Immunosuppression, immune escape, and immunotherapy in pancreatic cancer: focused on the tumor microenvironment

Abstract

Pancreatic ductal adenocarcinoma (PDAC), the most common type of pancreatic cancer, is characterized by poor treatment response and low survival time. The current clinical treatment for advanced PDAC is still not effective. In recent years, the research and application of immunotherapy have developed rapidly and achieved substantial results in many malignant tumors. However, the translational application in PDAC is still far from satisfactory and needs to be developed urgently. To carry out the study of immunotherapy, it is necessary to fully decipher the immune characteristics of PDAC. This review summarizes the recent progress of the tumor microenvironment (TME) of PDAC and highlights its link with immunotherapy. We describe the molecular cues and corresponding intervention methods, collate several promising targets and progress worthy of further study, and put forward the importance of integrated immunotherapy to provide ideas for future research of TME and immunotherapy of PDAC.

Introduction

Pancreatic cancer is widely known for its high risk and mortality. If not examined regularly, about 50% of the patients are diagnosed at its locally advanced stage or even late stage with metastasis because of the insidious early symptoms of most pancreatic cancers [1, 2]. According to the American Cancer Society statistics, pancreatic cancer accounted for 3% of all types of cancer in 2020, but its 5-year survival rate is only 9% [3]. Although the morbidity and mortality of pancreatic cancer vary significantly across countries with varying levels of socio-economic development, the overall 5-year survival rates are low, ranging from 5 to 10% worldwide [4]. Therefore, pancreatic cancer is regarded as one of the most dangerous and challenging tumors in clinical practice [5], and its relatively poor prognosis has drawn worldwide attention. Since pancreatic ductal adenocarcinoma (PDAC) accounts for about 90% of pancreatic tumors, most studies and clinical practices of pancreatic cancer focus on PDAC [2, 6].

Currently, surgery is the most effective clinical treatment for most PDAC patients. However, surgery may not benefit patients with late or locally advanced diseases. Other treatments include adjuvant chemotherapy, radiotherapy, targeted therapy, and immunotherapy. In adjuvant chemotherapy, the current FOLFIRINOX, gemcitabine (GEM) combined with cisplatin, and GEM plus albumin paclitaxel regimens have shown specific efficacy. Although drugs targeting point mutations or genotypes have provided certain benefits, chemotherapy and targeted therapy for PDAC are not as effective as expected, and the progress is limited [7–10].

Immunotherapy includes immune checkpoint blockade (ICB, by immune checkpoint inhibitor, ICI), vaccines, adoptive cell therapy like chimeric antigen receptor T cell (CAR-T), and others [11]. Although immunotherapy has produced remarkable results in tumors such as melanoma, NSCLC, and renal cell carcinoma since the ICI ipilimumab in 2011, the effect in the treatment of PDAC has been disappointing. Even though 11 ICIs and 2 CAR-T products have been approved to treat 16 types of malignant diseases and one tissue-agnostic indication, the only preferred or recommended immunotherapy for PDAC is pembrolizumab, and only for patients with the Microsatellite Instability-High/Deficient Mismatch Repair molecular subtype (MSI-H/dMMR, an indicator or predictor for immunotherapy efficacy. The condition is known as dMMR if one or more proteins of the key genes dependent on the MMR system are not expressed or are dysfunctional. Otherwise, it is known as mismatch repair proficient, pMMR. While MSI results from the inactivation of MMR genes and MMR protein dysfunction, the tumor is classified as MSI-H if two or more repeats at the MSI diagnostic site are altered. If only one, it is classified as MSI-L. Otherwise, the tumor is classified as having microsatellite stability MSS) [7, 9–16]. Unfortunately, the MSI-H subtype only has a prevalence of 0–2.00% in PDAC [17], which means an extremely low number of patients who may have the chance to be sensitive to the extant immunotherapy, and the possible low therapeutic response. Therefore, it is vital to study immunotherapy against PDAC, and the current focus is on researching the mechanism of immunotherapy resistance and improving the therapeutic effect through the TME of PDAC.

Apart from the fact that the majority of tumors are already in the advanced stage at the time of diagnosis and treatment, certain factors that interfere with immunotherapy have been raised [5, 18], such as the unique genomic landscape of PDAC shaped by oncogenic drivers, which promotes immune suppression from the earliest stages of tumor inception to subvert adaptive T cell immunity [19]. Furthermore, PDAC metastasis, which would create additional immune barriers along the cancer-immune incline, could obscure signs of drug activity [20]. Last but not least, mechanisms of the tumor microenvironment through which the tumor progresses and escape from immune attack are also worth noting [21].

This article will review the progress of tumor microenvironment research and the progress of PDAC immunotherapy.

Tumor microenvironment of PDAC

Stephen Paget once proposed the famous "seed-soil theory" regarding tumors and the microenvironment. Tumor cells are the "seeds," and the tumor microenvironment is the "soil" The interaction between tumor cells and the tumor microenvironment contributes to tumor progression [16, 22, 23]. The theory and definition of TME have gradually been perfected through numerous pieces of research.

Tumor microenvironment (TME) refers to the environment in which tumor cells initiate and grow. It also includes microvessels and the surrounding cells: immune cells, fibroblasts, glial cells, endothelial cells, neurons, the nearby matrix, and various biomolecules such as growth factors, chemokines, and various proteolytic enzymes. Together, they constitute a micro internal environment with a persistent immuno-inflammatory response and hypoxia, low pH, and high pressure. These characteristics are conducive to tumor growth and invasion, angiogenesis, therapeutic resistance, and promoting the occurrence and development of tumors [24].

Immuno factors play an essential role in the tumor microenvironment, including immune cells, cytokines, and related inflammatory reactions. Among them, immune cells such as granulocytes, monocytes, macrophages, and dendritic cells play an essential role in cancer cell recognition, inflammatory initiation, and anti-tumor response [25, 26] and are the present key to research on the regulation of TME. In addition, different tumor microenvironments differ in the degree of cellular infiltration, composition, and inflammatory response [22].

Notably, the tumor microenvironment is inextricably linked to the hallmarks of tumor. Not only are most of the classic six hallmarks in 2000 (sustaining proliferative signaling, evading growth suppressors, resisting cell death, enabling replicative immortality, inducing/accessing vasculature, activating invasion and metastasis) correlate with the TME, but the previously compromised two hallmarks in 2011 (deregulating cellular metabolism and avoiding immune destruction) are also obviously relevant to the TME as well as its immune factors. Four new parameters (unlocking phenotypic plasticity, non-mutational epigenetic reprogramming, polymorphic microbiomes, and senescent cells) are reviewed in the most recent update in 2022, all of which are closely related to the dynamic balance and heterogeneity of the TME [27]. Therefore, there is continuously increasing and widely appreciated evidence demonstrating that the tumor microenvironment plays an integral role in tumorigenesis and malignant progression, let alone in the effect of immunotherapy. Immunotherapy has limited benefits in PDAC due to the tumor microenvironment.

Lack of immunogenicity and low immune response, which indicates insufficient immune activation and excess immune suppression, is a significant feature of the TME in PDAC [5, 13, 19, 28, 29]. The complex and unique mechanisms in the local environment of PDAC can directly influence the biological behavior of tumor cells. Tumor cells further affect the myeloid and lymphoid systems and the stroma. Under unique regulation, various components of TME interact with each other, resulting in connective tissue proliferation and immunosuppression, directly and indirectly, which creates a highly dynamic and hypovascular TME [24]. This leads to different immune evasion mechanisms and further promotes the primary tumor's invasion, progression, and metastasis [30]. Compared with other immunotherapy-responsive tumors, the dominant defect in the TME of PDAC is insufficient activation because it is less antigenic and has lower immune activation "potential energy" [20].

Another crucial feature is the high heterogeneity of TME among PDACs. Collisson et al. classified PDAC into three subtypes with different local characteristics: classical, quasi-mesenchymal, and exocrine-like in 2011 [31]. Bailey et al. defined a 4-subtype classification based on genomic expression analysis, including squamous, pancreatic progenitor, immunogenic, and aberrantly differentiated endocrine exocrine (ADEX) [32]. Many other classifications have been based on genetic expression, histology, pathology, and other factors [33, 34]. The currently most accepted classification divides pancreatic cancer into two major subtypes: classic and basal-like, with different histological features, treatment response, and overall survival [35]. Francesco et al. considered TME more directly by microenvironment factors in 2018 and divided PDAC into five clinical subtypes: Pure-basal-like, Stroma-activated, Desmoplastic, Pure-classical, and Immune-classical. Those four subtypes have different immune cell infiltration rates, stromal components, and immune checkpoint expression, indicating that their TMEs are subtype-specific and vary in response to treatment [24, 36].

The third important feature of the PDAC microenvironment is its rich and dense desmoplastic stroma, accounting for 70–90% of the tumor volume [37]. It is a complex network that consists of a variety of extracellular matrix (ECM) and stromal cells, including cancer-associated fibroblasts (CAFs, described in detail below), vascular-associated smooth muscle cells, pericytes, endothelial cells, mesenchymal stem cells, immune cells, neurons and blood vessels [38–40]. Fibroblasts respond to carcinogenic signals early in the pancreas, so low-grade precancerous pancreatic intraepithelial neoplasia (PanIN) is already surrounded by fibrotic tissues [24, 41]. When the tumor forms, the surrounding stroma has already wrapped it into a physically hard tumor [24]. Even at the metastatic sites, there can be a significant increase in connective tissue proliferation-promoting stromal components such as hyaluronic acid and collagen [42]. The stroma can promote the occurrence and development of PDAC by regulating various factors in the TME, including ECM remodeling, angiogenesis, metabolic competition, inflammation, and immune response. Its primary impact is the protective physical and chemical barrier formed by the stroma accumulated by excessive fibrosis. The stromal pressure around the tumor compresses and collapses the blood and lymphatic vessels, resulting in a nutrition-deficient, hypoxic, and acidic microenvironment. Concurrently, the collapsed blood vessels and specifically increased interstitial fluid pressure (IFP) in the TME hinder drug delivery and diffusion [42]. As a result, the dynamic balance in TME is disrupted, which further promotes ECM remodeling, local immune cell dysfunction, and inhibition of immune infiltration, resulting in the formation of a microenvironment in tumor tissue that promotes tumor growth, angiogenesis, tumor invasion, metastasis, and immunosuppression and drug resistance [1, 13, 39, 40, 43]. The stroma also has other characteristics, such as the abundant proteolytic enzymes in the ECM, which can support frequent and complex matrix remodeling in tumorigenesis and progression and intra-tumor heterogeneity of the stroma [42]. Therefore, the PDAC stroma is often considered one of the critical reasons for insufficient response to ICB therapy. Accumulated studies have documented the anti-stroma therapy, such as transforming growth factor-β (TGF-β), focal adhesion kinase (FAK) signaling, glutamine metabolism, etc., and pointed out that the synergistic effect of anti-stroma therapy and immunotherapy may improve the therapeutic efficacy [44]. Therefore, the stroma is a potential target for alleviating or reversing the immunosuppressive and tumor-promoting TME [13].

In summary, to understand the TME of PDAC, we must recognize that it has the general TME characteristics of solid tumors, clarify its dynamic changes with the tumor progression, and acknowledge the heterogeneity of the TME of PDAC (Fig. 1).

Fig. 1.

A glance into the components and mechanisms of PDAC tumor microenvironment

The tumor microenvironment of PDAC is a complex in which tumor cells are infiltrated with various kinds of immune cells and molecules, and are wrapped with dense stroma. The cell components include tumor cells, T cells, macrophages, neutrophils, MDSCs, mast cells and stromal cells like CAFs. The hyper heterogeneity of TME is represented in the populations of each kind of immune cells. Different subtypes of immune cells are recruited and induced by various factors, and may demonstrate diverse functions in immune or stroma regulation by cytokines, exosomes or other ways of crosstalk. The complexity of the TME is made up of these mechanisms of the components.

Immune cells in the PDAC microenvironment

Before summarizing immunotherapy, it is vital to understand the cells in the TME of PDAC. The PDAC microenvironment is in a state of "local inflammation" [45], in which tumor cells interact with infiltrating immune cells, including neutrophils, mast cells, tumor-associated macrophages (TAM), myeloid-derived suppressor cells (MDSC), T cells and many other immune cells [16].

In addition to immune cells, fibroblasts in stromal components, known as CAFs, are an important component and role-player of TME. Fibroblasts also regulate the immune activities of TME by diverse mechanisms, which are closely related to immunosuppression and tumor immune escape.

T cells

T lymphocytes play an important role in adaptive immunity and are crucial in tumor immunity. T cells play anti-tumor and pro-tumor immunity in the PDAC microenvironment [25]. T cells in the TME of PDAC include CD8⁺ cytotoxic T cells (CTL), also known as effector T cells (Teff), CD4⁺ T cells such as helper T cells (Th) Th1, Th2, Th17, regulatory T cells (Treg), and natural killer (NK) T cells [46].

