havcr/kim-1 activates the il-6/stat-3 pathway in clear ......molecular and cellular pathobiology...

Molecular and Cellular Pathobiology

HAVCR/KIM-1 Activates the IL-6/STAT-3 Pathway in ClearCell Renal Cell Carcinoma and Determines TumorProgression and Patient Outcome

Thaïs Cuadros1, Enric Trilla3, Eduard Sarr�o1, Maya R. Vil�a1, Jordi Vilardell5, In�es de Torres4, Mayte Salcedo4,Joan L�opez-Hellin1, Alex S�anchez2, Santiago Ram�on y Cajal4, Emilio Itarte5, Juan Morote3, andAnna Meseguer1,6,7

AbstractRenal cell carcinoma (RCC), the third most prevalent urological cancer, claims more than 100,000 lives/year

worldwide. The clear cell variant (ccRCC) is the most common and aggressive subtype of this disease. Whilecommonly asymptomatic, more than 30% of ccRCC are diagnosed when already metastatic, resulting in a 95%mortality rate. Notably, nearly one-third of organ-confined cancers treated by nephrectomy develop metastasisduring follow-up care. At present, diagnostic and prognostic biomarkers to screen, diagnose, and monitor renalcancers are clearly needed. The geneencoding the cell surfacemoleculeHAVCR1/KIM-1 is a suggested susceptibilitygene for ccRCC and ectodomain shedding of this molecule may be a predictive biomarker of tumor progression.Microarray analysis of 769-P ccRCC-derived cells where HAVCR/KIM-1 levels have been upregulated or silencedrevealed relevant HAVCR/KIM-1–related targets, some of which were further analyzed in a cohort of 98 ccRCCpatients with 100month follow-up.We found that HAVCR/KIM-1 activates the IL-6/STAT-3/HIF-1A axis in ccRCC-derived cell lines, which depends on HAVCR/KIM-1 shedding. Moreover, we found that pSTAT-3 S727 levelsrepresented an independent prognostic factor for ccRCC patients. Our results suggest that HAVCR/KIM-1upregulation in tumors might represent a novel mechanism to activate tumor growth and angiogenesis and thatpSTAT-3 S727 is an independent prognostic factor for ccRCC. Cancer Res; 74(5); 1416–28. �2014 AACR.

IntroductionRenal cell carcinoma (RCC) is the third most prevalent

urological cancer. It accounts for approximately 3% of all newcancer cases and the incidence rates for all stages have beenrising steadily over the last 3 decades (1–3). The clear cell RCC(ccRCC) is the most common subtype and accounts for approx-imately 80% of all renal cancers. Commonly asymptomatic, onethird of ccRCCs are diagnosed when they are already meta-static, resulting in a 95% mortality rate. Moreover, one third oforgan-confined cancers treated by nephrectomy developmetastasis during the follow-up (4). Although ccRCC is resistantto chemotherapy, development of new therapies has improvedthe median survival period of patients with advanced ccRCC,which is about 26 months now (5).

Hepatitis A virus receptor/kidney injury molecule 1(HAVCR/KIM-1) is overexpressed in ccRCC tumors and blocksexpression of epithelial differentiation markers when over-expressed in 769-P cells (6). The HAVCR1 gene codes for atype I transmembrane glycoprotein that contains an extracel-lular immunoglobulin-like domain topping a long mucin-likesequence, which is shed and released by metalloproteinases.HAVCR, initially described in primate kidney cells (7), repre-sents the founding member of the HAVCR/KIM/TIM family (8,9). HAVCR/KIM-1 ectodomain detection in urine represents adiagnostic marker for acute kidney injury (10, 11) and earlydetection of ccRCC (12, 13). HAVCR/KIM-1 ectodomain shed-ding has been recently correlated with a more invasive phe-notype in vitro and more aggressive tumors in vivo (14).

To examine the biologic function ofHAVCR/KIM-1 effects inccRCC, microarray assays on 769-P ccRCC–derived cells, withupregulated or silenced HAVCR/KIM-1 levels, were conductedand relevant HAVCR/KIM-1 targets further analyzed inpatients with ccRCC. Results presented in this article providerelevant data to better understand the role of HAVCR/KIM-1on renal cancer development and progression.

Patients and MethodsA complete description is given in Supplementary Patients

and Methods.

