REVIEWS IN BASIC AND CLINICALGASTROENTEROLOGY AND HEPATOLOGY
Robert F. Schwabe and John W. Wiley, Section Editors
Role of the Microenvironment in the Pathogenesis and Treatment ofHepatocellular CarcinomaVIRGINIA HERNANDEZ–GEA,1,2,* SARA TOFFANIN,1,3,4,* SCOTT L. FRIEDMAN,1,3 and JOSEP M. LLOVET1,2,3,5
1Division of Liver Diseases, Mount Sinai School of Medicine, New York, New York; 2HCC Translational Research Laboratory, Barcelona-Clínic Liver Cancer Group,Institut d’Investigacions Biomèdiques August Pi i Sunyer, Liver Unit, Hospital Clinic, University of Barcelona, Catalonia, Spain; 3Mount Sinai Liver Cancer Program
4
Divisions of Liver Diseases, Tisch Cancer Institute), Mount Sinai School of Medicine, New York, New York; Gastrointestinal Surgery and Liver Transplantation Unit,ational Cancer Institute, IRCSS Foundation, Milan, Italy; and 5Institució Catalana de Recerca i Estudis Avançats, Barcelona, Catalonia, Spain
rpta
raTe
Hepatocellular carcinoma (HCC) is the most commonprimary liver tumor and the third greatest cause ofcancer-related death worldwide, and its incidence isincreasing. Despite the significant improvement inmanagement of HCC over the past 30 years, there areno effective chemoprevention strategies, and only onesystemic therapy has been approved for patients withadvanced tumors. This drug, sorafenib, acts on tumorcells and the stroma. HCC develops from chronicallydamaged tissue that contains large amounts of inflam-mation and fibrosis, which also promote tumor pro-gression and resistance to therapy. Increasing our un-derstanding of how stromal components interact withcancer cells and the signaling pathways involved couldhelp identify new therapeutic and chemopreventivetargets.
Hepatocellular carcinoma (HCC) is the most commonprimary form of liver cancer and the third most
eadly type of cancer globally, following lung and stom-ch cancers.1 With more than 750,000 new cases diag-
nosed every year worldwide, HCC is the sixth most com-mon neoplasm.2 Unlike other carcinomas, its incidence issteeply increasing, mainly due to the increasing prevalenceof advanced hepatitis C virus (HCV) infection. HCC com-monly arises in the setting of cirrhosis (�80% of cases),appearing 20 to 30 years after the initial insult to the liver.The use of antivirals and vaccination has successfullydiminished the incidence of hepatitis B virus (HBV)-re-lated HCC, although there are no effective chemopreven-tive strategies to attenuate the development of canceronce cirrhosis is established.3 HCC is diagnosed in mostpatients at advanced/symptomatic stages, when limitedtherapeutic options are available. The results of the ran-domized phase 3 SHARP (Sorafenib HCC Assessment
Randomized Protocol) trial showed that the multikinase
inhibitor sorafenib improved overall survival of patientswith advanced HCC,4 representing a breakthrough in theclinical management of this cancer.
The liver tumor microenvironment is a complex mix-ture of tumoral cells within the extracellular matrix(ECM), combined with a complex mix of stromal cells andthe proteins they secrete. Together, these elements con-tribute to the carcinogenic process. Cancer cells do notmanifest the disease alone, and the stroma is inappropri-ately activated in cancer to contribute to malignant char-acteristics of tumor cells. The tumor microenvironmentand the tumor cells create a complex cellular system withreciprocal signaling (Figure 1).
Stromal components of the microenvironment can bedivided into 3 subclasses: angiogenic cells, immune cells,and cancer-associated fibroblastic cells. There is growingevidence of the contribution of stromal cells to the hall-marks of cancer: sustaining proliferative signaling, evadinggrowth suppressors, resisting cell death, enabling replicativeimmortality, inducing angiogenesis, activating invasion andmetastasis, reprogramming energy metabolism, and evadingimmune destruction.5 Alterations within the microenvi-onment may favor tumor progression and play an im-ortal role in chemoresistance.6,7 Targeting stromal cellso abrogate their tumor-supporting role represents anttractive therapeutic strategy.
*Authors share co-first authorship.
Abbreviations used in this paper: CSC, cancer stem cell; EC, endo-thelial cell; ECM, extracellular matrix; EGF, epidermal growth factor;FGF, fibroblast growth factor; HGF, hepatocyte growth factor; HIF-1,hypoxia-inducible factor 1; IL, interleukin; MMP, metalloproteinase;NF-�B, nuclear factor �B; PDGF, platelet-derived growth factor; ROS,eactive oxygen species; TAF, tumor-associated fibroblast; TAM, tumor-ssociated macrophage; TIMP, tissue inhibitor of metalloproteinase;reg, regulatory T cell; TGF, transforming growth factor; VEGF, vascularndothelial growth factor.
March 2013 ROLE OF THE MICROENVIRONMENT IN HCC 513
The role of the microenvironment in tumor initiationand progression in HCC is critical. For instance, the statusof nontumoral tissue has an important role in predictingtumor recurrence, which affects 70% of patients afterresection or local ablation.8 Typically, there are 2 patternsof HCC recurrence: true metastasis of the primary tumor(generally within 2 years following resection/transplanta-tion, defined as “early recurrence”) and de novo tumor(after 2 years from treatment or “late recurrence”).9,10
Among these features, late recurrence is generally dictatedby the persistence of protumorigenic signals within thedamaged milieu of the fibrotic and cirrhotic liver11; dis-inct molecular subgroups of HCC have been identifiednd linked to poor prognosis.12–17 In another context, the
information encoded within the surrounding adjacent non-tumoral tissue is essential to predicting the outcome ofpatients at very early stages (ie, tumors less than 2 cmwithout vascular invasion or extrahepatic spread) and maybe even more relevant than the genomic profile of the tumoritself.10 These findings highlight the profound involvement
f a dynamic network of nontumoral cells, molecules, andoluble factors in the generation of a supportive and permis-ive environment for initiation and progression of HCC.
In this review, we provide an overview of current knowledgen the role of the tumor microenvironment in HCC and high-
Figure 1. Cellular componentsf the microenvironment andolecular mechanisms influenc-
ng tumor growth and progres-ion. Interactions among stro-al, inflammatory, and cancer
ight potential prognostic and therapeutic implications.
Importance of the TumorMicroenvironmentThe development and progression of HCC is a
multistage process. A chronic insult (eg, HCV, HBV, andalcohol) induces liver injury through reactive oxygen spe-cies (ROS) production, cellular DNA damage, endoplas-mic reticulum stress, and necrosis of damaged hepato-cytes. Most HCCs arise in the setting of chronic hepatitisinduced by HCV or HBV infection. HCV is a single-stranded RNA virus that cannot integrate into the hostgenome but triggers an immune-mediated inflammatoryresponse that promotes neoplastic transformation ofdamaged hepatocytes. Conversely, HBV can integrate intothe genome of infected hepatocytes and promotes hepa-tocarcinogenesis through sustained inflammatory dam-age, hepatocyte regeneration, direct oncogenic transfor-mation following integration of the viral genome intohost genes, and the transactivating potential of severalviral oncoproteins, especially HBx. The sustained dysregu-lation of the liver cell by HBV infection can ultimatelyaffect DNA repair mechanisms and promote mutationalevents, which contribute to malignant transformation ofhepatocytes.
The hepatic response involves the activation of he-
patic stellate cells and macrophages, which produce
aahrp
sb
tt
doaartoss
gtc
REV
IEWS
AN
DPER
SPEC
TIVES
514 HERNANDEZ–GEA ET AL GASTROENTEROLOGY Vol. 144, No. 3
components of the ECM and growth factors that pro-mote migration of endothelial cells, neoangiogenesis,and fibrosis. This process is associated with distortionof the parenchyma and vascular architecture character-ized by progressive capillarization, with reduction ofendothelial cell fenestrae size, and deposition of base-ment membrane components including collagen typeIV and laminin within the space of Disse. This process,in the context of inflammation and oxidative DNAdamage, favors the accumulation of mutations andepigenetic aberrations in preneoplastic hepatocytes orliver stem cells, thereby promoting the development ofdysplastic nodules and their malignant transformationto early HCC.18 Therefore, HCC is not just a mixture ofcells and ECM; it contains several cell types that inter-act with each other and the surrounding tissue, creat-ing a complex interaction network within a permissivemicroenvironment. The stromal components supporttumor growth and promote invasion through the stim-ulation of cancer cell proliferation, migration, and in-vasion and activation of angiogenesis, which togetherdetermine the phenotype of the tumor.
