Advances in understanding the molecular mechanism of pancreatic cancer metastasis

Yong-Xing Du， Zi-Wen Liu， Lei You， Wen-Ming Wu and Yu-Pei ZhaoBeijing， China

?

Advances in understanding the molecular mechanism of pancreatic cancer metastasis

Yong-Xing Du， Zi-Wen Liu， Lei You， Wen-Ming Wu and Yu-Pei Zhao
Beijing， China

BACKGROUND： Pancreatic cancer （PC） is usually diagnosed at the late-stage and therefore， has widespread metastasis and a very high mortality rate. The mechanisms underlying PC metastasis are not well understood. Recent advances in genomic sequencing have identified groups of gene mutations that affect PC metastasis， but studies elucidating their roles are lacking. The present review was to investigate the molecular mechanisms of PC metastasis.

DATA SOURCES： Relevant articles on PC metastasis were searched in MEDLINE via PubMed prior to April 2015. The search was limited in English publications.

RESULTS： PC metastatic cascades are multi-factorial events including both intrinsic and extrinsic elements. This review highlights the most important genetic alterations and other mechanisms that account for PC invasion and metastasis，with particular regard to epithelial-mesenchymal transition，inflammation， stress response， and circulating tumor cells.

CONCLUSIONS： Analyses of relevant gene functions and signaling pathways are needed to establish the gene regulatory network and to define the pivotal modulators. Another promising area of study is the genotyping and phenotyping of circulating tumor cells， which could lead to a new era of personalized therapy by identifying specific markers and targets. （Hepatobiliary Pancreat Dis Int 2016；15：361-370）KEY WORDS： pancreatic cancer；

metastasis；

molecular mechanism；

targeted therapy

Author Affiliations： Department of General Surgery， Peking Union Medical College Hospital， Chinese Academy of Medical Sciences and Peking Union Medical College， Beijing 100730， China （Du YX， Liu ZW， You L，Wu WM and Zhao YP）

? 2016， Hepatobiliary Pancreat Dis Int. All rights reserved.

Published online November 9， 2015.

Introduction

P ancreatic cancer （PC） is one of the most lethal malignancies and a leading cause of cancer-related death worldwide. According to the American Cancer Society， the annual new cases of PC were approximately 46 420 and the deaths were 39 590 in 2014.［1］Despite decades of considerable effort， its 5-year survival rate remains no more than 6%. More than 80% of the patients present with local or distant metastases at the time of diagnosis and thus are not eligible for potentially curative surgical treatment.［2］Even among patients who are fortunate enough to undergo surgical resection， 70% develop recurrence and metastatic disease within a year.［3］This dismal reality makes PC the “king of cancers” and the “hardest stone” to tackle with among cancers in the 21st century. Thus， a better understanding of the biology and genetics of PC metastasis is crucial for the discovery of more sensitive and specific markers or targets to combat this silent killer.

Metastasis is a non-random， biologically heterogeneous and multi-step pathophysiologic process that involves the dissemination of cancer cells from a primary tumor to anatomically distant organ sites and their subsequent adaptation to the foreign tissue microenvironments.［4， 5］Genetic and/or epigenetic alterations as well as non-neoplastic stromal cells help to deregulate both intrinsic and extrinsic signaling pathways， which drives the progression of each step in the metastatic cascade and allows initial micro-metastasis to occur. PC disseminates mainly via lymphatic and hematogenous routes， and this process can occur even before clinically detectable tumor formation， which suggests that metastatic propensity is rapidly acquired during pancreatic carcinogenesis and progression. However， the molecular characteristics of advanced-stage PC have not been well investigated as early pancreatic carcinogenesis， and thus our understanding of the mechanisms that underlie PC metastasis is quite limited. This review aimed to investigate the molecular mechanisms underlying PC metastasis.

Genetic basis of PC metastasis

In the last decade， genetic studies of PC have progressed markedly with the application of sequencing technology for histologically and clinically well-defined lesions. The discovery of the genes that are most frequently mutated in PC metastasis provides insight into the fundamental pathways that drive this malignancy and create new opportunities for targeted gene-based interventions.

Unfortunately， systematic studies of PC metastasis have been greatly limited by a lack of high-quality metastatic tissues. To address this problem， Embuscado et al［6］initially instituted a rapid autopsy program to collect high-quality tissues from patients with metastatic PC. Their sequencing analysis confirmed that 75% of the patients had the inactivation of DPC4 in metastatic PC， while the mutation rate of both KRAS2 and TP53 remained concordant with early-stage disease. This finding indicated that DPC4 continued to function during distant metastasis. Using DNA microarray， Kim et al［7］analyzed the changes of genes in the transcriptional profile. They found that 194 genes （including GABRP， SERPINB3 and CTNNB1） were either up- or down-regulated in lymph node-positive PC. These genes are related to cell proliferation， apoptosis， adhesion， motility and cell-cycle regulation. Using the same method， however， Campagna et al［8］failed to identify genes that were significantly deregulated between these matched primary and metastatic carcinomas， suggesting that the metastatic phenotype develops during tumor formation. Nonetheless， they identified 173 genes （including MXI1） with different expressions in advanced primary tumors with co-existent gross metastasis （pathologic stage pT4） compared with surgically resected primary disease （pathologic stage pT2/ pT3）. These findings indicated that additional changes in gene expression might occur within the primary site that correlated with progression to the advanced-stage disease，but did not rule out the presence of significant changes in gene expression during PC metastasis. Using human whole genome DNA microarrays， Stratford et al［9］discovered a six-gene signature （FOSB， KLF6， NFKBIZ， ATP4A，GSG1， SIGLEC11） associated with metastatic disease by comparing the gene expression profiles of primary PC at the extremes of disease： early （localized） and late （metastatic） stages.

