vaccines in PDAC have resulted in investigation of alternative vaccine platforms. Although PDAC tumors are typically considered poorly immunogenic and to produce little neoantigens, neoantigen-targeting vaccines have shown considerable promise for PDAC immunotherapy. Neoantigens are new proteins developed when intratumor mutations occur and are often recognized by the immune system as “non-self” antigens (69). Advances of strategy: PancVAX2 is a cancer vaccine developed through screening of Panc02 neoantigens to select those predicted to bind to MHCI and MHCII molecules, with the intention of triggering neoantigen-specific, CD8⁺ and CD4⁺ T-cell responses (70). In the Panc02 tumor model, PancVAX2–treated mice experienced significant reduction in tumor burden, compared to animals treated with vaccines targeting single CD8⁺ or CD4⁺ T-cell populations. The improved tumor outcomes from PancVAX2 were also accompanied by significant infiltration of CD8⁺ T cells expressing a characteristic effector cytotoxic signature of IFNγ+, IL2+, and GZMB+ (70). Concomitantly, PancVAX2 also induced a Th1 CD4⁺ T-cell response, enhancing IFNγ, IL2, and TNFα. Also of note, tumor-infiltrating CD8⁺ T cells expressed exhaustion marker PD-1 at a significantly lower level in PancVAX2–treated mice compared to the CD8⁺ vaccine (70). This may suggest that PancVAX2 elicits CD8⁺ and CD4⁺ T-cell cooperativity in which CD4⁺ T cells may indirectly “relieve” some burden from the CD8⁺ T cells. Given the potential for personalized tumor neoantigen targeting using this platform, this vaccine therapy holds great therapeutic promise for tumor types with significant intertumoral heterogeneity such as PDAC. Autogene cevumeran is also being investigated to accommodate the heterogeneous nature of PDAC neoantigens. Autogene cevumeran is a personalized neoantigen vaccine utilizing an mRNA-based delivery platform (71). This vaccine is also personalized in nature in that it comprises MHCI- and MHII-restricted neoantigens determined by sequencing of patient’s tumor and blood. These neoantigens are then delivered utilizing lipoplex nanoparticles. Results from a phase I clinical trial combining autogene cevumeran with anti–PD-L1 blockade and a modified FOLFIRINOX regimen revealed positive clinical outcomes in patients with PDAC (71). When stratified into treatment responders and nonresponders, responders to autogene cevumeran treatment did not reach a median relapse-free survival (RFS) whereas nonresponders had a median RFS of 13.4 months, effectively demonstrating the anticancer potential of this treatment (71). Expectedly, the improved patient outcomes were accompanied by expansion of T-cell clones specific for the neoantigens comprised in the autogene cevumeran vaccine. In responders, autogene cevumeran expanded these clones from undetectable levels to up to 10% of all T cells in the blood (71). Though these results are extremely encouraging as is, modifying the vaccine to effectively engage both CD8⁺ and CD4⁺ T-cell populations may reduce the proportion of treatment nonresponders even further and broaden the reach of this therapy. Limitations of strategy: Developing personalized cancer vaccines requires resource-intensive steps such as RNA sequencing of each individual’s tumor to identify suitable neoantigens for vaccine development. Therefore, this strategy may be difficult to scale to larger patient populations due its costly nature and lengthy manufacturing time.

The microbiome has been shown to have a seemingly dual role in dictating tumor fate for many cancer types. Several studies have indicated that enrichment of certain bacterial strains within the gut microbiome modulate response to anticancer treatments, including immunotherapy. Increasing evidence suggests that the gut microbiota exerts a significant influence on the metabolism of chemotherapeutic agents and the TME in pancreatic cancer. This interplay can ultimately impact the efficacy of conventional chemotherapy as well as immunotherapy interventions. Akkermansia municiphilia has been shown to both serve as a predictor and mediate therapeutic response to anti–PD-1 blockade in lung and kidney cancers (72, 73). Further, the byproducts of microbial metabolism, referred to as metabolites, can tailor antitumor immunity and enhancing efficacy of immunotherapies (74–76). In contrast, the microbiome can contribute to therapeutic resistance and progression of cancer, notably through downregulation of immune checkpoint ligands and driving chronic inflammation (77, 78).