Relevance of the Microenvironment inOther MalignanciesThe link between inflammation and generation of
a preneoplastic milieu has been reported in many diseases,such as in the development of colorectal and pancreaticcarcinomas in the context of inflammatory bowel diseaseand chronic pancreatitis, respectively.19 Once the cancerhas been established, the contribution of the microenvi-ronment to the regulation of tumor behavior has beenwell recognized for other malignancies, including breast,lung, and pancreatic carcinomas.5
Abnormal ECM production and altered physical prop-erties are frequently reported in malignancies. In breastcarcinoma, for example, the tumor stroma is 10 timesstiffer than in the normal breast, partially due to excessactivity of lysyl oxidase and accumulation of collagen andother ECM components.20 Similarly, in pancreatic ductal
denocarcinoma, the large amounts of ECM proteins,ctivated fibroblasts, stellate cells, and inflammatory cellsave been described as a “fortress-like” hypovascular bar-ier that impairs the delivery of chemotherapeutics andromotes aggressive neoplastic cell behavior.21
Different aspects of tumor biology, including develop-ment, progression, and response to therapy, can be af-fected by components of the tumor microenvironment. Inmice, the recapitulation of human breast tumor ortho-topic xenografts is largely determined by the presence ofhuman tumor-derived stromal fibroblasts. Accordingly,studies in human tissues showed that tumor-associatedfibroblasts (TAFs) isolated from breast carcinomas pro-moted the growth of breast cancer cells through theproduction of soluble factors such as colony-stimulatingfactor 1.22 In line with these findings, gene expressiontudies from the tumor microenvironment of human
reast carcinomas reported up-regulation of several fac-
ors, including chemokines with protumorigenic func-ion.23
The finding that the degree of activation of the stromaaffects tumor growth and progression led to the conceptof “stromal staging,” which has potential clinical utility.For example, increased production of ECM proteins in-cluding fibronectin, collagen IV, and tenascin C is associ-ated with poor prognosis in patients with small cell lungcarcinoma.24 This is related to the activation of prosur-vival and antiapoptotic pathways in neoplastic cells fol-lowing their adhesion to components of the ECM, forexample, by the stimulation of the PI3K/Akt pathwaydownstream of �1 integrins.25 Furthermore, the abun-
ance of specific stromal cells correlates with patientutcome. In breast cancer, the density of tumor-associ-ted macrophages (TAMs) is associated with poor survivalnd reduced response to chemotherapy.26 Furthermore,ecent findings indicate that the detection of p53 muta-ions in the stromal component increases the likelihoodf nodal metastasis, suggesting that mutation-bearingtromal cells can provide a favorable setting for tumorpread.27 Interestingly, the stroma may also mediate resis-
tance to molecular therapies by secreting growth factors(eg, hepatocyte growth factor [HGF]) that can stimulatesurvival responses and prevent apoptosis. This indicatesthat the tumor microenvironment actively favors the se-lection and expansion of cellular clones that are morelikely to survive and adapt to the changes induced bystromal cells.28
Indeed, chemotherapy induces the production of colo-ny-stimulating factor 1, a chemoattractant for macro-phages, which exacerbates tumor progression by promot-ing angiogenesis, invasion, and metastasis of neoplasticcells.29 In one study, the pharmacologic blockade of mac-rophage recruitment markedly improved the ability of thechemotherapeutic agent paclitaxel to slow the growth ofboth primary and metastatic tumors.26 Finally, a stromal
ene signature predicts resistance to neoadjuvant chemo-herapy in patients with estrogen receptor–negative breastancer.30
Biological Processes Involved in theTumor MicroenvironmentThe precancerous milieu of chronic liver disease is
characterized by neoangiogenesis, including several vascu-lar abnormalities such as arterialization and sinusoidalcapillarization, as well as inflammation and fibrosis. Thesebiological processes become more pronounced with pro-gression of liver failure, in which the incidence of cancerincreases exponentially (Figure 2). Synchronous eventsoccurring in this setting also include hypoxia, oxidativestress, and autophagy.
AngiogenesisAngiogenesis plays an important role in hepato-
carcinogenesis from its early stages.31 HCC is a highly
vascularized tumor; indeed, pathological angiogenesis is
csdapn
REV
IEW
SA
ND
PER
SPEC
TIV
ES
March 2013 ROLE OF THE MICROENVIRONMENT IN HCC 515
one of the main contributors to chronic liver diseases. Thehepatic wound-healing response due to chronic liver in-jury leads to fibrogenesis, a process that entails secretionof several proangiogenic factors by the stromal cells, es-pecially matrix metalloproteinases (MMPs), platelet-de-rived growth factor (PDGF), transforming growth factor(TGF)-�1, fibroblast growth factor (FGF), and vascularendothelial growth factor (VEGF). Moreover, ECM depo-sition and anatomic alterations during the fibrogenic pro-cess provoke resistance to blood flow that reduces meta-bolic exchange of oxygen, favoring hypoxia.
Indeed, gene expression of VEGF, the most criticalproangiogenic factor, is already induced in dysplastic nod-ules and further increases according to the progression ofHCC development.31 Once the tumor is established, thesurvival of neoplastic cells requires the formation of a newvascular network to provide nutrients and oxygen. Theangiogenic process in HCC is complex and tightly regu-lated, resulting from the balance between multiple angio-genic and antiangiogenic factors from the tumor and thehost cells. Growth of the tumor mass creates a nutrient-and oxygen-deprived environment, which induces the ac-tivation and proliferation of endothelial cells (ECs) tosprout new vessels from preexisting ones.32 ECs becomeproliferative and liberate enzymes to disrupt the basementmembrane, and they eventually migrate to their finallocation where they assemble to form a new vessel to-gether with the ECM.33
The expression of VEGF correlates with the aggressive-
Figure 2. Pathological featuresthat may be present in hepato-cellular carcinoma (HCC). (A)Poorly-differentiated HCC. Tu-mor cells have marked pleomor-phic nuclei and an inflammatoryinfiltrate consisting of neutro-phils. Ballooning degenerationand production of Mallory’s hya-lines are also noted. H&E originalmagnification 400X. (B) Poorly-differentiated HCC with tumorcells arranged in a solid pattern.A focus of lymphocytic inflam-matory infiltrate is present. H&E,original magnification 100X. (C)Well-differentiated multinodularHCC with dense fibrosis forminga wide septum that separate twoHCC nodules. H&E, originalmagnification 40X. (D) Increasedvascularization in HCC. Vesselsare highlighted by CD34 immu-nostaining. These vessels arenourishing the tumor. Originalmagnification 100X. Imagescourtesy of Dr M. Isabel Fiel,Mount Sinai School of Medicine.
ness of HCC.31 The effects of VEGF are transduced fol-
lowing binding to its receptors, VEGFR1 and VEGFR2, toactivate several signaling pathways involved in prolifera-tion, migration, and invasion of ECs.34 In addition, VEGFan function as a cytokine that directly affects hepatictellate cells, Kupffer cells, and hepatocytes35,36 and me-iates the dissolution of the vascular basement membranend the interstitial matrix.31 VEGF and angiopoietin 2lasma levels have been identified as independent prog-ostic biomarkers in patients with advanced HCC.37 An-
giopoietin 2 is frequently up-regulated in HCC and booststhe effect of VEGF on ECs.38 Moreover, the Tie-2 receptoris expressed by both ECs and stellate cells, further empha-sizing the complex orchestration of angiogenic regulationin liver tumors.
FGF is a member of the heparin-binding growth factorsand acts synergistically with VEGF to induce angiogene-sis, whereas PDGF is involved in cell migration and newvessel maturation. Cancer cells secrete PDGF, which actsthrough a paracrine mechanism that involves other celltypes such as ECs and fibroblasts and correlates withcancer progression.39 Other significant mediators in tu-mor neoangiogenesis are integrins and cadherins, whichmediate cell-matrix and cell-cell interactions, respectively,to establish contacts required for new vascular tube for-mation.33
InflammationHCCs arises in a diseased liver with a dynamic
inflammatory environment that predisposes to initiation
of cancer. Inflammation, an essential part of the wound-
rheaia
REV
IEWS
AN
DPER
SPEC
TIVES
516 HERNANDEZ–GEA ET AL GASTROENTEROLOGY Vol. 144, No. 3
healing response of the liver and undoubtedly beneficialin the short-term, perpetuates chronic injury. Chronicinflammation drives a maladaptive reparative reactionand stimulates liver cell death and regeneration, which iseventually associated with the development of dysplasticnodules and cancer (Figure 3).
Several inflammatory mediators have been implicatedin sustained inflammation and immunosuppression asso-ciated with development of HCC. Carcinogenesis is asso-ciated with persistent cytokine production that can stim-ulate many liver cell types with a variety of unique as wellas redundant interactions. Altered cytokine profiles havebeen described in HCC not only in tumor cells but also in
the surrounding tissue; however, the full portrait of theirmechanistic role remains unclear. A predominant role ofthe Th2-like (interleukin [IL]-4, IL-8, IL-10, and IL-5)cytokine compared with Th1-like (IL-1�, IL-1�, IL-2, tu-mor necrosis factor �) cytokine in the microenvironmenthas been associated with a more aggressive and metastaticHCC phenotype.40,41 IL-6 is an abundant cytokine in cir-hotic livers, produced by Kupffer cells in response toepatocyte damage, and a potent activator of STAT3, andlevated levels in serum are associated with risk of HCCnd poor prognosis.10,42 Moreover, modulation of thenflammatory microenvironment by suppression of HGFnd IL-6 production by estrogens represses metastasis of
Figure 3. Anatomic and cellularalterations leading to the devel-opment of HCC. (A) Normal liverparenchyma. Hepatocytes withmicrovilli and sinusoidal endo-thelial cells whose fenestrationsfavor metabolic exchange. Spaceof Disse with few quiescent stel-late cells containing lipid drop-lets. (B) Fibrotic liver. Uponchronic liver injury, hepatocyteslose their microvilli, sinusoid en-dothelial cells become defenes-trated, and stellate cells are acti-vated, losing lipid droplets andsecreting ECM. (C) HCC. Malig-nant transformation of hepato-cytes with uncontrolled growth.Infiltration of inflammatory cellsand cytokines with extensive fi-brosis and recruitment of TAFsand CSCs. (D) Development ofnew vessels (neoangiogenesis)
and distant metastases.