In addition， as a fruitful companion system for PC genetic research， mouse PC models have been shown to be an efficient alternative for the study of gene expression profiling of metastasis and tumor invasion. Niedergethmann et al［10］implanted a human PC cell line （MiaPaca-2）orthotopically in severe combined immunodeficiency （SCID） mice， and then conducted a transcriptome analysis of the primary tumor， the invasion front and liver metastasis with microarray technique. Hundreds of genes were identified statistically as metastasis-related. A total of 66 significantly regulated pathways were revealed， involving the apoptotic cascade， angiogenesis and cell interaction. Thus， an orthotopic pancreatic tumor model in SCID mice may serve as an efficient model to study expression profiling for early local invasion and liver metastases. Similarly， Suemizu et al［11］developed another reliable model system for quantitative assay of liver metastasis using NOD/SCID/γcnull（NOG） mice. They established a highly metastatic PC cell line， livermetastasized BxPC-3 （LM-BxPC-3） using an NOG mice model. By comparing the gene expression profile of LMBxPC-3 with the parental poorly metastatic cell line， they identified S100A4 as a key regulator of liver metastasis in PC， thus providing us with a tool for exploring new targeted anti-cancer therapy. Obtaining enough tumor cellularity in samples subjected to whole-exome or wholegenome sequencing is a known barrier in PC sequencing studies due to abundance of stromal and inflammatory cells. To obtain high-quality data and facilitate mutation detection， Jones et al［12］used patient-derived xenografts to remove non-neoplastic contaminants. This approach resulted in detection of an average of 48 genetic alterations per model sequenced. Recently， Boj et al［13］used organoid models of murine ductal PC to probe genes and pathways altered during disease progression. Although a number of genes showed significantly different expression between the primary and metastatic sites， little overlap was seen among the different gene sets， which suggests both complexity and a high level of genomic heterogeneity in PC metastasis.

Yachida et al［14］found that the metastatic clone populations are from the original parental， non-metastatic clone； but these clones are genetically evolved. Similarly，a phylogenetic analysis of ten matched primary and metastatic tumors by Campbell et al［15］supported the formation of genetic heterogeneity predominantly by tandem forces of genomic instability in metastasis-initiating cells and ongoing evolutionary selection. To elucidate the somatic mutations that drive clonal expansion of metastasis， we recently analyzed matched tumor samples of both primary and metastatic lesions and relevant normal control tissues from a patient with liver metastatic PC，by applying both genome-wide SNP array and exomecapture sequencing analyses.［16］We found that 12 genes，including KRAS and TP53， have higher allele frequencies of functional mutations in the metastatic tumor. Moreover， all of these candidate genes were expressed abnormally in PC tissues and functionally affected the migration， proliferation， and colony-formation abilities of PC cell lines. We later investigated the role of a novel gene，NOP14， in maintaining malignant features of metastatic tumor cells.［17］The results showed that inhibiting or upregulating NOP14 expression in PC cells could reduce or promote motility， proliferation and metastatic capacity in vivo. Additional studies are ongoing at our center to clarify the underlying mechanism.

Almost 80% of patients with PC are staged as having unresectable tumor at the time of diagnosis， predominantly due to metastasis， which underscores the importance of finding metastasis at a clinically relevant time. Yachida et al［14］estimated the average interval from initiated tumor cell to founder cell of the parental clone，metastatic subclones and distant metastasis formation as 11.7， 6.8 and 2.7 years， respectively， indicating a large window of time for diagnosis prior to metastasis. However， Haeno et al［18］recently used a mathematical and computational modeling approach to predict that even very small primary tumors frequently undergo microscopic metastasis prior to surgical removal. Simultaneously， Rhim et al［19］developed a sensitive method to tag pancreatic epithelial cells in a mouse model of PC for dynamic monitoring of their motility. They found that tagged cells invaded and entered the bloodstream unexpectedly early， even before clear malignancy could be detected by rigorous histological analysis. Although these new findings in PC may evoke dismay， they actually transform the way that we consider PC to evolve and provide the opportunity to refocus and prioritize our efforts toward the improvement of patient outcomes.［20］A primary “gene blueprint” associated with PC metastasis can be delineated based on the genetic studies described above （Table）， but further verification and investigation of the underlying mechanisms associated with most of the mutated genes are lacking. Detailed research in this area will be important for establishing the gene regulatory networks responsible for PC metastasis，and will contribute to the future identification of sensitive PC markers or targets. In this regard， Weissmueller et al［22］confirmed that mutant p53 strongly influenced the maintenance of the pro-metastatic phenotype in a mouse PC model by inducing its downstream mediator，platelet-derived growth factor （PDGF） receptor β via a cell-autonomous mechanism. Inhibition of PDGF receptor β signaling by RNA interference or small-molecule inhibitors prominently inhibited PC cell invasion in vitro and metastasis in vivo， which indicated that PDGF receptor β could be a prognostic marker and potential target for treating metastasis in p53 mutant tumors.

Table. Genetic studies of pancreatic cancer metastasis

Epithelial-mesenchymal transition

Epithelial-mesenchymal transition （EMT） is widely investigated. The epithelium is typically composed of a layer of thick epithelial cells with cell junctions and adhesions between neighboring cells that hold them tightly together and prevent the dissociation of individual cellsfrom the epithelial monolayer. The EMT process endows individual epithelial cells with mesenchymal attributes，including a spindle-like， fusiform morphology and overexpression of mesenchymal markers such as N-cadherin，fibronectin and vimentin.［23］For most epithelial tumors，progression toward malignancy is characterized by the loss of epithelial differentiation and the acquisition of a mesenchymal phenotype， leading to enhanced migration and invasion.