T cells are patchy in human PDAC tissues, usually located in the periphery of tumors, while the proportion of T cells in peripheral blood is relatively low. The number of T cells varies from patient to patient and changes during the progression of PDAC [46–48]. T cells were primarily found in lymphoid follicles near PDAC tissue, according to the resected specimens' immunohistochemical analysis, and most infiltrating lymphocytes were CD3⁺ T cells. Among the infiltrating CD3⁺ T cells, compared with CD8⁺ T cells, CD4⁺ T cells dominate. Among the CD4⁺ T cells infiltrating PDAC tissues, there are only a few T-bet⁺ Th1, while GATA3⁺ Th2 has the majority, and another proportion of those with higher numbers are Foxp3⁺ Treg cells [45]. However, only Th1 among the above-mentioned CD4⁺ T cells is associated with good prognoses besides Teff. The increase of Th2 (often expressed as Th1/Th2 ratio) and Tregs can induce immunosuppression, correlating with reduced infiltration of CD8⁺ T cells and poor prognoses such as short disease-free time and overall survival [16, 49, 50]. As mentioned above, there are many reasons for the lack of Teff in PDAC, such as poor tumor immunogenicity caused by the low frequency of neoantigens [51] and inhibition of T cell infiltration by the stroma in distinct ways, and the influence of other immune cells.

CD8⁺ T cells are the subtype that plays positive immune effects, so they have become the focus of PDAC TME and immunotherapy research. Many studies have pointed out that CD8⁺ T cells may be independent in evaluating the PDAC microenvironment or immunotherapy response. CD8⁺ T cells produce IFN-γ, TNF, and cytotoxic molecules against tumor cells [46]. The increase in the number, recruitment, infiltration, and activation of CD8⁺ T cells is related to effective treatment and improved survival rate [52, 53]. However, even if Teffs exist in the TME, they may be disabled or have poor cytotoxicity, supporting the characteristics of strong immunosuppression in the PDAC tumor microenvironment [54–56].

CD4⁺ T cells can recognize the polypeptides presented by MHC II molecules expressed on the surface of antigen-presenting cells such as macrophages or dendritic cells (DC) and differentiate from immature ones to helper T cells [57]. There is also a high degree of heterogeneity in this population. Th1 cells can produce IFN- and cytotoxic molecules with anti-tumor properties. Th2 cells can produce IL-4, IL-5, IL-3, and IL-13 and induce M2 polarization of tumor-associated macrophages, which is associated with immunosuppression and tumor progression [48]. Th17 produces pro-inflammatory cytokines 1L-17, IL-21, and IL-22, in which IL-17 induced by KRAS and IL-17RA expressed by PanIN can induce the characteristics of tumor stem cells through IL-6/pStat3 and NF-κB signals, and promote tumor progression [47].

Treg is a vital subtype of CD4⁺ T cells, whose most common feature is FOXP3. Tregs can maintain tolerance to self-antigens in healthy tissues. However, in pathological situations, IL-10 and TGF-β secreted by Tregs can decrease CD8⁺ and CD4⁺ T cell subtypes, and Tregs can compete to obtain the antigens presented by DCs, which finally create an immunosuppressive environment and promote tumor progression [47, 57]. Tregs seem more preferentially located in the stroma of PDAC than in tumor lesions, and the main distribution is in the associated lymph nodes around PDAC, as described above [51]. Tregs' count gradually increases from PanIN to high-level PDAC, and the increased count is associated with low survival. However, studies on the mechanisms of Tregs have also produced some exciting results. For example, Zhang et al. found that Treg deletion could not alleviate the microenvironmental immunosuppression of PDAC. On the contrary, the TME lost the vital source of TGF-β, which affected myofibroblastic CAFs (MyCAFs), an inhibitory subtype of CAFs, and increased myeloid cell recruitment due to the pathological increase of CCL3, CCL6, and CCL8, resulting in immunosuppression recovery and promotion of carcinogenesis [57, 58]. Therefore, the complexity of Treg warrants further investigations.

As the main immune effectors, T cells are not only regulated by the autoregulatory subtypes but also have complex crosstalks with other components of TME, which will be described in detail below. The tumor cells, other immune cells, and various components in the stroma, such as CAFs, have crosstalks with T cells. For example, FAK, TAMs, and T cells have a variety of interconnections. DCs have multiple abilities to regulate T cell subtypes. B cells can also regulate TME T cells, whose specific mechanism remains unexplored [51].

Immunotherapy targeting T cells is a long-term topic. At present, immunotherapy targeting T cells has been rather complex and in-depth. Emerged studies have focused on exploring the pathways from upstream or crosstalks of TME components, blocking multiple immune checkpoints, or improving the efficacy of adoptive cell therapy. For example, Louis et al. found that STAT1 induces gene expression in immunosuppressive T cells and myeloid cells. Ruxolitinib, which targets the Janus Kinase (JAK)/STAT signaling pathway, alleviates immunosuppression and improves the therapeutic effect of anti-PD-1 drugs [59]. Knudsen and colleagues systematically screened new combination therapies for PDAC based on live-cell imaging. They demonstrated that mitogen-activated protein kinase 1 (MEK) inhibitor trametinib and CDK4/6 inhibitor palbociclib significantly affects increasing T cell infiltration and myeloid population changes, enhancing the effect of anti-PD-L1 and regulating T cells and other immune cells by interfering with the cell cycle. This increases immune activity through CD8⁺ T-dependent and atypical Rb-dependent pathways [60, 61]. Cell engineering modification of T cells in adoptive cell therapy is also a hot spot. For example, according to Kobold et al., overexpression of CXCR6 in adoptive T cells to PDAC improves infiltration into tumor tissues, promotes adhesion to tumor cells, and improves therapeutic effect [62, 63].

As the most important effector cells in the PDAC TME, T cells have many subtypes with different functions in their population and numerous crosstalks with other components. Thus, T cells will continue to be the irreplaceable research object of various immunity mechanisms or targeted therapies (Fig. 2).

Fig. 2.

Mechanisms of T cells and related basic immune pathways

In the peripheral or peri-tumoral lymphoid tissue, the inactivated T cell surface receptor TCR is activated by MHC class I/II molecules brought by antigen-presenting cells. The activated CD4⁺ T cells can further secrete cytokines such as IL-2 to promote their proliferation and CD8⁺ T cells. When recruited and migrated into the tumor tissue, various immune cell infiltration can be observed in the TME. Among them, CD4⁺ Th1 cells can secrete cytokines such as IL-2, IFN-γ, and TNF-β to promote CD8⁺ T and M1 macrophages. Th2 cells maintain a dynamic balance with Th1 cells. In an immunosuppressive environment, Th2 cells tend to increase while Th1 cells decrease. Foxp3⁺ Treg cells have a suppressive regulatory effect on immune cells such as CD8⁺ T. The immunosuppression of tumor cells is mainly derived from the PD-L1 expression, which binds to PD-1 on the surface of effector T cells of their own and M2 TAMs. The PD-1/PD-L1 pathway (the CTLA-4/B7 pathway) generates signals to inactivate T cells, enabling immune escape and producing an immunosuppressive tumor microenvironment.

Macrophages and tumor-associated macrophages

Macrophages utilize phagocytosis to remove old or dangerous cells and cell fragments and produce cytokines, proteolytic enzymes, and metabolites. Compared with normal pancreatic tissue, macrophage infiltration in the PDAC microenvironment significantly increases. Macrophages are located near blood vessels, and their quantity is usually positively correlated with vascular density [64]. In addition to PDAC tissue, macrophages can be found in the fibrous adipose tissue around tumor cells [45]. There is even point of view by Carig Gibbs that macrophages can constitute as much as 50% of the mass of a tumor [65]. Therefore, macrophages are indisputably one of the tumor microenvironment's most abundant and essential immune cells.

Tumor cells recruit circulating monocytes to tumor tissue by secreting a series of cytokines such as CSF-1 and chemokines such as CC and CXC family [66, 67] and induce them to differentiate into macrophages. The infiltrating macrophages can be called tumor-associated macrophages (TAMs). TAM should play an anti-tumor role, but when taken as a whole, it is considered immunosuppressive and tumor-promoting in many types of tumors. Tumor cells not only have molecular camouflage and signaling mechanisms to escape and inhibit the phagocytosis of macrophages but also can induce TAMs to differentiate into immunosuppressing ones. In addition, TAMs can promote tumor growth by inducing new angiogenesis via angiogenic factors such as VEGF and exosomes such as miR-155-5p and miR-221-5p that target E2F2 as increasing microvessel density in tumor tissue [68]. Importantly, TAMs express immune checkpoint ligands, such as PD-L1, PD-L2, CD86, and CD80, to combine with immune checkpoint receptors, such as PD-1 and CTLA-4, to induce CD8⁺ T cell dysfunction. Moreover, TAMs release ARG-1, indoleamine 2-dioxygenase, IL-10, and TGF-β to reduce the activity of immune effector cells while recruiting other immunosuppressive cells such as Tregs. TAMs also aid tumor invasion and metastasis by remodeling the matrix with enzymes such as matrix metalloproteinases (MMP), cathepsin B, D, K, L, S, Z, and exosomes rich in Wnt. Notably, TAMs are also profoundly implicated in drug resistance, such as interfering with tumor vaccine efficacy by inhibiting the antigen presentation efficiency of DCs and affecting adoptive cell therapy by establishing a TME with a high degree of fibrosis and enriching blood vessels [46, 64, 69].

The heterogeneity within the TAM population has been proposed for a long time. The most common view is that TAMs can be functionally divided into M1 and M2 subtypes. The M1 subtype is classified as either a pro-inflammatory phenotype or a killing phenotype. Pathogen-related modes such as lipopolysaccharide, IFN-γ, GM-CSF, and TNF-α can induce M1 polarization. M1 TAMs are characterized by phagocytosis and production of high levels of pro-inflammatory cytokines such as IL-12, IL-23, TNF-α, etc., and can release exosome called M1 macrophage-derived exosome (M1Exo), to stimulate the immune activity of immune cells such as Th1 in the TME and kills tumor cells [70, 71]. M2 TAMs have been described as a tumor-promoting phenotype or a restorative phenotype. Injury-related modes such as IL-4 and IL-13 can induce M2 polarization, characterized by the production of TGF-β and IL-10, inducing Th2, Tregs, and other cells to promote immunosuppressive activity and promoting recurrence, invasion, metastasis, and drug resistance by regulating EMT, matrix repairing and remodeling. Therefore, M2 TAMs are associated with poor prognosis [71, 72]. The proportion of M2 TAMs in the PDAC tumor microenvironment is much higher than that of M1. Therefore, the above-mentioned tumor-promoting function of TAM mainly comes from M2. In addition, there is selectively maintaining positive feedback crosstalk between M2 TAM and Treg, which means that the tumor microenvironment of PDAC is often strongly immunosuppressed in a progressive manner [45, 64]. Intriguingly, there are also views that the current dichotomy of M1-M2 is too simple to describe the complex role of TAM in the TME since larger-scale techniques and data have already revealed more unknown macrophage subtypes [69].

In recent years, knowledge of TAM mechanisms in the TME has been tested. For example, Ernesto et al. found that tumor-derived sialic acid in PDAC induces the differentiation of monocytes to macrophages through Siglec-7 and Siglec-9 signaling. The triggering of Siglec-9 in macrophages reduces the inflammatory process and increases the expression of PD-L1 and IL-10 [73]. For immunosuppressive subtypes, inhibition of CSF-1/1R in PDAC leads to priority depletion of CD206^high, promotes antigen presentation ability of TAMs, and enhances anti-tumor T cell activity [fn71]. Similarly, Benoit et al. proposed that the inherent PI3KA signaling of PDAC accelerates metastasis and rearranges macrophage components. Inactivating PI3KA prevents the infiltration of CD206⁺ macrophages around high-grade lesions, reduces metastasis, and improves patient prognosis [74]. Zhang et al. proposed the immunomodulatory effect of fat, showing that E3 ubiquitin-conjugating enzyme FATS (fragile site-associated tumor suppressor) in the CFSs (common fragile sites) in tumor cells has anti-tumor effects. FATS inhibition promotes the TAM phenotypic transition from M2 to M1 in the TME, facilitating tumor regression [75]. Additionally, apolipoprotein E up-regulated by TAM promotes PDAC immunosuppression through an NF-κB-mediated production of CXCL1 and CXCL5, which further highlights the potential feasibility of targeting TAM and lipids [76]. In the crosstalk between TAM and stroma, MMP-9 secreted by macrophages in PDAC induces mesenchymal transformation through PAR1 activation, enabling tumor cells' survival [77].

