Keywords

The Immune Response in Pancreatic Cancer and Its Major Players

The immune system can be both harmful and beneficial during carcinogenesis and progression of pancreatic cancer (PC). The ability of both innate and adaptive immune cells to exert either tumor-suppressive or tumor-promoting properties yields a mosaic pattern of immune cell composition in the tumor microenvironment (TME) . Therefore, an understanding of the individual components of this mosaic is required to develop efficient therapeutics.

Chronic inflammation is an important characteristic of PC, which is maintained by a complex interplay of immune cells in the TME [1, 2]. The myeloid compartment has many components, undoubtedly the most important one of them being tumor-associated macrophages (TAM). TAMs are found as M1 or M2 macrophages, which are classified according to the cytokine profile and surface markers they express [3]. Both M1 and M2 macrophages derive from monocytes. M1 macrophages, as “good cops,” produce pro-inflammatory cytokines like TNF, IL12, IL-1β, and IFN-γ and show tumoricidal activity and induce an antitumor Th1 immune response. On the other hand, M2 macrophages, as the “bad cops,” produce anti-inflammatory tumor-promoting cytokines like TGFβ and IL-10 and stimulate a Th2 immune response [3]. Next to TAMs, myeloid-derived suppressor cells (MDSC) are produced from immature myeloid cells and are known to suppress adaptive immunity with the recruitment of regulatory T cells (Treg) to the TME and by reducing antitumor T cell activation [4, 5]. In line with this, the presence of immunosuppressive cells like M2 macrophages, MDSCs, and Treg cells in the pancreatic TME has been shown to negatively correlate with overall survival [6,7,8,9,10,11]. Although both pro- and antitumorigenic abilities of neutrophils are reported, the inhibition of neutrophil recruitment to the TME remains a promising option in preclinical studies [12,13,14].

In the adaptive immune system, antigen-presenting cells (APC) such as dendritic cells (DC) can prime naïve T cells broadly into functional CD4+ helper T cells (Th) or CD8+ cytotoxic T cells (CTL) [15]. Th cells are further mainly characterized as Th1, Th2, and Treg, and their coordination is highly deterministic for the type of tumor immune response [15]. Th1 cells as conductors of an antitumorigenic response promote antigen presentation on APCs and cytolytic activity of CD8+ T cells and boost M1 macrophages [16, 17]. However, Th2 and Treg cells are pro-tumorigenic since they can oppose the Th1 immune response and escalate T cell exhaustion. Their presence is correlated with reduced survival in PC patients [18,19,20,21,22,23]. CD8+ CTLs are the “best cops” in tumors, since they can directly recognize tumor cell-specific antigens and induce cancer cell death [15, 24].

Immunotherapy for PC: Obstacles and Potential Solutions

Boosting the adaptive immune response is one of the most attractive goals in cancer therapeutics: Other than generating a repertoire of T cells recognizing tumor-specific antigens, the ability of the adaptive immune system to form an immunological memory holds promise for long-term disease control [25]. Immunotherapeutic approaches, currently being established as a fourth pillar of cancer therapeutics (next to chemo-/targeted therapy, radiotherapy, and surgery), augment the antitumor adaptive immune response [26]. Immune checkpoint inhibitors are the best studied candidates in immunotherapeutic options so far. While checkpoint inhibitors like anti-CTLA-4 and anti-PD-1 antibodies showed very promising results in clinical studies for many solid tumors and hematologic malignancies, as single agents or in combination, they appear to be ineffective in PC [27,28,29,30,31,32,33,34,35,36]. Therefore, precise understanding of the immune cell network in PC is essential to explore ways to exploit immunotherapeutic approaches for treatment of patients with PC.

Immune Checkpoint Inhibition

CTLA-4 and PD-1 were the first immune checkpoint targets discovered and evaluated for cancer immunotherapeutics [37,38,39]. During APC:MHC molecule engagement with T cell receptor (TCR) on T cells, axes of co-stimulatory and co-inhibitory signals in T cells mediate T cell activity. These co-signaling pathways are essential for physiological homeostasis since an imbalance can cause either autoimmunity or disability to fight invaders. Tumors may evolve the ability to skew this balance by reducing co-stimulation and inducing co-inhibition to impair antitumor T cell activity [40]. CTLA-4 and PD-1 are such co-inhibitory molecules leading to T cell anergy and exhaustion [41,42,43,44]. Antibodies targeting CTLA-4 and PD-1 can impair such signaling pathways in T cells and boost an antitumor cytotoxic immune response in tumors.