The best characterized alteration of EMT in carcinoma cells involves the loss of E-cadherin， a key cellto-cell adhesion molecule that mediates intercellular junctions and helps to maintain the quiescence of the cells within the epithelial cell sheet. Increased expression of E-cadherin prevents/postpones tumor invasion and metastasis， whereas its reduction is known to potentiate these phenotypes.［24］von Burstin et al［25］demonstrated that genetic inactivation of E-cadherin in parental cells induces EMT and increases PC metastasis in vivo. Interestingly， mutations in the E-cadherin gene （CDH1） have been validated as the cause of hereditary diffuse gastric cancer and recommended in clinical guidelines for genetic screening.［26］However， whether loss of E-cadherin in hereditary diffuse gastric cancer is associated with EMT or the consequence of many EMT-independent cellular changes is unclear. EMT has been suggested to be a prerequisite for the invasion and dissemination by carcinoma cells.［24］This is supported by the clinical finding that increased expression of mesenchymal markers often correlates with poor survival for patients with pancreatic or breast cancer. In addition， many xenograft experiments with genetically modified mouse cancer models provided direct evidence for the importance of EMT in cancer initiation and progression. Rhim et al［19］recently used genetic lineage tagging of cancer cells in a genetically engineered PC mouse model to show that EMT could occur in precancerous pancreatic intraepithelial neoplasia lesions and contribute to unexpected early micro-metastasis in vivo before， or in parallel with，tumor formation at the primary site.

Accumulating evidence indicates that in undergoing EMT， cells acquire stem cell-like properties， with carcinogenic and metastatic potential.［19， 23， 27， 28］Normal stem cells and cancer stem cells may share a mesenchymal phenotype that enhances their ability to preserve stem cell characteristics， retain migratory properties， and respond to different stimuli during expansion and differentiation.［23］Rasheed et al［28］immunohistochemically analyzed aldehyde dehydrogenase （ALDH） expression， which can mark PC cells with stem cell and mesenchymal features，in paired primary tumors and metastatic lesions from eight PC patients in a rapid autopsy program. The results showed that six of the eight patients had ALDH-negative primary tumors， among whom four had matched metastatic lesions （located in the liver and lung） containing ALDH-positive cells. Moreover， further in vitro cell invasion assays and gene expression analyses indicated that ALDH-positive pancreatic tumor cells expressed genes consistent with a mesenchymal state and had a three-fold increase in potential to migrate and invade in vitro.

EMT can be triggered by the interplay of multiple cellular signaling pathways such as Hedgehog （HH），Notch， VEGF， Wnt， TGF-β， HIF and NF-κB，［23， 29］many of which are involved in pancreatic metastasis. Furthermore，cumulative evidence indicates that the vast majority of the above signaling pathways known to induce EMT converge at the induction of E-cadherin repressors， particularly the Snail genes.［23， 30］Xu et al［31］found that activation of sonic Hedgehog （SHH）-Gli1 inhibited E-cadherin expression in vitro and significantly increased liver and intra-splenic micro-metastasis in vivo. Moreover， they identified 278 up-regulated and 59 down-regulated genes in response to Gli1 expression in AsPC-1 cells. They further found that SHH-Gli1 signals promoted EMT by mediating a complex signaling network that included TGF-β， Ras， Wnt， PI3K/Akt and S100A4. Blockage of the HH signaling pathway with cyclopamine could profoundly inhibit metastasis in a spontaneously metastasizing xenograft model that confirmed the inhibition of the pharmacologic HH pathway as a plausible therapeutic strategy for PC.［32］Another transcriptional repressor is ZEB1， which could recruit histone deacetylase HDAC1 and HDAC2 to the CDH1 promoter to down-regulate E-cadherin expression in PC. Histone deacetylase inhibitors （HDACi） attenuate tumor cell migration and proliferation， suggesting that inhibiting HDACs may be an antitumor therapy for PC.［33］At our center， we first reported that a nucleoprotein， NOP14， promoted PC cell proliferation and migration in vitro and in vivo.［17］Our results showed that suppressing NOP14 significantly increased E-cadherin and slightly decreased vimentin expression in parallel with lower levels of MMP9 and Rho A， indicating that NOP14 might promote PC cell migration， possibly by inducing an EMT-like cell phenotype and up-regulating MMP9 and Rho A. The relationship between NOP14 and EMT is unclear， but worth further exploration. Potentially， NOP14 triggers EMT by upregulating growth factors or transcription factors such as Snail or ZEB1； this could provide novel targets and help develop innovative therapeutic regiments.

Inflammation

Valastyan and Weinberg［5］reported that successful me-tastasis depends on both the intrinsic properties of the tumor cells and factors derived from the tumor microenvironment. Inflammatory cells and mediators， including chemokines and cytokines， are the most important constituents. More figuratively， Balkwill and Mantovani［34］described genetic damage as the “match that lights the fire” of cancer， and subsequently some types of inflammation may provide the “fuel that feeds the flames.”Epidemiological studies also have shown that chronic pancreatitis predisposes individuals to PC.［35］Increasing evidence suggests that inflammation can promote tumor invasion and metastasis， leading some researchers to advocate cancer-related inflammation as the seventh hallmark of cancer.［36， 37］

Chemokine receptors and their ligands can direct the movement of cells in cancer by affecting cell motility， invasiveness and survival. In many cancers， including PC， the expression of chemokine receptors increases，thereby promoting communication between chemokines and their receptors for the migration， localization and survival of cancer cells in distant organs.［38］Studies have indicated that overexpression of CXCR4 in metastatic PC cell lines leads to increased cell invasion and metastasis through interaction with its ligand CXCL12.［39］Recently， Wen et al［40］found that extracellular DNA （ex-DNA）， a newly discovered component of inflammatory tissue states induced by mediators of inflammation （for example CXCL8）， contributed to the highly invasive and metastatic characteristics of PC. They detected exDNA on the surface of PC cells， where it is critical to metastatic behavior in vitro and in vivo. Further mechanistic investigations suggested the presence of a vicious circle in which exDNA promotes the expression of the inflammatory chemokine CXCL8， which in turn stimulates the production of exDNA by PC cells. Intriguingly， this loop can be terminated by DNase I treatment， thus inhibiting PC metastasis.