The research on therapeutic targets of TAMs has also made rapid and considerable progress in recent years. These targets include chemokine-chemokine receptor signaling, such as CXC, CC, CX3C, and C, and RTK signaling such as CSF-1/CSF-1R. Besides chemokines, metabolic signaling such as glycolysis, fatty acid metabolism, glutamine synthesis, tryptophan metabolism, arginine metabolism, receptors for advanced glycation end-product, etc., is also in the spotlight. In addition, strategies targeting extracellular signaling are emerging. For example, M1Exo-based specific drug delivery systems like M1Exo-loaded gemcitabine (GEM) and destiroxacin (DFX) effectively combat the GEM resistance in PDAC [64, 66, 70]. There are three main strategies for targeting TAMs, as summarized by Xiang et al.: macrophage elimination, recruitment inhibition, and reprogramming. Bisphosphonates such as clodronate and zoledronate can be used to eliminate macrophages, but they are not as specific [69]. In TAM recruitment inhibition and reprogramming, BLZ945, a clinically relevant CSF1R inhibitor, has emerged as a key player in recent years. For example, Magkouta et al. observed an abrogation of infiltration of TAMs by using BLZ945 in mesothelioma [78]. Huang et al. detected SPON2 as a promotor for TAM migration and infiltration and enhanced programming of M2-polarization with BLZ945 [79]. Akkari L. et al. combined radiotherapy with a continuous BLZ945 treatment in glioma preclinical trials and again demonstrated that CSF1R inhibition reprogrammed macrophages. The combined therapy targeting dynamic TAM populations enhanced therapeutic sensitivity and prevented recurrence [80]. Many other attempts, such as nanoparticles and recombinant protein treatment, have been studied in macrophage reprogramming to re-educate TAM into an immune-supportive phenotype, such as the nano-complex PLGA nanoparticles by Han et al. [81] and the recombinatnt N-terminus of Slit2 protein (rSlit2-N) by Ahirwar et al. [82] Enhancement in the function of M1 TAM and altered polarization from M2 to M1 were observed in both studies.

Neutrophils

Neutrophils are the most abundant immune cells in circulation, as well as the one that constitutes the first line of protection against pathogens [83]. Neutrophils have revealed several mechanisms, including phagocytosis, oxidative burst, degranulation, and production of neutrophil extracellular traps (NETs) [84].

Neutrophils are often rapidly recruited to the cells that reach the inflammatory site. In the tumor microenvironment of PDAC, neutrophils are observed to infiltrate more than normal pancreatic tissues and chronic pancreatitis tissues, mainly located near the dense tumor-cell area [85]. The proportion of neutrophils in all infiltrating immune cells is observed in varying degrees [45, 83]. It has been proposed that an increase in neutrophil infiltration in PDAC is linked to an undifferentiated cancer type and a poor prognosis [83]. Furthermore, the infiltration rate of neutrophils can be an independent prognostic factor, such as monitoring the treatment efficacy of local tumor control [86]. The neutrophil/lymphocyte ratio (NLR) is often a negative predictor of overall survival in PDAC patients [85].

The term "tumor-associated neutrophils" (TAN) has also appeared. They are neutrophils that are recruited and infiltrated into TME by tumors. Chemokines produced by tumor or stromal cells can induce neutrophil migration. Granulocyte–macrophage colony-stimulating factor (GM-CSF), granulocyte colony-stimulating factor (G-CSF), members of the CXC family and CCL family, IL-17 [87], and tumor secreted IL-1β and TGF-β have also been revealed to be able to recruit neutrophils [84, 85].

According to its functional phenotype, TAN is divided into two subtypes: the immunostimulatory N1 subtype and the immunosuppressive N2 subtype. N1 TAN demonstrates more potent cytotoxicity, produces pro-inflammatory and immunostimulatory chemokines, and induces the recruitment and activation of immune cells, which displays an anti-tumor function. In contrast, N2 TAN is induced by TGF-β and suppresses T cell function to promote immunosuppression and supports tumor growth, invasion, and EMT process [88, 89]. TAN may show anti-tumor effects in the early invasion stage but is more likely to express a tumor-promoting phenotype later [85]. Related research has made general progress. For example, Cui et al. found that the elastase Elane secreted by neutrophils shows remarkable selectivity, which induces the death of malignant cells in the tumor and metastatic sites in animal models without affecting the proximal healthy cells. This result complemented the phenotypic function of N1 [90]. Our team recently discovered that neutrophils with the negative purinergic receptor P2RX1 in PDAC liver metastases significantly increased the level of immunosuppressive molecules PD-L1 to participate in PD-1-induced CD8⁺ T cell failure and promoted mitochondrial metabolism. As a result, these neutrophils aggravate immunosuppression and promote tumor immune escape, reflecting N2 phenotypic function [91, 92]. Notably, the heterogeneity and typing of TAN remain controversial. N1-N2 classification alone may not be able to define all TANs in TME clearly, and TANs seem to be more likely to be polarized differently by signals from TME.

A neutrophil extracellular trap is an essential form of neutrophil functioning [93]. NET is a reticular structure composed of histone-binding nuclear DNA and neutrophils released cytotoxic particles. NETsis creates a physical and chemical barrier for pathogens and immune cells. In theory, NET may have a potential anti-tumor effect by killing tumor cells through cytotoxicity or activating the immune system. However, emerging evidence shows that NET has tumor-promoting effects [84]. As a physical barrier for T cell penetration, it can exclude CD8⁺ T cells from TME and specifically block the cytotoxicity of tumor cells mediated by NK cells or CD8⁺ T cells [85].

Both neutrophils and NET are involved in several steps of PDAC metastasis, such as 1) promoting EMT through E-cadherin degradation by elastase, loosening endothelial cell connection by VEcadherin, and promoting vascular leakage and tumor cell movement; 2) promoting endoderm metastasis, migration and invasion of PDAC cells through IL-1β/EGFR/ERK pathway [94]; 3) inducing the proliferation of dormant metastatic tumor cells through matrix metalloproteinase-9 (MMP-9); 4) forming liver pre-metastatic niche through crosstalk with CAF [83, 85]. Moreover, the presence of NET can be used as a negative prognostic index of survival time and a factor to detect chemotherapy efficacy in patients with PDAC and is considered an additional index of the TNM stage in PDAC [85].

Several studies have explored the feasibility of targeting TAN or NET. For example, lorlatinib inhibited the development and mobilization of neutrophils in PDAC mice and delayed tumor progression by regulating tumor-promoting TAN in TME [95]. In addition, Hafsa et al. found a crosstalk pathway with NET and CAF, in which CAF-derived amyloid protein Amyloid β can be targeted as an important potential site. Furthermore, PAD4 inhibitor GSK484 reduces NET formation, significantly inhibiting pancreatic tumor growth in xenograft mice [93].

Therefore, neutrophils and NET are essential for PDAC development and progression and are promising therapeutic targets in PDAC.

Myeloid-derived suppressor cell

Myeloid cells, defined as CD45⁺CD11b⁺ cells, originate from hematopoietic stem cells (HSCs) in the bone marrow, followed by common myeloid progenitor cells (CMPs). Myeloid cells differentiate into different populations, such as macrophages, granulocytes, mast cells, and dendritic cells. In TME, neutrophils and monocytes are usually in an immature state. These immature immunosuppressive cells that lack the molecular signature of mature myeloid cells are called myeloid-derived suppressor cells (MDSCs) [96].

Chemokines and their receptors guide the transport, expansion, and accumulation of MDSCs in inflammatory diseases, including tumors. The recruitment of MDSCs from bone marrow to the PDAC tumor microenvironment is induced by various chemokines such as GM-CSF, IL-6, IFN-γ, CCL2, and IL-1β. These cytokines can also inhibit MDSC differentiation in pathological conditions such as chronic inflammation and tumors, increasing the number of immature cells [97].

Compared with the normal pancreas, the number of MDSCs increases in PDAC and accounts for many immune cells in the TME. Additionally, MDSCs accumulate in peripheral blood and surrounding organs such as the liver, lung, and spleen. MDSCs in the primary tumor and circulation can also increase continuously with the clinical progress of the tumor [16, 46, 97].

The classification of MDSC is still controversial. It is believed that MDSC can be divided into two main subgroups: PMN-MDSC (neutrophil-MDSC) and M-MDSC (mononuclear-MDSC), also called G-MDSC and M-MDSC. Compared with M-MDSC, PMN-MDSC accounts for a more significant proportion of the MDSC in peripheral circulation and tumor microenvironment. However, M-MDSC plays a more vital immunosuppressive function in TME [96]. Although MDSCs differ from mature cells, the distinction between PMN-MDSCs and neutrophils is controversial, requiring unique markers. New markers for classification, such as CD13^high/low, are gradually recognized [97].

MDSCs exert immunosuppressive effects in TME, and the specific mechanisms vary in different MDSC subtypes and their sites [98]. MDSCs negatively mediate immune response by producing Arg-1, iNOS, COX-2, PD-L1, and ROS under pathological conditions [96]. It is also believed that MDSCs express B7-1 to produce T-cell inhibition [99]. Hypoxic TME can further enhance MDSC-mediated immunosuppression. MDSCs also promote tumor metastasis since they can create pre-metastatic niches before colonizing tumor cells. Several excellent works have focused on the study of targeting MDSC to reverse immunosuppression. For example, Shipra Das and colleagues pointed out that inhibition of the PGE2/p50/NO axis can prevent the inhibitory function of M-MDSC and restore the effect of anti-tumor immunotherapy [100]. Fouad et al. found that targeting CD200 that promotes MDSC amplification and transplantation significantly reduced the proportion of multidrug-resistant stem cells in the TME and greatly improved the therapeutic effect of PD-1 checkpoint antibodies [101]. Yang et al. proposed that radiation-induced lactic acid plays a vital role in improving the immunosuppressive phenotype of MDSC and regulating the immunosuppressive immune microenvironment of PDAC. Moreover, targeting lactic acid from tumor cells and HIF-1α signaling in MDSC may alleviate the radiation resistance of PDAC [102].

Therefore, further research on the classification and targeting feasibility of MDSC in the tumor microenvironment is of great significance in PDAC and other tumors.

Mast cell

Mast cells (MCs) are innate immune cells widely distributed throughout the body and reside especially in protective tissues such as skin, respiratory tract, and intestinal mucosa, acting as barriers against invading pathogens. Abundant secretory granules distinguish MCs in the cytoplasm, containing various immunomodulatory and vasoactive mediators such as growth factors, proteases, leukotrienes, cytokines, and chemokines. MCs mainly regulate biological processes and responses by secreting these molecules, shaping the cellular microenvironment, and participating in tissue remodeling, angiogenesis, pro-inflammatory/anti-inflammatory responses, immunosuppression, cell proliferation, survival, recruitment, maturation, and differentiation [103, 104].

The most common chemokine in MC recruitment in healthy tissues is stem cell factor (SCF), while the CCL15 and CXCL12-CXCR4 axis can induce MC recruitment in the TME [104]. The phenotype and function of MCs mainly depend on their origin, maturity, and the microenvironment in which they are present. MCs have high site-specific plasticity, and tissue-specific MCs may differ in particle content, cytokine expression patterns, and cell receptors [103]. MCs can be recruited early and infiltrated into many solid tumors. Once they appear, they are called tumor-associated mast cells (TAMCs) [105]. TAMC is regarded as one of the most controversial immune cell types in the TME. So far, the diversity of TAMC in various TMEs has not been adequately described [104]. On the one hand, TAMCs can enable different oncogenic processes that lead to tumor progressions, such as angiogenesis and lymphatic vessel formation, fibrosis, and metastasis. On the other hand, TAMCs release mediators that induce the recruitment of other immune cells to perform pro-tumor or anti-tumor functions. In short, there are complex crosstalks between TAMCs and tumor cells, other immune cells, and tumor stroma in the TME, which may lead to their polarization towards tumor-promoting or tumor-suppressing phenotypes.

In PDAC, the infiltrating TAMC is only a tiny part of the cell population within the TME, mainly located at the intra-tumoral border zone. It is debatable whether TAMC is distributed in the peritumor, but there is almost no distribution in the central area of the tumor tissues [104, 106]. Peritumoral TAMCs appear to be more cancer-promoting in PDAC, and only a small number of intratumoral TAMCs are associated with a favorable prognosis. TAMCs can inhibit the activation of T cells in a PD-L1-dependent manner and establish a connection with MDSCs through the CD40/CD40L axis, resulting in immunosuppression of TME [104]. However, there is growing evidence that TAMC can prevent cancer. For example, the IL-33/MC response has been observed to exert a tumor-inhibitory effect in PDAC [103]. In pancreatic neuroendocrine tumor (PanNEN), a malignant tumor of pancreatic tissue, low-density mast cells were associated with high-grade, non-insulinoma, and tumor progression. In contrast, the high degree of mast cell infiltration was related to the increase of CD4⁺ T cell and CD15⁺ neutrophil count, which acted as an independent predictor of prolonging progression-free survival, suggesting the anti-tumor effect of mast cells [107].

