Mechanistic Insights into Self-Reinforcing Processes Driving Abnormal Histogenesis During the Development of Pancreatic Cancer

Abstract

Pancreatic ductal adenocarcinoma, one of the most feared lethal and painful diseases, is increasing in incidence. The poor prognosis of pancreatic ductal adenocarcinoma–affected patients primarily is owing to our inability to develop effective therapies. Mechanistic studies of genetic, epigenetic, and cell-to-cell signaling events are providing clues to molecular pathways that can be targeted in an attempt to cure this disease. The current review article seeks to draw inferences from available mechanistic knowledge to build a theoretical framework that can facilitate these approaches. This conceptual model considers pancreatic cancer as a tissue disease rather than an isolated epithelial cell problem, which develops and progresses in large part as a result of three positive feedback loops: i) genetic and epigenetic changes in epithelial cells modulate their interaction with mesenchymal cells to generate a dynamically changing process of abnormal histogenesis, which drives more changes; ii) the faulty tissue architecture of neoplastic lesions results in unsynchronized secretion of signaling molecules by cells, which generates an environment that is poor in oxygen and nutrients; and iii) the increased metabolic needs of rapidly dividing cells serve as an evolutionary pressure for them to adapt to this adverse microenvironment, leading to the emergence of resistant clones. We discuss how these concepts can guide mechanistic studies, as well as aid in the design of novel experimental therapeutics.

CME Accreditation Statement: This activity (“ASIP 2013 AJP CME Program in Pathogenesis”) has been planned and implemented in accordance with the Essential Areas and policies of the Accreditation Council for Continuing Medical Education (ACCME) through the joint sponsorship of the American Society for Clinical Pathology (ASCP) and the American Society for Investigative Pathology (ASIP). ASCP is accredited by the ACCME to provide continuing medical education for physicians.

The ASCP designates this journal-based CME activity (“ASIP 2013 AJP CME Program in Pathogenesis”) for a maximum of 48 AMA PRA Category 1 Credit(s)™. Physicians should only claim credit commensurate with the extent of their participation in the activity.

CME Disclosures: The authors of this article and the planning committee members and staff have no relevant financial relationships with commercial interests to disclose.

The incidence of pancreatic ductal adenocarcinoma (PDAC) is increasing with more than 44,000 predicted new cases in the United States and 65,000 in Europe,^1,2 with a 5-year survival of less than 5%. PDAC arises from epithelial cells through an accumulation of genetic and epigenetic alterations in oncogenes and tumor suppressors,^3,4 which contribute to form precursor lesions^5,6 known as pancreatic intraepithelial neoplasias (PanINs) (Figures 1 and 2). Less frequently, PDAC may progress from two types of cystic lesions: mucinous cystic neoplasms and intraductal papillary mucinous neoplasms. In this process, tumor cells proliferate and secrete molecules that drive their communication with surrounding cells. In the fashion of a self-reinforcing loop, surrounding cells also proliferate and secrete new substances, which initiate new communications among themselves, with other noncancer cell types within the tumor (Figure 3).

Figure 1.

Self-reinforcing processes that drive abnormal histogenesis during the development of pancreatic cancer. Diagrammatic representation of positive feedback loops that contribute to pancreatic carcinogenesis involves the progressive genetic and epigenetic changes in epithelial cells, which modulate their interaction with mesenchymal cells to generate a dynamically changing process of abnormal histogenesis to further drive more changes. The faulty tissue architecture of neoplastic lesions results in unsynchronized secretion of signaling molecules by cells, which generates an oxygen- and nutrient-poor environment as a result of aberrant angiogenesis. Finally, the increased metabolic needs of rapidly dividing cells serve as an evolutionary pressure for them to adapt to this adverse microenvironment, leading to the emergence of resistant clones.

Figure 2.

Histologic correlates of abnormal histogenesis during pancreatic cancer progression. Neoplastic pancreatic tissue from p48-cre/Kras^G12D transgenic animals were stained using the Masson trichromic method, in which epithelial cells are labeled in red and the extracellular matrix is labeled in blue. This series of micrographs show that from the beginning, pancreatic cancer development involves the tight interaction between epithelial cells and its surrounding mesenchyma. A: PanIN1A lesion formed by cells with normal-shaped nuclei but showing incipient nuclear piling up and increased cytoplasm. B: PanIN1B lesion showing papillary projections formed by cells with normal-looking nuclei with a mucin-containing cytoplasm that displaces nuclei to the base of the lesion. C: PanIN2 lesion with abnormally shaped nuclei and typical papilar projections occupying the duct lumen. D: PanIN2 to 3 lesion showing the typical piling up of nuclei with incipient atypia showing anisokaryosis, poikilokaryosis, and papilar projections. Note that a ring of extracellular matrix surrounds the duct-like structure. E: PanIN3 lesion with extensive atypia showing the surrounding ECM as a dense ring that deforms the ductular structures. F: Multifocal cancerous lesions embedded in a dense desmoplasia.