Another important potential mechanism that links inflammation and cancer is NF-κB-mediated signaling. NF-κB， as a transcription factor， plays a crucial role in tumorigenesis， apoptosis， angiogenesis， chemoresistance，invasion and metastasis in PC by regulating a series of relevant genes.［41， 42］NF-κB can regulate EMT through the NF-κB/Snail1 pathway，［43］serving as an important component of invasion and the development of metastases. Wu et al［44］proposed that Snail could be stabilized by inflammatory cytokine TNF-α-induced activation of NF-κB. Moreover， they demonstrated that knockdown of Snail expression reduced cell migration and invasion mediated by TNF-α， thus providing a plausible mechanism for inflammation-induced metastasis. Similar findings were also demonstrated by Maier et al.［45］Even more compelling， they suggested that NF-κB activation by TNF-α could induce EMT even in PC cells with defective TGF-β signaling. In contrast， TGF-β-induced EMT relies on intact NF-κB signaling and could be blocked in pancreatic cells that express the NF-κB inhibitory protein， IκB-α. These insights are consistent with the observation that EMT and invasiveness are most abundant at inflammatory foci， and the induction of pancreatitis can increase the number of circulating pancreatic cells. Conversely， treatment with the anti-inflammatory agent dexamethasone may attenuate this dissemination.［19］

The inflammatory microenvironment of tumors is characterized by the presence of host leukocytes， both in the supporting stroma and in tumor areas.［36］On the one hand， these infiltrating leukocytes can release multiple chemokines to guide cancer cell migration. On the other hand， tumor cells also have paracrine effects on stromal cells to induce a more favorable microenvironment for metastasis. Kurahara et al［46］discovered that M2-polarized tumor-associated macrophages （TAMs）facilitate nodal lymphangiogenesis by producing vascular endothelial growth factor C， thus promoting lymphatic metastasis in PC. Neyen et al［47］demonstrated that activated macrophages could up-regulate expression of the pattern recognition receptor scavenger receptor A （SR-A）via co-culture of macrophages with tumor cells， resulting in increased PC cell migration in vitro and in vivo. Interestingly， co-culturing of macrophages with tumor cells also could activate the expression of Toll-like receptor 4 （TLR4） on macrophages and promote EMT in PC cells through TLR4/IL-10 signaling.［48］In addition， Singh et al［49］reported that the CXCR4/CXCL12 protein signaling axis up-regulates SHH， depending on the activation of downstream Akt and ERK signaling pathways. Both of these pathways cooperatively promote nuclear accumulation of NF-κB by inducing the phosphorylation and destabilization of its inhibitory protein IκB-α. The results of their study identified a complicated regulatory network of bidirectional tumor-stroma interactions involving CXCR4/CXCL12， SHH， Akt， ERK and NF-κB.

Another seemingly paradoxical aspect of inflammation in cancer is immune surveillance， which was originally proposed by Burnet and Thomas more than 50 years ago. Strong evidence from genetic studies of mouse models indicates that cells of the adaptive immune system could act as sentinels in recognizing and eliminating nascent transformed cells.［50］Innate immune responses，which manifest as inflammation， are crucial for the initiation of adaptive immune responses. In fact， extensive evidence indicates that this surveillance function of immunity is only the beginning of the story. The immune system—tumor interaction is a dynamic and sophisticat-ed process， characterized as “cancer immunoediting”.［51］The phrase， “elimination， equilibrium and escape”， is sometimes used to sum up this process whereby the immune system controls tumor outgrowth and shapes tumor immunogenicity. A central principle of cancer immunoediting is T-cell recognition of tumor antigens，which drives the immunological elimination or sculpting of a developing cancer. These neo-antigens arise from tumor-specific mutations； their recognition is a major factor in the activity of clinical immunotherapies. Therefore， genetic sequencing may also provide a method of identifying tumor-specific mutational antigens， thus laying a foundation for development of individualized cancer immunotherapies.

Stress response

Metastasis formation is conditioned by the capacity of cancer cells to overcome various concomitant stresses， to survive and flourish within their new and often very hostile environments. Many intracellular signaling pathways are implicated in this stress response， including NF-κB，［42， 50］Notch［51， 52］and Hedgehog，［53， 54］among which NF-κB generally represents “a central mediator of the human stress response”. Activation of these signaling pathways regulates relevant target gene expression， which can evoke effective responses against the stress and ensure that the cell does not succumb during the process.

A pivotal factor in the pancreatic stress response is the stress protein p8， a small chromatin protein encoded by Nupr1， which was identified more than one and a half decades ago during the acute phase response of induced pancreatitis. Cano et al［55］noted that p8 is increased in response to most common， even very mild， stresses including starvation， hypoxia， and growth-inhibiting and apoptosis-inducing signals. The increased p8 promotes tumor metastasis and cell-cycle arrest， and prevents apoptosis. Interestingly， the Nupr1/p8 promoter contains three Smad-responsive elements and an NF-κB-specific motif， suggesting a possible role for p8 in Smad-mediated transduction and the NF-κB signaling pathway.［56］In their subsequent studies， Sandi et al［57］confirmed that p8 expression controls PC cell migration， invasion and adhesion—three key processes required for metastasis. Their cDNA microarray data showed that CDC42 is a major regulator of cytoskeleton organization. Silencing of p8 provokes the overexpression of CDC42， resulting in increased cell adhesion and decreased migration and invasion， whereas CDC42 knockdown almost completely reverses the effects of p8 silencing in vitro. In addition，they found that activation of Nupr1 protects cells against stress-induced death by inhibiting apoptosis through a pathway that is dependent on the transcription factor RelB and immediate early response 3.［58］In contrast， inactivation of Nupr1 promotes homotypic cell cannibalism， which suppresses metastasis and results in death in PC cells.［59］A recent study further confirmed the protective role of Nupr1 against metabolic stress-induced autophagy-associated cell death in PC cells.［60］The findings of this study revealed a process that crucially underlies PC tumorigenesis and metastasis formation， providing new avenues for more efficient therapeutic targeting.