Hbae
trpcl
pp
Sirrdpa
(
clt
irw
hs
REV
IEW
SA
ND
PER
SPEC
TIV
ES
March 2013 ROLE OF THE MICROENVIRONMENT IN HCC 517
HCC.43 High levels of IL-22 have been detected in theCC microenvironment, leading to tumor growth, inhi-
ition of apoptosis, and promotion of metastasis due toctivation of STAT3.44 Up-regulation of IL-10 is also pres-nt in HCC tumors45,46 as well as in their microenviron-
ment38 and confers a high risk of progression after resec-ion. However, its specific risk in development of HCCemains unknown. Higher levels of IL-2 and IL-15 ineritumoral liver tissue are also associated with a de-reased rate of intrahepatic tumor recurrence and pro-onged overall survival.47
Chemokines (eg, CXCL12, CX3CL1, CCL20) are cyto-kine-like molecules with chemotactic properties criticalto cell trafficking into and out of the tumor microen-vironment. They orchestrate the inflammatory responsethrough their binding to 4 families of receptors (CCR,CXCR, CX3CR, and XCR) found mainly in inflammatory,endothelial, and epithelial cells. Chemokines have been im-plicated in many key steps of cancer development, includingevasion of the immune system, angiogenesis, invasion, anddissemination.48 The CXCL12-CXCR4 axis is especially im-
ortant in regulation of angiogenesis and is highly ex-ressed in HCC compared with cirrhosis.49 CXCL12 binds
CXCR4 in endothelial cells and promotes migration, pro-liferation, and development of new vessels, acting syner-gistically with VEGF.50 It has also been implicated ingrowth, invasion, and metastasis of HCC.51,52
Another important axis in regulation of HCC is CCL20-CCR6, which mediates the recruitment of circulating reg-ulatory T cells (Tregs) into the tumor microenvironment.Its up-regulation is associated with promotion of tumorgrowth, a low level of differentiation, and the presence ofintrahepatic metastasis.53 Nuclear factor �B (NF-�B) and
TAT3 are signaling pathways involved in the hepaticnflammatory response to injury that is critical for liveregeneration with overlapping target genes. NF-�B plays aole in hepatocarcinogenesis; however, its function variesepending on the mouse model and type of injury ap-lied. In humans, proinflammatory stimuli such as hep-titis viruses54 and free fatty acids55 activate NF-�B, which
might initiate and promote HCC in the inflamed liver.STAT3 remains inactive in nonstimulated cells and be-comes rapidly activated through phosphorylation by cy-tokines and growth factors produced within the tumormicroenvironment. Active STAT3 has been detected inHCC specimens and is associated with a more aggressivephenotype and poor prognosis.56
The gut microbiota also plays a role in the pathogenesisof HCC. Chronic liver disease is often associated withtranslocation of the intestinal bacteria, and gut-derivedlipopolysaccharide via Toll-like receptor 4 can amplify thetumorigenic response of the liver to promote HCC.57–59
Several growth factors regulate the immune and in-flammatory response in the HCC microenvironment, inparticular TGF-�, HGF, and epidermal growth factorEGF). TGF-�, a tumor suppressor in normal and prema-
lignant cells, acts as an oncogenic growth factor in cancer
cells.60 It is expressed mostly in stromal cells rather than
malignant epithelial cells and is markedly increased inHCC.61 Reduced expression of TGF-� receptor II has beenorrelated with poor prognosis in HCC, as defined byarger tumor size, poor differentiation, intrahepatic me-astasis, and shorter recurrence-free survival.61 FGF and
HGF control proliferation and invasion of HCC cells.62,63
Overexpression of the HGF receptor c-Met has been de-tected in several human tumors64 including HCC, where its associated with a poor outcome. Similarly, a c-Met–egulated expression signature defines a subgroup of HCCith a poor prognosis and aggressive phenotype.65 EGF
receptor also plays an important role in tumor progres-sion and tumor-associated angiogenesis via regulation ofseveral angiogenic factors, with a direct effect on tumorand endothelial cells.66
FibrosisThe ECM is essential to support the architecture of
the liver and constantly interacts with the environment,allowing signal transduction and changes in gene expres-sion.67 In disease, the activity of the ECM remodelingenzymes is deregulated, leading to a fibrotic microenvi-ronment characterized by increased stiffness and abun-dance of growth factors that contribute to tumorigene-sis.67 An excess of ECM production together with areduced ECM turnover characterizes liver fibrosis. Dereg-ulation of collagen cross-linking and ECM stiffness playsa causative role in the pathogenesis of cancer by enhanc-ing integrin signaling.68 This situation leads to an exces-sive deposition of fibrillar collagen types I and II andfibronectin in the liver. There is also enhanced growth,survival, and proliferation of tumoral cells through regu-lation of the integrin family. Integrins �1�1 and �2�1
ave also been implicated in progression and cell inva-ion,69 because their inhibition reduces migration of liver
cancer cells induced by several growth factors (TGF-�1,EGF, or basic FGF).
Deregulation of ECM homeostasis directly affects epi-thelial cells and leads to cellular transformation and me-tastasis. Tumor growth requires the breakdown of preex-isting boundaries and rearrangement of liver tissue, aprocess mainly regulated by MMPs and tissue inhibitor ofmetalloproteinases (TIMPs). Overexpression of MMPs cancompromise the basement membrane barrier and facili-tate tissue invasion by cancer cells. HCC is associated withhigher proteolytic activity and high MMP2 levels. More-over, an imbalance between MMP2 and TIMP2 correlateswith the occurrence of metastasis, leading to a poor out-come.70 Linear and thick collagen fibers are often found inareas with active tissue invasion71 and vascularization,72
and several studies have shown that tumor cells migrateon collagen fibers.72
The ECM is also essential for tumor angiogenesis. Toinitiate vascular branching, the basement membrane mustbe removed mainly by MMPs. The ECM is also involved invessel lumen formation, tubulogenesis, and deposition ofa supportive basement membrane. Notably, tumor new
vasculature is more porous and leaky than normal,73,74
uH
algfocttmpn(
rtec
a
cmtp
piwI
u
REV
IEWS
AN
DPER
SPEC
TIVES
518 HERNANDEZ–GEA ET AL GASTROENTEROLOGY Vol. 144, No. 3
facilitating immune cell infiltration, metastasis, and tu-mor progression. The ECM also modulates activation ofimmune cells and can regulate T-cell activation75 andimmune cell differentiation, for example, by impairing thenormal maturation of T-helper cells.67,76
ECM stiffness also plays an important role in develop-ment of HCC. Lysyl oxidase 2, an enzyme able to modifyECM stiffness via promoting cross-linking of fibrillar col-lagen 1, is involved in the creation of a pathologic stromaable to promote tumor growth and metastasis.77
Other Biological ProcessesHypoxia. Although HCC is a highly vascularized
tumor, neoplastic vessels are functionally abnormal andareas of hypoxia are common. Reduced oxygen availabilityinduces expression of hypoxia-inducible factor 1 (HIF-1),a major transcription factor that regulates the expressionof several genes with critical roles in angiogenesis, im-mune evasion, invasion, and metastasis.78 Hypoxia stim-
lates growth and blocks apoptosis of HCC, and levels ofIF-1 correlate with a worse prognosis.79
Oxidative stress. Cancer cells, besides generatingoxidative stress intrinsically, are also exposed to a pro-oxidant environment generated by several stromal com-ponents.80 Overproduction of ROS provokes nitrosative
nd oxidative stress through interaction with DNA, RNA,ipid, and proteins, leading to an increase in mutations,enomic instability, epigenetic changes, and protein dys-unction. Fibroblast activation is profoundly affected byxidative stress and produces several mediators impli-ated in tumor progression. TAMs can generate ROS dueo activation of NOX2 and inducible nitric oxide syn-hase, which promote tumor progression, invasion, and
etastasis. Moreover, tumoral conditions such as hypoxiaroduce oxidant species that promote DNA mutations ineoplastic cells. Recently, mutations in specific genes
RPS6KA3-AXIN1 and NFE2L2-CTNNB1) that alter Wnt/�-catenin signaling have been associated with oxidativestress and metabolism that cooperate in liver carcinogen-esis.81 Moreover, altered oxidative stress pathways in non-cancerous human liver tissue can predict recurrence ofHCC. High levels of ROS promote invasiveness of hepatictumor cells82 and contribute to tumor invasion via pro-duction of MMP.
Autophagy. Autophagy, a catabolic process up-egulated under metabolic stress conditions, is induced inhe tumor microenvironment. Stromal components arexposed to oxidative stress conditions induced by cancerells that together with hypoxia induce autophagy.83 Au-
tophagy in the tumor stroma acts as a prosurvival mech-anism that generates energy able to fuel cancer cells,alleviating the metabolic imbalance and promoting sur-vival.84 It has been postulated as one of the escape mech-
nisms for cancer cells during antiangiogenic treatment.85
Although the role of autophagy in the setting of HCC isstill under development, defects in autophagic genes(BECN1, ATG7) have been described in HCC cells.86,87
Autophagy modulation has been identified as a promising
therapeutic strategy in combination with molecular tar-geted therapy.88
Cellular Components of the HCCMicroenvironmentHCC usually arises in a severely perturbed mi-
croenvironment that hastens dysfunction of epithelialcells and malignant transformation. Targeting the com-ponents of the microenvironment therefore emerges as arational preventive strategy (Figure 1). Here we describethe main cellular components in the microenvironmentand identify potential molecular targets for therapies.