Advances of strategy:The role of the microbiome in PDAC, however, has been singularly implicated with antitumor immune responses. Long-term survivors of PDAC were observed to have a microbiome “signature” that is not present in short-term survivors (79). The tumor microbiome of long-term survivors had greater tumor microbial diversity than short-term survivors, as well as enrichment of Pseudoxanthomonas, Saccharopolyspora, Streptomyces, and Bacillus clausii taxa. The presence of these taxa was also associated with infiltration and activation of GZMB+ CD8⁺ T cells (79). This tumor microbiome “signature” may function as a future predictor of response to immunotherapy in PDAC; however, further investigation is needed. Accordingly, microbial metabolites produced from the microbiome have also contributed to improved therapeutic efficacy in PDAC. Microbial metabolites trimethylamine N-oxide rendered tumors sensitive to anti–PD-1 blockade through reprogramming TAMs to an immunostimulatory MHCII⁺ CD86⁺ phenotype (80). Pentanoate, a short-chain fatty acid, also enhanced the antitumor efficacy of ROR1-targeting CAR-T (81). Beyond immunotherapies, tryptophan metabolite 3-indole-3-acetic acid improved PDAC sensitivity to FIRINOX treatment through ROS accumulation and reduced tumor cell autophagy (82). The transplantation of fecal microbiota from long-term survivors of PDAC resulted in the restoration of a treatment-friendly immune microenvironment in tumor-bearing mice (79). These findings shed light on the potential of microbiota as a promising therapeutic intervention for PDAC as depicted in Fig. 2. Limitations of strategy: Defining a microbiome-based biomarker to predict therapeutic response in cancer has been largely hindered due to a lack of consensus on which microbial species or metabolites must be present. In addition, efficacy of microbiome-targeting treatments may be hindered when administered concomitantly with antibiotic treatment.

Despite the advent of various therapies for PDAC, including immunotherapy—a promising approach in numerous cancer types—its efficacy in PDAC treatment has been limited due to resistance mechanisms that hamper its effectiveness. The TME of PDAC plays a pivotal role in these resistance mechanisms, fostering an immunosuppressive milieu characterized by a dense desmoplastic stroma composed of CAFs, pancreatic stellate cells, ECM components, and immunosuppressive cell populations. Genetic alterations in KRAS and immunosuppressive factors further exacerbate this environment, leading to diminished numbers of CD8⁺ T cells and reduced expression of activation markers such as IFNγ. Consequently, there is a pressing need to explore novel strategies capable of overcoming these resistance mechanisms. Approaches such as modulating the TME through blockade of receptors and stromal molecules implicated in resistance, employing genetic manipulation targeting specific regions like microRNAs, and modulating extrinsic and intrinsic factors associated with T cells represent promising avenues for enhancing PDAC therapy efficacy. Here we elucidate the primary mechanisms underlying immunotherapy resistance in PDAC and propose innovative strategies to manipulate these processes (Fig. 3), ultimately leading to more effective therapies for patients with cancer and a reduction in the lethality of this aggressive disease.

There is mounting evidence suggesting that genetic and epigenetic factors could contribute to the development of immunotherapy resistance in PDAC, although the existing literature offers limited insights into this aspect. Several studies have hinted at the association between mutations in the p53 gene and modifications in the innate immune response, which could potentially drive tumorigenesis and foster resistance to immunotherapy in PDACs (83). Specifically, the Trp53R172H mutation observed in PC cells has been identified as a facilitator of neutrophil accumulation, a phenomenon that could bolster resistance mechanisms against immunotherapeutic interventions. These findings underscore the intricate interplay between genetic alterations and immune responses in shaping the resistance landscape of PDAC, shedding light on potential avenues for targeted therapeutic interventions (84).

