A comprehensive review of pancreatic cancer and its therapeutic challenges

Genetic alterations during pancreatic tumorigenesis

Biological and biochemical analyses have demonstrated high genomic complexity in pancreatic cancer, in which mutations are frequently detected in somatic cells, and the tumors present highly heterogeneous with various genetic changes [14, 15]. These studies reveal high frequencies of genetic alterations in the tumors, which are noticed as changes of expressions of tumor suppressor genes and oncogenes, such as K-ras, T53, p16^INK4A/ARF, MLH1, LKB1, PRSS1, and BRCA2 [16–18]. Among these genetic changes, we here focus on describing the oncogenic K-ras and, tumor suppressors’ p53 and p16^INK (CDKN2, MTS1).

Mutant K-ras is detected in more than 90% of pancreatic cancer patients and considered as one of the possibly early elements in pancreatic cancer pathogenesis [16, 19–21]. K-ras is located on chromosome 12 and its encoded protein belongs to the Ras family of GTP-associating proteins. Ras family consists of three major forms, K-, H. and N-Ras and the active form of Ras transmits signals to promote various cellular activities, including cell growth, differentiation, survival and apoptosis under certain circumstances. In most of pancreatic cancer, the single mutation is detected at the position G12 of the amino acid sequence of K-Ras, which is substituted glycine to aspartic acid or valine [22, 23]. Less frequent missense mutations of K-ras are at the codons of 13, 59, 61 or 63 [24–26]. Balletic mutations of this oncogene are also revealed by deep sequencing of exomes [27]. Studies reveal the peri-ductal lymphocyte infiltration and gastric mucous neck cell hyperplasia in the pancreases of the genetically modified mutant K-ras mice [28, 29]. Furthermore, the specific mice are generated by crossing the knock-in mice in which a Cre-activated Kras^G12D is knocked into the endogenous K-ras locus with mice expressing Cre recombinase that is expressed by a Pdx1 (pancreatic islet specific) or Ptf1-p48 (pancreatic acinar specific) [30–32]. These transgenic mice express Cre recombinase under the control of the mouse Pdx1 (pancreatic and duodenal homeobox 1) promoter. Mosaic Cre recombinase activity is detected in the pancreatic epithelium, antral stomach and duodenum in neonates and in pancreatic beta islet cells in adults. Specifically, early lesions in the mice are detected, in which the Notch pathways is noticeably activated with over-expressions of cyclooxygenase-2 (COX-2) or MMP-7, sometimes accompanied with metastasis. Thereafter, approximately one year after the development of PanIN, some of these genetic engineered pancreatic cancer mice develop pancreatic cancer (mainly PDAC), in which the origin of the tumors appears from acinar cells or acinar precursors. The spontaneous knockout or mutations of p53 or p16 permits a full penetrance of the cancer in the animals. Thus, these studies suggest the close connection among mutant K-ras, PanIN and pancreatic tumors.

Mutations in K-ras prohibit its encoded onco-protein to associate with and further be inhibited by GTPase-activating proteins (GAPs), leading to mutant Ras staying in a persistently or constitutively active status. As the result, multiple downstream effector pathways of Ras, such as Raf/ERK1/2/MEK or PI3K/Akt, are activated for promoting uncontrollable cell growth, desensitizing cell death, remodeling cellular metabolism, escaping from immune surveillance and increasing cell invasion. Studies showed that pancreatic tumor cells can be shed and further circulate in the blood stream. Therefore, mutant K-ras can be detected in circulating tumor cells, which has been used to facilitate clinical diagnostic imaging analyses [33].