Immune CellsHCC is rich in immune cells. Tumor-infiltrating
lymphocytes are the primary immune component in solidtumors and comprise a host antitumor reaction.89 Mosttumor-infiltrating lymphocytes are CD4� (helper or Tregells). Treg cells have a detrimental effect on the develop-ent of cancer, because they promote immune tolerance
o neoplastic cells. Treg cells usually infiltrate HCC, and aredominance of Treg cells over TCD8� cells is associated
with a worse prognosis90; additionally, Treg levels havebeen correlated with HCC stages.91 Myeloid-derived sup-
ressor cells also play a role in T-cell regulation andnduction, favoring a suppressive immune responseithin the microenvironment.92 Increased secretion of
L-17 by CD4� lymphocytes in HCC also correlates withincreased postoperative recurrence following resection.93
Myeloid-derived suppressor cells and Treg cells are bothimportant in the establishment and promotion of im-mune suppression. Dendritic cells are decreased and dys-functional in patients with HCC and contribute to theinsufficient immune antitumoral response. Dendritic cellvaccination has even been proposed as an antitumor ther-apy in HCC but would first require a better understandingof the hepatic microenvironment for its full develop-ment.94
Fibroblasts and MacrophagesTAFs are the major source of collagen in the
HCC stroma; however, their origin is still a matter ofdebate. They differ from normal fibroblasts in theirability to secrete high levels of stromal cell– derivedfactor 1 and CXCL12 and promote tumor growth andangiogenesis.95 There is a complex cross talk betweenTAFs and tumor cells. For instance, both can secretePDGF and TGF-�, which leads to stellate cell activationand consequently ECM deposition, but they also en-hance growth and migration of cancer cells.96 TAFs alsointeract with the microvasculature by secreting VEGFand MMPs as well as several hepatocyte proliferationfactors such as HGF.97 TAFs also secrete immune-mod-
latory cytokines (IFN-�, IL-6, and tumor necrosis fac-tor) that can mobilize cytotoxic T lymphocytes, naturalkiller cells, and TAMs.98 TAMs, the most abundant cell
component, represent a subset of myeloid CD11b� tu-
tcEts
ea
ptti
amt
oe
p
vt
iDin
t
H
caemeb
REV
IEW
SA
ND
PER
SPEC
TIV
ES
March 2013 ROLE OF THE MICROENVIRONMENT IN HCC 519
mor-infiltrating cells characterized by the expression ofthe Tie-2 angiopoietin receptor.99 TAMs suppress anti-umor immunity in HCC, and their density has beenorrelated with a poor prognosis.100 TAMs also releaseGF, chemokines, MMP, and VEGF, which regulate
umor growth, ECM remodeling, angiogenesis, inva-ion, and metastasis.101 Recently, it has been shown
that c-Myc controls the activation102 of TAMs, as wellas impairing VEGF signaling and infiltrating inflamma-tory cells.103 There is a bidirectional cross talk, andnvironmental conditions such as hypoxia can alsoffect Myc signaling.104
Cancer Stem CellsCancer stem cells (CSCs), defined by their self-
renewing and differentiation capacity, have been proposedas the clonogenic core of several tumors. In HCC, liverCSCs can be isolated based on their expression of severalcell markers (EpCAM, CD133, CD90, CD44, CD24, CD13,and OV6).105 Additionally, signaling pathways identifiedin liver cancer are active in isolated liver CSCs (eg, Wnt,Notch, TGF-�, Hedgehog, and PI3K/AKT/mTOR), sup-porting the idea that CSCs contribute to the molecularheterogeneity of HCC.106,107
Although the clinical relevance of CSCs remains elusive,there is growing evidence supporting a role in initiatingand sustaining primary tumors and facilitating metasta-sis.108 –111 Recent data support a strong association be-tween a hepatic progenitor cell origin of the tumor andprognosis in HCC.19 One therapeutic approach may be totarget not only the signaling pathways involved in stemcell fate (self-renewal and multilineage differentiation po-tential), reproduction, and proliferation (Notch, Wnt,Hedghog)112 but also the stem cell niche.113 This ap-
roach may undermine CSC self-renewal and reproduc-ion. However, they are a complicated target because ofheir chemoresistance and radioresistance and their abil-ty to stimulate angiogenesis.114
Animal Models to Study the TumorMicroenvironmentLiver carcinogenesis is a multistep process with
several cellular and mechanical deregulations that even-tually lead to malignant transformation of hepatocytes.Numerous mouse models successfully produce HCC;however, not all of them mimic the pathogenic se-quence of human HCC that starts with fibrosis, cirrho-sis, angiogenesis, and preneoplastic nodules beforeHCC develops. There are 4 main categories of murineHCC models: chemically induced, oncogene driven,xenograft, and genetically modified. Chemically in-duced models (N-nitrosodiethylamine) are among themost commonly used in HCC research.115,116 When
ssociated with CCl4, the N-nitrosodiethylamine modelimics the sequence of injury/fibrosis/malignant
ransformation that occurs in humans.57 Conditional
overexpression of the oncogenic protein Myc in which
the expression of human Myc can be regulated in mu-rine liver will induce HCC, whereas Myc inactivationresults in tumor regression.117 Additionally, depletion
f AMPK-related kinase 5 in mice with deregulatedxpression of MYC and HCC prolongs survival.118
In xenograft models, human cancer cells are injectedinto immune-deficient mice. Orthotopic implantationof tumor cells in the liver is preferable to subcutaneousxenograft models, because it better replicates the tumormicroenvironment.119 To improve the reproducibilityfurther, tumor cells can be injected after fibrosis isestablished by either CCl4 or thiocetamide injection.
Genetically modified mice are engineered to mimicathophysiological and molecular features of HCC.120
There are a huge number of genetically modified mice(overexpression of myc, �-catenin and HRAS, TGF-�, de-ficiency of PTEN, among others), which have been re-viewed elsewhere.119 –122 Several animal models can repro-duce the human stepwise development of HCC. PDGF-Ctransgenic mice develop steatosis and activation of stellatecells that progresses into bridging fibrosis, angiogenesis,and tumorigenesis.123
Multidrug resistance gene 2 (Mdr2) knockout mice area well-established model of inflammation-associatedHCC. These mice lack a liver-specific P-glycoprotein andmimic human intrahepatic cholestasis very well.124,125 De-elopment of HCC is also preceded by chronic inflamma-ion in mice overexpressing lymphotoxins � and � in
hepatocytes.126
To date, the exact role of NF-�B signaling in hepato-carcinogenesis is not totally understood and may dependon the mouse model and injury used. Several componentsof the NF-�B canonical pathway have been manipulatedn a range of models, often yielding conflicting results.
eletion in hepatocytes of either NEMO, a regulatorynhibitor of the iKK pathway, or TAK1 leads to sponta-eous steatohepatitis and HCC.127 Conditional liver-spe-
cific deletion of IKK2 increases formation of liver tumorsin N-nitrosodiethylamine–treated mice,116 whereas inhibi-ion of the NF-�B signal in Mdr2-KO mice128 and trans-
genic overexpression in hepatocytes of lymphotoxins �and �130 resulted in chronic hepatitis at 9 months and
CC at 12 months.High-throughput technology has made possible the
haracterization of tumors at the gene expression levelnd has revolutionized our understanding of HCC. Genexpression signatures obtained from experimental ani-als and hepatic cells can be integrated into the gene
xpression patterns for human HCC, thus identifying theest-fit mouse models to study human cancer.123
Prognostic Relevance of the NontumorAdjacent Tissue in Patients With HCCGenomic studies have shown the relevance of the
tumor microenvironment in predicting outcome in pa-tients with HCC (Table 1). A 36-gene signature originat-ing from the surrounding non-neoplastic liver tumor was
reported to predict multicentric occurrence or late recur-
aien(orawa
amp
stopsf
tofost
w
iit
a V a
REV
IEWS
AN
DPER
SPEC
TIVES
520 HERNANDEZ–GEA ET AL GASTROENTEROLOGY Vol. 144, No. 3
rence in patients with HCV-related HCC.129 We identifiedpoor prognosis signature driven by late recurrence orig-
nating from the adjacent cirrhotic tissue in patients witharly HCCs using 2 different patient cohorts.10 The sig-ature reflected the presence of a protumorigenic milieu
“field effect”) with promoting effects on the developmentf metachronous tumors independent from the primaryesected HCC. Interestingly, the same gene signature wasble to predict the development of HCC in 216 patientsith HCV-associated cirrhosis who were followed up forbout 10 years.130 This signature included genes involved
in inflammation (IL-6, NF-�B signaling) and EGF. Therole of the EGF pathway is certainly important in themolecular pathogenesis of HCC due to its known onco-genic activity and the availability of molecular inhibitorstargeting this cascade. Recently, the presence of a specificpolymorphism within the EGF gene (EGF 61*G) wascorrelated with a high risk of developing HCC in patientswith chronic hepatitis C, advanced fibrosis, and cirrho-sis.131,132 The presence of this polymorphism is linked to
prolonged half-life of EGF messenger RNA, which pro-otes sustained EGF signaling in the damaged preneo-
lastic tissue, thereby promoting hepatocarcinogenesis.In addition to gene expression studies, a signature con-
tituted by 19 microRNAs derived from the adjacent non-umoral tissue of patients with HCC with different etiol-gies was proposed to accurately identify patients with aoor prognosis.133 Similar to other studies based on tran-criptomic data,10 microRNA profiling from the tumorailed to predict patient outcomes, however.
Interaction between tumor cells and stromal and endo-helial cells can also have a profound effect on the capabilityf tumor cells to migrate and invade the ECM and the newlyormed vascular vases, thereby promoting the developmentf metastasis. Expression profiles from livers bearing meta-tatic HCC were different from livers without metastaticumors.38 These investigators generated a 17-gene signature
predictive of metastasis development using surroundingnontumoral tissue from HBV-related HCC samples. The
Table 1. Molecular and Cellular Markers from the Tumor Micr
Includes etiologies not related to alcohol, hemochromatosis, and HB
signature was enriched in Th2-dominant cytokines and dif- a
fered considerably from the signature of the primary tumor.Importantly, both the poor prognosis signature from theadjacent tissue and the metastatis-related signature wereindependent of the global inflammation status of the liver,suggesting that specific changes within the microenviron-ment affected progression of HCC.10,41 Furthermore, bothsignatures highlighted the involvement of inflammatory andimmune components in the pathogenesis of this disease.Conversely, the presence of a 14-immune gene signatureincluding CXCL10, CCL5, and CCL2, which attract CD8� Tand natural killer cells, was associated with better prognosisin patients with early HCC.134 Importantly, CD8� T andnatural killer cells display antitumor activity, reflected in theenhanced activated caspase-3 expression in cancer cells.