The question is though, why checkpoint inhibitors are not effective in PC as opposed to other solid tumor entities. PC owes this to its extreme immune-privileged nature [45]. Immune privilege is the ability to retain the production of antigens, without creating an anti-tumor immune response [46]. Normally, during carcinogenesis, tumor cells produce unique antigens (de novo mutations, re-expression of embryonic stage proteins), which may be recognized by the immune system, potentially leading to tumor cell elimination. During the immunosurveillance process (a hypothesis developed by Paul Ehrlich), the immune system continuously inspects the body for any malignant transformation [47,48,49]. However, some transformed cells have the ability to escape detection in a process called immunoediting. Immunoediting proposed by Schreiber and colleagues comprises three phases (triple E): elimination, equilibrium, and escape [50]. During the elimination phase, most of the transformed somatic cells die due to immunosurveillance, while the remaining survivors in the equilibrium step no more respond to immune reaction. Through a Darwinian-like selection, these clones proliferate and expand within the escape phase. While many tumors undergo the triple E of immunoediting process, PC holds a unique state [51, 52].

PC carcinogenesis is different in terms of the immunoediting process compared to many other solid tumors. With the use of genetically engineered mouse models (GEMMs), PC was shown to have an immunosuppressive microenvironment and a scarcity of antitumor T cells already during the carcinogenesis process [45]. Due to immunosuppression, the adaptive immune system is not educated toward recognition of any tumor-specific antigens, bypassing the elimination phase of triple E. With this rather immune quiescence-like phenotype, PC limits the entry of antitumor immune cells into the microenvironment maintaining its immune privileged status [51].

Overall, an approach to augment T cell entry and activity in the PC microenvironment may have the ability to render PC cells responsive toward immune checkpoint inhibitors. The factors which will determine such responsiveness are (1st) antigenicity of cancer cells and (2nd) immunogenicity of the tumor in general [53].

Antigenicity is the degree to which tumor cells produce and present neoantigens to generate an antitumor adaptive immune response [53]. These antigens can be divided into tumor-specific antigens (TSA) and tumor-associated antigens (TAA). TSAs are produced upon tumor-specific mutations of genes or reactivation of genes for embryonic development, which are not occurring in healthy somatic cells, while TAAs are wild-type proteins but expressed higher in tumor cells compared to somatic ones [54]. Production and MHC-mediated presentation of such antigens determine the level of antigenicity of tumors [53, 54].

Tumors carrying a high mutational burden generally respond better to checkpoint inhibition since they have a diverse tumor-antigen responsive T cell repertoire [55,56,57]. PC on the other hand doesn’t carry such mutational load, compared to other entities [58, 59]. However, a subgroup of PC patients, representing around 1% of a patient cohort, carry mutations leading to mismatch repair (MMR) deficiency and microsatellite instability (MSI) and may profit from checkpoint inhibitors [60, 61]. As a result, anti PD-1 immunotherapy is approved by FDA for solid tumors including PC with MMR deficiency and MSI [62]. Moreover, one study identified long-term survivors in a PC patient cohort based on their ability to express good quality neoantigens, but not quantity [63]. Most importantly, a decrease in neoantigen quality of metastatic tumors compared to their respective primaries implied the importance of immunosurveillance in cancer metastasis and its implication in therapeutics [63]. Other than antigen production, presentation of these antigens via MHC molecules has been shown to be reduced in PC through the activation of oncogenic drivers like RAS [64,65,66]. Also, reduced MHC expression in disseminated PC cells appears to be an important driver of metastasis [67]. Since a correlation between antigenic load and immune checkpoint inhibition efficacy is absent in PC, as opposed by other solid tumor entities, in addition, factors determining immunogenicity of PC require exploitation.