p53 is located on chromosome 17 and its encoded protein serves as a tumor suppressor for protecting the integrity of the genome of cells. This tumor suppressor is often inactivated or mutated in about 70% of pancreatic cancer and most of mutations are the missense mutations, especially at the locus of R248, R273 or R175 [34]. Using animal models, studies demonstrate that the loss of heterozygosity (LOH) of p53 is an important factor for driving pancreatic cancer progression [35, 36]. Specifically, loss of p53 appears to cooperate with oncogenic K-ras-induced pancreatic cancer initiation and progression, by perturbing cell cycle progression, impairing DNA damage repair, augmenting survival activities and hindering apoptosis in cells [35, 36]. p53 mutations are more commonly detected in mutant K-ras pancreatic tumors than that expressing wild-type p53, suggesting that K-ras mutations developed in early stages of pancreatic tumorigenesis, creates a genetic background favoring p53 mutations. The cooperation of these two oncogenes or proteins promotes uncontrollable cell growth, cell cycle progression, improper damage repair and establishment of genetic instability, which promotes pancreatic tumorigenesis [35, 36]. p16^INK (CDKN2, MTS1) is another tumor suppressor and its inactive form is found in a large number of pancreatic cancer patients [16, 37]. This suppressor gene is anchored on chromosome 9 and the encoded protein regulates the cell cycle by preventing cells from improperly entering the S phase through inhibiting cyclin-dependent kinase (CDK) 4/6. Due to the inactivation of p16^INK by the promoter methylation, missense mutation and deletion, its related cell cycle checkpoint is perturbed, which allows pancreatic cancer cells to improperly progress from G₁ to S phase without repairing potential damages. As the result, risks of genetic instability in cells are increased. This is further demonstrated by the animal study that the knockout of p16^INK causes the deregulation of the cell cycle transition and rapid advance of pancreatic tumorigenesis a K-ras transgenic mouse strain [38]. Overall, mutations of p16^INK and p53 are frequently observed in pancreatic tumors and the linear relationship of these two tumor suppressors in pancreatic tumorigenesis remains unclear.

Epigenetic abnormalities that alter DNA methylation, histone modification or microRNA expression are other factors to change gene functions in driving and promoting pancreatic tumorigenesis [32]. In some pancreatic cancers, tumor suppressor or DNA repair genes (such as CDKN2A, CDH1 and MLH1) are found to be silenced by methylation [16]. The over-expressions of microRNAs in pancreatic cancer have also been revealed, which seemed to participate in pancreatic neoplastic development [39–44].

Pathological precursors of pancreatic cancer and its subtypes

During the development of cancer in the pancreases, the conventional model of the progression of pancreatic cancer suggests that early genetic changes initiate tumor-prone activities in a cell or few cells that then undergo clonal expansion to achieve a full transformation. Another scenario is that pancreatic cancer cells disseminate early and then undergo transformation independently [38]. It is also being suggested that pancreatic cancer is originated from acinar cells that undergo the process of the acinar-to-ductal metaplasia (ADM), during which K-ras mutations are acquired [39–41]. Overall, genetic-initiated alterations are considered as the bases for the classification of different subtypes of pancreatic tumors. In recent years, using advanced biological techniques, such as genomic, transcriptomic and proteomic assays that enable to identify different characteristic clusters of tumor cells with different gene expressions and mutations, studies demonstrated that pancreatic tumors can be divided into different subtypes [39–41]. In particular, global gene expression analysis reveals three subtypes of PC as: classical, quasi-mesenchymal, and exocrine-like subtypes [39]. It also pointed out that the prognosis and therapeutic responses of these subtypes are different. The classical subtype of pancreatic cancer expresses high levels of adhesion-associated and epithelial genes. The quasi-mesenchymal subtype shows an augmented expression of mesenchyme-related genes. The exocrine-like subtype contains the tumor cells that express high levels of digestive enzyme genes. Three metabolic subtypes were also identified in pancreatic cancer patient samples by metabolomics analysis, as slow proliferating, glycolytic and lipogenic subtypes [42]. The classifications of the subtypes of the cancer by these methods correlate well with each other. In addition, the analyses of the genomic sequencing plus copy number variation measurement demonstrated the mutation landscapes of the cancer and four subtypes are accordingly classified as stable, locally rearranged, scattered and unstable [43]. The tumor-specific and stroma-specific subtypes of pancreatic cancer were also classified by RNA sequencing and computational analysis [44]. It is noticeable that various immunological features of the tumors are recently included for determining the tumor subtypes [41]. Taken together, the uses of current available modern techniques permit better understandings of pancreatic cancer subtypes at molecular levels and more accurate predictions for outcomes of treatments.