As a member of the immune cells in the PDAC tumor microenvironment, MC demands more attention and research to determine whether it has therapeutic potential.

Cancer-associated Fibroblast

As previously described, a major feature of the PDAC tumor microenvironment is the extremely rich and dense stroma (including fibroblasts, immune cells, and ECM). Stroma accounts for up to 90% of the PDAC tumor volume [108]. In addition, tumor or cancer-associated fibroblasts (CAF) account for a large part of the surrounding dense stroma, help form highly heterogeneous stromal cells, and play an essential role in the composition and alteration of TME [38, 109].

CAF definitions are still inconsistent. CAF are fibroblasts found in or near tumors. [109], while scholars of tumor biology hold that the activated fibroblasts in the tumor are usually called tumor-associated fibroblasts [39]. There is also a view that the cells with a slender shape lack tumor mutation, and negative epithelial, endothelial, and leukocyte markers in the tumor tissue can be regarded as CAFs [110]. However, it can be confirmed that CAF crosstalk directly or indirectly with PDAC cells and the infiltrating immune cells, which plays a vital role in the initiation and progression of PDAC [38, 110]. Activated CAF is linked to a poor prognosis, drug resistance, and disease recurrence in various tumors. [109].

CAFs originate from a variety of events. In PDAC, it is generally believed that CAFs have two origins. Some of them are from the activation of local fibroblasts and pancreatic stellate cells [108, 110], while the others come from the recruitment and differentiation of bone marrow mesenchymal stem cells (BMSCs) and epithelial-mesenchymal transformation [38, 111]. In addition, different CAF subtypes in PDAC are caused by different normal fibroblast progenitor cells [112].

Normal fibroblasts are not only producers of ECM but also play a key role in the dynamic balance, repair, inflammation, and fibrosis of normal tissue and communications with many other cell types [41, 110, 112]. When associated with tumors, as an integral component of TME, CAF can promote tumor growth in various ways, including secretion of ECM proteins, enzymes such as matrix metalloproteinases (MMPs), and secretory factors such as VEGF, GAS6, HGF, exosomes, deoxycytidine, etc. These factors promote stroma remodeling and hardness, induce inflammation, and increase angiogenesis. Ultimately, this feedback loop promotes tumor growth, invasion, metastasis, and the transformation from epithelium to mesenchymal, and changes the response to treatment, such as resistance to chemotherapy and radiotherapy through physical, chemical, and biological ways. Second, CAF affects tumor cell signal transduction, changes tumor cell metabolism and epigenome, increases the incidence of tumor-initiating cells, and improves tumor cell stemness. Third, CAF provides metabolites such as amino acids, fatty acids, and lactic acids for tumors. Moreover, CAF secretes IL-6, CXCL9, and TGF-β to regulate the immune response. These cytokines induce differentiation of immune cells, such as promoting the M2 polarization of TAM and damaging Teffs’ function, leading to an immunosuppressive TME [39, 108–112]. In addition, CAF is reportedly involved in regulating systemic effects, such as cachexia, anemia, and immunosuppression [41]. Interestingly, both pro-tumor and anti-tumor effects of CAF have been revealed [113]. Therefore, there is functional heterogeneity among the cell population of CAF.

CAF is heterogeneous in terms of cell source and function, and CAF classification is still being updated. The markers of CAF include α-smooth muscle actin (α-SMA), fibroblast activating protein (FAP), fibroblast specific protein 1 (FSP1), platelet-derived growth factor receptor (PDGFR)-α/β, and vimentin [111]. The most classical point of view is that according to the phenotype of α-SMA expression, the CAF in PDAC is divided into subtypes: myofibroblastic CAF (myCAF) and inflammatory CAF (iCAF). MyCAF represents the myofibroblast type, a high α-SMA and matrix-producing contraction phenotype driven by TGF-β. MyCAF is primarily located near tumor cells and is considered tumor-suppressive. iCAF represents the inflammation-regulatory type, showing the phenotype of low α-SMA and immunomodulatory secretions driven by IL-1, which secretes IL-6 and CXCL12. iCAF is mainly found near the tumor’s edge or further away. STAT3 activation promotes the development of PanIN precursor lesions and the invasive capacity and colony formation of PDAC cells. Moreover, iCAF plays an immunosuppressive role in the TME of PDAC and thus is considered the tumor-promoting type [109, 110, 112, 114]. In PDAC, several studies have found that the IL-1 and TGF-β secreted by tumors antagonize each other, defining the formation of iCAF and myCAF, respectively. The proliferation rate of iCAF is lower than that of myCAF, or there is almost no proliferation within iCAF [112]. The heterogeneity of CAF has plasticity, and different subtypes can be transformed into each other [38]. For example, the Hedgehog (HH) signaling pathway is one of the factors regulating these two CAF subtypes. HH signaling in myCAF is significantly higher than in iCAF, indicating that it is a critical signaling pathway for the maintenance of myCAF subtypes. Inhibiting the HH pathway can decrease the number of myCAFs while increasing the number of iCAFs, resulting in a shift in the proportion of myCAF/iCAF in PDAC. This change results in a decrease in CTL, an increase in Treg, and changes in the composition of fibroblasts and immune infiltration in the PDAC microenvironment, which finally enhances the immunosuppression in TME [115]. The same as the sonic hedgehog (SHH) signaling pathway, which is also an essential regulatory pathway and target [13].

The third new CAF subtype widely accepted in recent years is antigen-presenting CAF (apCAF), which may be derived from mesothelial cells and is characterized by MHC II⁺. apCAF has the function of antigen presentation to CD4⁺ T cells and is immunomodulatory. apCAF may be immunomodulatory in PDAC and breast cancer [13, 111, 112]. Recently, it has been suggested that a new subtype, mscCAF, which are bone marrow mesenchymal stem cells (MSCs) present in the TME of some solid tumors. mscCAF accounts for about 7.3% of the total CAF in vitro. In immunodeficient mouse models, mscCAF can promote the metastasis of multiple subcutaneously transplanted breast cancer cell lines in a CCL5-dependent manner, so mscCAF may be related to the enlargement metastasis and collagen deposition of primary tumors [42].

In addition to the classical classification, Neuzillet et al. have classified the CAFs of PDAC into at least four subtypes: CAF-A (which secrets POSTN and is related to tumor proliferation, invasion, and metastasis), CAF-B (which secrets MYH11 and is related to lymphatic metastasis, and poor prognosis), CAF-C (which secrets PDPN and is related to immune promotion and good prognosis), and CAF-D (with no appropriate markers yet and with relatively poorest prognosis) [111, 116]. Hutton et al. classified pancreatic fibroblasts into two cell lines based on CD105 expression, with CD105⁺ CAF promoting cancer and CD105⁻ CAF inhibiting tumor growth depending on acquired immunity. MHC II, a marker of apCAF, is mainly expressed in CD105⁻ CAFs. The CD105^± CAFs have different responses to fibroblast regulatory signalings such as TGF-β and IL-1 and can be induced to express characteristic genes of both myCAF and iCAF in vitro [117].

As a unique fibroblast in pancreatic tissue, pancreatic stellate cells (PSCs) are also worthy of notice. Normal PSCs exist around acini, ducts, and blood vessels, store vitamin A-rich lipid droplets, and express α-SMA protein. PSCs have the general characteristics of fibroblasts and can transform into myCAF or iCAF under different induction [42, 118]. In the TME, PSCs are mostly observed to be tumor-promoting. It is suggested that PSCs can significantly increase the number of Treg, M2 TAM, and MDSC and reduce the number of CD8⁺ T, CD4⁺ T, M1 TAM, and NK T cells. In addition, PSCs affect immune cells on migration and adhesion and regulate T cells through mechanisms such as the CXCL12/CXCR4 axis. Consequently, an immunosuppressive TME is formed, and tumor growth, proliferation, and metastasis are induced [38, 111, 119].

Research progress of PDAC immunotherapy

As mentioned above, the T cells in the TME have complex subtypes and different functions. T cells are important effectors of immune responses and regulators of immunosuppression. In addition, there are diverse crosstalks between T cells and almost all components in the PDAC TME. These facts make T cells a non-negligible or necessary target for immunotherapy research. This part will focus on T cells research progress related to PDAC immunotherapy.

The initiation of T cells requires the recognition of processed tumor antigens presented by antigen-presenting cells (APC), such as monocytes and dendritic cells, through unique T cell receptors (TCR) that bind to major histocompatibility complex (MHC) molecules and tumor antigens, which mutated or non-mutated tumor-associated proteins may produce. The process of initiation usually takes place in lymphoid tissues. First, CD4⁺ T cells aid CD8⁺ T cell initiation through cytokines. Then both CD4⁺ and CD8⁺ T cells are activated through costimulatory pathways, such as CD28-B7-1 (CD80) and CD28-B7-2 (CD86), which allow them to proliferate, secrete inflammatory cytokines, acquire cytolytic properties, and migrate to sites of antigen expression to form tumor immunity [15, 120].

However, as the unavailable barrier of PDAC to immunotherapy, the most significant resistance to tumor immunity with PDAC is immune tolerance [121]. Tolerance can be produced at any of the above stages, such as initiation, proliferation, differentiation, capacitation, and recruitment, and then maintained or even enhanced via a variety of mechanisms, including crosstalks between TME cells, immunosuppressive cytokines, and chemokines, as well as down-regulation of immune checkpoint pathways [13, 28]. Therefore, immune checkpoint blockade with inhibitors, tumor vaccination, and other attempts have been made to overcome the tolerance in the PDAC TME and to gain considerable therapeutic response marked as increased Teff activity (Fig. 3).

Fig. 3.

Mechanisms of immunotherapy

Anti-immune checkpoint therapy such as anti-CTLA-4 can regulate immunity from the stage of T cell activation in peripheral lymphoid tissues, preventing CTLA-4 from being bound and producing inhibitory signals to promote the activation and proliferation of T cells, and thus promote tumor immune infiltration. On the other hand, anti-PD-1 and anti-PD-L1 mainly act at the immune-tumor interface and bind to the corresponding receptors/ligands, preventing tumor cells or TAMs from inactivating effector T cells through the PD-1/PD-L1 pathway, thereby promoting tumor immunity. TGF-β signaling pathway inhibitors can prevent effector cells from being dysfuctional, and can remodel the ECM like FAK signaling pathway inhibitors. Metabolic checkpoint inhibitors reprogram the cells within the TME. Tumor vaccines (taking GVAX as an example) provide specific antigens and stimulate the activation of immune cells. Nanoparticles target immune cells such as TAMs and reprogram them into the tumor-suppressive M1 subtype. Adoptive cell therapy like CAR-T directly supplied specific Teffs.

Immune checkpoint blockade

The effect of immunotherapy in many solid tumors mainly comes from the emergence of ICIs (immune checkpoint inhibitors). Immune checkpoint blockade (ICB) therapy can produce significant anti-tumor effects on most tumors. However, in PDAC, compared with other tumors, immunotherapy, including ICB, has limited achievement, and most of the effects and results are still far from satisfactory [122].

To discuss ICB, first, it is necessary to recognize critical immune checkpoint pathways, such as the PD-1/PD-L1 signaling pathway and CTLA-4/B7 signaling pathway, two classic and important pathways that mediate T cell immune checkpoint inhibition. In addition to these two pathways, there are dozens of other immunomodulatory receptor-ligand interactions at the interface between tumor cells and host immune cells, which can be targeted clinically by monotherapy or combined therapy. For example, pathways like GM-CSF/CSF1R and SIRPα/CD47 are mainly related to TAMs and are also important [16, 70], but this will not be introduced in detail time in this review. Instead, other notable therapeutic targets include TGF-β, FAK, and many aspects of the microenvironmental metabolism will be discussed below.

Classic immune checkpoint pathways

The programmed cell death protein 1/PD1 ligand 1 pathway is a dominant immune checkpoint pathway that plays an essential role in the TME. Unlike its normal function in controlling immune homeostasis, PD-1/PD-L1 signaling is reprogrammed in tumor tissue by tumor cells to escape immune attack [123]. PD-1 is expressed on T cells, B cells, and NK cells. PD-L1 and PD-L2 are ligands found on activated APCs, hematopoietic and parenchymal cells in the inflammatory microenvironment, placenta, and tumor cells (a variety of solid and hematological tumors) [15]. When bound to a tumor-expressed ligand (PD-L1 or PD-L2), PD-1 prevents CD8⁺ T cells from binding to target cells, and inhibition of this pathway allows pre-existing and properly localized anti-tumor T cells to participate in the immune attack and destroy their target cells [28].