The genetic landscape of PDAC is extensively characterized, with mutations commonly occurring in genes such as CDKN2A, MLH1, BRCA2, ATM, KRAS, and BRCA1. Among these, the mutation of KRAS stands out as the most prevalent oncogenic alteration observed in PC cells. Additionally, several tumor suppressor pathways are frequently inactivated genetically, including the INK4a/ARF (p16), TP53, and DPC4/Smad4 pathways. These genetic aberrations play pivotal roles in driving the initiation and progression of PC by dysregulating critical cellular processes such as cell cycle control, DNA repair, and signaling pathways associated with tumor suppression. Understanding the intricate genetic landscape of PC is crucial for elucidating its pathogenesis and developing targeted therapeutic approaches aimed at combating this deadly disease.

Canonically, mutant KRAS is considered a primary genetic driver of PDAC tumorigenesis, given its involvement in pathways that regulate cell growth, proliferation, and survival (85). More recently, however, mutant KRAS has been characterized as possessing an immunomodulatory quality, notably as a suppressor of tumor immunity. When challenged with KRAS knockout KPC cells, >80% of mice failed to develop tumors over a span of 3 months, suggesting the tumors are unable to evade immune detection without intact KRAS (86). The lack of immunosuppression in KRAS knockout tumors was further confirmed through an increase in tumor infiltrating CD3⁺, CD4⁺, CD8⁺, and CD45R⁺ T and B cells, further indicating the tumor is no longer immunologically “cold.” KRAS-mediated tumorigenesis and immunosuppression was partially recovered through the introduction of oncogenes Myc and Braf^V600E, warranting further investigation into the immunomodulatory properties of these cancer-driving genes as well. The immunosuppressive nature of KRAS has also been defined utilizing chemical inhibition of KRAS^G12D. MRTX1133 is a highly selective, noncovalent inhibitor of KRAS^G12D and is currently being investigated in phase I/II clinical trials. Single-cell RNA sequencing of KPC689 and KPPC tumors following MRTX1133 treatment revealed an increase in tumor-infiltrating naïve CD4⁺ and CD8⁺ T cells in addition to effector CD8⁺ T cells (87). Further analysis of the CD8⁺ T-cell population, showed enhanced T-cell activation, with increased expression of Ifng, Prf1, Tbx21, and Gzmb in MRTX1133-treated tumors. From this analysis, it can be gathered that active KRAS in PDAC renders the tumor less hospitable to these effector cell populations, highlighting its role in sustaining a suppressive TME (87). The epithelial-to-mesenchymal phenotype is perhaps a less-surprising contributor to immunotherapy in PDAC. To investigate the connection between EMT and immunotherapy resistance in PDAC, cell lines were isolated from murine tumors that experienced therapeutic relapse following checkpoint inhibitor therapy and were subsequently inoculated in naïve mice. As anticipated, these tumors were also resistant to checkpoint inhibitor therapy and presented a highly enriched EMT gene signature (88). Gain-of-function studies in 4662 PDAC tumors revealed that overexpression of EMT-associated genes Zeb1 and Snail abrogated immunotherapy response and worsened survival. Bulk RNAseq of these tumors revealed overexpression of Zeb1 and Snail resulted in transcriptional silencing of IRF6, which is responsible for T-cell–mediated, TNFα-induced apoptosis of tumor cells.