The cross talk between tumor cells and stroma is mutual.Indeed, HCC cells might promote the recruitment and acti-vation of immune cells to the tumor niche. Oncogenic�-catenin signaling was found to promote an inflammatoryprogram in hepatocytes that involved direct transcriptionalcontrol by �-catenin and activation of the NF-�B pathway,which exacerbated the aggressiveness and metastasis ofHCC.135 In addition, HCC cells can produce IL-8, levels of
hich have been associated with poor survival,136 throughactivation of p38 MAPK, ERK, and PI3K/Akt signaling path-ways.137
Targeting the Tumor Stroma: APromising Challenge for New TherapiesIn recent years, the tumor stroma has emerged as a
critical target for therapy in patients with preneoplasticconditions or established HCC (Figure 4). Modulators ofdifferent biological processes, including inflammation, fi-brosis, angiogenesis, and signals of proliferation and sur-vival, might be effective in the prevention and primarytreatment of early HCC. Due to the implications of in-flammatory pathways (eg, IL-6, NF-�B) and EGF signalingn cirrhotic patients at high risk for developing HCC andn subjects with early HCC, strategies interfering withhese networks might be effective in chemoprevention
nd primary treatment. The protumorigenic role of the
c
nbyhptadc
awtecpdt
REV
IEW
SA
ND
PER
SPEC
TIV
ES
March 2013 ROLE OF THE MICROENVIRONMENT IN HCC 521
EGF pathway at preneoplastic stages is further strength-ened by the evidence that cirrhotic subjects with the 61*Gpolymorphism within the EGF gene have an increased riskof developing HCC compared with other cirrhotic pa-tients. Therefore, inhibitors of the EGF pathway such astyrosine kinase inhibitors of the intracellular domain ofEGF receptor (eg, erlotinib, gefitinib) are promisingagents to explore in the chemoprevention mode. One ofthe key problems with chemoprevention studies is thatthe targeted population is so broad that the studies wouldrequire thousands of patients and long-term follow-up toshow any clinical benefit. These issues can be overcome bytargeting patients at high risk for development of HCC,for instance by selecting patients presenting with the poorprognosis signature, which is present in �20% ofases,10,117 or in those harboring the G/G phenotype
within the EGF gene. Certainly, all these approaches haveot yet reached early clinical studies and thus are far fromeing tested in pivotal trials for regulatory approval. Be-ond inhibitors of the EGF cascade, preclinical studiesave shown that sorafenib might be effective as a chemo-reventive agent. Sorafenib reduced liver fibrosis in ratsreated with thioacetamide and decreased portal pressure,s well as yielding a remarkable improvement in liveramage, intrahepatic inflammation, and angiogenesis of
Figure 4. Schematic represen-tation of therapeutic opportunitiesand application of prognostic bio-markers in the management ofpatients with HCC and preneo-plastic conditions.
irrhotic rats.138
Because the tumor microenvironment plays a pivotalrole in the natural history of HCC, there is a strongrationale for modulating the dynamic cross talk betweenthe tumor and the stroma as primary treatment of thisdisease. An important advantage of altering the tumormicroenvironment is underscored by the fact that thetarget cells are genetically stable and therefore less likelyto develop resistance.139 Because angiogenesis is a hall-mark of HCC, therapies blocking the growth of newvessels or normalizing the tumor vasculature system rep-resent key strategies to block tumor dissemination. In thissetting, sorafenib simultaneously acts on the tumor vas-culature (by targeting VEGFR2, VEGFR3, and PDGFR-�)
nd tumor cells (by inhibiting the Ras/MEK/ERK path-ay), thereby blocking angiogenesis and tumor prolifera-
ion.140 In particular, antiangiogenic agents might be ben-ficial in patients subjected to transcatheter arterialhemoembolization; high levels of VEGF have been re-orted after this procedure due to the high hypoxic con-itions induced by the interruption of blood flow into theumor.141 Several antiangiogenic agents are currently un-
der investigation in phase 2/3 clinical trials with patientswith HCC (Table 2). Most of them are small moleculeinhibitors targeting molecular mediators of angiogenesisand growth factor receptors (eg, VEGF receptor, PDGF
receptor, FGF receptor). Others are specific monoclonal
gibtsp
rmbwhc
atrapti
o
cor
REV
IEWS
AN
DPER
SPEC
TIVES
522 HERNANDEZ–GEA ET AL GASTROENTEROLOGY Vol. 144, No. 3
antibodies, such as bevacizumab, which targets VEGF andhas been approved by the Food and Drug Administrationfor the treatment of several cancers, including metastaticcolon cancer, and ramucirumab, a monoclonal antibodyagainst VEGFR2.142 Despite the initial excitement aboutthe use of antiangiogenic therapies for the treatment ofpatients with HCC, several concerns about their adverseeffects (eg, gastrointestinal bleeding, thromboembolicevents, hypertension) have emerged. A phase 3 trial eval-uating the efficacy of sunitinib compared with sorafenibhas been prematurely halted due to severe adverse eventsand futility related to the administration of sunitinib.143
Furthermore, the acquisition of resistance to antiangio-genic agents through activation of alternative pathwaysrepresents a threat that undermines the clinical manage-ment of patients with HCC.31 Alternative strategies tar-
eting the tumor microenvironment are currently undernvestigation. A considerable number of clinical trialsased on immunotherapy have been performed in pa-ients with HCC. Nevertheless, the conclusions of thesetudies have been unsatisfactory and there are not enough
Table 2. Molecular Therapies Assessed or Under Investigatio
Inflammation/Immune system OPB-31121 STAT3Licartin HAb18G/
Invasion/metastasis PI-88 Endo-�-glu
ositive clinical data supporting their efficacy in HCC.
This lack of efficacy can be partly explained by theedundancy of the immune components in the tumor
icroenvironment. Nevertheless, new strategies need toe designed given the importance of inflammatory path-ays and immune regulation to HCC. Early positive dataave been reported with the STAT3 inhibitor in preclini-al models with altered TGF-� signaling.
Similarly, a monoclonal antibody designed to boostantitumor immune response by binding and stimulatingT lymphocytes (anti-CTLA4, tremelimumab) is under in-vestigation in patients with HCV-related HCC (Clinical-Trials.gov identifier: NCT01008358).144 These and other
pproaches targeting the HCC microenvironment will beested in advanced clinical stages in the near future. Cur-ently, a few agents modulating inflammatory pathwaysre under clinical evaluation in HCC. Among these, ahase 1/2 evaluation of OPB-31121, an orally adminis-ered STAT3 inhibitor, in patients with progressive HCCs ongoing.
Modulators of signaling that control ECM remodelingr inhibitors of metastasis (eg, TGF-�, HGF/c-Met,
March 2013 ROLE OF THE MICROENVIRONMENT IN HCC 523
targeting the tumor microenvironment. Nevertheless,clinical trials with MMP inhibitors have shown no efficacyin patients with advanced stages of cancer and producedsome intolerable side effects.145 Of note, the TGF-� inhib-tor LY2109761 showed promising preclinical results in aenograft model of HCC.146
Recently, a randomized controlled phase 2 trial in pa-tients with unresectable HCC who did not respond orwere intolerant to first-line therapy showed that tivantinib(ARQ197), a specific inhibitor of Met, increases overallsurvival and time to progression of patients whose tumorsexpressed high levels of Met.147 Furthermore, cabozan-tinib (XL184), a dual c-Met/VEGFR2 inhibitor, has shownearly evidence of antitumor activity in a randomized dis-continuation phase 2 study.148
Finally, a successful strategy might require a combina-tion of therapies targeting both the microenvironmentand the tumor itself. In this context, depletion of TAMusing zoledronic acid significantly improved response tosorafenib in a xenograft model of HCC.149
Conclusion and Future ProspectsHCC commonly arises in a damaged organ featured
by extensive inflammation and fibrosis. Different players,including immune cells, hepatic stellate cells, and macro-phages, react to liver injury by producing cytokines andcomponents of the ECM, which promote angiogenesis, andsurvival of damaged hepatocytes or cancer stem cells. Thisregenerative response favors the accumulation of mutationsand epigenetic aberrations, which lead to malignant trans-formation of preneoplastic nodules. The interaction betweenstromal and tumor cells is dynamic and dramatically altersthe behavior and aggressiveness of HCC, particularly at earlystages of disease. Recent studies have highlighted the role ofEGF and inflammatory pathways in the developmentof HCC in cirrhotic patients as well as in the likelihood ofrecurrence in patients with early HCC undergoing surgicalresection. These findings point out new targets for chemo-prevention and primary treatment. Although several antian-giogenic and antiproliferative agents are currently underinvestigation in phase 2/3 clinical trials, there is still a sig-nificant lack of studies on modulators of the ECM compo-nents and inhibitors of inflammatory pathways. Indeed, con-sidering the pivotal implication of immune cells andsignaling in HCC, a therapeutic reprogramming of the im-mune microenvironment in tumors might represent a prom-ising strategy for improving the efficacy of standard antican-cer treatments (eg, sorafenib). These strategies should aim tobolster antitumor immunity, for example, by decreasing thenumber of Treg cells and reversing the imbalance of bothimmune/inflammatory cytokines and immune cells. In thiscontext, the development of animal models mimicking thenatural changes of the HCC microenvironment representsan unmet need for the preclinical evaluation of such com-binations.
Although there has been much progress in understanding
the alterations within the tumor microenvironment in HCC,
validated biomarkers of poor prognosis and response totherapy from the tumor stroma are still lacking. Nonethe-less, the recent advent of next-generation sequencing tech-nology represents a powerful and promising technology touncover novel alterations with potential clinical relevance forthe treatment of cirrhosis and early HCC.
References
1. Llovet JM, Burroughs A, Bruix J. Hepatocellular carcinoma. Lan-cet 2003;362:1907–1917.
2. Jemal A, Bray F, Center MM, et al. Global cancer statistics. CACancer J Clin 2011;61:69–90.
3. Villanueva A, Llovet JM. Targeted therapies for hepatocellularcarcinoma. Gastroenterology 2011;140:1410–1426.
4. Llovet JM, Ricci S, Mazzaferro V, et al. Sorafenib in advancedhepatocellular carcinoma. N Engl J Med 2008;359:378–390.
5. Hanahan D, Coussens LM. Accessories to the crime: functions ofcells recruited to the tumor microenvironment. Cancer Cell 2012;21:309–322.
6. Trimboli AJ, Cantemir-Stone CZ, Li F, et al. Pten in stromalfibroblasts suppresses mammary epithelial tumours. Nature2009;461:1084–1091.
7. Bhowmick NA, Chytil A, Plieth D, et al. TGF-beta signalling infibroblasts modulates the oncogenic potential of adjacent epithe-lia. Science 2004;303:848–851.