Figure 3.

Epithelial–mesenchymal interactions during the development of pancreatic cancer. The functional co-evolution between pancreatic epithelial cells and their stromal counterparts from early preneoplastic to frank neoplastic lesions is shown. Several growth factors are secreted at abnormal amounts, times, and places to generate abnormal signaling cascades that drive the communication between both the epithelial compartment and the tissue microenvironment. As explained in the text, different growth factors exert their function either in a paracrine or autocrine manner to help form the tumors, desmoplasia, and generate an oxygen- and nutrient-poor tumor bed, impacting the pathobiology of pancreatic cancer and contributing to its resistance and aggressiveness. FGF, fibroblast growth factor; MMP, matrix metalloproteinase; PDGF, platelet-derived growth factor; SDF1, serum derived factor-1; SHH, Sonic hedgehog; VEGF, vascular endothelial growth factor.

This extended cellular network generates a dynamic tumor microenvironment that influences genetically heterogeneous tumor cells and selects for highly proliferative and resistant clones. Epithelial and mesenchymal cells each contribute to remodeling the stroma into a dense fibrotic tissue (desmoplasia) enriched in fibrillar collagens, stromal cells, and other migratory cell populations. Accordingly, each component of the developing tumor takes an active role in the process of carcinogenesis. Thus, dissecting the temporal and spatial sequence of events that drives these processes should provide new ways for therapeutically transforming pancreatic cancer from a rapidly fatal to a chronic and treatable, or even curable, disease.

Stromal Cells Are Key Players in Pancreatic Cancer Development

The stroma co-evolving with tumor cells during PDAC progression is composed of several cell types,^7–9 including fibroblasts, pancreatic stellate cells (PSCs), endothelial cells, bone marrow–derived cells, as well as inflammatory and immunoregulatory populations. Although fibroblasts contribute to pancreatic fibrosis, PSCs also contribute to this process by secreting fibrillary collagens (type 1, type 3) and fibronectin.¹⁰ PSCs are located between acini and endothelial cells, where they influence the differentiation and function of acinar cells or work as pericytes. PSCs represent less than 5% of cells within the normal gland¹⁰ but in tumors they proliferate to outnumber malignant epithelial cells.

These cells normally exist in a quiescent state marked by the presence of vitamin A–rich lipid droplets, desmin, and glial-fibrillary-acidic protein. On tissue stress and injury, PSCs become activated myofibroblast-like cells, lose their droplets, express α-smooth muscle actin, and secrete extracellular matrix (ECM) proteins.^8,11 The proliferation of these α-smooth muscle actin–positive cells and the deposition of collagen in human PanIN and PDAC compared with normal pancreas is observed clearly in micrographs of human patient samples and mice models. The idea that activated myofibroblasts in the pancreas arise from PSCs has many proponents but still remains under debate¹² because a few studies suggest that they can originate from circulating bone marrow–derived stem cells.^13,14 Nevertheless, the currently available data make it impossible to ignore the fact that stromal cells, both resident and migratory, are as important as epithelial cancer cells in the formation and pathobiology of PDAC tumors.

Epithelial and Mesenchymal Interactions Are Critical in the Histogenesis of PanINs