Case selectionClinical and pathologic data from 168 patients with renal

masses treated with radical or partial nephrectomy for RCC

Authors' Affiliations: 1Fisiopatología Renal, CIBBIM; 2Statistics andBioinformatics Unit (UEB), Vall d'Hebron Institute of Research; 3Serviciode Urología, Hospital Vall d'Hebron; 4Servicio de Anatomía Patol�ogica,Hospital Vall d'Hebr�on; 5Departament de Bioquímica i Biologia Molecular,Unitat de Bioquímica, Facultat de Bioci�encies; 6Departament de Bioqui-mica i Biologia Molecular, Unitat de Bioquímica de Medicina, UniversitatAut�onoma de Barcelona, Bellaterra; and 7Instituto Reina Sofía deInvestigaci�on Nefrol�ogica, Fundaci�on Renal�I~nigo �Alvarez de Toledo, Spain

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

CorrespondingAuthor:AnnaMeseguer, Vall d'HebronResearch Institute,Pg. Vall d'Hebron, 119-129, 08035 Barcelona, Spain. Phone: 34-934-894-061; Fax: 34-934-894-015; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-13-1671

�2014 American Association for Cancer Research.

CancerResearch

Cancer Res; 74(5) March 1, 20141416

on September 7, 2021. © 2014 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst January 3, 2014; DOI: 10.1158/0008-5472.CAN-13-1671

http://cancerres.aacrjournals.org/

between January 2002 and December 2005 at Vall d'HebronHospital were reviewed. Standard tumor registry data andoutcome information were retrospectively collected by tumorregistrars for only ccRCC (n¼ 98). Pathologic information wasbased on primary tumor histology on re-review of surgicalsamples to determine histologic subtype using the WorldHealth Organization (WHO) 2004 criteria. Anatomic extent ofthe tumor was classified using the 2010 tumor-node–metas-tasis system, and clinical and follow-up information was basedon physician reports. Prognostic stratification of patientsaffected by ccRCC was scored using the University of Califor-nia, Los Angeles (UCLA) Integrated Staging System–UnionInternationale Contre le Cancer (UICC) nomogram.

Tumor microarray designTissue microarray (TMA) was constructed from a total of 98

ccRCC, as described in ref. 14.

ImmunohistochemistryImmunohistochemistry was performed as described in ref.

14 using mouse monoclonal antibody (mAb) against pSTAT-3Y705 and rabbit mAb against pSTAT-3 S727 (Cell SignalingTechnology). Two pathologists, blinded to clinicopathologicvariables, evaluated TMA sections.

Statistical analysesAssociations between pSTAT-3 S727 and pSTAT-3Y705

expression and clinicopathologic parameters were evaluatedwith the nonparametric Mann–Whitney U test. Disease-freesurvival was calculated as the date of surgery to the date ofloco-regional or distant recurrence. Subsequently, Kaplan–Meier survival estimates were compared using the log-ranktest. Multivariate analysis was performed using a Cox regres-sion model to estimate the independent prognostic impor-tance of clinicopathologic parameters. Statistical analysis wasperformed with the Statistical Package for Social Sciences,version 12 software (SPSS).

Cell cultureHuman renal adenocarcinoma cell lines, 769-P(CRL-1933)

and 786-O (CRL-1932), and the human embryonic kidneycell line, HEK293T (CRL-11268), obtained from AmericanType Culture Collection, were cultured as described in ref.14. The 769-P and 786-O cell lines were authenticated byLGC Standards.

Establishment of stable cell linesFor HAVCR/KIM-1- and interleukin (IL)-6–silenced cell

lines, three short hairpin RNAs (shRNA) of the human HAVCRor IL-6 genes were used (MISSIONshRNA; Sigma) and intro-duced into the 769-P cell line by lentiviral infection. Transfec-tion with a MISSION nontarget shRNA control vector wasused as negative control. For HAVCR/KIM-1 overexpression in769-P cell line, HAVCR/KIM-1 cDNA (ID: 26762) was clonedinto the pBIG-2r vector (kindly provided by Dr. J. Madrenas,University of Western Ontario, London, Ontario, Canada)including an human influenza hemagglutinin (HA) epitopeand stably transfected in 769-P cells using Lipofectamine and

Plus reagent (Invitrogen). Overexpression of HAVCR and itssheddingmutants (FUW246-274 and FUW267-278) in the 786-O cell line was previously described (14).

