Most pancreatic tumors (around 95%) manifest from the exocrine parenchyma of the gland, arising from connective tissue, acinar cells, or the ductal epithelium [1]. Pancreatic duct adenocarcinoma (PDAC) is the most common, accounting for more than 80% of pancreatic cancer cases ^[1][2][1,2]. Pancreatic cancer is associated with poor survival and the incident cases and number of deaths have reportedly doubled from 1990 to 2017 [3] and remain on a trending increase [4]. New treatment strategies are clearly necessary to enhance patient outcomes.

Immunotherapy, and in particular immune checkpoint inhibitors (ICIs), enable the activation of T cells to clear tumor cells. This is mainly by blocking inhibitory signals arising from the interaction between programmed death receptor 1 (PD1) and the cytotoxic T lymphocyte-associated protein 4 (CTLA4) with their ligands, programmed death ligand 1 (PD-L1) and B7 ligands, respectively [5]. ICIs have significantly enhanced the survival of treatment-refractory patients with metastatic melanoma [6] and non-small cell lung cancer [7], but have thus far yielded dismal responses in pancreatic cancer, even when applied in combination with chemotherapy ^[5][8][5,8]. The only patient subset with acceptable responses is those having microsatellite instability (MSI) or deficient mismatch repair (dMMR); in addition, high tumor mutational burden (TMB) has been significantly associated with improved overall survival after ICI treatment [9]. In patients suffering from PDAC, however, only 1–2% are mismatch repair deficient and most cases present with low TMB (<10 mutations/megabase) ^[10][11][10,11]. Other biomarkers of response, namely PD-L1 expression and infiltrating CD8+ T cells, have not had an established role in selecting patients with PDAC for ICI therapy [5]. An improved understanding of tumor evolution and the PDAC tumor microenvironment (TME) is necessary to bring forth more pertinent makers of responses in this disease and better combination treatments.

2. Role of Hypoxia in PDAC Carcinogenesis

Findings from various studies have been recently combined to propose an updated evolutionary model for PDAC, where often a simple KRAS-activating mutation transforms a ductal epithelial cell, contributing to low-grade pancreatic intraepithelial neoplasia (PanIN). Subsequent incursion of loss of tumor suppressors involved in the cell cycle control, namely TP53, CDKN2A and/or SMAD4, promote cell growth and the progression of the lesion to high-grade PanIN. An unknown trigger, which could be telomere loss, promotes complex mitotic errors manifesting as polyploidy in most cases, and chromothripsis in others, enables faster acquisition of structural alterations and copy number variations (CNVs). The result is rapid proliferation, heterogenous driver and pathway alterations, a compendium of transcriptional subtypes, KRAS allelic imbalance, invasion, and dissemination [12]. On the other hand, in KRAS wildtype (WT) PDACs, which represent 10.7% of cases, a recent report determined that TP53 is the most frequently mutated gene closely followed by BRAF [13], while gene amplification and fusion events occur in 10% and 21% of KRAS WT cases, respectively [13]. In addition, a small fraction of PDACs present with single base substitution (SBS) signatures of homologous recombination deficiency (HRD) (10%) or dMMR-related signatures (1–2%). HRD and dMMR are primarily due to germline alterations in BRCA1, BRCA2, PALB2 or RAD51C involved in the homologous recombination repair (HRR) pathway, or MLH1, MSH2, MSH6 or PMS2 involved in MMR. Germline alterations are followed by a second somatic mutation, leading to biallelic inactivation ^[14][15][16][14,15,16]. CNVs and structural variants have also been surveyed in PDAC genomes distributing them into four subtypes. This included an “unstable” subtype, which harbors greater than 200 variants [17], and was found to be associated with the HRD signature. Of interest, a recent study correlated readouts from transcriptome-based hypoxia gene signatures across tumor types with copy number signature attributes. The authors reported a significant positive correlation between hypoxia as determined by these signatures, and attributes related to HRD and aneuploidy [18]. Indeed, hypoxia has been associated with the increase in genomic instability and TMB ^{[19][20][21][22][23][24]}[19,20,21,22,23,24]. In that respect, hypoxia could potentially increase neoantigen load, thus giving rise to the emergence of cancer clones that could potentially be recognized as non-self and be eliminated by the immune response. For that to be achieved however, anti-tumor immune cells need to be within the tumor mass, which is a rare scenario in PDACs that are branded as being immune-cold tumors characterized by a highly desmoplastic, hypovascularized, and hypoxic TME ^[25][26][25,26]. The adaptive response to hypoxia confers PDAC malignancy by promoting more aggressive and treatment-refractory phenotypes [26]. The primary activator of the hypoxia-mediated response is the hypoxia-inducible factor-1α (HIF-1α). The stability of this transcription factor is determined by the presence of oxygen since the protein responsible for initiating its degradation, prolyl hydroxylase (PHD), is activated in an oxygen-dependent manner. In the TME, the drop in oxygen levels associated with hypoxia will stabilize HIF-1α, enabling its translocation to the nucleus where it interacts with the HIF-1β subunit giving rise to a transcriptionally active heterodimer that induces the expression of more than a hundred genes. This is achieved by engaging the hypoxia response elements (HRE) in their promotor regions. Among the activated genes are those necessary for sustaining energy production in the cells, which mainly occurs via the activation of glycolysis, inhibition of oxidative phosphorylation, and by increasing the expression of glucose transporters to enable the higher uptake of this nutrient [27]. The increased intracellular acidity resulting from the hypoxia-promoted metabolic switch is counteracted though the HIF-1α induction of factors responsible for shuttling the excess lactate and hydrogen ions outside the cell. The cumulative effect is a nutrient-deprived and highly acidic TME that is hostile to the function of tumor-antagonizing cells such as cytotoxic T-cells, while being highly favorable to regulatory T cells (Tregs) and M2-polarized macrophages, which support tumor growth [28]. In addition, the leaky, haphazardly structured, and disorganized blood vessels that are instigated by hypoxia act as a physical barrier to the recruitment of immune cells [29]. A second canonical inducible activator of the response to hypoxia is the HIF-1α homologue, HIF-2α ^[30][31][30,31]. While there is an overlap between the responses triggered by HIFs, each isoform has a specific set of target genes ^[30][31][30,31]. With respect to pancreatic cancer, there have been controversial reports on the involvement of HIF-2α; however, the regulation of β-catenin by HIF-2α was found to be critical for the formation of early pancreatic lesions [32]. A more in-depth study further confirmed the crosstalk between HIF-2α and Wnt/β-catenin signaling by showing that the interaction between the two proteins increased the activity of β-catenin, while enhancing the stability of HIF-2α [33]. Furthermore, HIF-2α was shown to promote pancreatic tumor cell proliferation, metabolic shift and stemness features. It was also correlated with markers of epithelial-to-mesenchymal transition (EMT) in vivo; and with a worse prognosis in patients with pancreatic cancer [32]. In a more recent study in PDAC mouse models, HIF-2α expression in cancer-associated fibroblasts (CAFs) was found to play a key role in tumor progression and growth, as well as the recruitment of immunosuppressive immune cells to the TME [34]. Interestingly, treating PDAC mice with a HIF-2α inhibitor reduced immunosuppression and enhanced the response to immunotherapy [34]. Similar findings have been reported when combining other hypoxia-targeted approaches with ICIs, as discussed later. Indeed, relieving hypoxia could transform immune-cold tumors to immune-hot and potentiate responses to immunotherapy.