Galectins are a family of β-galactose–binding lectins. These proteins have a number of functions, from modulating intracellular signaling and immune responses to promoting inflammation and fibrosis. Galectin-9 (Gal-9) is associated with the tumor occurrence or metastasis (89–91). Gal-9, which, unlike other galectins, plays a role in both promoting and inhibiting tumor growth, depending on interactions with T cells, antigen-presenting cells, or receptors on tumor cells (92). Moreover, accumulating evidence has demonstrated that galectins can directly influence both innate and adaptive anticancer immunity by glycan receptors (93, 94). Gal-9 specifically interacts with one of its receptors, T-cell immunoglobulin and mucin-domain containing-3 (Tim-3), and induces CD8⁺ T-cell apoptosis (95, 96). High levels of Gal-9 expression have been observed in human PDAC, with concomitant elevation detected in the immune cells present in patients’ bloodstream. Notably, the administration of neoadjuvant chemotherapy has been shown to significantly diminish Gal-9 expression on pancreatic carcinoma cells (97). Moreover, patients with longer survival durations (>12 months) exhibited notably lower serum Gal-9 levels compared to short-term survivors. Gal-9 has been implicated in driving the polarization of macrophages toward the M2 phenotype, while concurrently suppressing the secretion of pro-inflammatory cytokines such as TNFα and INFγ by T cells, thereby potentially facilitating tumor growth. Interaction between Gal-9 and its receptor, Dectin 1—expressed abundantly on macrophages within PDAC—leads to the induction of tolerogenic macrophage programming and adaptive immunosuppression (98). Conversely, blockade of Gal-9 has demonstrated efficacy in inducing tumor regression and prolonging survival rates. Furthermore, studies evaluating the efficacy of CAR-T in PDAC have revealed that combining CAR-T with a biological Gal-9 inhibitor significantly enhances CAR-T cytotoxicity against PDA cells, highlighting the immunosuppressive role of Gal-9. Additionally, infiltrating γδ T cells within PDA, a subset of CD3⁺ CD4⁻ CD8⁻ lymphocytes, have been found to release immunosuppressive cytokines like IL10 and IL17, while also expressing Gal-9 and PD-L1, thereby establishing a potent immunosuppressive microenvironment conducive to PDA development and progression (99). Interestingly, exogenous recombinant human Gal-9 (rh-Gal-9) exhibits an opposing effect to endogenous Gal-9, potentially inducing apoptosis through cytochrome release, and modulating miRNA expression to inhibit tumor cell proliferation. These findings suggest that exogenous rh-Gal-9 may hold promise as a novel therapeutic agent for pancreatic cancer (100).

In human PDAC tumors, high expression of galectin-4 was localized to the stroma surrounding the tumor and was also negatively associated with patient survival. Accordingly, galectin-4–competent mice experienced reduced survival and increased tumor size compared to galectin-4 knockdown tumors (101). Additionally, galectin-4–competent mice had increased inflammatory CAF populations compared to galectin-4 knockdown tumors. Given that galectin-4 was primarily localized to the PDAC stroma, and its ability to alter CAF subtypes, this suggests galectin-4 may contribute to therapeutic resistance through stromal modulation. The direct role of galectin-4 on PDAC stroma maintenance, however, has yet to be elucidated. Other members of the galectin family have also been implicated in PDAC tumorigenesis. Akin to galectin-4, high expression of Gal-9 in PDAC tumors was associated with reduced survival in human PDAC (98). In preclinical models, Gal-9 has been described as a key driver of immunosuppressive cell populations in nasopharyngeal carcinoma tumors. When ligated with C-type lectin dectin-1, the dectin-1–Gal-9 axis was responsible for drastic suppression of T-cell immunity in KPC tumors (98). The immunosuppressive role of the dectin-1–Gal-9 axis was confirmed with depletion of both components of the axis leading to increased effector memory CD44⁺ CD4⁺ T cells. In both KPC and orthotopic PDAC models, the blockade of Gal-9 alone resulted in improved survival and reduced tumor size. Importantly, Gal-9 blockade rendered orthotopic KPC-derived tumors sensitive to anti–PD-1 therapy, further establishing the role of Gal-9 as a driver of immunosuppression and therapeutic resistance in this context (98).