As introduced earlier, CD8⁺ T cells can be used as predictive biomarkers because of their positive immune effects, and pathways like PD-1/PD-L1 are essential mechanisms of immunosuppression. Therefore, they are often used in combination for evaluation. Danilova et al. concluded that PDAC patients with PD-L1⁻/CD8^high had a better prognosis, while patients with PD-L1⁺/CD8^high had the opportunity to block the PD-1 pathway to reactivate CD8⁺ T cell responses to improve prognosis. Therefore, it is proposed that the combined evaluation of PD-L1 expression and CD8⁺ T cell density can better assess the immune status of the PDAC microenvironment than detecting PD-L1 response expression alone [36]. On the other hand, multivariate markers may have higher predictive ability than univariate analysis. For example, factors such as tumor mutation load and CD8⁺ T cell density are functionally associated with PD-L1 expression and associated with each other [15]. Tumeh et al. cited more factors, such as the proximity of PD-1⁺ to PD-L1⁺ cells, CD8⁺ T cell activation, and IFN-γ pathway signaling markers, demonstrating the increased predictive value of multiple factors [124] and combining multi-biomarkers and multi-factors can more accurately immunophenotype PDAC microenvironment, allowing for developing potential immunotherapy strategies based on the corresponding precise subtypes.

Studies of immunotherapy are not confined to a single pathway now. The combined targeting of PD-1/PD-L1 and various cytokines or RTKs is a long-term hot topic. Daniele et al. studied chemokine receptor CXCR4 and pointed out that CXCR4 stimulated by CXCL12 inhibited directional migration mediated by CXCR1, CXCR3, CXCR5, CXCR6, and CCR2. Transcriptional analysis of metastatic biopsies from patients with microsatellite stable colorectal cancer and PDAC revealed that inhibition of CXCR4 by small molecule inhibitor AMD3100 (Plerixafor) induced an integrated immune response since combined with PD-1/PD-L1 blockade can induce T cell infiltration and anti-tumor response in mice and human PDAC [125]. Meggy et al. described a successful OPERA trial of anti-CXCL12 (NOX-A12) in patients with advanced metastatic colorectal cancer and PDAC (NCT03168139), in which a low incidence of high-grade adverse reactions (grade I/II 83.8%, grade III 15.5%, grade V 0.7%) and prolonged treatment time was observed. A systematic serial biopsy discovered clinical responses were related to Th1-like tissue reactivity after CXCL12 inhibition, with T cell aggregation and directed movement towards tumor cells in responding tissues. Thus, the safety and feasibility of the strategy of combining CXCL12 with PD-1 checkpoint inhibitors are proposed [126]. Wang et al. reported that Cancer-FOXP3 (C-FOXP3), co-expressed with tumor cell PD-L1, promotes the immune escape of PDAC by up-regulating CCL5 to recruit Tregs into PDAC. Anti-PD-L1 enhances the anti-tumor effect of CCL5 blocking in high-level-C-FOXP3 xenotransplantation and orthotopic mouse models, thus suggesting the potential feasibility of combined blocking of CCL5 and PD-L1 [127]. Ma et al. evaluated the effect of anti-PD-1 combined with OX40 (which provides survival signals for activated T cells) agonists in orthotopic PDAC mouse models. This combined strategy reduced the Treg proportion in PDAC and increased the number of memory CD4⁺ and CD8⁺ T cells, eradicating all the detectable tumors [128]. Li et al. verified MET, a PDAC-specific RTK that interacts with PD-L1 and promotes tumor growth in orthotopic and subcutaneous PDAC mouse models. Combinational blockade of MET and PD-L1 achieved significant benefits in the models, indicating the importance of MET as an auxiliary target for ICB [129]. Michael et al. evaluated the efficacy of Bruton tyrosine kinase (BTK) inhibitor acalabrutinib alone or in combination with anti-PD-1 pembrolizumab in patients with advanced PDAC (NCT02362048) in phase II, multicenter, open-label, randomized (1:1) clinical trial. Lower therapeutic toxicity (monotherapy group vs. combination therapy group: 14.3% vs. 15.8% in the incidence of grade III-IV adverse events associated with treatment) and responsiveness (monotherapy vs. combination therapy: 0% vs. 7.9% in total effective rate, 14.3% vs. 21.1% in disease control rate) were observed. As a result, despite a continuous decrease in M-MDSC, the clinical efficacy of the combined therapy strategy targeting BTK was limited [130]. More intriguingly, Deng et al. revealed that glucocorticoid and glucocorticoid receptors (GR) could activate the expression of PD-L1 and inhibit the expression of MHC-I in tumor cells in PDAC mouse models. Targeting GR overcomes the resistance to ICB therapy and is expected to be involved in the immunotherapy of PDAC [131].

Tuyen et al.’s study correlates TAM regulation with T cell differentiation when associated with TAM. In the mouse model of PDAC liver metastasis treated with GEM combined with anti-PD-1, they observed an increase in the infiltration of Th1 and M1 macrophages in the tissue of metastatic tumor, a significant increase of M1 macrophages in the spleen, an increase in T cells in peripheral blood cells and innate immune signals. These changes enhanced immune responses and survival time associated with CD8⁺ T cells [132]. Similarly, anti-IL-20 monoclonal antibody (7E) targeting cytokine IL-20 prolonged the survival time and reduced PD-L1 expression in tumor tissues of transgenetic KPC mouse (Kras^{+/LSL−G12D−}; Trp53^{+/LSL−R172H−}; Pdx1-Cre, a successfully established spontaneous PDAC model mouse which has many characteristics similar to human PDAC, such as pancreatic endothelial cell tumor formation, strong immune response and high rate of liver and lung metastases.) and the orthotopic PDAC models. The treatment inhibited the M2-like polarization of TAM and tumor growth and alleviated cachexia symptoms in the cancer-associated cachexia (CAC) model, showing good fitness with anti-PD-L1 [133]. Notably, osteopontin (OPN) has received attention as a factor associated with monocytes. Anti-PD-L1 combined with CSF1R inhibitors such as PLX3397 can induce potent antitumor activity and prolong the survival time of OPN^high tumor-bearing mice, with increased CD8⁺ T cell infiltration and Th1/Th2 ratio, decreased TAMs, and cytokine balance in the TME [134]. Lu et al. have also reported that OPN and its receptor CD44 up-regulated the expression of H3K4me3 through its promoter in the PDAC mouse model. Inhibiting the epigenetic axis of WDR5-H3K4me3 effectively controls OPN expression in tumor cells and M-MDSCs, reduces the immune escape of PDAC, and improves the anti-tumor efficacy of anti-PD-1 immunotherapy in vivo [135].

Some progress has also been made in researching the pathways related to PDAC stroma. For example, Koikawa et al. blocked multiple tumor-promoting pathways and destroyed the desmoplastic and immunosuppressive TME. In addition, complete elimination of continuous remission of invasive PDAC in different models was observed by targeting proline isomerase Pin1, which is overexpressed in tumor cells and CAFs, with anti-PD-1 and GEM [136]. When it comes to MEK and microvascular remodeling mentioned above, Marcus et al. induced an RB protein-mediated senescence-associated secretory phenotype (SASP) through the combination of MEK and CDK4/6 inhibitors in preclinical PDAC mice, which promoted tumor vascular remodeling to increase drug delivery, stimulated CD8⁺ T cell infiltration and made tumors sensitive to PD-1, thus inhibiting the proliferation of PDAC [137].

The cytotoxic T lymphocyte-associated antigen 4 (CTLA-4)/ B7 pathway acts differently from the PD-1/PD-L1 pathway. CTLA-4 immune checkpoint primarily functions early in the immune response cycle, enhancing Treg immunosuppressive activity. Therefore, it has a comprehensive and extensive impact on the immune system. CTLA-4 is expressed on CD4⁺ T and CD8⁺ T cells but mainly on CD4⁺ T cells, especially Tregs. The ligands for CTLA-4, B7-1, and B7-2 are found on APCs, activated T and B lymphocytes, and malignant T or B tumor cells [15]. CTLA-4 pathway inhibits T cell activation by competitively inhibiting the CD28 costimulatory receptor, while inhibition of CTLA-4 renders peripheral T cells activated by APCs.

PD-1/PD-L1 is relatively tumor-specific, while CTLA-4 plays an overall role in tumor tissues and peripheral circulation. Recently, there has been speculation that the PD-1/PD-L1 pathway affects primary tumor tissues and peripheral T cells during the recruitment process [138]. CD8⁺ T effector cells are considered the main immune cell type affected by the PD-1 immunosuppressive checkpoint pathway, while CTLA-4 mainly regulates the activity of the effector and regulatory subtype of CD4⁺ T cells, which in turn affect CD8⁺ T cells. Tumor-specific PD-1⁺CD8⁺ T cells bound to the ligand PD-L1 will be inactivated in the TME, while CTLA-4 expressed by Tregs enhances their ability to suppress effector cells from cell-dependent cytokine production and directly killing towards tumor cells. Thus, immunotherapies targeting the immune checkpoint CTLA-4 can prevent immunosuppression and restore T cells' ability to eliminate antigen-expressing tumor cells.

Based on the success in melanoma, the anti-CTLA-4 antibody ipilimumab was the first ICI evaluated in PDAC. Although ICB was effective enough to produce adverse reactions in the Phase II trial in 2010, no objective therapeutic response was observed [122]. Three follow-up phase Ib dose-escalation studies (NCT01473940, NCT01473940, and NCT01473940) of ipilimumab and GEM in patients with advanced (stage III and IV) PDAC displayed minimal benefits compared with gemcitabine alone [139]. In contrast, current attempts on CTLA-4 are not as impactful as on PD-1, and more CTLA-4-related immunotherapies tend to appear in studies combined with PD-1/PD-L1 blockade and other treatments. For example, Wang et al. found in colorectal cancer and PDAC animals that the anti-PD-1/L1 combination introduced before MAPKi optimizes response persistence by promoting pro-inflammatory polarization of macrophages and clonal expansion of the IFN-γ^hi ones and clonal expansion of CD8⁺ T cells with high expression of cytotoxic and proliferative genes rather than CD4⁺ Tregs. As a result, short-term anti-PD-1/L1 + anti-CTLA-4 therapy before combined MAPKi therapy is recommended to prevent therapeutic resistance [140]. Oliver et al. found that anti-CTLA-4 combined with anti-PD-1 immunotherapy was clinically and objectively effective through the clinical trial of ipilimumab and nivolumab in treating rare and advanced as 33% of patients with atypical bronchial carcinoids and 43% of patients with pancreatic neuroendocrine tumors. The 4.8 months of progression-free and 14.8 months of total survival time were observed. However, 66% of patients reported immune-related toxicity, and 34% experienced level three or four events [141]. Recently, Jonathan et al. reported an open, multicenter, randomized, phase II trial of durvalumab (anti-PD-L1) combined with tremelimumab (anti-PD-4) or combined with low-dose or hypofractionated radiotherapy for metastatic NSCLC that had not responded to previous PD(L)-1 therapy (NCT02888743). Unfortunately, this trial was discontinued due to its invalidity evaluated in the mid-term analysis. For NSCLC patients resistant to PD(L)-1, radiotherapy did not improve the response to the combination of PD-L1 and CTLA-4 [142]. A phase II randomized, open trial (NCT02866383) evaluated the safety and efficacy of nivolumab and ipilimumab in combination with high-dose radiotherapy in patients with metastatic PDAC, biliary cancer, or previous GEM intolerance. Although the results are expected by the end of 2021, they have not yet been released [139].

Therefore, in PDAC, the main obstacles of multi-target combined immunotherapy, such as anti-CTLA-4 and PD-1, are the low treatment response and the treatment-related toxicity. ICB research on classic pathways still has a long way to go. The future direction may be focused on more predictive biomarkers in treatment, the mode of administration or dose in treatment regimens, or cross-immunotherapy, such as ICB + vaccination or ICB + ACT, rather than the simple superposition of multiple ICIs or ICB strategy combined with traditional radiotherapy and chemotherapy.

Transforming growth factor β (TGF-β) signaling

Transforming growth factor β (TGF-β) has significant effects on the immune cells and stroma in the PDAC microenvironment, and it is also an important immune target worthy of discussion.