8. Llovet JM, Schwartz M, Mazzaferro V. Resection and liver trans-plantation for hepatocellular carcinoma. Semin Liver Dis 2005;25:181–200.
9. Sherman M. Recurrence of hepatocellular carcinoma. N EnglJ Med 2008;359:2045–2047.
10. Hoshida Y, Villanueva A, Kobayashi M, et al. Gene expression infixed tissues and outcome in hepatocellular carcinoma. N EnglJ Med 2008;359:1995–2004.
11. Tao Y, Ruan J, Yeh SH, et al. Rapid growth of a hepatocellularcarcinoma and the driving mutations revealed by cell-populationgenetic analysis of whole-genome data. Proc Natl Acad Sci U S A2011;108:12042–12047.
12. Chiang DY, Villanueva A, Hoshida Y, et al. Focal gains of VEGFAand molecular classification of hepatocellular carcinoma. CancerRes 2008;68:6779–6788.
13. Boyault S, Rickman DS, de Reynies A, et al. Transcriptomeclassification of HCC is related to gene alterations and to newtherapeutic targets. Hepatology 2007;45:42–52.
14. Hoshida Y, Nijman SM, Kobayashi M, et al. Integrative transcrip-tome analysis reveals common molecular subclasses of humanhepatocellular carcinoma. Cancer Res 2009;69:7385–7392.
15. Yamashita T, Forgues M, Wang W, et al. EpCAM and alpha-fetoprotein expression defines novel prognostic subtypes of hep-atocellular carcinoma. Cancer Res 2008;68:1451–1461.
16. Lee JS, Heo J, Libbrecht L, et al. A novel prognostic subtype ofhuman hepatocellular carcinoma derived from hepatic progenitorcells. Nat Med 2006;12:410–416.
17. Villanueva A, Hoshida Y, Toffanin S, et al. New strategies inhepatocellular carcinoma: genomic prognostic markers. ClinCancer Res 2010;16:4688–4694.
18. Seitz HK, Stickel F. Risk factors and mechanisms of hepatocar-cinogenesis with special emphasis on alcohol and oxidativestress. Biol Chem 2006;387:349–360.
19. Farrow B, Evers BM. Inflammation and the development of pan-creatic cancer. Surg Oncol 2002;10:153–169.
20. Lu P, Weaver VM, Werb Z. The extracellular matrix: a dynamicniche in cancer progression. J Cell Biol 2012;196:395–406.
21. Levental KR, Yu H, Kass L, et al. Matrix crosslinking forces tumorprogression by enhancing integrin signalling. Cell 2009;139:891–906.
22. Orimo A, Gupta PB, Sgroi DC, et al. Stromal fibroblasts present
in invasive human breast carcinomas promote tumor growth and
REV
IEWS
AN
DPER
SPEC
TIVES
524 HERNANDEZ–GEA ET AL GASTROENTEROLOGY Vol. 144, No. 3
angiogenesis through elevated SDF-1/CXCL12 secretion. Cell2005;121:335–348.
23. Allinen M, Beroukhim R, Cai L, et al. Molecular characterizationof the tumor microenvironment in breast cancer. Cancer Cell2004;6:17–32.
24. Sethi T, Rintoul RC, Moore SM, et al. Extracellular matrix proteinsprotect small cell lung cancer cells against apoptosis: a mecha-nism for small cell lung cancer growth and drug resistance invivo. Nat Med 1999;5:662–668.
25. Nicholson KM, Anderson NG. The protein kinase B/Akt signallingpathway in human malignancy. Cell Signal 2002;14:381–395.
26. Denardo DG, Brennan DJ, Rexhepaj E, et al. Leukocyte complex-ity predicts breast cancer survival and functionally regulatesresponse to chemotherapy. Cancer Discov 2011;1:54–67.
27. Patocs A, Zhang L, Xu Y, et al. Breast-cancer stromal cells withTP53 mutations and nodal metastases. N Engl J Med 2007;357:2543–2551
28. Yates LR, Campbell PJ. Evolution of the cancer genome. Nat RevGenet 2012;13:795–806.
29. Pollard JW. Trophic macrophages in development and disease.Nat Rev Immunol 2009;9:259–270.
30. Farmer P, Bonnefoi H, Anderle P, et al. A stroma-related genesignature predicts resistance to neoadjuvant chemotherapy inbreast cancer. Nat Med 2009;15:68–74.
31. Zhu AX, Duda DG, Sahani DV, et al. HCC and angiogenesis:possible targets and future directions. Nat Rev Clin Oncol 2011;8:292–301.
32. North S, Moenner M, Bikfalvi A. Recent developments in theregulation of the angiogenic switch by cellular stress factors intumors. Cancer Lett 2005;218:1–14.
33. Coulon S, Heindryckx F, Geerts A, et al. Angiogenesis in chronicliver disease and its complications. Liver Int 2011;31:146–162.
34. Liu Y, Poon RT, Li Q, et al. Both antiangiogenesis- and angiogen-esis-independent effects are responsible for hepatocellular car-cinoma growth arrest by tyrosine kinase inhibitor PTK787/ZK222584. Cancer Res 2005;65:3691–3699.
35. LeCouter J, Moritz DR, Li B, et al. Angiogenesis-independentendothelial protection of liver: role of VEGFR-1. Science 2003;299:890–893.
36. Lichtenberger BM, Tan PK, Niederleithner H, et al. AutocrineVEGF signalling synergizes with EGFR in tumor cells to promoteepithelial cancer development. Cell 2010;140:268–279.
37. Llovet JM, Pena CE, Lathia CD, et al. Plasma biomarkers aspredictors of outcome in patients with advanced hepatocellularcarcinoma. Clin Cancer Res 2012;18:2290–2300.
38. Yoshiji H, Kuriyama S, Noguchi R, et al. Angiopoietin 2 displaysa vascular endothelial growth factor dependent synergistic effectin hepatocellular carcinoma development in mice. Gut 2005;54:1768–1775.
39. Bronzert DA, Pantazis P, Antoniades HN, et al. Synthesis andsecretion of platelet-derived growth factor by human breast can-cer cell lines. Proc Natl Acad Sci U S A 1987;84:5763–5767.
40. Budhu A, Wang XW. The role of cytokines in hepatocellularcarcinoma. J Leukoc Biol 2006;80:1197–1213.
41. Budhu A, Forgues M, Ye QH, et al. Prediction of venous metas-tases, recurrence, and prognosis in hepatocellular carcinomabased on a unique immune response signature of the livermicroenvironment. Cancer Cell 2006;10:99–111.
42. Tilg H, Wilmer A, Vogel W, et al. Serum levels of cytokines inchronic liver diseases. Gastroenterology 1992;103:264–274.
43. Wang YC, Xu GL, Jia WD, et al. Estrogen suppresses metastasisin rat hepatocellular carcinoma through decreasing interleukin-6and hepatocyte growth factor expression. Inflammation 2012;35:143–149.
44. Jiang R, Tan Z, Deng L, et al. Interleukin-22 promotes humanhepatocellular carcinoma by activation of STAT3. Hepatology
2011;54:900–909.
45. Chia CS, Ban K, Ithnin H, et al. Expression of interleukin-18,interferon-gamma and interleukin-10 in hepatocellular carci-noma. Immunol Lett 2002;84:163–172.
46. Beckebaum S, Zhang X, Chen X, et al. Increased levels of inter-leukin-10 in serum from patients with hepatocellular carcinomacorrelate with profound numerical deficiencies and immaturephenotype of circulating dendritic cell subsets. Clin Cancer Res2004;10:7260–7269.
47. Zhou H, Huang H, Shi J, et al. Prognostic value of interleukin 2and interleukin 15 in peritumoral hepatic tissues for patientswith hepatitis B-related hepatocellular carcinoma after curativeresection. Gut 2010;59:1699–1708.
48. Roussos ET, Condeelis JS, Patsialou A. Chemotaxis in cancer.Nat Rev Cancer 2011;11:573–587.
49. Li W, Gomez E, Zhang Z. Immunohistochemical expression ofstromal cell-derived factor-1 (SDF-1) and CXCR4 ligand receptorsystem in hepatocellular carcinoma. J Exp Clin Cancer Res 2007;26:527–533.
50. Kryczek I, Lange A, Mottram P, et al. CXCL12 and vascularendothelial growth factor synergistically induce neoangiogenesisin human ovarian cancers. Cancer Res 2005;65:465–472.
51. Sutton A, Friand V, Brule-Donneger S, et al. Stromal cell-derivedfactor-1/chemokine (C-X-C motif) ligand 12 stimulates humanhepatoma cell growth, migration, and invasion. Mol Cancer Res2007;5:21–33.
52. Liu H, Pan Z, Li A, et al. Roles of chemokine receptor 4 (CXCR4)and chemokine ligand 12 (CXCL12) in metastasis of hepatocel-lular carcinoma cells. Cell Mol Immunol 2008;5:373–378.
53. Uchida H, Iwashita Y, Sasaki A, et al. Chemokine receptor CCR6as a prognostic factor after hepatic resection for hepatocellularcarcinoma. J Gastroenterol Hepatol 2006;21:161–168.
54. Kim HR, Lee SH, Jung G. The hepatitis B viral X protein activatesNF-kappaB signalling pathway through the up-regulation of TBK1.FEBS Lett 2010;584:525–530.
56. He G, Yu GY, Temkin V, et al. Hepatocyte IKKbeta/NF-kappaBinhibits tumor promotion and progression by preventing oxidativestress-driven STAT3 activation. Cancer Cell 2010;17:286–297.
57. Dapito DH, Mencin A, Gwak GY, et al. Promotion of hepatocellu-lar carcinoma by the intestinal microbiota and TLR4. Cancer Cell2012;21:504–516.
58. Yu LX, Yan HX, Liu Q, et al. Endotoxin accumulation preventscarcinogen-induced apoptosis and promotes liver tumorigenesisin rodents. Hepatology 2010;52:1322–1333.