Although advanced technology helps us obtain the molecular insight into pancreatic cancer, it still remains enigma about the association of PanINs to the onset of pancreatic cancer, as inflammatory lesions are frequently detected in pancreatic cancer patients in early stages of the malignancy, but not in all PanIN patients. It is known after pancreatic inflammation or injury, acinar cells in pancreatic ducts start to lose their properties gradually and form lesions with changes around the ductal, which can be categorized pathologically in four grades (PanIN 1A, 1B, 2 and 3) [16, 38, 45]. The lesions of PanIN 1 are recognized by consisting of columnar epithelial cells with basally aligned nuclei [46]. The lesions are flat as PanIN 1A or papillary as PanIN 1B. PanIN 2 lesions have more changes in the nuclei manifested as loss of nuclear polarity, nuclear pleomorphism, hyperchromasia, or pseudo-stratification. In PanIN 3 lesions, a large degree of dysplasia exists, which alter architectures, such as formation of papillae or clusters of cells from the epithelium invading into the lumen of the duct, accompanied with various nuclear changes. Therefore, current research approaches for obtaining better defining subtypes, mutations of the cancer, together with external elements, certainly provide better distinguish of benign PanINs from those inflammation-associated neoplasia or tumorigenesis.

Overall, it is clear that through precursor lesions, pancreatic tumors in some PanIN patients are evolved [16, 38, 45, 46]. Along with the progression of a clone carrying neoplastic precursors to malignancy, mutations of K-ras and tumor suppressors drive tumorigenesis, which worsens with arising genomic instability or genetic heterogeneity for malignant advancing [47–49]. A better understanding of molecular alterations in pancreatic cancer development should provide early diagnostic opportunities for recognizing primary cancer lesions as well as early onset periods for effective clinic interventions. Various growth-related signaling pathways driven by oncogenes and mutated genes are involved in promoting the formation of invasive pancreatic tumors. The major aberrant signaling pathways driven by these mutant factors in pancreatic cancer initiation and progression are summarized in Figure 1.

Figure 1. Pro-pancreatic tumorigenic, aberrant genes/proteins and pathways. Signals mediated by mu-K-ras, mu-p53 and loss of p16^INK promote various uncontrollable cell growth, cell cycle progression and etc.

Pancreatic cancer-associated fibroblasts

One of the features of pancreatic cancer is that the stroma is extremely desmoplastic infiltrative, which appears to account for its being highly resistant to currently available chemotherapies [50]. Stellate cells in the pancreas are myo-fibroblast-like cells or cancer-associated fibroblasts (CAFs) and the main component in tumor microenvironment that is composed of an extensive remodeling of the extracellular matrix (ECM) and cancer associated changes in the stroma [51, 52]. The tumor stroma is fibrotic and mainly contains fibrillar type I collagen and hyaluronan (HA). CAFs are a major player in building up fibrotic ECM. CAFs organize collagen fibrils to cause epithelial to mesenchymal transition (ETM) and further provide paths for cancer invasion. An increase of HA levels functions as an instructor directing cancer initiation and progression [53, 54].