TGF-β1, TGF-β2, and TGF-β3 are transforming growth factor β subfamilies. They signal through the receptor complex of serine/threonine kinase TGFBR1 and TGFBR2, which phosphorylate the Smad family and then transduce multiple signaling pathways [143, 144]. Normally, TGF-β signaling regulates the dynamic balance of tissues and cell adhesion, proliferation, and survival, ensuring that epithelial cells are in a static state with a low increment rate, similar to tumor-suppressive activity [143]. TGF-β signaling is essential for the dynamic balance of epithelial cells, interstitial cells, and immune cells in the liver, pancreas, and gastrointestinal system. In tumors, high-frequency changes can occur in the TGF-β signaling pathway encoded protein. The changes happen in up to 40% of the whole, 38% of liver tumors, and 50% in PDAC [144], with the common Smad4 mutation in PDAC becoming an important link in this pathway [145]. Within the TME, various cells produce and respond to TGF-β, thus forming a complex network of epithelial cells, tumor cells, immune cells, and stromal fibroblasts. At this time, TGF-β signaling can interact with multiple immunosuppressive pathways, such as the classic PD-1/PD-L1 pathway, induce Tregs, promote the elimination of CD8⁺ T cells, and hinder the acquisition of Th1 phenotype, thus driving immune dysfunction within the TME. The resulting immunosuppression seriously limits the efficacy of ICB and other immunotherapies. In addition, TGF-β promotes the elongation and hardness of CAFs and the expression of integrins such as IRGD5 [146] and induces fibrosis and remodeling in the stroma. These all drive the formation of characteristic PDAC TME [147, 148]. TGF-β has the dual effect of suppressing and promoting tumors, with immunosuppression and stroma remodeling primarily promoting tumors in PDAC.

Many studies have been carried out to overcome the T-cell inhibition induced by the TGF-β signaling pathway. Tauriello and colleagues found that PD-1/PD-L1 suppression had a poor response in mouse models, while the addition of TGF-β inhibition had an effective response to tumor cytotoxic T cells, meaning blocking TGF-β signaling transduction made the tumor sensitive to PD-1/PD-L1 therapy [149]. Ravi et al. integrated the two classic pathways with TGF-β and designed bifunctional antibody-ligand traps (Y-traps) for CTLA-4 + TGF-β and PD-L1 + TGF-β, α-CTLA-4-TGF-β RIIecd and α-PD-L1-TGF-β RIIecd. Compared to a single CTLA-4 or PD-L1 antibody, the Y-trap counteracts Treg differentiation and immune tolerance mediated by TGF-β, which is characterized by the decrease of infiltrating Tregs in TME, the increase of IFN-γ, and the improvement of tumor response of CD4⁺ T and CD8⁺ T cells [150]. Principe DR. proved that CD8⁺ T cell activation could not be achieved by gene ablation or pharmacological inhibition of the overall reduction of the TGF-β signaling. PDAC with TGF-β suppression depicted a compensatory increase in the expression of PD-L1, resulting in no significant effect on anti-tumor immunity. Based on these conclusions, targeting both TGF-β and PD-L1 receptors on the genetic model of advanced PDAC determined that the concomitant inhibition of TGF-β and PD-L1 receptors can promote PDAC clearance and survival improvement by CD8⁺ T cells [151]. These results strongly suggest that TGF-β and PD-L1 pathway inhibitors may cooperate in PDAC, which is worthy of further exploration as a clinical immunotherapy method. The same team later found that long-term GEM intervention increased the synthesis of CCL, CXCL, and TGF-β signaling in mouse and xenograft models. However, the GEM + α-PD-1 combination did not intervene in the disease progression. Simultaneously, the heredity or drug ablation of TGF-β signaling resulted in the tumor’s significant response to combined therapy, such as a strong CD8⁺ T response, reduced tumor load, and a significant increase in overall survival rate. Given the lack of third-line treatment options, this regimen has promising clinical significance for GEM-resistant PDAC patients. In the follow-up, what is worth exploring is whether there is a similar synergistic mechanism in the first-line treatment regimen based on FOLFIRINOX [152, 153].

Several research teams further explored the mechanisms and ways to overcome the TGF-β signaling in PDAC. For example, Toshiro et al. found in the epidermal model that the competition between T cells for active TGF-β represents a selective pressure, promoting the accumulation and persistence of antigen-specific Trms (resident memory T cells) in epidermal niches [154]. Huang et al. demonstrated that PDAC developed with an intact TGF-β signaling pathway could avoid apoptosis caused by the synergistic action of TGF-β and RAS signaling pathway through ID1. This result confirms that ID1 is a key node related to TGF-β and one of the potential therapeutic targets [155]. Linara et al. depicted that the cholesterol pathway is a possible tumor-promoting dependent way of TGF-β signaling and can also be an intervention target [156].

As a result, while there are few ongoing clinical trials, multidrug combination therapy that includes TGF-β inhibitors may be a more effective immunotherapy strategy for PDAC than current immune checkpoint inhibitors.

Focal adhesion kinase (FAK)

FAK is a non-receptor cytoplasmic protein tyrosine kinase, a key regulator of ECM signaling mediated by integrin and growth factor receptors. FAK is involved in various cell signaling pathways, controls cell migration, apoptosis, and cell survival, and regulates stroma mainly through cell adhesion [157–160].

FAK is usually overexpressed in tumor tissues, and its phosphorylation activation is linked to increased type I collagen deposition and immunosuppression. The high expression of phosphorylated FAK1 in patients is closely correlated with decreased tumor-infiltrating CD8⁺ cytotoxic T lymphocytes, increased collagen deposition, and granulocyte infiltration. FAK1 is the primary driver of connective tissue proliferation and immunosuppressive TME in PDAC [38, 48]. For example, Itoh et al. conducted a retrospective analysis and cell line immunoblotting test and pointed out that overexpressed FAK could promote tumor progression, especially in vascular invasion. Therefore, FAK was raised as a new therapeutic target and prognostic marker [159]. Duxbury et al. verified the hypothesis that FAK gene silencing could promote apoptosis and reverse the apoptosis resistance of human PDA cells through a preclinical model of PDAC, indicating that FAK is a promising anti-metastasis target [160]. Serrels et al. found that FAK promotes tumor immune escape by inducing an immunosuppressive microenvironment. In addition, FAK can regulate the transcription of CCL5, which drives the recruitment of tumor-associated Tregs, thus promoting tumor growth and metastasis by inhibiting the activity of cytotoxic CD8⁺ T cells. Concurrently, the depletion of FAK or kinase inhibition causes the regression of tumor cells. FAK activity in CAF is an independent marker of PDAC and a driving factor of drug availability because the inactivation of FAK in CAF reduces fibrosis in primary tumors (decreased ECM expression and CAF deposition) and the number of immunosuppressive cells (affecting the polarization and migration of M2 TAMs) [161]. Moreover, FAK regulates chemokines such as CCL1 and CCL2 to enhance the glycolysis of malignant cells [162] and the transcription of CCL5 that drives the recruitment of tumor-associated Tregs. Simultaneously, the depletion of FAK or kinase inhibition causes the regression of tumor cells [157], again showing FAK's feasibility as a therapeutic target.

Subsequently, recent studies have made more attempts and discussions on immunotherapy targeting FAK. Currently, the results of clinical studies on single-drug FAK inhibitors are limited, but several Phase I/II trials of FAK inhibitors and anti-PD-1 or anti-MEK and RAF are underway, with a variety of FAK inhibitors under development [163]. For example, an ongoing randomized Phase II study (NCT03727880) investigates the efficacy and safety of a combination of standard PDAC chemotherapy and pembrolizumab with or without FAK inhibitor defactinib before and after surgery in high-risk resectable PDAC patients. Concurrently, defactinib is also included in the phase II MATCH trial [139]. In addition, Stokes et al. explored the role of PF-562,271, a potent small molecular inhibitor of FAK and related tyrosine kinase PYK2, in PDAC TME. PF-562,271 therapy inhibits the growth of PDAC cells either directly or by destroying the migration of tumor cells and recruitment of CAFs and monocytes, which again supports the role of anti-FAK therapy in PDAC [158]. Finally, Jiang et al. reported the resistance of PDAC to FAK inhibition. They found that FAK inhibitor treatment in KPC mice resulted in stroma depletion characterized by decreasedα-SMA⁺ fibroblasts and collagen density and down-regulation of the TGF-β/Smad signaling pathway by activating the STAT3 signal pathway, finally resulting in treatment resistance. In contrast, the combined targeting of FAK + Janus kinase/STAT3 played a synergistic effect and provided a new strategy for PDAC therapy [164, 165].

In summary, the conclusions of these studies suggest that FAK kinase inhibitors have the prospect of clinical application. Furthermore, targeting the pleiotropic cellular function of FAK may broadly impact regulating the immunosuppressive TME, thus guiding integrated immunotherapy of PDAC.

Metabolic checkpoints

Metabolism is an unavoidable topic in tumors, which plays a vital role in tumor biological behavior, and it is also a significant direction of tumor-targeted therapy. However, metabolism in the PDAC tumor microenvironment is of much mystery. In hypoxia, low nutrition, high pressure, and acidity, PDAC cells undergo a series of metabolic reprogramming, scavenging recycling, and metabolic crosstalks. As a result, the rewired glucose, amino acid, and fat metabolism and these metabolic crosstalks contribute to the development of PDAC. In addition, these reprogrammings promote the occurrence, growth, and metastasis of PDAC through epigenetic regulation and are closely related to drug resistance to chemotherapy, radiotherapy, and immunotherapy [166, 167].

The notion that tumor cells would consume glucose in the TME to fuel nutritional competition as a metabolic mechanism of immunosuppression may be falsified. Recent studies have confirmed that across a range of tumor models, myeloid cells have the greatest ability to absorb glucose in the tumor, followed by T cells and then tumor cells. Conversely, tumor cells have the highest uptake of glutamine (Gln). Furthermore, the cell-intrinsic programs drive immune cells and tumor cells to prioritize glucose and glutamine, respectively. These results indicate that glucose is not the real limiting metabolic factor in TME [168]. However, the importance of glutamine in the metabolism of immune cells does been observed. For example, MYC, the key metabolic reprogramming factor in T cells, is dose-dependent and overexpressed as one of the mechanisms for tumor cells to escape immune surveillance by increasing the expression of CD47 and PD-L1 and is heavily dependent on glutamine catabolism [169–172]. Therefore, this review will discuss the latest progress in metabolic checkpoint therapies with glutamine as a representative site of metabolic intervention in PDAC.

Glutamine is a non-essential amino acid that plays numerous functions in cell metabolism. It is necessary for tumor cell survival because it is the primary source of carbon and nitrogen, which contributes to the macromolecular synthesis and redox balance [166, 167]. Because of the difference in glutamine metabolism between tumor cells and effector T cells, glutamine can be targeted as a metabolic checkpoint in tumor immunotherapy, combined with immune checkpoint inhibitors to enhance the anti-tumor efficacy of immunotherapy. At present, a considerable number of teams have been focusing on glutamine.

Regarding how glutamine promotes tumor progression or drug resistance, Maira et al. demonstrated the nutritional stress caused by glutamine deprivation in PDAC, proving that glutamine deficiency regulates EMT by up-regulating Slug depending on MEK/ERK signaling and ATF4 [173]. Ralph et al. found that NetG1⁺ CAF supported the survival of PDAC through glutamate/glutamine metabolism mediated by NetG1 through three-dimensional in vitro co-culture assays and orthotopic murine models. Neutralizing antibody-blocking fibroblast protein NetG1 reversed the tumor-promoting characteristics of CAF and inhibited tumorigenesis in vivo [113]. Hu et al. found that SIRT5 gene ablation promoted the occurrence and progression of tumor events in PDAC mouse models. SIRT5, a key tumor suppressor and potential target in PDAC, increased non-canonic glutamine use via GOT1, and SIRT5 deletion promoted tumorigenesis. Selective SIRT5 activator MC3138 displayed an anti-tumor effect and promoted GEM sensitivity in human PDAC cells [174]. Koelina et al. found that human and mouse primary PC tumors with high expression of MUC5AC showed higher resistance to GEM because of the increased use of glutamine and nucleotide biosynthesis. In contrast, β-catenin and glutamine decomposition inhibitors combined with GEM could eliminate MUC5AC-mediated drug resistance in mouse and human tumors [175].

Glutamate-related inhibitors, such as DON (a glutamine analog, 6-diazo-5-oxo-1-norleucine) and JHU083, significantly improved anti-tumor effects, especially when combined with ICB, such as anti-PD-1. Other strategies target the Gln metabolism pathway factors, such as the Gln synthetase [176]. Adaptive reprogramming of CD8⁺ T cells decreased production and recruitment of MDSCs, and remodeling of ECM are observed in these studies, which led to the suppression of the growth and metastasis of tumor cells, sensitivity to ICB, and prolonged survival [169, 176, 177].