Extremely low oxygen tension exists in tumor tissues，especially PC， and has been shown to increase tumor invasiveness.［51］Onishi et al［53］investigated the contribution of the HH pathway to hypoxia-induced invasiveness and proposed that hypoxia activates the HH pathway in PC cells by increasing the transcription of Smo in a ligand-independent manner. More recently， Morifuji et al［54］conducted reoxygenation experiments using chronic-hypoxia-resistant PC cells that were newly generated under hypoxic conditions for 3-6 months. They found that reoxygenation markedly up-regulates Gli expression，resulting in significantly increased proliferation， invasiveness， and tumorigenicity in PC cells. Furthermore，inhibition of sonic HH and Smoothened counteracts this phenomenon. The findings of these studies suggest that the HH signaling pathway may be a therapeutic target for refractory PC in a metastatic process induced by chronic hypoxia or reoxygenation.

Circulating tumor cells

Circulating tumor cells （CTCs） are tumor cells that circulate through normal vessels and capillaries， and through neovessels formed by tumor-induced angiogenesis. The presence of CTCs has been shown to correlate with distant metastases and unfavorable prognosis in patients with biliary and PCs.［61］Further characterization of CTCs leads to the isolation of subpopulations of tumor initiating cells， also called cancer stem cells， which facilitate both diagnosis and treatment of early-stage cancer， and discovery of novel treatments for late-stage PC.［62］Li et al［63］first identified PC stem cells using an xenograft model. They identified a highly tumorigenic subpopulation of PC cells with a CD44+/CD24+/ESA+phenotype； this subpopulation accounted for only 0.2%-0.8% of the PC cells but exhibited stem cell attributes such as self-renewal， differentiation and migration. Similarly， Rhim et al［19］proposed that a portion of circulating pancreatic cells exhibits stem cell properties and seeds the liver， while inflammation increases blood entry and enriches circulating pancreatic cells to further contribute to tumor dissemination. In fact， the capability ofCTCs or cancer stem cells to survive and metastasize is determined by both their intrinsic self-renewal potential and by their interactions with the niche microenvironment， which controls the balance between the quiescence and proliferation of CTCs through complex signaling pathways. As aforementioned， EMT can endow cancer cells with stem cell-like characteristics， thereby increasing tumor invasion and metastasis. Therefore， the premetastatic niche plays a significant role in determining the fate of cells. Elucidating the interactions in this niche will be crucial for PC prevention and treatment. Because CTCs are typically present at low concentrations in cancer patients， the technologies required for their isolation and characterization are the greatest challenge for research in this field.［62， 64］However， with the emergence of new technologies and approaches， CTC genotyping and phenotyping can be predicted to play an increasingly prominent role in personalized therapy and the identification of novel therapeutic targets. They will also facilitate the discovery of effective biomarkers for monitoring the course and status of the disease. More recently， Sheng et al［65］described a remarkable technique for high-efficiency （＞90%）， high-purity （＞84%） collection of CTCs. This technique is named a geometrically enhanced mixing chip， which applies the principles of geometrically optimized micromixer structures to enhance transverse flow and flow folding， maximizing interactions between CTCs and antibody-coated surfaces. More importantly，using a combination of trypsinization and high flow-rate washing， CTCs were successfully released and isolated as intact cells that were not subjected to the capture and release process in 17 of 18 blood samples （＞94%） from PC patients. A new era of personalized therapies and realtime monitoring of cancer patients appears to be developing as a result of CTC research.

Most worthy of mention is CD47， which was identified as a biomarker expressed in a subset of circulating luminal breast cancer cells with metastasis-initiating capacity but is rarely detected in non-metastatic primary tumor cells.［66］In a small cohort of patients with metastases， increased EPCAM+/CD44+/CD47+/MET+CTCs，but not bulk EPCAM+CTCs， correlated with lower overall survival and increased metastasis. CD47 is an antiphagocytic signal that can help cancer cells to evade macrophage-mediated destruction. Blockage of CD47 has a remarkable anti-metastasis effect in animal models.［67］Targeting strategies are being actively developed for clinical use. Recently， a high-affinity antagonist of human CD47 was reported to exhibit a prominent synergistic effect with other tumor-specific monoclonal antibodies by increasing phagocytosis in vitro and enhancing antitumor responses in vivo， thereby providing a universal method for augmenting the efficacy of therapeutic anticancer antibodies.［68］

Fig. Putative schematic views of the biological and genetic cascades from the primary tumor to metastatic clones. EMT： epithelial-mesenchymal transition； CTC： circulating tumor cell.

Conclusions and perspectives

The most devastating aspect of cancer is the emergence of metastases in organs distant from the primary tumor. As the culprit behind most cancer related death， metastasis constitutes the ultimate challenge in human anticancer wars since it often happens in an unexpected earlystage and conventional therapeutic strategies only have a limited efficacy after “the horse has left the barn”. For PC， this is extremely the truth， but the precise underlying mechanisms have not been defined. These mechanisms are thought to involve numerous mutations that allow primary tumor cells to complete a sequence of histological and molecular changes （Fig.）. Thus， understanding the potential molecular mechanisms that underlie this complex process is a critical topic in cancer research. Recent advances in high-resolution sequencing technologies have enabled identification of families of genes， not just single genes， with higher allele frequencies that are involved in metastasis， and these techniques can elucidate the central events in the metastatic cascade. The next step is to determine the timing of these genetic alterations， and the specific mechanisms by which they regulate PC metastasis. Further analysis of gene function and relevant signaling pathways will help establish the regulatory networks that are responsible for PC metastasis and lay a foundation for the development of genomics based targeted therapy. Agents targeting oncogenic mutations， such as amplified ERBB2 in breast cancer （Trastuzumab）， mutant EGFR in NSCLC （Gefitinib）and BRAF mutation in melanoma （Vemurafenib）， have achieved acclaimed success in controlling advanced tumor or even metastatic disease. However， treatment of metastatic disease like PC is more complicated by the dif-ferential expression， activity or both of oncogenes in primary tumors and metastases， and the dynamic evolution of cancer genomes during progression. Therefore， over the coming years， much work remains to complete the mutational atlas in both primary tumors and metastases，and what is more， to create a functional encyclopedia of altered pathways and acquired vulnerabilities correspond to each cancer genome.［69］Cancer genomics is a comprehensive way to know about the enemy， as Garraway and Lander［69］described， though alone not a guarantee of victory， it is an essential part of any overall attack. As clinical oncology moves toward personalized cancer medicine，another promising field is the genotyping and phenotyping of CTCs， which can be used to identify additional specific markers and targets， and could lead to a new era of personalized therapy for patients with PC. CTCs are currently included as biomarkers in a large number of clinical trials.［70］With more advanced technologies applied to detection and characterization of CTCs， a breakthrough in cancer diagnostics and drug development can be expected in the near future. The complex interaction between the microenvironment and the tumor in terms of EMT， the stress response and cancer-related inflammation are also important. Detailed investigations in these areas will facilitate our understanding of tumor cell disassociation from the primary tumor， escape from immune system attack， adaptation to a hostile microenvironment and formation of metastatic clones.