Pancreatic stellate cells express vimentin, α-smooth muscle actin (αSMA) and desmin [51]. Via secretion of various growth factors, cytokines and chemokines, these CAFs act as safety guarders for cancer cell proliferation and invasion [55]. Pancreatic cancer stroma consisting of CAFs, ECM, vascular endothelial cells and immune suppressive regulatory T cells, functions as a barrier to protect cancer lesions and prevent drug penetration. It was demonstrated by that the co-injection of non-invasive pancreatic cancer cells with CAFs into rats, the cancer cells became metastatic [56]. Furthermore, the expressions of the matrix metalloproteinases (MMPs) in the surrounding of pancreatic cancer lesions are increased, which accelerate the degradation of collagen IV and laminin for ECM remodeling [57, 58]. Although roles of CAFs in pancreatic cancer development have not been fully understood, studies using cell lines and mouse models demonstrated that putative cells carrying CAF markers cooperated with cancerous cells, which played important roles in establishing anti-neoplastic drug resistance [54]. These CAF cells built up resistant barriers via secreting various cytokines, changing ECM structures, establishing cancer-associated metabolisms and inducing epigenetic and genetic instabilities [59]. Thus, it can certainly assume that cross talks between CAFs and pancreatic cancer cells are essential in the establishment of the tumor-prone microenvironment not only for facilitating the expansion of tumor masses, but also for preventing therapeutic interventions or establishing resistance to anti-pancreatic cancer treatments.

Tumor suppressive lymphocytes

Another feature of pancreatic cancer is characterized by a highly immunosuppressive microenvironment that is built up by lacking intra-tumor effector T lymphocytes, existing mutant K-ras-driven oncogenic-prone inflammation, infiltrating suppressive immune cells and establishing the dense desmoplastic stromal reactions [16, 60, 61]. In order to build such cancer-prone environment, numbers of the effective immune cells [such as natural killer (NK) cells and CD8⁺ lymphocytes] are declined in pancreatic cancer lesions. On the contrary, regulatory T cells (Tregs), myeloid-derived suppressor (MDSCs) and tumor-associated macrophages (TAMS), together with CAFs, are significantly increased to seal the cancer cells from encountering the immune system.

T regulatory lymphocytes (Tregs) are suppressive T lymphocytes and can be recognized by their surface expressions of CD4⁺CD25⁺Foxp3⁺. Under normal conditions, Treg cells, via expressing CTLA-4 and secreting IL-10 or TGF-β, play important roles in inducing immune tolerance against auto-antigens, which prevents the occurrence of autoimmunity [59–61]. During tumorigenesis, Treg cells interfere with effector T cells to block immune responses against tumors. Pancreatic cancer cells, through chemokines or ligands/receptors, attract Tregs to translocate to their surrounding areas. For example, high levels of the ligands for chemokine receptor 5 (CCR5) were secreted by pancreatic cancer cells, which attract Tregs that express a fair amount of CCR5 [62, 63]. The interference of the association of the ligand and CCR5 prevented Treg migration to tumor lesions. Furthermore, TGF-β was shown to be required for recruiting Tregs to pancreatic tumor microenvironment [64, 65]. Higher populations of Tregs in pancreatic cancer lesions are strongly associated with poorer prognosis of the cancer patients.

Cancer-associated macrophages

Macrophages are the phenotypically different innate immune cells derived from circulating monocytes in the blood stream and exist abundant in all tissues of the body. The terms of M1 and M2 of macrophages are described as their functional states. M1 macrophages are tumor destroyers that can be stimulated by Th1-associated factors or bacteria and express high amounts of IL-12. In contrast, M2 macrophages that are responsive to Th2-related cytokines and have high expression levels of IL-10 are immunosuppressive and in favor of tumor promotion. These two kinds of macrophages with the extremities of the polarization can be detected in a tumor lesion and exhibited high heterogeneities. Increases of M2 cells represent the progression of tumors, which often are accompanied with declining CD8⁺ and effector CD4⁺ T lymphocytes [66, 67]. By secreting matrix proteins and protease such as matrix metalloproteinases (MMPs), the tumor-associated macrophages promote metastasis of pancreatic cancer cells. In addition, by producing angiogenic factors, for example, vascular endothelial growth factor (VEGF) or Cox-2, and immunosuppressive lymphokines such as IL-10, the tumor-associated macrophages promote further development of an immunosuppressive environment [68, 69]. Tumor-associated macrophages express the programmed death-ligand 1 (PD-L1) on their surface, through which they interact with and induce effector T lymphocytes to undergo apoptosis for achieving their immunosuppressive function.