Even at the PanIN stage, abnormal epithelial cells exert influence on their surrounding microenvironment. PSCs are found within PanINs from human pancreatic samples and mouse models^15,16 where they are stimulated by signals from other cells, including epithelial populations^17,18(Figures 1 and 3). These signals include platelet-derived growth factor, fibroblast growth factor, and transforming growth factor β1 (TGF-β1). TGF-β stimulates collagen I, collagen III, and fibronectin biosynthesis, whereas fibroblast growth factor and platelet-derived growth factor induce PSC proliferation.^17–19 When orthotopically injected into mouse pancreata, TGF-β1–overexpressing PDAC cells²⁰ increase desmoplasia, showing that secreted factors from epithelial cells can induce the mesenchymal reaction characteristic of pancreatic cancer. Sonic hedgehog, which is expressed during development and silenced in mature cells, becomes re-expressed again in PanINs, increasing in concentration during the progression of PanIN2 and 3 to PDACs.²¹ Expression of Sonic hedgehog in mice using the pdx1 promoter leads to PanIN formation.²¹ In contrast, elastase 1–driven expression of this factor in pancreatic acinar cells does not result in PDAC but causes an expansion of nestin-positive mesenchymal cells.²² Together, these studies provide relevant examples of how communication between both malignant epithelial cells and their nonmalignant mesenchymal counterparts are established de novo and aid the progression from a preneoplastic into a frank neoplastic phenotype.

Other stromal responses include the effect of angiogenic factors secreted by tumor cells that act on inflammatory cells.^23–26 In exchange, these inflammatory cells trigger a reinforcing positive feedback loop by producing IL-6, thus inducing proliferation of premalignant cells and their protection from apoptosis.²⁷ These protumoral effects are synergized by granulocyte-macrophage colony-stimulating factor,²⁸ which is produced by both murine and human PanIN cells.²⁸ Granulocytes, mast cells, and macrophages located within human PDAC produce late protumoral factors, such as TGF-β²⁹ and vascular endothelial growth factor. Therefore, an intense exchange of information between tumor cells, stromal cells, and immune cells contribute to arm PDAC with its aggressive nature.³⁰ Interestingly, however, few studies directly have investigated whether stromal cells from PanINs secrete growth factors and cytokines. Serum-derived factor-1 is secreted by PSCs from both human and mouse PanINs, although the receptor for this factor, CXCR4, is expressed in PanINs and cancer epithelial cells.^31–33 Thus, mesenchymal serum-derived factor-1 stimulates CXCR4 on PDAC cancer cells, leading to signals that promote their survival, proliferation, and migration.^32–34 Combined, this knowledge supports the conclusion that paracrine signaling from PSCs to PDAC cells stimulates the malignancy of PDAC cells beginning at the PanIN stage.

In this regard, complementary in vitro studies allow us to draw mechanistic information that explains the behavior of stromal cells at the PanIN stage. Conditioned media from PSCs, for instance, stimulates tumor cell proliferation and migration.^31,35,36 When co-injected into nude mice, PSCs also mediate protumoral effects on epithelial cells through the secretion of epidermal growth factor, platelet-derived growth factor, and fibroblast growth factor, whereas other factors such as TGF-β1, IL-1, and IL-6 act in an autocrine manner to exacerbate PSC activation.^37,38 Activated PSCs produce several types of collagen^9,39 and diffusible ECM components that also help to stimulate PDAC cell proliferation and migration.^40–42 However, this desmoplasia is not homogeneous in composition, which likely affects PanINs differently depending on whether they are located near areas of densely packed fibers or juxtaposed to looser tissue layers. This heterogeneously remodeled ECM results from the localized secretion of powerful proteolytic machinery. Matrix metalloproteinases, the best-studied ECM proteases, are secreted by many cells from the tumor, including PSCs, tumor epithelial cells, and leukocytes, to both degrade fibers and release growth factors sequestered in the ECM.⁴³

All these studies revealed an exquisite communication network that mediates reciprocal signals among most cells within the tumor as well as messages going from cells to the ECM and back. Such findings guide us to the idea that drugs that inhibit stromal cell activation or stromal/tumor cell communication may have promising therapeutic value in PDAC.

Exploring the Concept of PDAC as a Disease of Abnormal Tissue Morphogenesis

One of the striking characteristics of cancer is the tendency to recapitulate, in part, the morphologic features of the organ from which they originate. In many cases, these features are pathognomonic so that a well-trained pathologist can predict the site of a primary tumor by looking at a metastasis. The fact that cancers do not display an unlimited variety of histologic profiles but rather present with similar phenotypes indicates that our genes and the epigenome, which regulates their temporal and spatial pattern of expression, not only give rise to the normal phenotype, but also shape its neoplastic counterpart. This idea is supported by the fact that pancreatic epithelial cells display characteristics of progenitor populations found in early development of pluripotent stem cells with similar features to those described in other forms of cancer.^44–46 For example, a number of signaling pathways (eg, hedgehog, Notch, and Wnt) and transcription factors (eg, Gata4, Nanog) associated with embryonic development are expressed in cancer cells.^47–50 Cooperation between pancreatic tumors and surrounding tissue that leads to undesirable outcomes such as hypoxia and chemoresistance can be seen as resulting in part from a faulty morphogenetic cascade that tries but fails to recapitulate normal histogenesis.