Cell lysate preparation and Western blot analysisCell lysates from HAVCR/KIM-1–overexpressing or

HAVCR/KIM-1–silenced 769-P cells and Western blot assayswere performed as described in ref. 14.

Proliferation assayCell proliferation assays were carried out by labeling 769-P

cells with the amine reactive fluorescent probe 5(6)-Carboxy-fluoresceindiacetate N-succinimidyl ester (CFSE; Sigma-Aldrich). Fluorescence wasmeasured using FACSCalibur (Bec-ton Dickinson). Data were analyzed using the FCS Express 4software. Parent cells correspond to cells labeled and analyzedimmediately after labeling.

Wound healing769-P cells were cultured to confluence and artificial wound

was created by scratching cells with a 200-mL pipette tip.Closure-denuded area was monitored using the Live Cell-RStation (Olympus). Digital images were obtained every 30minutes. Width of wounds at every time point was measuredusing the ImageJ software. Results are representative of threeindependent experiments.

Microarray expression analysisTotal RNA was extracted from cells either overexpressing

HAVCR/KIM-1 or with HAVCR/KIM-1 silenced from threeseparate experiments. Twelve independent microarrays wereperformed using the Human Exon 1.0 arrays (Affymetrix-Genechip array) at Vall d'Hebron Institute of Research Geno-mics facility. Statistical microarray analyses are described inSupplementary Patients and Methods. Minimum InformationAbout a Microarray Gene Experiment (MIAME) guidelineshave been followed. Accession number for datasets is E-MTAB-2071.

Reverse transcriptase quantitative PCRWeused the 7500Real-TimePCRSystem(AppliedBiosystems)

with the following TaqMan probes: SNAI2 (Hs00950344_m1),SMARCA1 (Hs00161922_m1), ANPEP (Hs00174265_m1), IL-6(Hs00985639_m1), HIF-1A (Hs0015153_m1), SLC2A1 (Hs0089-2681_m1), VEGFA (Hs00900055_m1), and PPIA (Hs99999904_m1) genes. PPIA probe was used as endogen control. TriplicatePCR amplifications were performed for each sample.

IL-6 ELISAIL-6 levels in cell culture supernatants were measured using

the commercially available quantitative assay from R&D Sys-tems (cat# D6050).

Indirect immunofluorescence assayCells grown on 24-well chamber slides from Nunc (Fisher

Scientific) were processed at confluence, as described in ref. 14.Antibodies used include mouse mAb HAVCR/KIM-1 antibody(R&D Systems), rabbit mAb gp130 (Santa Cruz Biotechnology),

HAVCR/KIM-1 Activates the IL-6/STAT-3 Pathway

www.aacrjournals.org Cancer Res; 74(5) March 1, 2014 1417




Cuadros et al.

Cancer Res; 74(5) March 1, 2014 Cancer Research1418




rabbitmAbpSTAT-3 S727, andmousemAbpSTAT-3 Y705 (CellSignaling Technology). Alexa 468–labeled anti-mouse andAlexa 488–labeled anti-rabbit (Invitrogen) were used as sec-ondary antibodies. Cells were observed by confocal fluores-cence microscopy.

ResultsOverexpression and silencing of HAVCR/KIM-1 in the769-P ccRCC cell lineEffects of HAVCR/KIM-1 overexpression or silencing were

studied in 769-P cells because they exhibit endogenousHAVCR/KIM-1 levels allowing overexpression or downregula-tion in the same cell context. HAVCR/KIM-1 levels in 769-Pcells overexpressing HAVCR/KIM-1(HK-16R) or control cells(CR4) are shown in Fig. 1A. For HAVCR/KIM-1 silencing, 769-Pstable cell pools obtained from three different si-HAVCR/KIM-1 RNAs and their corresponding siRNA control cells weretested for HAVCR/KIM-1 protein expression (Fig. 1B). Fromthem, sh851 showed to be the most effective and was chosenfor further experiments. Immunocytochemistry showed thatsubcellular location of HAVCR/KIM-1 is not affected by itsoverexpression or silencing (Fig. 1C). Effects of HAVCR/KIM-1in proliferation andmigration were analyzed in 769-P cells. Weobserved that HAVCR/KIM-1 overexpression or silencing pro-motes statistically significant augmented or decreased cellproliferation index, respectively (Fig. 1D), indicating thatHAVCR/KIM-1 induces cell growth. Cell migration assaysperformed by wound-healing experiments showed thatHAVCR/KIM-1 overexpression induced a slight delay onmigration at short times that was nonstatistically significantat later times. Contrarily, cells migrated faster than controlswhen HAVCR/KIM-1 levels were below constitutive levels (Fig.1E). Overall, results suggest that tumor cells might control cellgrowth and migration by modulating HAVCR/KIM-1 expres-sion levels.