In general, the profiles of tumor-associated microenvironment (TAM) are much more complex than we expect. TMA is infiltrated by dense stroma and functions pivotally in attracting cancer-related, various types of cells. Although tumor lesions exhibit different TAM phenotypes and are much more complicated, we here in a simplified way summarize a pancreatic tumor surrounded by a passive microenvironment barrier (Figure 2). Indeed, the tremendous plasticity of macrophages and TAMs provide a great therapeutic challenge, but also emphasize the urgent needs for developing TAM-targeted therapies.

Figure 2. Pancreatic cancer microenvironment. Pancreatic cancer lesions mainly include pancreatic cancer cells, suppressive lymphocytes, cancer-associated macrophages/fibroblasts and cancer-related cytokines/growth factors.

Overall, from many of in vivo, in vitro and in patient studies, the classification of pathological lesions in conjunction with changes in genome, epigenome and microenvironment provides vast information to construct the accepted scheme of pancreatic tumor initiation and development from normal pancreases to PanIN, tumors in situ and further metastatic ones (Figure 3). Importantly, mutant K-ras activated by single point mutations can be detected during PanIN stages and appears a crucial element in promoting the onset of pancreatic cancer. Notably, some studies indicate that not all PanIN patients expressing mutant K-ras develop the cancer, which may be due to lack of occurrences of or coordination from other genetic or epigenetic changes [16, 70]. Therefore, more thorough investigations will shed lights into the role of mu-K-ras in pre-cancer stages and identify its cooperators in promoting pre-cancer cells to cancerous status.

Figure 3. Pathological changes in the progression of pancreatic tumors.

Surgery

Surgery is currently the only potentially curative treatment for pancreatic cancer patients, but only about twenty percent of the patients appear to reach a questionable “cure”. Most of pancreatic tumors often progress to either metastatic condition or to locally advanced situation for further radical resection [71, 72]. The availability to classify the tumors as resectable, borderline resectable, locally advanced pancreatic cancer (LAPC) or metastatic, significantly improves outcomes of the surgical treatment and helps make decisions for other alternative treatments [72].

The conventional surgery using in removing pancreatic tumors is the pancreatoduodenectomy by resecting the distal stomach, the pancreas head, the bile duct plus the duodenum. For patients with lesions in the pancreatic body or tail, distal pancreatectomy is employed, which is frequently in conjunction with splenectomy. When the tumor is extensively growing in the pancreas, total pancreatectomy can be necessary to achieve a radical resection. Overall, the 5-year survival rate following surgery is dismal with less than 20% and however, when one of the adjuvant chemo-treatments is administered, the survival rate can be somewhat improved [72].

Pancreatic cancer chemotherapeutics

With a surgery aiming for cure, the recurrence rates of pancreatic cancer still are very high. Therefore, chemotherapy is the next, inevitable choice for patients after the operations. Nevertheless, the overall prognosis of patients undergoing adjuvant chemo-treatments remains dismal due to the low vasculature and building-up immunosuppressive cancer-associated microenvironment around the pancreas. Such barriers prohibit effective drug delivery or penetration, which appears less concern for patients receiving post-operative adjuvant treatments [72, 73].

Gemcitabine and its combination with other drugs are often used in pancreatic cancer chemotherapy. Gemcitabine is the first agent approved by the FDA for treating pancreatic cancer patients [74]. The known anti-cancer mechanisms of gemcitabine are primarily via generating the metabolite dFdCDP that inhibits ribonucleoside reductase and blocks DNA synthesis. This chemo-drug was initially used for treating virus infection and has then been implemented in anti-cancer therapy for solid tumors, including pancreatic cancer. Although its effect on overall survival rates is modest, its clinical beneficial response appears more positive than other drugs. However, the lack of the achievement of its survival benefits can be attributed to the establishment and development of cancer microenvironment in progressive cancerous lesions, which reduces or blocks persistent drug penetration [74].