During normal development, a precise series of gene expression events generate signals (eg, angiogenic factor) and outcomes (eg, angiogenesis) in a controlled manner at the appropriate level, place, and time. In contrast, during carcinogenesis, many molecules are produced simultaneously, leading to confusing morphogenic instructions that prevent normal programs from occurring or, if they do occur, they begin and terminate at the wrong time and place. This lack of synchronization of morphogenetic cascades gives rise to a distorted architecture, which causes epithelial and mesenchymal cells to miscommunicate. Thus, once abnormal morphogenesis has begun, it causes additional architectural problems, behaving as another of the self-reinforcing mechanisms that are critical to the formation of PDAC. Numerous studies have shown the re-emergence of mesenchymal/epithelial developmental signaling pathways in PDAC. Early in development, for instance, hedgehog is absent from the dorsal and ventral pancreatic endoderm,⁵¹ but its forced expression in this tissue influences the stroma.⁵¹ Sonic hedgehog expression in PDAC cells also stimulates the formation and remodeling of the tumor microenvironment in a manner that is abnormal. Thus, HH expression in PDAC may be thought of as a faulty recapitulation of an early developmental program of tissue morphogenesis. Similar examples are found with fibroblast growth factors 1, 7, and 10, serum-derived factor-1, and many members of the TGF-β/activin/nodal/inhibin family of cytokines. Thus, together, this knowledge serves as evidence that PDAC presents as a process of histogenesis that progresses abnormally as a result of signals arising and terminating inappropriately.

This concept is worthy of being compared and contrasted with previous theories of carcinogenesis. Indeed, until the 1970s, most informed investigators agreed that pancreatic cancer recapitulated developmental programs, and in that sense our considerations are not new. However, most of these early theories viewed neoplastic lesions as arising from dedifferentiated cells, but current considerations in this regard are different because dedifferentiated cells imply that an adult population has reverted its phenotype to that found in an early developmental stage. If that was the case, neoplastic cells would have all of the instructions necessary for not only initiating the formation, but also the completion, of an organ, develop normal architectural relationships within the organ, organize an efficient system of nutrient supply (vessels), and appropriate disposal of its products (secretory ducts). In addition, their genes would lack mutations and be expressed in a sequential manner as needed. Obviously, this is not the case with pancreatic cancer, leading some investigators to alternatively propose that PDAC arises through a process of transdifferentiation. Examples of this phenomenon in the pancreas are found in benign intraductal papillary mucinous neoplasms where the normal ductular cells transdifferentiate into those resembling the gastrointestinal epithelium. Interestingly, this type of differentiation also has been proposed for pancreatic ductal adenocarcinoma, primarily based on studies in animal models, although the evidence supporting transdifferentiation in human PDAC still remains controversial. Other investigators have invoked abnormal differentiation of stem cells as the origin for this type of cancer. It is noteworthy, however, that what we call stem cells in solid tumor are different from those found in embryonic development with a differentiation potential, which is currently unclear.

Thus, these considerations lead us to propose that pancreatic cancer cells neither go back in differentiation nor do they cross the differentiation barrier to another cell type. Rather, pancreatic cancer cells appear to have their own differentiation pathway, which in most part uses overlapping gene networks with those turned on and off during development to produce structures that are phenotypically similar yet distinct from those found in embryonic development. In this context, we suggest that this neoplasia develops as a sui generis case of abnormal morphogenesis. This inference should help to develop a holistic view of the disease as a process in which unsynchronized molecular and cellular events taking place in most of the structures that form the tumor, and not exclusively in the epithelial cells, self-reinforce a path toward progression. When, where, and how to intervene with these processes, although a significant challenge, should be aided by the conceptual framework discussed in this article.