Apart from glutamine metabolism in the PDAC microenvironment, numerous attempts for other therapeutic targets have been explored. One of the cases involves the tumor's inherent M⁶A demethylase FTO. FTO can limit the activation and effective state of CD8⁺ T cells and increase glycolysis regulators JunB and C/EBPβ in an M⁶A-dependent manner. Liu et al. used a small molecular compound, Dac51, to inhibit FTO activity and found the enhancement of immunotherapy based on PD-L1 blocking therapy. Targeting FTO can inhibit the accumulation of these glycolysis regulatory factors and the transcription of glycolysis-related genes in tumor cells, followed by removing metabolic barriers on T cell activation, promoting anti-tumor function of infiltrating T cells, and the inhibition of tumor growth in vivo [53]. Moreover, Zhu et al. conducted a CRISPR screening of PDAC cells cultured in vitro or in vivo and found that the loss of heme synthesis independent of tissue-derived or immune cells inhibited tumor growth. Furthermore, tumor cells escaped cell death induced by TNF-α from CD8⁺ T cells through metabolic autophagy, which nominates potential anti-tumor targets through this metabolic dependence [178].

These prove several potential targets in metabolism-related checkpoints in the PDAC tumor microenvironment. Checkpoint blockade on metabolism is expected to become an important part of combined immunotherapy, which calls for further exploration and attempts.

Tumor vaccines

Vaccination is one of the options for tumor immunotherapy that has developed rapidly in the past decade and achieved significant results. There have been several kinds of tumor vaccines, including whole-cell, peptide-based, DC, DNA vaccine (plasmid vaccine, virus-based vaccine, bacterial vector, and yeast-based recombination vaccine), and mRNA vaccine [179].

GM-CSF gene transduced autologous tumor vaccine (GVAX) is a whole-cell vaccine expressing and secreting granulocyte–macrophage colony-stimulating factor (GM-CSF). GVAX vaccination can improve the specific anti-tumor immune response by stimulating the infiltration of anti-tumor T cells and activating myeloid cells. Many teams have explored the mechanism and effect of GVAX in PDAC.

High-throughput T cell receptor Vβ sequencing (HTTCS) analysis displayed that increased responses from CD4⁺ and CD8⁺ T cells were observed in PDAC patients who received immune checkpoint therapy (anti-PD-1 and anti-CTLA-4) combined with GVAX, as evidenced by more significant TCR library diversity, and a certain degree of clinical response [28]. Furthermore, Laheru et al. combined two GM-CSF-secreting PDAC cell lines (GVAX) CG8020/CG2505 with CTX and found enhanced TME mesothelin-specific CD8⁺ T cell responses could be detected in some patients in the cohort, which may be related to their prolonged survival. Notably, the toxicity of GVAX treatment is minimal, whether used alone or in combination, indicating the feasibility and safety of clinical application [180]. In preclinical mouse models of PDAC, the feasibility of GVAX combined with anti-PD-1 therapy was also investigated. GVAX alone induced high expression of CSF-1R in lymphoid aggregates, resulting in immunosuppression and vaccine resistance. In contrast, the combined GVAX + anti-PD-1 + anti-CSF-1R immunotherapy effectively increased CD4⁺ and CD8⁺ T cells co-expressing PD-1 and CD137 and PD-1⁺OX40⁺CD4⁺ T cells in TME and made the PD-1⁺CD137⁺CD8⁺ T cells with high IFN-γ expression. Thus, adding myeloid-targeting drugs to vaccine-based tumor immunotherapy is suggested to reverse immunotherapy resistance or hypopotency [181]. Therefore, these studies further support the benefits of integrated immunotherapy combined with tumor vaccine in PDAC.

More in-depth research on GVAX has recently been conducted, and more clinical trials for transformation are planned. For example, Zheng et al. conducted a clinical trial of neoadjuvant therapy in patients with resectable PDAC (NCT00727441) and observed that GVAX alone or in combination with two forms of low-dose cyclophosphamide (Cy) neoadjuvant therapy was safe and feasible. Patients who received neoadjuvant and adjuvant GVAX had a longer overall survival (35.0 months vs. 24.8 months in controls) and higher tumor TLA density [182]. Additionally, Takahiro et al. conducted a clinical trial on GVAX + Cy + CRS-207 (a live attenuated Listeria monocytogenes vaccine) with or without nivolumab in patients with metastatic PDAC previously treated (NCT02243371). The combination group showed a greater improvement in OS than the non-treatment group. Unfortunately, the primary endpoint of this study was not met, as the OS improvement in the combination group was comparable to the standard treatment. It is worth noting that significant changes in the TME can be noticed in long-term survivors of the combination group, including an increase in CD8⁺ T cells and a decrease in CD68⁺ myeloid cells [183]. Other similar clinical trials, such as the Phase II study of GVAX combined with 5-FU in patients with resectable PDAC (NCT00084383), the trial of GVAX combined with ipilimumab in patients with metastatic PDAC treated with FOLFIRINOX (NCT01896869), and the safety of nivolumab combined with BMS-813160 (double antagonist of CCR2/CCR5) and GVAX in patients with locally advanced PDAC after chemotherapy and radiotherapy (NCT03767582) are ongoing [139].

Peptide-based, or polypeptide, is another important type of vaccine, which can induce immune responses by recognizing and activating T cells and MHC I. In 1995, Gjertsen et al. tried mutant ras peptide as a vaccine for pancreatic cancer and successfully induced ras-specific T-cell response in some cases. However, no significant tumor remission was observed then [184]. Today, polypeptide vaccines such as synthetic RAS and T-WIN that targets both tumor cells and the TME are still being updated and explored [185].

As one of the antigen-presenting cells, DCs can effectively deliver antigens to CD4⁺ and CD8⁺ T cells and secrete cytokines such as IL-15, IL-12, IFN-γ, and TNF to promote the activation of cytotoxic CD8⁺ T cells by transforming the type of immune response into type 1 response. Recently, it has been observed from lung adenocarcinoma that in tumors, the repository of tumor antigen-specific TCF-1⁺CD8⁺ T cells that drive the response to immune checkpoints blockade is maintained in the tumor-draining lymph nodes by conventional type I dendritic cells (cDC1s). Therefore, decreasing cDC1 with tumor progression will lead to anti-tumor immunity failure [186]. Similarly, DC deficiency in the PDAC microenvironment can lead to immune surveillance dysfunction, while the restoration of cDCs can enhance the activity of CD8⁺ T cells and Th1 to control the tumor [187]. These results also prove the theoretical feasibility of the DC vaccine.

DC vaccine is activated by tumor antigen during tumor therapy, inducing cytotoxic T-cell response. The inoculation effect of the DC vaccine is significantly related to the number of vaccine-DC migrated to draining lymph nodes. In PDAC, intraperitoneal injection is frequently used to improve inoculation efficacy [179]. Recently, a Phase I/II trial (UMIN000027279, registered in Japan) evaluated the safety and efficacy of WT1 and MUC1 polypeptide-loaded DC vaccine combined with AG regimen or FOLFIRINOX combined with chemotherapy regimen in the treatment of advanced or relapsed PDAC. The authors discovered that the DC vaccine improved tumor-specific immunity and clinical outcomes, implying that the DC vaccine combined with a conventional chemotherapy regimen is safe and profitable in patients with advanced PDAC [188]. Yang et al. verified the hypothesis that irreversible electroporation (IRE, a local ablation method) could overcome the immunosuppression of TME and improve the efficacy of the DC vaccine through an orthotopic PDAC mouse model (median survival time: IRE + DC 77 days, DC 49 days, IRE 44 days, sham operation control 35 days), which supports the combined application of IRE ablation and DC vaccine [189]. Furthermore, Shangguan et al. investigated the efficacy of prophylactic DC vaccination as a route of administration of DC vaccine to control the growth of PDAC. It was found that the tumor volume of the prophylactic DC vaccination group decreased, the survival time was prolonged, and the apoptosis rate, CD8⁺ T cell increase, and collagen deposition were observed histologically compared with the control group in mouse models, proposing a new application approach of DC vaccine [190].

In recent years, tumor vaccines derived from induced pluripotent stem cells (iPSC) have also emerged. Ouyang et al. studied the anti-tumor effect of the iPSC-based tumor vaccine on PDAC. The iPSC-based tumor vaccine is composed of autologous iPSCs and CpG. iPSC vaccine can stimulate cytotoxic anti-tumor CD8⁺ T cell effect and memory response, induce cancer-specific humoral immune responses, reduce immunosuppressive CD4⁺ Tregs, and prevent tumor formation in 75% of the PDAC mice [191]. Lu et al. created a new personalized preventive and therapeutic vaccination regimen called Virus-Infected, Reprogrammed Somatic cell-derived Tumor cell (VIReST). KRasG12D and p53R172H tumor driver mutations edited iPSCs were pre-infected with oncolytic Adenovirus (AdV) as prime or Vaccinia virus (VV) as a boost to improve vaccine immunogenicity. In PDAC mouse models, the VIReST regimen induced a tumor-specific T-cell response, delayed disease emergence, and progression, and significantly increased survival time. Notably, the regimen was tolerable and non-toxic [192].

Personalized immunotherapy based on neoantigens is emerging as the next generation of sequencing advances. The multi-omics data of tumor biopsies from patients are analyzed to predict which mutations would produce tumor-specific neoantigens to carry out the personalized design of the neoantigen vaccine. For example, Huang et al. proposed that ADAM9, EFNB2, MET, TMOD3, TPX2, and WNT7A are effective antigens for developing anti-PDAC mRNA vaccine [193]. Through the mouse lung adenocarcinoma model expressing tumor-specific neoantigens, therapeutic vaccination eliminated the newly discovered TCF1⁺ progenitor cell subsets and improved the therapeutic response to ICB [194], which once again supports the importance and feasibility of targeting tumor neoantigens. Similarly, Chen et al. administered the personalized peptide vaccine Ineo-Vac-P01 based on neoantigens to patients with advanced PDAC and low TMB (NCT03645148). It was observed that patients with relatively long OS had higher titers of IFN-γ and CD4⁺ or CD8⁺ effect and memory T cells in peripheral blood after vaccination, with a significant increase in the abundance of antigen-specific TCR clones from 0 to nearly 100%. The results depicted that the neoantigen vaccine could induce specific T cell subtypes with no serious adverse effects related to vaccination [195].

In other tumor types, such as in B16 melanoma, the combination of anti-PD-1 + anti-CTLA-4 + Fvax has provided beneficial results [99]. Therefore, it can be considered that multi-target blockade combined with tumor vaccine, vaccine personalization, and optimizing the mode of vaccine administration may be the possible effective direction and suggests future research focus on combined immunotherapy. Because of its designable diversity and relatively high safety, vaccination will likely become a powerful approach in PDAC combined immunotherapy.

Emerging focuses

In addition to the signaling pathways related to ICB, metabolic pathways and vaccinations described above, immunotherapy for PDAC remains to be explored. Some other treatments for PDAC that deserve further attention are listed below, which are expected to be part of PDAC immunotherapy.

First, CD40 is a cell surface member of the tumor necrosis factor receptor superfamily, expressed in DC cells and other myeloid cells such as B cells, monocytes, or tumor cells. CD40 activation allows DCs to promote the activation of anti-tumor T cells and induce macrophages to differentiate into the tumor-suppressive and matrix-destructive subtypes by binding with its ligand CD40L (CD154) [122, 196]. However, related cytokine-mediated therapy on CD40 is not entirely optimistic in clinical research, requiring more treatments to be attempted [139]. The DC vaccine mentioned above has been tried in combination with CD40 agonists. Lau et al. combined tumor lysate-loaded DC vaccine with CD40 agonist FGK45 in PDAC murine models. Although anti-CD40 monotherapy did improve the infiltration of CD8⁺ T cells, these essential effector cells displayed hallmarks of exhaustion, including PD-1, TIM-3, and NKG2A. In comparison, the combination therapy with the DC vaccine induced a significant change in tumor transcriptome and mitigated the expression of inhibitory markers on CD8⁺ T cells [197]. This result also provides a potential direction for subsequent clinical trials with CD40-targeted therapy.

The IL-10-related pathway is one of the focuses of the research and development of targeted therapy. The functional characterization of IL-10 shows its direct regulation of MHC II antigens and induction of T cell, B cell, and mast cell growth. Furthermore, increasing IL-10 can promote T-cell-mediated tumor rejection [198]. Mumm et al. found that pegylated IL-10 (pemetrexed) induces the expansion of tumor-specific activated CD8⁺ T cells expressing IL-10 receptor (IL-10Ra) in peripheral blood and TME and tumor rejection mediated by Teffs [199]. Naing et al. further supported the view that pemetrexed can induce the activation of CD8⁺ T cells in the whole body and TME to improve infiltration. They found that pemetrexed combined with anti-PD-1 induced the expansion of LAG-3⁺PD-1⁺CD8⁺ T cells in TME, the increase of Th1 and Th2, the significant increase of IL-7, and the decrease of Foxp3⁺ Tregs and the phosphorylation of STAT3 in CD8⁺ T cells. These changes led to an objective remission of the tumor [200]. These findings provide strong support for pegylated IL-10 to participate in combined immunotherapy. A recent randomized phase III study (NCT02923921) of SEQUOIA's second-line treatment with FOLFOX alone or in combination with pegilodecakin, a pegylated recombinant human IL-10 (PEG) in metastatic PDAC patients who progressed after GEM treatment failed to achieve satisfactory and expected results in clinical efficacy. However, results consistent with the immunostimulatory signal of the IL-10R pathway in pharmacodynamics were observed [198]. Therefore, therapies targeting the IL-10 pathway must be optimized to achieve clinical breakthroughs.