Contributors： DYX wrote the manuscript. LZW， YL and WWM critically reviewed the manuscript. ZYP is the guarantor.

Funding： This study was supported by grants from the National Natural Science Foundation of China （81272767 and 81201734）. Ethical approval： Not needed.

Competing interest： No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

References

1 Siegel R， Ma J， Zou Z， Jemal A. Cancer statistics， 2014. CA Cancer J Clin 2014；64：9-29.

2 Vincent A， Herman J， Schulick R， Hruban RH， Goggins M. Pancreatic cancer. Lancet 2011；378：607-620.

3 Sohn TA， Yeo CJ， Cameron JL， Koniaris L， Kaushal S， Abrams RA， et al. Resected adenocarcinoma of the pancreas-616 patients： results， outcomes， and prognostic indicators. J Gastrointest Surg 2000；4：567-579.

4 Fidler IJ. The pathogenesis of cancer metastasis： the ‘seed and soil' hypothesis revisited. Nat Rev Cancer 2003；3：453-458.

5 Valastyan S， Weinberg RA. Tumor metastasis： molecular insights and evolving paradigms. Cell 2011；147：275-292.

6 Embuscado EE， Laheru D， Ricci F， Yun KJ， de Boom Witzel S，Seigel A， et al. Immortalizing the complexity of cancer metastasis： genetic features of lethal metastatic pancreatic cancer obtained from rapid autopsy. Cancer Biol Ther 2005；4：548-554.

7 Kim HN， Choi DW， Lee KT， Lee JK， Heo JS， Choi SH， et al. Gene expression profiling in lymph node-positive and lymph node-negative pancreatic cancer. Pancreas 2007；34：325-334.

8 Campagna D， Cope L， Lakkur SS， Henderson C， Laheru D，Iacobuzio-Donahue CA. Gene expression profiles associated with advanced pancreatic cancer. Int J Clin Exp Pathol 2008；1：32-43.

9 Stratford JK， Bentrem DJ， Anderson JM， Fan C， Volmar KA，Marron JS， et al. A six-gene signature predicts survival of patients with localized pancreatic ductal adenocarcinoma. PLoS Med 2010；7：e1000307.

10 Niedergethmann M， Alves F， Neff JK， Heidrich B， Aramin N， Li L， et al. Gene expression profiling of liver metastases and tumour invasion in pancreatic cancer using an orthotopic SCID mouse model. Br J Cancer 2007；97：1432-1440.

11 Suemizu H， Monnai M， Ohnishi Y， Ito M， Tamaoki N， Nakamura M. Identification of a key molecular regulator of liver metastasis in human pancreatic carcinoma using a novel quantitative model of metastasis in NOD/SCID/gammacnull （NOG）mice. Int J Oncol 2007；31：741-751.

12 Jones S， Zhang X， Parsons DW， Lin JC， Leary RJ， Angenendt P，et al. Core signaling pathways in human pancreatic cancers revealed by global genomic analyses. Science 2008；321：1801-1806.

13 Boj SF， Hwang CI， Baker LA， Chio II， Engle DD， Corbo V， et al. Organoid models of human and mouse ductal pancreatic cancer. Cell 2015；160：324-338.

14 Yachida S， Jones S， Bozic I， Antal T， Leary R， Fu B， et al. Distant metastasis occurs late during the genetic evolution of pancreatic cancer. Nature 2010；467：1114-1117.

15 Campbell PJ， Yachida S， Mudie LJ， Stephens PJ， Pleasance ED，Stebbings LA， et al. The patterns and dynamics of genomic instability in metastatic pancreatic cancer. Nature 2010；467：1109-1113.

16 Zhou B， Irwanto A， Guo YM， Bei JX， Wu Q， Chen G， et al. Exome sequencing and digital PCR analyses reveal novel mutated genes related to the metastasis of pancreatic ductal adenocarcinoma. Cancer Biol Ther 2012；13：871-879.

17 Zhou B， Wu Q， Chen G， Zhang TP， Zhao YP. NOP14 promotes proliferation and metastasis of pancreatic cancer cells. Cancer Lett 2012；322：195-203.

18 Haeno H， Gonen M， Davis MB， Herman JM， Iacobuzio-Donahue CA， Michor F. Computational modeling of pancreatic cancer reveals kinetics of metastasis suggesting optimum treatment strategies. Cell 2012；148：362-375.

19 Rhim AD， Mirek ET， Aiello NM， Maitra A， Bailey JM， McAllister F， et al. EMT and dissemination precede pancreatic tumor formation. Cell 2012；148：349-361.

20 Tuveson DA， Neoptolemos JP. Understanding metastasis in pancreatic cancer： a call for new clinical approaches. Cell 2012；148：21-23.

21 Tan X， Zhou L， Wang W， Wang B， Egami H， Baba H， et al. Genomic analysis of invasion-metastasis-related factors in pancreatic cancer cells. Exp Ther Med 2010；1：211-216.