HAVCR/KIM-1 overexpression or HAVCR/KIM-1silencing affects gene expression in ccRCC cells:microarray assaysTo explore the effects of HAVCR/KIM-1 up or downregula-

tion on global gene expression, cDNAmicroarray analyseswere

performed using the 769-P–modified cell lines from above. Agene was called differentially expressed if the adjusted P valuewas less than 0.01 and the log-fold change was at least 1 (foldchange of at least 2). These analyses result in two sets of genesthat are represented in volcano plots as those modulated byHAVCR/KIM-1 silencing (Fig. 2A, left) or overexpression (Fig.2A, right). From 466 genes significantly regulated by HAVCR/KIM-1 overexpression and 236 genes regulated by HAVCR/KIM-1 silencing, 112 genes were altered in both situations.Indeed, all of the 112 genes that were commonly regulated inboth experiments underwent changes in opposite directions(Fig. 2B). Figure 2C depicts a heatmap of expression profiles ofdifferentially expressed genes. Ingenuity Pathway Analysis(IPA) was used to identify pathways that were affected by thegenes selected and also to represent the relations appearingbetween genes. These analyses revealed highly significantalterations mainly in cancer but also in gastrointestinal, hema-tologic, infectious, and cardiovascular disease–related mole-cules, upon HAVCR/KIM-1 overexpression or silencing. Geneexpression, cell death and survival, cell proliferation, celldevelopment, and cell cycle were the most significant molec-ular and cellular functions correlating with HAVCR/KIM-1expression levels. Reverse transcriptase quantitative PCR(RT-qPCR) data of randomly selected genes confirmed andvalidated the microarray results (Fig. 2D).

HAVCR/KIM-1 protein levels and ectodomain sheddingcontrol IL-6 mRNA levels in ccRCC-derived cell lines

Among the 112 genes altered by both up or downregulationofHAVCR/KIM-1, we decided to study IL-6 and its downstreampathway because it has been described that the majority ofpatients with metastatic RCC and poor survival exhibitincreased IL-6 serum levels. Moreover, it was observed thatprimary ccRCC cultures express IL-6 (15–18). IPA diagramsrepresented in Fig. 3A show that essential components ofthe IL-6 signaling pathway are altered by HAVCR/KIM-1overexpression (left) or silencing (right). RT-qPCR assaysconfirmed that 769-P cells overexpressing HAVCR/KIM-1exhibited increased mRNA IL-6 levels, whereas HAVCR/KIM-1–silenced 769-P cells showed diminished IL-6 levels (Fig.3B, left). RT-qPCR performed in independent HAVCR/KIM-1–