Aberrant activation of intracellular signaling pathways is characteristic in cancer, as constitutive activation of these pathways is needed to sustain the pro-survival, invasive, and proliferative nature of these cells. Given that KRAS mutations are present in 85% of PDAC tumors, constitutive activation of the Ras-Raf-MEK-ERK signaling pathway is a key driver of cancer progression and resistance to ICB (85). Targeting this signaling pathway with MEK inhibitor trametinib in combination with multi-kinase inhibitor nintedanib (T/N) resulted in significant reduction in tumor volume of up to 40% in murine orthotopic models of classical and mesenchymal PDAC (102). Further, this combination therapy improved overall survival in both PDAC subtypes. Notably, (T/N) combination therapy sensitized mesenchymal PDAC tumors to PD-L1 blockade, such that this combination induced tumor regression of up to ∼80%, highlighting its ability to reverse MEK signaling–mediated therapeutic resistance. Another established characteristic of cancer is unregulated cell proliferation. Polo-like kinase 1 (Plk1) is a serine/threonine kinase that is recognized for its role in mitosis, specifically for inducing mitotic entry and spindle assembly. In the context of cancer, overactive Plk1 signaling has been recognized to contribute to driving excessive cell proliferation. Inhibition of Plk1 resulted in increased PD-L1 expression on Panc1 cells, through promoting nuclear translocation of NF-κB (103). Importantly, both anti–PD-L1 blockade and treatment with Plk1 inhibitor BI2536 as monotherapies had minimal effects on tumor progression. However, the combination of anti–PD-L1 blockade and BI2536 acted synergistically, significantly reducing orthotopic KPC tumor volume and weight. In addition, this combination treatment increased CD4⁺ T cells, CD8⁺ T cells, and DCs within the tumor (103). Taken from these findings, abrogation of Plk1 signaling renders PDAC tumors sensitive to ICB.

DC dysfunction can lead to impaired immune surveillance against cancer cells, which can accelerate tumor progression. In mouse models, pancreatic tumors have fewer and less-active DCs than lung tumors (104). Overcoming conventional DC (cDC) deficiency in early-stage PDAC leads to disease restraint, while restoration of cDC function in advanced PDAC restores tumor-restraining immunity and enhances responsiveness to radiation therapy. Without DCs, other immune cells in the tumors do not recognize cancer cells as a threat. A recent study in murine pancreatic cancer demonstrates that DC paucity can lead to dysfunctional immune surveillance against an engineered model neoantigen, accelerating neoplastic progression (104). Human patients with PDAC have low numbers of DCs that become rarer with tumor progression (105, 106). cDC2s are a heterogeneous population that can have tumor-suppressive roles based on the inflammatory context (107–109). CD11b⁺ TAMs/DCs that skew immunity toward T-helper 2 (Th2) responses have been described both in the PDAC TME and metastases (110, 111). Recent studies have highlighted the tumor-permissive functions of CD11b+ DCs mediated by FOXP3+ Tregs or FOXP3-negative regulatory TR1 cells (112, 113). While our focus has primarily been on the cDC subsets, cDC1 and cDC2, emerging research indicates a greater diversity and adaptability within these populations (114). It is imperative to comprehensively map the distinct distribution and localization patterns of cDCs and investigate their implications on the pathogenesis and therapeutic responses of PDAC.

Investigations into cDC1s within the genetically engineered mouse model (GEMM) of melanoma driven by B-Raf/PTEN^−/− mutations have shed light on intrinsic mechanisms within cancer cells that suppress and exclude cDC1s, such as β-catenin signaling pathways (115, 116). Our examination delves into the dysregulation of cDC1s at the onset of carcinogenesis and its impact on T-cell activation. Dysfunctional or absent DCs can exacerbate nonproductive Th cell responses (117, 118). Consistent with this notion, the mobilization of cDCs during precursor lesions of pancreatic intraepithelial neoplasia stages has been associated with decreased pathogenic Th17 activity in the pancreas, thus impeding disease progression and highlighting a protective Th1 response during early stages (119, 120). The increased influx of cDCs observed in our study coincided with a reduction in collagen deposition, potentially enhancing antigen sampling and trafficking to draining lymph nodes [Hugues (121)]. These findings suggest that the TME of PDAC retains the capacity for Th1 and CTL activity but is hindered by inadequate cDC presence. In summary, augmenting the quantity and/or functionality of the limited DC population in PDAC represents a promising avenue for enhancing the efficacy of cytotoxic or immune-based treatments, which often exhibit limited effectiveness when used in isolation against this particular malignancy.