22 Weissmueller S， Manchado E， Saborowski M， Morris JP 4th，Wagenblast E， Davis CA， et al. Mutant p53 drives pancreatic cancer metastasis through cell-autonomous PDGF receptor β signaling. Cell 2014；157：382-394.

23 Thiery JP， Acloque H， Huang RY， Nieto MA. Epithelial-mesenchymal transitions in development and disease. Cell 2009；139：871-890.

24 Hanahan D， Weinberg RA. Hallmarks of cancer： the next generation. Cell 2011；144：646-674.

25 von Burstin J， Eser S， Paul MC， Seidler B， Brandl M， Messer M，et al. E-cadherin regulates metastasis of pancreatic cancer in vivo and is suppressed by a SNAIL/HDAC1/HDAC2 repressor complex. Gastroenterology 2009；137：361-371， 371.e1-5.

26 Syngal S， Brand RE， Church JM， Giardiello FM， Hampel HL，Burt RW， et al. ACG clinical guideline： Genetic testing and management of hereditary gastrointestinal cancer syndromes. Am J Gastroenterol 2015；110：223-263.

27 Mani SA， Guo W， Liao MJ， Eaton EN， Ayyanan A， Zhou AY， et al. The epithelial-mesenchymal transition generates cells with properties of stem cells. Cell 2008；133：704-715.

28 Rasheed ZA， Yang J， Wang Q， Kowalski J， Freed I， Murter C， et al. Prognostic significance of tumorigenic cells with mesenchymal features in pancreatic adenocarcinoma. J Natl Cancer Inst 2010；102：340-351.

29 Yang J， Weinberg RA. Epithelial-mesenchymal transition： at the crossroads of development and tumor metastasis. Dev Cell 2008；14：818-829.

30 Nishioka R， Itoh S， Gui T， Gai Z， Oikawa K， Kawai M， et al. SNAIL induces epithelial-to-mesenchymal transition in a human pancreatic cancer cell line （BxPC3） and promotes distant metastasis and invasiveness in vivo. Exp Mol Pathol 2010；89：149-157.

31 Xu X， Zhou Y， Xie C， Wei SM， Gan H， He S， et al. Genomewide screening reveals an EMT molecular network mediated by Sonic hedgehog-Gli1 signaling in pancreatic cancer cells. PLoS One 2012；7：e43119.

32 Feldmann G， Dhara S， Fendrich V， Bedja D， Beaty R， Mullendore M， et al. Blockade of hedgehog signaling inhibits pancreatic cancer invasion and metastases： a new paradigm for combination therapy in solid cancers. Cancer Res 2007；67：2187-2196.

33 Aghdassi A， Sendler M， Guenther A， Mayerle J， Behn CO，Heidecke CD， et al. Recruitment of histone deacetylases HDAC1 and HDAC2 by the transcriptional repressor ZEB1 downregulates E-cadherin expression in pancreatic cancer. Gut 2012；61：439-448.

34 Balkwill F， Mantovani A. Inflammation and cancer： back to Virchow？ Lancet 2001；357：539-545.

35 Yadav D， Lowenfels AB. The epidemiology of pancreatitis and pancreatic cancer. Gastroenterology 2013；144：1252-1261.

36 Mantovani A， Allavena P， Sica A， Balkwill F. Cancer-related inflammation. Nature 2008；454：436-444.

37 Mantovani A. Cancer： Inflaming metastasis. Nature 2009；457：36-37.

38 Balkwill F. Cancer and the chemokine network. Nat Rev Cancer 2004；4：540-550.

39 Saur D， Seidler B， Schneider G， Algül H， Beck R， Senekowitsch-Schmidtke R， et al. CXCR4 expression increases liver and lung metastasis in a mouse model of pancreatic cancer. Gastroenterology 2005；129：1237-1250.

40 Wen F， Shen A， Choi A， Gerner EW， Shi J. Extracellular DNA in pancreatic cancer promotes cell invasion and metastasis. Cancer Res 2013；73：4256-4266.

41 Holcomb B， Yip-Schneider M， Schmidt CM. The role of nuclear factor kappaB in pancreatic cancer and the clinical applications of targeted therapy. Pancreas 2008；36：225-235.

42 Fujioka S， Sclabas GM， Schmidt C， Frederick WA， Dong QG，Abbruzzese JL， et al. Function of nuclear factor kappaB in pancreatic cancer metastasis. Clin Cancer Res 2003；9：346-354.

43 Strippoli R， Benedicto I， Pérez Lozano ML， Cerezo A， López-Cabrera M， del Pozo MA. Epithelial-to-mesenchymal transition of peritoneal mesothelial cells is regulated by an ERK/NF-kappaB/Snail1 pathway. Dis Model Mech 2008；1：264-274.

44 Wu Y， Deng J， Rychahou PG， Qiu S， Evers BM， Zhou BP. Stabilization of snail by NF-kappaB is required for inflammationinduced cell migration and invasion. Cancer Cell 2009；15：416-428.

45 Maier HJ， Schmidt-Strassburger U， Huber MA， Wiedemann EM， Beug H， Wirth T. NF-kappaB promotes epithelial-mesenchymal transition， migration and invasion of pancreatic carcinoma cells. Cancer Lett 2010；295：214-228.

46 Kurahara H， Takao S， Maemura K， Mataki Y， Kuwahata T，Maeda K， et al. M2-polarized tumor-associated macrophage infiltration of regional lymph nodes is associated with nodal lymphangiogenesis and occult nodal involvement in pN0 pancreatic cancer. Pancreas 2013；42：155-159.

47 Neyen C， Plüddemann A， Mukhopadhyay S， Maniati E，Bossard M， Gordon S， et al. Macrophage scavenger receptor a promotes tumor progression in murine models of ovarian and pancreatic cancer. J Immunol 2013；190：3798-3805.