Angiogenesis, Hypoxia, and Nutrient Deprivation Behave as a Positive Feedback Loop Driving the Selection of Cells with High Malignant Potential

Abnormal architectural relationships among structures forming a defined tissue are one of the most obvious consequences of abnormal morphogenesis either during normal development (eg, thalidomide toxicity) or in cancer. In this section, we discuss how an increased demand of nutrients and oxygen by cells from the PDAC tumor bed is met by a relative decrease in their supply imposed by both architectural and functional changes that occur during the remodeling of preneoplastic lesions as they become malignant. In particular, we recognize that a faulty architectural design results in another positive feedback loop that leads to the establishment of more aberrant structures. The law of supply and demand, imposed by a decrease in the offering of nutrients and an increase in energetic needs by the pancreatic cancer, lends itself as a good example of this positive feedback loop. Indeed, by the time PDAC is diagnosed, most patients have invasive tumors with extensive desmoplasia, which are unexpectedly hypovascularized^52,53 and poorly perfused, thereby generating low oxygen and nutrient availability at tumor sites. Reduced vascularity also is found in chronic pancreatitis, suggesting that hypoxic conditions begin early during the development of PDAC and only increase in frank adenocarcinoma.⁵⁴ These features, along with the increased metabolic needs of rapidly proliferating cells, function as a selection pressure toward their adaption to conditions that would be suboptimal to normal cells, thus promoting tumor cell survival and their characteristic resistance to chemotherapy.⁵⁵

Hypoxia Plays a Pivotal Role in the Selection of Aggressive Pancreatic Cancer Cell Clones

The effect of hypoxia on pancreatic cancer cells has been modeled extensively in vitro. For instance, cell lines surviving hypoxia show enhanced invasive and migratory propensities,^56,57 indicating that this condition favors cancer progression. Although the precise mechanisms by which hypoxia affects this process remain incompletely understood, changes in oxygen tension switch metabolism, modulate autophagy-mediated survival and death, modulate genomic instability, and promote additional cell responses that further facilitate malignant clonal selection. Mechanistically, oxygen deprivation triggers the stabilization of hypoxia-inducible factor 1α (HIF-1α), which dimerizes with HIF-1β, undergoes nuclear translocation, and binds to a multitude of hypoxia-responsive elements present in the cancer genome. These events activate a complex genetic–epigenetic program that seeks to counteract the deleterious impact of decreased oxygen tension.⁵⁸ HIF-1α is overexpressed in PDAC,⁵⁹ where it regulates gene networks involved in autophagy, glycolysis, angiogenesis, epithelial-to-mesenchymal transition, and metastasis.^58,60 This is a good example of how transcription and chromatin-induced changes in gene expression contribute to pancreatic carcinogenesis. Another example is revealed by studies on the small chromatin binding protein, nuclear protein 1. Nuclear protein 1 is overexpressed in PDAC,^61,62 where it is induced by hypoxia, glucose deprivation, and other stresses.^63,64 Genetic inactivation of nuclear protein 1 sensitizes PDAC cells to hypoxia and nutrient deprivation and induces their death by autophagy.⁶⁴ In this manner, nuclear protein 1 works as a sensor of tumor cell stresses that, when activated, protects tumor cells against damage caused by a hostile environment. Thus, in summary, PDAC cells become primed for survival under hypoxic environments by deploying a defined gene expression network, which ultimately results in a more resistant phenotype, illustrating another positive feedback loop that drives PDAC progression.

A Metabolic Switch Is a Defining Feature of Highly Aggressive Pancreatic Cancer

Hypoxia not only affects epithelial cells, but also stimulates the activation of mesenchymal cells that secrete angiogenic factors such as vascular endothelial growth factor in an attempt to increase O₂ through enhanced vascular permeability and angiogenesis.^65,66 Despite this mechanism, PDAC still fails to increase its vascularity as robustly as other tumors, suggesting that the function of proangiogenic factors is counterbalanced by antiangiogenic molecules such as endostatin.⁶⁵ Thus, both angiogenic and antiangiogenic substances secreted in different amounts at various times during the development of pancreatic cancer ultimately lead to a functional equilibrium characterized by the hypovascularity that is typical of this malignancy.