5-Fluorouracil (5-FU) is another important anti-cancer drug. As an antimetabolite pyrimidine analogue, 5-FU promotes the incorporation of fluorouridine triphosphate into RNAs, of fluorodeoxyuridine triphosphate into DNAs as well as suppression of thymidylate synthase, resulting in severe genomic damages for eliminating cancer cells [75]. The treatment of this drug alone shows diverse ranges of the responses in pancreatic cancer patients [75, 76]. 5-FU is frequently used in combinations with other anticancer-drugs, including gemcitabine, cisplatin, doxorubicin or others [77]. Like gemcitabine, outcomes of the combination treatment did not show significant benefits for the patients as the single treatment. The clinical studies showed the relatively encouraging results, in which both the response rate of the co-treatment of 5-FU with cisplatin and the survival period are modestly improved in patients suffered from metastatic pancreatic cancer.

Folfirinox is a relatively effective, but aggressive combination treatment and a standard therapy for pancreatic cancer patients. This combined treatment is composed of 4 existing anticancer-drugs: 5-FU, irinotecan, oxaliplatin and leucovorin and offers an objective response rate of 32% with a median progression-free survival of 6 months for pancreatic cancer patients [78, 79]. However, this treatment has some side effects presented as fatigue, bone marrow suppression with neutropenia, gastrointestinal disorder and sensory neuropathy. Currently, modified Folfirinox has been implemented in clinic only for those patients who are after surgery and relatively fitted or tolerable to the treatment.

Immunotherapy

As mentioned above, a solid fibrotic and cancerous microenvironment is rapidly formed around pancreatic cancer lesions, which prevents drug penetration and at the same time, suppresses immune reactions. Immunosuppressive cytokines (such as IL-6 or others) that are produced by surrounding tumor stroma are augmented in the tumor lesions [74, 80]. These secreted cytokines, including arginase-I, reactive oxygen species (ROS) and suppressive cytokines (IL-10, TGF-β and others), stimulate further expansion of immunosuppressive lymphocytes and other cancer associated populations to antagonize anticancer responses. Under such suppressive influences, the body antitumor immune-responses are further mitigated and at the same time, inhibitory dendritic cells (DC), Tregs and cancer-associated macrophages are accumulated. With these vicious feedback and feedforward stimulations, a solid tumor associated microenvironment is built to keep chemo-drugs from reaching pancreatic cancer lesions [74]. Notably, immunotherapy is encouraging, but more complex than we anticipate. Thus, it is imminent for developing strategies to target immune checkpoints that are perturbed by cancerous cells. Programmed death-1 (PD-1) and cytotoxic T lymphocyte antigen-4 (CTLA-4) are well known immune checkpoints among others [80, 81]. Ipilimumab is an anti-CTLA-4 antibody and the first FDA-approved, antibody-based treatment [82]. The drug is shown to improve the overall survival of pancreatic cancer patients. Nivolumab and pembrolizumab (the PD-1 inhibitors) are used clinically for treating melanoma and still under clinical investigations for treating pancreatic cancer [83]. In the animal study, it was also shown that the combined inhibition of IL-6 and PD-1 enhanced numbers of tumor infiltrating T lymphocytes [84]. Therapeutics against potentially mitigating immunosuppressive cells to reach pancreatic tumor microenvironment appeared to enhance efficacy of immune-based therapies. Because pancreatic tumors possess the immune-advantaged nature at early stages of the initiation, which equip them to escape immune surveillance, currently existing immunotherapies have not demonstrated breaking-through results in this aspect yet.

Other approaches for pancreatic cancer therapy