Autophagy plays a dual role in the development and progression of tumors. First, autophagy promotes immune escape in PDAC by selectively degrading MHC-I [201]. A certain amount of therapeutic drug resistance comes from the mechanism of apoptosis. Therefore, inducing non-apoptotic cell death may provide an alternative strategy for killing PDAC cells. In various forms of non-apoptotic cell death, the lipid peroxidation-mediated and iron-dependent ferroptosis induced by CD8⁺ T cells has recently attracted much attention as a potential anti-tumor strategy [202]. Glutathione peroxidase 4 (GPX4) and acyl-CoA synthase long-chain member 4 (ACSL4) can affect the sensitivity of cells to ferroptosis. Wang et al. found that the combination of PD-L1 blocker and Cyst(e)inase, an engineering enzyme that degrades cystine and cysteine, significantly increased lipid peroxidation in tumor cells in vivo the percentage of CD8⁺ T and CD4⁺ T cells in the TME [203]. Daniel et al. used pharmacological methods to show that targeting cysteine, glutathione, or lipid antioxidants can cause ferroptosis in cells with the cytoplasmic aspartate aminotransferase GOT1 gene knockdown. GOT1 inhibition represses mitochondrial metabolism and promotes a catabolic state, which enhances labile iron availability through autophagy and potentiates the activity of ferroptosis stimuli [204]. Dai et al. found that the administration of Liproxstatin-1 (a ferroptosis inhibitor), clophosome-mediated macrophage depletion, or pharmacological and genetic inhibition of the 8-OHG/TMEM173 pathway can suppress Kras-driven pancreatic tumorigenesis in mice [205]. In summary, targeting the tumor iron transport pathway and combining anti-tumor therapy with ferroptosis could be a novel approach to PDAC immunotherapy.

Other features of the PDAC tumor microenvironment, such as high-degreed perineural invasion (PNI), are also worthy of discussion. For example, our group found that in the primary tumor microenvironment of PDAC with PNI, acetylcholine impairs the ability of PDAC cells to recruit CD8⁺ T cells via HDAC1-mediated suppression of CCL5. Moreover, acetylcholine directly inhibits IFN-γ production by CD8⁺ T cells in a dose-dependent manner and favors Th2 over Th1 differentiation. Furthermore, hyperactivation of cholinergic signaling enhanced tumor growth by suppressing the intratumoral T-cell response in an orthotopic PDAC model [50]. Therefore, the effect of PNI and neurotransmitters on the PDAC microenvironment also leads to thinking about whether neurotransmitter blockers or hijacking cancer-nerve crosstalk in combination with immunotherapy are feasible for treating of PDAC.

Other research directions are to more accurately define more cell subtypes in the TME, classify heterogeneous cell subsets in immune cells, CAFs, and other cells, or target specific cell subsets in therapies such as through tumor neoantigen-based vaccines. John et al., for example, explained the presence of group 2 innate lymphoid cells (ILC2) in tumor tissue. They found that both PD-1⁺ TILC2 and PD-1⁺ T cells are present in most human PDAC, and ILC2s are identified as anti-tumor immune cells for PDAC immunotherapy. More broadly, ILC2s emerge as tissue-specific enhancers of tumor immunity that amplify the efficacy of anti-PD-1 immunotherapy. Furthermore, as ILC2s and T cells co-exist in tumors and share stimulatory and inhibitory pathways, ILC2s can also be targeted by PDAC immunotherapy to amplify the effect of anti-tumor ICB [206].

As a major immunotherapy program, adoptive cell therapy (ACT), like ICB and vaccines, has achieved remarkable progress in PDAC treatment in recent years. Currently, the strategy based on TCR and CAR is the most advanced ACT method, and CAR-T therapy for hematological malignant tumors has achieved clinical success. However, the clinical breakthrough of ACT in solid tumors has been hindered. Solid tumors have stroma barriers and hypoxic zones that prevent immune cells like T cells from infiltrating. Meantime, the large antigen load of solid tumors promotes the depletion of T cells [62, 63]. To overcome this obstacle, some researchers improved the ACT of PDAC through CXCR6-modified T cells [62]. In addition, some used short-chain fatty acids (SCFA), valeric acid, and butyric acid to enhance the anti-tumor activity of CTLs and CAR T cells through metabolic and epigenetic reprogramming [199]. Others found that continuous antigen exposure leads to the transformation of CD8⁺ T-to-NK-like T cells in CAR T cells, and knocking out ID3 and Sox4 related to CAR T cell fatigue can slow down T cell dysfunction and improve the efficacy of CAR-T [207]. As a result, to use ACT in PDAC, the potent immunosuppression and complex interference of TME must be overcome. Therefore, finding strategies for combination therapy or cell engineering is necessary to prevent adoptive cell dysfunction (Table 1).

Table 1.

Representative clinical trials of integrated immunotherapy on PDAC

Drug	Combination	Phase	Status	Primary outcome	ID
Anti-PD-1/Anti-CTLA4
Anetumab Ravtansine	Gemcitabine Hydrochloride,Ipilimumab, Nivolumab	1/2	Recruiting	MTD	NCT03816358
ONC-392 (Anti-CTLA4)	Pembrolizumab	1/2	Recruiting	DLT, MTD, RP2D, TRAE	NCT04140526
Nivolumab, Ipilimumab	Influenza vaccine, Radiation	2	Recruiting	ORR	NCT05116917
Anti-TGF-β
BCA101	Pembrolizumab	1	Recruiting	Safety, Tolerability, DLT	NCT04429542
SHR-1701 (PDL1/TGF-β)	Gemcitabine, Albumin Paclitaxel	1/2	Active	RP2D, ORR	NCT04624217
NIS793	PDR001 (Anti-PD-1)	1	Completed	TRAE, DLT	NCT02947165
FAK inhibitor
Defactinib	MR-guided stereotactic body radiation therapy	2	Recruiting	PFS	NCT04331041
Defactinib	VS-6766 (RAF/MEK inhibitor)	1	Recruiting	RP2D, AE	NCT03875820
GSK2256098	Trametinib (Anti-MEK1 and MEK2)	2	Active	RR	NCT02428270
Metabolic checkpoint inhibitor
L-glutamine	Gemcitabine, Nab-paclitaxel	1	Recruiting	RP2D, DLT	NCT04634539
SBP-101 (Diethyl Dihydroxyhomospermine)	Nab-paclitaxel, Gemcitabine	2/3	Recruiting	OS	NCT05254171
Metformin	Gemcitabine, Erlotinib	2	Completed	Survival after 6 months	NCT01210911
Tumor Vaccine
iNeo-Vac-P01 (Personalized Neoantigen Cancer Vaccine)	GM-CSF	1	Recruiting	TRAE, RFS	NCT04810910
GVAX	Cyclophosphamide, Pembrolizumab, Radiation	2	Active	DMFS	NCT02648282
GVAX	CRS-207 (Listeria-Mesothelin)	2	Completed	OS	NCT01417000
Other trials of interest
RO70097890 (Anti-CD40)	Nab-paclitaxel, Gemcitabine	1	Completed	TRAE	NCT02588443
Hydroxychloroquine (Autophagy inhibitor)	Gemcitabine, Nab-paclitaxel, Paricalcitol	2	Recruiting	Tumor size	NCT04524702
Bethanechol (mAChR agonist)	Gemcitabine, Nab-paclitaxel	2	Recruiting	R0%	NCT05241249

AE: Adverse Event; DLT: Dose Limiting Toxicity; DMFS: Distant Metastasis Free Survival; MTD: Maximum tolerated dose; ORR: Objective Response Rate; OS: Overall Survival; PFS: Progression Free Survival; SAE: Serious Adverse Event; R0%: R0 resection rate; RFS: Relapse Free Survival; RP2D: Recommended Phase 2 Dose; RR: Response Rate; TRAE: Rate of treatment related adverse events

Data are available from https://clinicaltrials.gov/

Summary and discussion

In summary, the late diagnosis of PDAC and poor response to treatment lead to poor prognosis and low survival rate of patients. In treating advanced PDAC, traditional treatments such as surgery, chemotherapy, and radiotherapy can only achieve limited benefits. Therefore, immunotherapy has made breakthroughs in other cancer types and has become a new direction of PDAC treatment research. However, due to the low cell infiltration, potent immunosuppression, high heterogeneity, and high stroma density of the TME, PDAC is defined as a "cold tumor." Furthermore, its response to most immunotherapy is unsatisfactory. Therefore, one of the current significant endeavors in PDAC immunotherapy is to transform the "cold tumor" into a targetable "hot tumor." However, this process requires sufficient targets, such as immune checkpoint molecules and high-quality tumor antigens (like ovalbumin and OVA [208]). Besides this, more in-depth explorations of the PDAC TME are required to enhance the efficacy of immune checkpoint inhibitors by combined methods, find more systematic and complete immune regulation mechanisms, and improve clinical relevance to guide the immunotherapy of PDAC to strive for a better prognosis for patients.

However, the current understanding of the TME of PDAC is far from complete, so further descriptions of the TME are necessary for immunotherapeutic research. The overall state of TME, such as inflammation, acidosis, hypoxia, nutrient deficiency, etc., can affect the immune response. TME hypoxia can accelerate tumor metastasis through the HIF-1α/CCL20/IDO axis by inducing EMT and establishing an immunosuppressive TME, and hypoxia can also promote tumor drug resistance [209, 210]. The nutrient status of the microenvironment influences tumor proliferation, apoptosis, and autophagy [211]. The degree of metastasis may also determine the presence of T-cell-depleting immunosuppression [212]. As previously stated, the characteristics of the cells within the PDAC TME distinguish each type, necessitating customized strategies. For example, the superiority of certain subtypes within the T-cell population in the TME, the biased polarization of macrophages when associated with the tumor that causes changes in the levels of cytokines and cell functions, the altered function of NETs in the pathological condition of the tumor, and the unneglectable member who contributes a lot to the heterogeneity of the TME, CAF. More importantly, we should recognize that the TME components interact as a network. As a result, the TME may be much more complicated and intertwined, with an unknown number of potential sites yet to be explored.

The multi-target combination may be the key research direction in PDAC immunotherapy. Currently, the theoretical feasibility and experimental success of integrated immunotherapy, including ICB, tumor vaccination, and ACT, raise the prospect of overcoming the immunosuppression barriers. By conducting adequate clinical trials, now is the time to turn these into clinical applications. This article reviews ICB’s integrated immunotherapies, such as anti-PD-1/PD-L1 and anti-CTLA-4, TGF-β inhibitors, FAK-related inhibitors, and metabolic checkpoint inhibitors like glutamine-related inhibitors in PDAC combined with other treatments like tumor vaccines. In some trials, increased CD8⁺ T infiltration is associated with an enhanced immune effect, and clinical remission of the tumor is also observed. Furthermore, immunotherapies combined with traditional cytotoxic drugs or radiotherapy have also shown effects in hepatobiliary and pancreatic system tumors [52]. More importantly, tailored therapies will need to take into account molecular/immune subtypes in addition to knowing combination therapies. In the future, personalized multi-drug combinations with ICIs, vaccination, ACT, or even cross-platform multi-regimen integrated immunotherapy targeting specific subtypes of PDAC may be the solution for the disease.

Authors’ contributions

Conception and outline design: YW. Sun, S.H. Jiang, D.J. Liu, Y.M. Huo; Acquisition and organization of references: Y.H. Zhu, J.H. Zheng, Q.Y. Jia, Z.H. Duan; Summarizing and reviewing of references: Y.H. Zhu; Verification and inspection: J.H. Zheng, Q.Y. Jia, Z.H. Duan; Revision of the manuscript: S.H. Jiang, D.J.Liu, Y.M. Huo. All authors read and approved the final manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (grant number 81874175 to Y.W. Sun; 81902377 to D.J. Liu;) and Project 2021RJLY-007 to Y.M. Huo.

Availability of supporting data

Not applicable.

Declarations

Ethical Approval and Consent to participate

Not applicable.

Consent for publication

All authors declare that they agree to submit the article for publication.

Competing interests

The authors declare that they have no competing interests.