48 Liu CY， Xu JY， Shi XY， Huang W， Ruan TY， Xie P， et al. M2-polarized tumor-associated macrophages promoted epithelialmesenchymal transition in pancreatic cancer cells， partially through TLR4/IL-10 signaling pathway. Lab Invest 2013；93：844-854.

49 Singh AP， Arora S， Bhardwaj A， Srivastava SK， Kadakia MP，Wang B， et al. CXCL12/CXCR4 protein signaling axis induces sonic hedgehog expression in pancreatic cancer cells via extracellular regulated kinase- and Akt kinase-mediated activation of nuclear factor κB： implications for bidirectional tumorstromal interactions. J Biol Chem 2012；287：39115-39124.

50 Pahl HL. Activators and target genes of Rel/NF-kappaB transcription factors. Oncogene 1999；18：6853-6866.

51 Sahlgren C， Gustafsson MV， Jin S， Poellinger L， Lendahl U. Notch signaling mediates hypoxia-induced tumor cell migration and invasion. Proc Natl Acad Sci U S A 2008；105：6392-6397.

52 Cheng ZX， Wang DW， Liu T， Liu WX， Xia WB， Xu J， et al. Effects of the HIF-1α and NF-κB loop on epithelial-mesenchymal transition and chemoresistance induced by hypoxia in pancreatic cancer cells. Oncol Rep 2014；31：1891-1898.

53 Onishi H， Kai M， Odate S， Iwasaki H， Morifuji Y， Ogino T， et al. Hypoxia activates the hedgehog signaling pathway in a ligandindependent manner by upregulation of Smo transcription in pancreatic cancer. Cancer Sci 2011；102：1144-1150.

54 Morifuji Y， Onishi H， Iwasaki H， Imaizumi A， Nakano K， Tanaka M， et al. Reoxygenation from chronic hypoxia promotes metastatic processes in pancreatic cancer through the Hedgehog signaling. Cancer Sci 2014；105：324-333.

55 Cano CE， Iovanna JL. Stress proteins and pancreatic cancer metastasis. ScientificWorldJournal 2010；10：1958-1966.

56 Kallwellis K， Grempler R， Günther S， P?th G， Walther R. Tumor necrosis factor alpha induces the expression of the nuclear protein p8 via a novel NF kappaB binding site within the promoter. Horm Metab Res 2006；38：570-574.

57 Sandi MJ， Hamidi T， Malicet C， Cano C， Loncle C， Pierres A，et al. p8 expression controls pancreatic cancer cell migration，invasion， adhesion， and tumorigenesis. J Cell Physiol 2011；226：3442-3451.

58 Hamidi T， Algül H， Cano CE， Sandi MJ， Molejon MI， Riemann M， et al. Nuclear protein 1 promotes pancreatic cancer development and protects cells from stress by inhibiting apoptosis. J Clin Invest 2012；122：2092-2103.

59 Cano CE， Sandí MJ， Hamidi T， Calvo EL， Turrini O， Bartholin L， et al. Homotypic cell cannibalism， a cell-death process regulated by the nuclear protein 1， opposes to metastasis in pancreatic cancer. EMBO Mol Med 2012；4：964-979.

60 Hamidi T， Cano CE， Grasso D， Garcia MN， Sandi MJ， Calvo EL， et al. NUPR1 works against the metabolic stress-induced autophagy-associated cell death in pancreatic cancer cells. Autophagy 2013；9：95-97.

61 Uchikura K， Takao S， Nakajo A， Miyazono F， Nakashima S，Tokuda K， et al. Intraoperative molecular detection of circulating tumor cells by reverse transcription-polymerase chain reaction in patients with biliary-pancreatic cancer is associated with hematogenous metastasis. Ann Surg Oncol 2002；9：364-370.

62 Cen P， Ni X， Yang J， Graham DY， Li M. Circulating tumor cells in the diagnosis and management of pancreatic cancer. Biochim Biophys Acta 2012；1826：350-356.

63 Li C， Heidt DG， Dalerba P， Burant CF， Zhang L， Adsay V， et al. Identification of pancreatic cancer stem cells. Cancer Res 2007；67：1030-1037.

64 Bhat K， Wang F， Ma Q， Li Q， Mallik S， Hsieh TC， et al. Advances in biomarker research for pancreatic cancer. Curr Pharm Des 2012；18：2439-2451.

65 Sheng W， Ogunwobi OO， Chen T， Zhang J， George TJ， Liu C， et al. Capture， release and culture of circulating tumor cells from pancreatic cancer patients using an enhanced mixing chip. Lab Chip 2014；14：89-98.

66 Baccelli I， Schneeweiss A， Riethdorf S， Stenzinger A， Schillert A，Vogel V， et al. Identification of a population of blood circulating tumor cells from breast cancer patients that initiates metastasis in a xenograft assay. Nat Biotechnol 2013；31：539-544.

67 Chao MP， Weissman IL， Majeti R. The CD47-SIRPα pathway in cancer immune evasion and potential therapeutic implications. Curr Opin Immunol 2012；24：225-232.

68 Weiskopf K， Ring AM， Ho CC， Volkmer JP， Levin AM， Volkmer AK， et al. Engineered SIRPα variants as immunotherapeutic adjuvants to anticancer antibodies. Science 2013；341：88-91.

69 Garraway LA， Lander ES. Lessons from the cancer genome. Cell 2013；153：17-37.

70 Parkinson DR， Dracopoli N， Petty BG， Compton C， Cristofanilli M， Deisseroth A， et al. Considerations in the development of circulating tumor cell technology for clinical use. J Transl Med 2012；10：138.

Accepted after revision August 28， 2015

10.1016/S1499-3872（15）60033-9

Yu-Pei Zhao， MD， Department of General Surgery， Peking Union Medical College Hospital， Chinese Academy of Medical Sciences and Peking Union Medical College， Beijing 100730， China （Tel： +86-10-69156007； Fax： +86-10-65124875； Email： zhao8028@263.net）

March 23， 2015