Figure 1. Overexpression or silencing of HAVCR/KIM-1 in the ccRCC-derived 769-P cell line. Proliferation and migration assays. A, 769-P cells weretransfected with wild-type HAVCR/KIM-1 cDNA fused to the HA epiotpe for detection of exogenous protein expression (HK16R) or plasmid control(CR4). Total HAVCR/KIM-1 expression was evaluated by Western blot analysis in cell lysates using specific antibodies against HAVCR/KIM-1 (hHAVcr-1)and exogenous protein expression detected with an antibody against HA epitope. B, for HAVCR/KIM-1 silencing, three different shRNAs targetingdifferent exons of the human HAVcr-1 gene were introduced into the 769-P cell line by lentiviral infection. Transfection with shRNA control vector(shCV, SHC002) served as a negative control. HAVCR/KIM-1 expression was evaluated by Western blot analysis in cell lysates using specific antibodiesagainst HAVCR/KIM-1 (hHAVcr-1). Actin was used as protein loading control. C, 769-P cells expressing wild-type HAVCR/KIM-1 (HK16R) andplasmid control (CR4) or expressing silencing HAVCR/KIM-1 vector (sh851) and vector control (shCV) were subjected to indirect immunofluorescence assaysusing anti-HAVCR/KIM-1 antibody (left). Cells were stained with Hoescht 33323 dye (middle) and observed by confocal fluorescence microscopy.D, CFSE staining of 769-P cells expressingwild-type HAVCR/KIM-1 (HK16R) and plasmid control (CR4) or expressing silencingHAVCR/KIM-1 vector (sh851)and vector control (shCV) was measured by flow cytometry, and data were analyzed using the FCS express software. Histograms showing the proliferatinggenerations (left) are representative of three independent experiments. Vertical line, undivided cells peak channel. Means of proliferation index of threeindependent experiments are shown (right). ��, P < 0.01. E, contrast phase micrographs of 769-P cells, expressing wild-type HAVCR/KIM-1 (HK16R) andplasmid control (CR4) or expressing silencing HAVCR/KIM-1 vector (sh851) and vector control (shCV), migrating into the denuded area of the scratchwound. One representative experiment illustrates wound closure after 10 hours compared with the correspondingwound at time 0 hour (left). Graphs indicateprogress of migration in mm2 (y axis) at different time frames, each one corresponding to 30 minutes (x axis; right). Migration was calculated by subtractingwidth of the wound at every time point from that at 0 hour. Black dots, control CR4 and shCV cells; gray dots, HK16R and sh851 cells. Results are themean ofthree independent experiments.






Figure 2. Microarray assay results in 769-P cells with HAVCR/KIM-1 overexpression or silencing. A, volcano maps indicate fold change in gene expression inHAVCR/KIM-1–overexpressed (right) or HAVCR/KIM-1–silenced (left) 769-P cells, compared with control cells. Genes with an adjusted P value lessthan 0.01 and fold change of at least 2 were considered differentially expressed. B, Venn diagrams showing the number of genes up- or downregulated byHAVCR/KIM-1 overexpression (right circle) or HAVCR/KIM-1 interference (left circle). Genes modified in both situations are placed in the intersection.C, heatmap analyses reveal a global view of genes up- and downregulated in overexpressed (HK16) or silenced (sh851) in relation to their respective controls(CR4 and shCV). Triplicates for each experimental situation were performed. D, RT-qPCR of random selected genes to validate the microarray results (top).�, P < 0.05; ��, P < 0.01). Log2 fold-change values from microarrays are indicated in each case (bottom).

Cuadros et al.





overexpressing 769-P clones produced the same results (Sup-plementary Fig. S1). Augmented IL-6 mRNA levels correlatedwith increased IL-6 in conditioned media of HAVCR/KIM-1–overexpressing 769-P cells (Fig. 3B, right). To investigate theimplication of HAVCR/KIM-1 shedding on IL-6 production, weused 786-O cells that show barely detectable endogenousHAVCR/KIM-1 levels. As in 769-P cells, overexpression ofHAVCR/KIM-1 in 786-O induced IL-6 expression (Fig. 3C, left).Moreover, expression of the HAVCR/KIM-1 mutant with con-

stitutive shedding (FUW 246-274) resulted in increased IL-6mRNA levels, whereas shedding-defective (FUW 267-268) wasunable to induce IL-6 expression (Fig. 3C, right).

HAVCR/KIM-1 activates the IL-6/STAT-3 signalingpathway in 769-P cells

IL-6 exerts its biologic effects through binding to its ligand-binding receptor gp80 and thereupon to the transducingreceptor gp130 (19). Phosphorylated tyrosines in gp130 form

Figure 3. Augmented IL-6 production byHAVCR/KIM-1 and effects of HAVCR/KIM-1 shedding impairment. A, IPA pathways showing up- and downregulationof the IL-6 pathway in HAVCR/KIM-1–overexpressed (left) and HAVCR/KIM-1–silenced cells (right). IL-6 and gp130 are encircled on each diagram. Redand green indicate up- and downregulation, respectively. B, RT-qPCR to determine IL-6 levels in 769-P cells with HAVCR/KIM-1 overexpression (HK16R)or silencing (sh851; left) and CR4 and shCV control cells. Concentrations of IL-6 in cell culture supernatants of HAVCR/KIM-1 overexpression andcontrol cell lines (right). C, RT-qPCR to determine IL-6 levels in 786-O cells with HAVCR/KIM-1 overexpression (FUW hHAVcr-1) and 786-O control cell line(FUW; left). Effect of HAVCR/KIM-1 shedding defective mutant (FUW 267-278) on IL-6 mRNA levels (right). �, P < 0.05; ��, P < 0.01.