Tumors with better ability to induce an adaptive immune response are considered immunogenic. This ability can be modulated both at the tumor cell level and at the level of cross talk of tumor cells with cells of the TME [53]. Transcriptomic analyses revealed an immunogenic subtype of PC, showing higher cytolytic T cell activity, antigen presentation, and CTLA-4 and PD-1 signatures [68]. Signatures as those may help to predetermine the prognostic value of checkpoint inhibitor therapy in the context of “personalized medicine” [69].

Tumor cell-specific immunogenicity can be decreased upon co-inhibitory checkpoint ligand expression in tumor cells, such as PD-L1. In various solid tumors, PD-L1 expression by tumor cells is increased due to oncogenic signaling pathways like PI3K, Hippo, Myc, and JAK-STAT [70,71,72,73,74]. In PC, the myeloid compartment was shown to induce EGFR-dependent MAPK signaling, leading to an increase of PD-L1 production in tumor cells [75]. An imbalance of autophagic modulation in mitochondrial iron homeostasis also may induce PD-L1 expression by pancreatic cancer cells [76].

Reprogramming the Tumor Microenvironment

Even if specific cancer cells are sufficiently antigenic and immunogenic, they may still not respond well to checkpoint inhibition due to an overall impaired immunogenicity mediated by the corresponding tumor tissue. The immunosuppressive TME is the main player in this context. An understanding of the responsible TME compartments, and of their cross talk with antitumor adaptive immune cells, is essential to reveal options for boosting immune checkpoint inhibitor response (Fig. 14.1).

Fig. 14.1
figure 1

The good and the bad cops of the tumor microenvironment and how to target them to boost a favorable immune response in PC. M1: M1 macrophages, M2: M2 macrophages, MDSC: myeloid-derived suppressor cells, CTL: cytotoxic T lymphocytes, DC: dendritic cells, Treg: regulatory T cells, CAF: cancer-associated fibroblasts

Full size image

Cancer-associated fibroblasts (CAFs) are the leading actors regarding the characteristic desmoplastic stroma formation in PC. Various studies revealed a binary action of stromal cells in the immunogenicity of PC. One study revealed a positive correlation between type-I collagen production and CTL infiltration in tumor specimens of PC patients, whereas another showed the inhibition of CTL activity by αSMA+ CAFs [77]. Other studies demonstrated an inhibitory action of CAFs toward CD8+ T cell infiltration [78]. While most of the research so far implies the prognostic value of “stromal remodeling” in PC, an understanding of CAF action heterogeneity in the TME may provide options to improve the efficacy of immune checkpoint inhibitors. For example, with the use of preclinical mouse models, impairment of CXCR4 or IL-6 signaling in CAFs was shown to be synergistic with anti PD-L1 therapy [79, 80]. Stromal remodeling with FAK inhibitors reduced the immunosuppressive milieu in the TME, increasing chemotherapy-checkpoint inhibitor combination therapy efficacy [81]. Previous studies showed the benefit of hyaluronan depletion and vitamin D receptor activation in stromal remodeling [82,83,84,85]. Here, a combination therapy with immune checkpoint inhibitors may have therapeutic impact.

The myeloid compartment is a double-edged sword as also mentioned earlier. Years of research dissected the complex roles of individual components in PC. Studies focusing on CD40 agonist treatment of PC actually revealed the quite unique properties of PC. Treatment of preclinical mouse models with a CD40 agonist (acting on APCs increasing their capability to prime CTL) in combination with gemcitabine created an only mild response by remodeling the stroma and reprogramming immunosuppressive myeloid cells inside the TME [86]. However, this regimen was not enough to create an adaptive immune response in tumors. The subsequent studies identified a subtype of immunosuppressive macrophages (Ly6Clow F4/80+), accumulating in the tumor periphery. These macrophages were shown to prevent CTL migration into the TME [87]. Finally, a combination therapy of nab-paclitaxel with gemcitabine and CD40 agonist revealed a synergism allowing penetration of active CTLs [88].

Re-education of neutrophils, MDSCs, and TAMs can also be achieved via various inhibitors targeting CSF1R, CXCR2, or RIPK1, which demonstrated synergism with immune checkpoint inhibitors in preclinical studies [14, 89, 90]. Other than directly targeting the myeloid compartment, inhibition of B cell-specific Bruton’s tyrosine kinase (BTK) reprogrammed tumor resident macrophages indirectly, increasing the antitumor immunity [91].