59. Zhang HL, Yu LX, Yang W, et al. Profound impact of gut homeo-stasis on chemically-induced pro-tumorigenic inflammation andhepatocarcinogenesis in rats. J Hepatol 2012;57:803–812.
60. Massague J. TGFbeta in cancer. Cell 2008;134:215–230.61. Abou-Shady M, Baer HU, Friess H, et al. Transforming growth
factor betas and their signalling receptors in human hepatocel-lular carcinoma. Am J Surg 1999;177:209–215.
62. Amann T, Bataille F, Spruss T, et al. Reduced expression offibroblast growth factor receptor 2IIIb in hepatocellular carci-noma induces a more aggressive growth. Am J Pathol 2010;176:1433–1442.
63. Neaud V, Faouzi S, Guirouilh J, et al. Human hepatic myofibro-blasts increase invasiveness of hepatocellular carcinoma cells:evidence for a role of hepatocyte growth factor. Hepatology1997;26:1458–1466.
64. Gherardi E, Birchmeier W, Birchmeier C, et al. Targeting MET incancer: rationale and progress. Nat Rev Cancer 2012;12:89–103.
65. Kaposi-Novak P, Lee JS, Gomez-Quiroz L, et al. Met-regulatedexpression signature defines a subset of human hepatocellularcarcinomas with poor prognosis and aggressive phenotype. J Clin
Invest 2006;116:1582–1595.
1
1
REV
IEW
SA
ND
PER
SPEC
TIV
ES
March 2013 ROLE OF THE MICROENVIRONMENT IN HCC 525
66. De Luca A, Carotenuto A, Rachiglio A, et al. The role of the EGFRsignalling in tumor microenvironment. J Cell Physiol 2008;214:559–567.
67. Lu P, Weaver VM, Werb Z. The extracellular matrix: a dynamicniche in cancer progression. J Cell Biol 2012;196:395–406.
68. Schrader J, Gordon-Walker TT, Aucott RL, et al. Matrix stiffnessmodulates proliferation, chemotherapeutic response, and dor-mancy in hepatocellular carcinoma cells. Hepatology 2011;53:1192–1205.
69. Yang C, Zeisberg M, Lively JC, et al. Integrin alpha1beta1 andalpha2beta1 are the key regulators of hepatocarcinoma cellinvasion across the fibrotic matrix microenvironment. Cancer Res2003;63:8312–8317.
70. Giannelli G, Bergamini C, Marinosci F, et al. Clinical role ofMMP-2/TIMP-2 imbalance in hepatocellular carcinoma. Int J Can-cer 2002;97:425–431.
71. Provenzano PP, Eliceiri KW, Campbell JM, et al. Collagen reorga-nization at the tumor-stromal interface facilitates local invasion.BMC Med 2006;4:38.
72. Condeelis J, Segall JE. Intravital imaging of cell movement intumours. Nat Rev Cancer 2003;3:921–930.
73. Hashizume H, Baluk P, Morikawa S, et al. Openings betweendefective endothelial cells explain tumor vessel leakiness. Am JPathol 2000;156:1363–1380.
74. Hewitt RE, Powe DG, Morrell K, et al. Laminin and collagen IVsubunit distribution in normal and neoplastic tissues of colorec-tum and breast. Br J Cancer 1997;75:221–229.
75. Adler B, Ashkar S, Cantor H, et al. Costimulation by extracellularmatrix proteins determines the response to TCR ligation. CellImmunol 2001;210:30–40.
76. Bollyky PL, Wu RP, Falk BA, et al. ECM components guide IL-10producing regulatory T-cell (TR1) induction from effector memoryT-cell precursors. Proc Natl Acad Sci U S A 2011;108:7938–7943.
77. Barry-Hamilton V, Spangler R, Marshall D, et al. Allosteric inhibi-tion of lysyl oxidase-like-2 impedes the development of a patho-logic microenvironment. Nat Med 2010;16:1009–1017.
78. Semenza GL. Oxygen sensing, homeostasis, and disease. N EnglJ Med 2011;365:537–547.
79. Wu XZ, Xie GR, Chen D. Hypoxia and hepatocellular carcinoma:The therapeutic target for hepatocellular carcinoma. J Gastroen-terol Hepatol 2007;22:1178–1182.
80. Fiaschi T, Chiarugi P. Oxidative stress, tumor microenvironment,and metabolic reprogramming: a diabolic liaison. Int J Cell Biol2012;2012:762825.
81. Guichard C, Amaddeo G, Imbeaud S, et al. Integrated analysis ofsomatic mutations and focal copy-number changes identifies keygenes and pathways in hepatocellular carcinoma. Nat Genet2012;44:694–698.
82. Chung JS, Park S, Park SH, et al. Overexpression of romo1promotes production of reactive oxygen species and invasive-ness of hepatic tumor cells. Gastroenterology 2012;143:1084–1094 e7.
83. Lisanti MP, Martinez-Outschoorn UE, Chiavarina B, et al. Under-standing the “lethal” drivers of tumor-stroma co-evolution:emerging role(s) for hypoxia, oxidative stress and autophagy/mitophagy in the tumor micro-environment. Cancer Biol Ther2010;10:537–542.
84. Eng CH, Abraham RT. The autophagy conundrum in cancer: in-fluence of tumorigenic metabolic reprogramming. Oncogene 2011;30:4687–4696.
85. Migneco G, Whitaker-Menezes D, Chiavarina B, et al. Glycolyticcancer associated fibroblasts promote breast cancer tumorgrowth, without a measurable increase in angiogenesis: evi-dence for stromal-epithelial metabolic coupling. Cell Cycle 2010;9:2412–2422.
86. Gong K, Chen C, Zhan Y, et al. Autophagy-related gene 7 (ATG7)and reactive oxygen species/extracellular signal-regulated ki-nase regulate tetrandrine-induced autophagy in human hepato-
87. Ding ZB, Shi YH, Zhou J, et al. Association of autophagy defectwith a malignant phenotype and poor prognosis of hepatocellularcarcinoma. Cancer Res 2008;68:9167–9175.
88. Shi YH, Ding ZB, Zhou J, et al. Targeting autophagy enhancessorafenib lethality for hepatocellular carcinoma via ER stress-related apoptosis. Autophagy 2011;7:1159–1172.
89. Qin LX. Inflammatory immune responses in tumor microenviron-ment and metastasis of hepatocellular carcinoma. Cancer Mi-croenviron 2012;5:203–209.
90. Fu J, Xu D, Liu Z, et al. Increased regulatory T cells correlate withCD8 T-cell impairment and poor survival in hepatocellular carci-noma patients. Gastroenterology 2007;132:2328–2339.
91. Shen X, Li N, Li H, et al. Increased prevalence of regulatory Tcells in the tumor microenvironment and its correlation with TNMstage of hepatocellular carcinoma. J Cancer Res Clin Oncol2010;136:1745–1754.
92. Zamarron BF, Chen W. Dual roles of immune cells and theirfactors in cancer development and progression. Int J Biol Sci2011;7:651–658.
93. Zhang JP, Yan J, Xu J, et al. Increased intratumoral IL-17-produc-ing cells correlate with poor survival in hepatocellular carcinomapatients. J Hepatol 2009;50:980–989.
94. Chen S, Akbar SM, Tanimoto K, et al. Absence of CD83-positivemature and activated dendritic cells at cancer nodules frompatients with hepatocellular carcinoma: relevance to hepatocar-cinogenesis. Cancer Lett 2000;148:49–57.
95. Orimo A, Weinberg RA. Stromal fibroblasts in cancer: a noveltumor-promoting cell type. Cell Cycle 2006;5:1597–1601.
96. Zhang DY, Friedman SL. Fibrosis-dependent mechanisms ofhepatocarcinogenesis. Hepatology 2012;56:769–775.
97. Friedman SL. Hepatic stellate cells: protean, multifunctional, andenigmatic cells of the liver. Physiol Rev 2008;88:125–172.
98. Kalluri R, Zeisberg M. Fibroblasts in cancer. Nat Rev Cancer2006;6:392–401.
99. Lucas T, Abraham D, Aharinejad S. Modulation of tumor associ-ated macrophages in solid tumors. Front Biosci 2008;13:5580–5588.
100. Ding T, Xu J, Wang F, et al. High tumor-infiltrating macrophagedensity predicts poor prognosis in patients with primary hepatocel-lular carcinoma after resection. Hum Pathol 2009;40:381–389.
101. Wu SD, Ma YS, Fang Y, et al. Role of the microenvironment inhepatocellular carcinoma development and progression. CancerTreat Rev 2012;38:218–225.
102. Pello OM, De Pizzol M, Mirolo M, et al. Role of c-MYC in alterna-tive activation of human macrophages and tumor-associatedmacrophage biology. Blood 2012;119:411–421.
103. Sodir NM, Swigart LB, Karnezis AN, et al. Endogenous Mycmaintains the tumor microenvironment. Genes Dev 2011;25:907–916.
104. Gordan JD, Thompson CB, Simon MC. HIF and c-Myc: siblingrivals for control of cancer cell metabolism and proliferation.Cancer Cell 2007;12:108–113.
105. Ji J, Wang XW. Clinical implications of cancer stem cell biology inhepatocellular carcinoma. Semin Oncol 2012;39:461–472.
106. Mishra L, Banker T, Murray J, et al. Liver stem cells and hepa-tocellular carcinoma. Hepatology 2009;49:318–329.
107. Marquardt JU, Thorgeirsson SS. Stem cells in hepatocarcinogen-esis: evidence from genomic data. Semin Liver Dis 2010;30:26–34.
108. Lee TK, Castilho A, Cheung VC, et al. CD24(�) liver tumor-initiatingcells drive self-renewal and tumor initiation through STAT3-medi-ated NANOG regulation. Cell Stem Cell 2011;9:50–63.