Figure 4. The IL-6/gp130/STAT-3 pathway is activated by HAVCR/KIM-1. A, effects of HAVCR/KIM-1 overexpression (HK16R) and HAVCR/KIM-1 silencing(sh851) in gp130, pSTAT-3 Y705, and pSTAT-3 S727 levels in 769-P cells when compared with CR4 and shCV control cells (right and left, respectively).B, 769-P cells expressing wild-type HAVCR/KIM-1 (HK16R) and plasmid control (CR4) were subjected to indirect immunofluorescence usinganti-gp130 (right), anti-pSTAT-3 Y705 (left), and anti-pSTAT-3 S727 antibodies (bottom). Cells were stained with Hoescht 33323 dye (middle) andobserved by confocal fluorescence microscopy. C, pSTAT-3 Y705 and pSTAT-3 S727 levels in IL-6–silenced HAVCR/KIM-1 overexpressing 769-P cells(HK16RshIL6) compared with cells carrying silencing empty vector overexpressing HAVCR/KIM-1 (HK16RshCV) measured by Western blot (left) andimmunohistochemistry (pSTAT-3 S727; right).

Cuadros et al.





docking sites for activator of transcription STAT-3 (20), whichbecomes activated through phosphorylation at tyrosine 705(21). As for IL-6, HAVCR/KIM-1 overexpression in the HK16Rclone correlated with increased gp130, pSTAT-3 Y705, andpSTAT-3 S727 levels (Fig. 4A). Immunohistochemistry alsoshowed the presence of pSTAT-3 Y705 andpS727 in the nucleusof HAVCR/KIM-1–overexpressing cells (Fig. 4B, left and bot-tom). The relationship between pSTAT3 S727 and HAVCR/KIM-1 levels was further demonstrated in 786-O cells (Sup-plementary Fig. S2). Moreover, knocking-down IL-6 expressionin cells overexpressing HAVCR/KIM-1 prevented HAVCR/KIM-1–induced STAT3 Y705 and S727 phosphorylation (Fig.4C). Overall, these results indicate that upregulation of IL-6levels by HAVCR/KIM-1 favors phosphorylation of STAT-3 atY705 and S727 residues in ccRCC-derived cells.

The IL-6/STAT-3 activation by HAVCR/KIM-1 controlsHIF-1A mRNA and protein expressionSTAT-3 transcriptionally regulates genes involved in tumor

proliferation, apoptosis inhibition, and angiogenesis. Amongthose genes, HIF-1A is a key element in promoting hypoxia-induced angiogenesis (22). Results from the microarray assaysindicated that HIF-1A was one of the top 10 upregulated genesby HAVCR/KIM-1 overexpression [log2 fold change (FC), 3.29;adj. P, 4.00E�12] and one of the most downregulated whenHAVCR/KIM-1 is silenced (log2 FC,�2.14; adj. P, 6.27E�07). IPAdiagrams show that components of the HIF-1A signaling path-way are also altered by HAVCR/KIM-1 overexpression (Fig. 5A,left) or silencing (Fig. 5A, right). RT-qPCR assays confirmed that769-P cells overexpressing HAVCR/KIM-1 exhibited increasedmRNA HIF-1A levels, whereas HAVCR/KIM-1–silenced 769-Pcells showed diminished HIF-1A levels (Fig. 5B, left). Accord-ingly, HIF-1A protein levels were also increased in HK16R1 cells(Fig. 5B, right). Because 769-P cells express barely detectablelevels ofHIF-1A, the effect ofHAVCR/KIM-1 silencingonHIF-1Aprotein was less apparent. To explore whether HIF-1A proteinlevels correlated with expression of HIF-1A target genes regu-lated by KIM-1 overexpression, we analyzedmRNA levels of twoclassic HIF-1A-regulated genes, GLUT-1 and VEGFA, and founda good correlation of GLUT-1 with HAVCR/KIM-1 overexpres-sion (Fig. 5C, right), but not with VEGFA (Fig. 5C, left). Thesedata fit well with results obtained in the microarray, in whichstatistical significance increments were obtained for GLUT-1(SLC2A1; log2 FC, 0.91; adj. P, 0.003), but not for VEGFA (log2 FC,�0.22; adj P, 0.449) upon HAVCR/KIM-1 overexpression.To investigate whether HIF-1A upregulation by HAVCR/