PDAC often presents with metastasis, making the cancer unresectable at diagnosis. For the minority eligible for surgery, it remains the only curative option. While conventional treatments like chemotherapy extend survival, PDAC still has a dismal 5-year survival rate (38, 116, 122, 123). Mono-immunotherapy has largely disappointed in PDAC due to its complex and heterogeneous TME (124). This field of active study is still in its early stages, with much yet to be understood on a molecular and clinical level. PDAC lags behind other solid tumors in unlocking the benefits of immunotherapy primarily due to gaps in understanding disease mechanisms and the intricate interplay between immunotherapy, chemotherapy, and the TME (125). Additionally, identifying predictive and prognostic biomarkers is critical for patient stratification and effectively targeting therapies across multiple pathways (126). Without this knowledge, the clinical application of precision therapies like immunotherapy in PDAC will be limited by the existing knowledge gap (127). Despite these challenges, ongoing research efforts and academic perseverance are gradually laying the groundwork for a future where PDAC may be managed more effectively. Biomarkers hold promise as tools for predicting prognosis and treatment response. However, their exploration is hindered by the difficulty in obtaining adequate tissue samples for analysis. Despite these challenges, there is a pressing need for large multicenter studies to identify biomarkers that can accurately determine therapeutic efficacy in PDAC. As we continue to expand our understanding and refine therapeutic approaches, there is hope for improved outcomes and a shift towards viewing PDAC as a manageable disease state. Given the close relationship between the immune landscape and the intratumoral microbiome in PDAC, it would be of great interest to explore the role of microbiome biomarkers and the efficacy (or lack of) of immunotherapeutics in trials (128, 129). Less-invasive approaches, such as analyzing peripheral TCR repertoires, offer potential insights into treatment response without the need for invasive tissue sampling. Furthermore, understanding the complex interplay between the tumor immune landscape and the intratumoral microbiome could reveal novel biomarkers for predicting therapy response. The field of immunotherapy in PDAC is still in its early stages and there are still many significant discoveries to be made. It is crucial to have a comprehensive understanding of clinical, immune, and molecular levels to untangle the complexities of PDAC and improve treatment strategies. Recent advancements in understanding pancreatic cancer have shown that the overall prognosis is still poor. Unlike other solid cancers that have been greatly impacted by immunotherapy, PDAC continues to present challenges due to its immunologically cold nature and low mutational load. The limited success of immunotherapy in PDAC can be attributed to various factors, including the obstacles presented by the TME, which is characterized by immunosuppression and dense desmoplastic reactions (130–132). Furthermore, the late stage at which PDAC is usually diagnosed with metastasis adds another layer of complexity by creating additional barriers for the immune system. Understanding how immunotherapy responds in different subpopulations of PDAC remains a significant challenge (133, 134). The heterogeneous nature of the TME plays a crucial role in determining tumor behavior, so it is essential to uncover the underlying mechanisms of response. While some subsets of patients show promising outcomes with immunotherapy, it is still unclear why some individuals respond while others do not. Looking ahead, personalized approaches like mRNA vaccines that target neoantigens demonstrate the growing sophistication in PDAC therapeutics (135, 136). Although combinations of immunotherapy offer promise, it is important to address logistical challenges, manage toxicity, and optimize treatment regimens to fully realize the potential of immunotherapy in PDAC.

No disclosures were reported.