09. Yamashita T, Ji J, Budhu A, et al. EpCAM-positive hepatocellularcarcinoma cells are tumor-initiating cells with stem/progenitorcell features. Gastroenterology 2009;136:1012–1024.
10. Yang ZF, Ho DW, Ng MN, et al. Significance of CD90� cancerstem cells in human liver cancer. Cancer Cell 2008;13:153–
166.
REV
IEWS
AN
DPER
SPEC
TIVES
526 HERNANDEZ–GEA ET AL GASTROENTEROLOGY Vol. 144, No. 3
111. Ma S, Chan KW, Hu L, et al. Identification and characterization oftumorigenic liver cancer stem/progenitor cells. Gastroenterology2007;132:2542–2556.
112. Baek HJ, Lim SC, Kitisin K, et al. Hepatocellular cancer arisesfrom loss of transforming growth factor beta signalling adaptorprotein embryonic liver fodrin through abnormal angiogenesis.Hepatology 2008;48:1128–1137.
113. Adams GB, Martin RP, Alley IR, et al. Therapeutic targeting of astem cell niche. Nat Biotechnol 2007;25:238–243.
114. Yao Z, Mishra L. Cancer stem cells and hepatocellular carci-noma. Cancer Biol Ther 2009;8:1691–1698.
115. Schiffer E, Housset C, Cacheux W, et al. Gefitinib, an EGFRinhibitor, prevents hepatocellular carcinoma development in therat liver with cirrhosis. Hepatology 2005;41:307–314.
116. Maeda S, Kamata H, Luo JL, et al. IKKbeta couples hepatocytedeath to cytokine-driven compensatory proliferation that promoteschemical hepatocarcinogenesis. Cell 2005;121:977–990.
117. Shachaf CM, Kopelman AM, Arvanitis C, et al. MYC inactivationuncovers pluripotent differentiation and tumour dormancy in hep-atocellular cancer. Nature 2004;431:1112–1117.
118. Liu L, Ulbrich J, Müller J et al. Deregulated MYC expressioninduces dependence upon AMPK-related kinase 5. Nature 2012;483:608–612.
119. Li Y, Tang ZY, Hou JX. Hepatocellular carcinoma: insight fromanimal models. Nat Rev Gastroenterol Hepatol 2012;9:32–43.
120. Frese KK, Tuveson DA. Maximizing mouse cancer models. NatRev Cancer 2007;7:645–568.
121. Heindryckx F, Colle I, Van Vlierberghe H. Experimental mousemodels for hepatocellular carcinoma research. Int J Exp Pathol2009;90:367–386.
122. Fausto N, Campbell JS. Mouse models of hepatocellular carci-noma. Semin Liver Dis 2010;30:87–98.
123. Campbell JS, Hughes SD, Gilbertson DG, et al. Platelet-derivedgrowth factor C induces liver fibrosis, steatosis, and hepatocel-lular carcinoma. Proc Natl Acad Sci U S A 2005;102:3389–3394.
124. Katzenellenbogen M, Mizrahi L, Pappo O, et al. Molecular mech-anisms of liver carcinogenesis in the mdr2-knockout mice. MolCancer Res 2007;5:1159–1170.
125. Mauad TH, van Nieuwkerk CM, Dingemans KP, et al. Mice withhomozygous disruption of the mdr2 P-glycoprotein gene. A novelanimal model for studies of nonsuppurative inflammatory cholan-gitis and hepatocarcinogenesis. Am J Pathol 1994;145:1237–1245.
126. Haybaeck J, Zeller N, Wolf MJ, et al. A lymphotoxin-driven path-way to hepatocellular carcinoma. Cancer Cell 2009;16:295–308.
127. Luedde T, Schwabe RF. NF-kappaB in the liver—linking injury,fibrosis and hepatocellular carcinoma. Nat Rev GastroenterolHepatol 2011;8:108–118.
128. Pikarsky E, Porat RM, Stein I, et al. NF-kappaB functions as atumour promoter in inflammation-associated cancer. Nature2004;431:461–466.
129. Okamoto M, Utsunomiya T, Wakiyama S, et al. Specific gene-expression profiles of noncancerous liver tissue predict the riskfor multicentric occurrence of hepatocellular carcinoma in hepa-titis C virus-positive patients. Ann Surg Oncol 2006;13:947–954.
130. Hoshida Y, Sangiovanni A et al. Gene expression signaturepredicts outcome of liver cirrhosis (abstr). Hepatology 2009;50:312A.
131. Abu Dayyeh BK, Yang M, Fuchs BC, et al. A functional polymor-phism in the epidermal growth factor gene is associated with riskfor hepatocellular carcinoma. Gastroenterology 2011;141:141–149.
132. Tanabe KK, Lemoine A, Finkelstein DM, et al. Epidermal growthfactor gene functional polymorphism and the risk of hepatocellular
carcinoma in patients with cirrhosis. JAMA 2008;299:53–60.
133. Jiang J, Gusev Y, Aderca I, et al. Association of MicroRNA ex-pression in hepatocellular carcinomas with hepatitis infection,cirrhosis, and patient survival. Clin Cancer Res 2008;14:419–427.
134. Chew V, Chen J, Lee D, et al. Chemokine-driven lymphocyteinfiltration: an early intratumoural event determining long-termsurvival in resectable hepatocellular carcinoma. Gut 2012;61:427–438.
135. Anson M, Crain-Denoyelle AM, Baud V, et al. Oncogenic beta-catenin triggers an inflammatory response that determines theaggressiveness of hepatocellular carcinoma in mice. J Clin In-vest 2012;122:586–599.
136. Ren Y, Poon RT, Tsui HT, et al. Interleukin-8 serum levels inpatients with hepatocellular carcinoma: correlations with clinico-pathological features and prognosis. Clin Cancer Res 2003;9:5996–6001.
137. Wang Y, Wang W, Wang L, et al. Regulatory mechanisms ofinterleukin-8 production induced by tumour necrosis factor-alphain human hepatocellular carcinoma cells. J Cell Mol Med;16:496–506.
138. Mejias M, Garcia-Pras E, Tiani C, et al. Beneficial effects ofsorafenib on splanchnic, intrahepatic, and portocollateral circu-lations in portal hypertensive and cirrhotic rats. Hepatology2009;49:1245–1256.
139. Mueller MM, Fusenig NE. Friends or foes—bipolar effects ofthe tumour stroma in cancer. Nat Rev Cancer 2004;4:839–849.
140. Wilhelm SM, Adnane L, Newell P, et al. Preclinical overview ofsorafenib, a multikinase inhibitor that targets both Raf and VEGFand PDGF receptor tyrosine kinase signalling. Mol Cancer Ther2008;7:3129–3140.
141. Sergio A, Cristofori C, Cardin R, et al. Transcatheter arterialchemoembolization (TACE) in hepatocellular carcinoma (HCC):the role of angiogenesis and invasiveness. Am J Gastroenterol2008;103:914–921.
142. Hurwitz H, Fehrenbacher L, Novotny W, et al. Bevacizumab plusirinotecan, fluorouracil, and leucovorin for metastatic colorectalcancer. N Engl J Med 2004;350:2335–2342.
143. Villanueva A, Llovet JM. Second-line therapies in hepatocellularcarcinoma: emergence of resistance to sorafenib. Clin CancerRes 2012;18:1824–1826.
144. Melero I, Sangro B, Riezu-Boj J, et al. Antiviral and antitumoraleffects of the anti-CTLA4 agent tremelimumab in patients withhepatocellular carcinoma (HCC) and chronic hepatitis C virus(HCV) infection: results from a phase II clinical trial. Presentedat: 103rd Annual Meeting of the American Association for CancerResearch; March 31 to April 4, 2012; Chicago, IL. Abstract4387.
145. Xu J, Xu HY, Zhang Q, et al. HAb18G/CD147 functions in inva-sion and metastasis of hepatocellular carcinoma. Mol CancerRes 2007;5:605–614.
146. Mazzocca A, Fransvea E, Lavezzari G, et al. Inhibition of trans-forming growth factor beta receptor I kinase blocks hepatocellu-lar carcinoma growth through neo-angiogenesis regulation.Hepatology 2009;50:1140–1151.
147. Borbath I PC, Rimassa L, Daniele B, et al. Tivantinib in Met�pretreated hepatocellular carcinoma (HCC): a randomized con-trolled phase 2 trial (RTC). Presented at: 6th Annual Conferenceof the International Liver Cancer Association; September 14–16,2012; Berlin, Germany. Abstract 0-023.
148. Zhu AX. Molecularly targeted therapy for advanced hepatocellu-lar carcinoma in 2012: current status and future perspectives.Semin Oncol 2012;39:493–502.
149. Zhang W, Zhu XD, Sun HC, et al. Depletion of tumor-associatedmacrophages enhances the effect of sorafenib in metastaticliver cancer models by antimetastatic and antiangiogenic ef-
fects. Clin Cancer Res 2010;16:3420–3430.
REV
IEW
SA
ND
PER
SPEC
TIV
ES
March 2013 ROLE OF THE MICROENVIRONMENT IN HCC 527
Received September 26, 2012. Accepted January 7, 2013.
Reprint requestsAddress requests for reprints to: Josep M. Llovet, MD, Division of
Liver Diseases, Box 1123, Mount Sinai School of Medicine, 1425Madison Ave, Room 11-70, New York, New York 10029. e-mail:[email protected]; fax: (212) 849-2574.
Conflicts of interest
The authors disclose no conflicts.
Funding
J.M.L. is supported by grants from the US National Institute ofDiabetes and Digestive and Kidney Diseases (1R01DK076986),European Commission Framework Programme 7 (HEPTROMIC,proposal 259744), the Samuel Waxman Cancer ResearchFoundation, the Spanish National Health Institute (SAF-2010-16055),and the Asociación Española Contra el Cáncer. S.L.F. is supported byNational Institutes of Health grants RO1DK56621, K05AA01840,
![[Micro] pathogenesis](https://static.documents.pub/doc/80x56/55a726df1a28ab7e5e8b45a7/micro-pathogenesis.jpg)