KIM-1 depends on IL-6–induced STAT-3 activation, wesilenced IL-6 in HK16R1 cells (Fig. 5D, left) and found thatIL-6 knockdown diminishes HIF-1A mRNA induced byHAVCR/KIM-1 (Fig. 5D, right).

pSTAT-3 is activated in patients with ccRCC and pSTAT-3S727H-score constitutes a novel independent prognosticfactor for patient outcomeThe relevance of the pSTAT-3 activation in patients with

ccRCC was studied in the same TMA previously used todetermine expression of HAVCR/KIM-1 (14). It includes 98patients with ccRCC with the following characteristics: 57.1%

men and 42.9% women; median age, 64 years; range, 25 to 86.Tumor was on the right side in 59.1% and left in 40.8% ofthe cases. Incidental presentation was found in 52.5% of thepatients and 46.9% were symptomatic. Of note, 87.9% of thepatients presented with localized tumors and 11.1% withmetastasis. A total of 92.9% patients underwent radicalnephrectomy, whereas 6 underwent nephron-sparing surgery.Median tumor size was 6.4 cm (range, 1.5–16). Fuhrman gradewas 1 in 20.4%, 2 in 41.8%, 3 in 24.4%, and 4 in 13.2% of thepatients. The most frequently observed pathologic classifica-tion of primary tumor size (pT) stages were pT1a in 26.5% andpT1b in 27.5%. Lymphovascular invasion was present in only5.1% of patients. Finally, theUICC risk groupwas 1 in 72.4% and2 in 27.5% of the patients studied.

Representative images of ccRCC tumors and unaffectednormal kidney counterparts stained with specific antibodiesagainst pSTAT-3 Y705 and pSTAT-3 S727 are shown in Fig. 6A.Tumors showed positive staining for both pSTAT-3 Y705 andpSTAT-3 S727mainly located in the nuclear compartment thatwas not found in normal tissues. Upon signal intensity eval-uation using H-score, results were correlated with clinicaloutcome. H-score value of 100 was arbitrarily set up as thethreshold value to separate patients with high or low pSTAT-3staining. Kaplan–Meier diagrams showing specific survivalrates in relation to patient follow-up (months) showed nosignificant correlation with pSTAT-3 Y705 H-score (notshown). However, Kaplan–Meier estimates of mortalityshowed statistically significant differences in overall survivalrates between patients with high versus low pSTAT-3 S727 H-score among individuals included in the same clinic andpathologic group (Fig. 6B–D).

In this study, we found significant statistical associationbetween pSTAT-3 S727 and recurrence-free survival and pro-gression. Univariate [P¼ 0.015; OR, 5; 95% confidence intervals(CI),1.3–19] andmultivariate analyses (P¼ 0.016; OR, 3.132; 95%CI, 1.247–7.870) indicate that expression levels of pSTAT-3 S727represent an independent prognostic factor with respect toclassical clinicopathologic features, which demonstrates thatpSTAT-3S727 is an independent factor that is clinically relevant.

DiscussionThis work reveals a potent control of IL-6 expression by

HAVCR/KIM-1 in ccRCC cell lines, mainly mediated byHAVCR/KIM-1 shedding. Increased IL-6 levels in serum andaugmented expression of IL-6 receptor in tumors weredescribed in patients with metastatic RCC and poor survival,indicating the presence of an IL-6 autocrinemechanism inRCCthat correlates with malignancy (15–18).