Immunosuppressive immune cells impair immunosurveillance not only via cytokine-chemokine release but also through generation of a metabolite-restricted TME. Arginine depletion via arginase-1 produced by TAMs and MDSCs limits T cell activity [92, 93]. Further, the immunosuppressive metabolite kynurenine is produced from tryptophan as a by-product of indoleamine 2,3-dioxygenase (IDO-1) enzymatic activity. IDO-1 expression from cancer cells, TAMs, and MDSCs not only limits tryptophan availability for antitumor T cells but also increases inhibition of T cell activity by kynurenine [94]. Adenosine production by Treg cells and prostaglandin E2 production from TAMs and MDSCs are also responsible for antitumor T cell activity impairment [95, 96].

Immunotherapeutic Properties of “Classical” Treatment Approaches

Other than targeted inhibitors, chemotherapeutic agents and radiotherapy also have the ability to convert nonresponsive, “immunologically cold,” tumors to responsive, “immunologically hot,” tumors. Chemo- and radiotherapy can boost both, antigenic properties of cancer cells due to their mutagenic effect and also immunogenicity of the tumor due to the induction of immunogenic cell death and subsequently enhanced inflammation [97, 98]. Next to their direct effect on cancer cells, such treatments may also alter the composition of immunosuppressive immune cells in the TME [88, 99]. Strikingly, immune checkpoint inhibition in cancer may not only enhance the response to radiation therapy in primary tumors but also has the potential for an abscopal response in metastatic sites [100, 101]. In conclusion, while chemotherapy and radiotherapy still are the gold standard therapies for cancer treatment, their combination with checkpoint inhibitors may be the next step to both increase the treatment response and T cell memory for long-term disease control, even for PC. Essentially, analysis of respective clinical trials may inform about dosing, sequence of treatment, and specific subgroups profiting most from the expected synergism.

Other Strategies for Boosting the Antitumor Immune Response

Immunotherapeutic approaches are not only limited to immune checkpoint inhibitors.

Oncolytic viruses (OV) can be designed to only target tumor cells, but not healthy somatic ones. This specificity can be achieved at multiple levels [102]. At the physiological level, OVs are not equipped to win a combat against healthy cells. Tumor cells, however, already may have imbalanced interferon signaling and increased cellular metabolism coupled with proliferation making them vulnerable towards viral infection. OVs can also be designed to take advantage of tumor-specific expression of cell entry receptors or transcription factors, limiting their action on healthy cells.

Cancer vaccines aim to boost adaptive immune response in the host against tumors. They can be produced as either whole cell (e.g. GVAX) or antigen-specific vaccines. GVAX is composed of pancreatic cancer cells genetically engineered to secrete GM-CSF with the aim to convert “cold” tumors to “hot” ones, and these cells are irradiated to prevent further proliferation [103]. Listeria vaccine is an engineered bacterial strain to secrete TAAs such as human mesothelin, boosting antitumor CTL activity. An approach with total cell followed by antigen-specific vaccine may recapitulate a “prime and boost” scenario [104].

Chimeric antigen receptor T cells (CAR-T) are genetically designed to express a receptor construct comprising an antibody-like ectodomain targeting TSAs and a TCR-like endodomain, bypassing the need for MHC engagement [105]. Upon antigen recognition they exert their cytotoxic properties. CAR-T cell therapy requires adoptive T cell transfer (ATC), in which patient’s T cells have to be isolated, expanded, and genetically engineered. Without a genetic manipulation, in vitro induction and expansion of TILs (TIL-ATC) is also a valuable approach to exploit tumor targeting not only by a single antigen but a pool of them [106, 107].

Currently Ongoing Clinical Trials for Immunotherapy of Patients with PC

An overview of clinical trials based on abovementioned preclinical studies is given in Table 14.1. Overall, these studies reveal that PC is actually antigenic enough to create an antitumor adaptive immune response. However, the main barrier to be exceeded is the immunosuppressive microenvironment, which blocks the antitumor T cell priming and infiltration. One important factor is that many of these studies for PC are still in their early stages. Thorough analysis of each of these trials will pave the way to dissect individual rationales for combination therapies.

Table 14.1 Selected clinical trials aiming to induce an antitumor immune response in pancreatic cancer
Full size table