One of the major consequences of intratumoral hypoxia is the metabolic switch from the primary use of oxidative phosphorylation to reliance on the glycolysis that occurs to meet the requirements of tumor proliferation under low oxygen and low nutrient supply.⁶⁷ Although normal cells rely primarily (90%) on oxidative phosphorylation, approximately 50% of cellular energy is produced by glycolysis in tumor cells, with the remainder being generated in the mitochondria, even in the presence of ample oxygen to fuel mitochondrial respiration, a phenomenon known as the Warburg effect.⁶⁸ This metabolic transformation is a consequence of acquired mutations that lead to cancer, coupled with hypoxic and nutrient-deficient conditions and faulty architectural relationships among cells and vessels, all of which contribute to the selection of tumor clones able to survive this hostile environment. Congruently, PDAC tumors in vivo show alterations in pathways involved in this metabolic switch,⁶⁹ which even at the PanIN2 and 3 stages lead to increased expression of the glucose transporter 1 (GLUT1), a protein that is critical for glucose uptake by tumor cells.⁷⁰ Subsequently, hypoxia also functions as a key mechanism that increases glycolysis. For example, the hypoxia-induced stabilization of HIF-1α not only regulates angiogenesis, but also directly up-regulates several genes involved in this process, such as GLUT1, hexokinase 1 and 2, lactate dehydrogenase A, and lactate transporter monocarboxylate transporter 4.⁷¹

Defined genetic alterations, such as loss of p53 function, also stimulate glycolysis at several levels.⁷¹ Similarly, activation of Ras inhibits pyruvate kinase and stimulates GLUT1 translocation to the plasma membrane.⁷¹ Consequently, turning off the expression of an inducible Kras^G12D oncogene in PDAC cells leads to reduced glucose uptake, decreased expression of rate-limiting glycolytic enzymes, decreased lactate production, and reduced channeling of glucose metabolites into nonoxidative anabolic pathways such as hexosamine biosynthesis.⁷² These changes in metabolic programming by Kras^G12D are mediated, at least in part, by mitogen-activated kinases and Myc.⁷² Therefore, shifting substrates from energy production to molecular synthesis seems to provide tumor cells with advantages that help to account for its increased survival and aggressiveness. Glycolytic pathways produce ATP and pyruvate that generate bioproducts, which enter the pentose phosphate pathway to generate ribose-5-phosphate and NADPH, key intermediates in nucleotide biosynthesis. During glycolysis, conversion of glucose to fructose provides an essential substrate for the nonoxidative pentose phosphate pathway. Pyruvate, another major intermediate, is converted to lactate by lactate dehydrogenase and secreted into the extracellular environment. Lactate secreted by tumor cells could be absorbed by stromal cells and converted to pyruvate that is used either for oxidative phosphorylation or secreted again and then taken up by tumor cells for use as glycolytic fuel.^71,73 A recent study investigated the expression of glycolytic proteins in PDAC tumors and associated stromal cells to show that both primary PDAC tumors and metastases express high levels of proteins involved in glycolysis (eg, GLUT1 and hexokinase 2) compared with normal tissue.⁶⁹ GLUT1 also is increased in tumor-associated stromal cells, although to a lesser degree than in the tumor cells.⁶⁹ In totality, the epithelial cell and tumor microenvironment relationship not only regulates morphogenesis by paracrine/autocrine mechanisms and drives angiogenesis, but also actively participates in the recycling of metabolites within the tumor. This recycling of metabolites, as in the case of the lactate cycle, is critical to maintain the energetic needs of the tumor, leading to the prediction that a better understanding of these interactions between cancer cells and the surrounding populations can be exploited for diagnostic (eg, labeled metabolites) or therapeutic efforts (eg, metabolic manipulations).

Conclusions

The particularly bad prognosis that is observed for patients with PDAC is at least in part owing to the pancreatic transformed cells that develop in cooperation with surrounding stromal cells. Activated PSCs and immune cells may be recruited in response to pancreatic injury and/or mutant epithelial cells, and initially may provide growth factors and cytokines that nurture preneoplastic epithelial cells, selecting for clones with key mutations, such as KRAS. Because PSCs proliferate and secrete ECM proteins, an adverse desmoplastic microenvironment is formed, characterized by hypoxia and nutrient starvation as a consequence of its hypovascularization, which selects for aggressively invasive, chemotherapy-resistant tumor cells. Hypovascularization-induced hypoxia activates HIF-1α accumulation in tumor cells, which together with Kras activation increases autophagy and induces a metabolic switch to glycolysis. The low vascularization of tumors also impedes the delivery of chemotherapeutic drugs. On the other hand, the presence of high intratumoral concentrations of antiapoptotic cytokines also enhances tumor cell survival. Tumor cells show some characteristics of embryonic cells and, through interactions with stromal cells, participate in a warped recapitulation of tissue morphogenesis. Together, these findings suggest that the development of therapeutic approaches that target stromal cells, stromal/tumor communication, hypoxia, and/or glycolytic metabolism may bring novel treatments for PDAC.