The response of renal cancer cells to IL-6 is most likelymediated by STAT-3 activation, which is, in turn, triggeredby IL-6–transducing receptor gp130 (23). Our results showthat HAVCR/KIM-1 regulates gp130 levels and activatespSTAT-3 at Y705 and S727. pSTAT-3 activation in RCC,especially in metastatic disease, has been previouslyreported (24). pSTAT-3 Y705 has been associated withtumors that carry poor prognosis like ccRCC and papillarytumors but not with tumors with a benign outcome, likeoncocytomas (25). This pattern of expression is the same as






Cuadros et al.





the one reported for HAVCR/KIM-1 (14). Therefore, it seemsthat effects of HAVCR/KIM-1 on pSTAT-3 Y705 observed in769-P cells might also occur in vivo (14). Albeit the well-established association between pY705 and tumor malig-nancy, significant correlations between pY705 and patientoutcome have not yet been described. No significant corre-lations were found between pY705 and clinical outcome but,for the first time, we described that STAT-3 S727 phosphor-ylation significantly correlates with prognosis and patientsurvival. Therefore, we propose pSTAT-3 S727 as a noveltissue biomarker for better stratification and follow-up ofpatients classified under the same stage and clinical score,assisting on clinical decisions.Because HAVCR/KIM-1 is not an independent factor of

survival (14), no statistical association was found betweenHAVCR/KIM-1 levels and pSTAT3 S727 in tumors, in relationto recurrence-free survival and progression [univariate anal-yses: P¼ 0.314; risk ratio, 0.6; 95%CI, 0.28–1.45), when analyzedin the same cohort of patients. Because our in vitro datacorrelate HAVCR/KIM-1 with pSTAT3 S727 levels only whenIL-6 is produced and we have also demonstrated that IL-6production depends on HAVCR/KIM-1 shedding, we postulatethat, in patients with ccRCC, pSTAT-3 S727 levels mightcorrelate betterwithHAVCR/KIM-1 ectodomain levels in urineand IL-6 levels in plasma than with HAVCR/KIM-1 levels intumors. Accordingly, we hypothesize that IL-6 and HAVCR/KIM-1 altogether might constitute useful surrogated biomar-kers for tumor progression and patient follow-up assessmentupon tumor removal.STAT-3 is a potent regulator of HIF-1A activity (22). Aber-

rantly enhanced HIF target genes, such as VEGF and GLUT-1,linked to persistent activation of STAT-3 and HIF-1A upregu-lation, contribute to increased tumor growth and metastaticspread of renal carcinomas and correlate with worse prognosisin ccRCC (26, 27). In this report, we have proved that over-expression of HAVCR/KIM-1 activates STAT-3 and increasesHIF-1A levels. Steady-state levels of HIF-1A protein mightreflect the balance between transcriptional induction promot-ed by HAVCR/KIM-1 and posttranslational degradation ofHIF-1A in normoxic conditions. HIF-1A protein levels corre-lated with GLUT-1 upregulation but not VEGFA, which mightrequire other transcription factors and/or coactivators likelynot present in the experimental conditions used in thisresearch.Biologic implications of pSTAT-3S727 in ccRCC, according

to its correlation with poor patient outcome, might be relatedwith enhanced transcriptional activation of a distinct subset ofSTAT-3 targets, which could be especially relevant in ccRCCprogression. In this report, we demonstrate that HAVCR/KIM-1, first described as one of the most upregulated genes in

ischemic rat kidney (8), is a potent activator of HIF-1A tran-scription, even in normoxic conditions. Therefore, constitutiveexpression of HAVCR/KIM-1 in normal parenchyma ofpatients with ccRCC, together with its further upregulationin tumors, might represent a novel mechanism to activatetumor growth and angiogenesis through activation of the IL-6/gp130/STAT-3/HIF-1A pathway.

HAVCR/KIM-1 has been mapped in 5q33.3 cytogeneticlocation (6). Because amplification of 5q chromosome is thesecondmost frequent cytogenetic alterations in ccRCC tumors(67%; ref. 28), we postulated that HAVCR/KIM-1 overexpres-sion observed in tumors could be related to the 5q amplifica-tion (6). Recently, The Cancer Genome Atlas (TCGA) ResearchNetwork refined the region of interest to the 5q35 region (28);thus, excluding the HAVCR/KIM-1 gene. This indicates thatother mechanisms rather than gene amplification mustaccount for HAVCR/KIM-1 overexpression in ccRCC. Pathwayand integrated analyses performed by the TCGA Networkhighlighted the importance of the well-known VHL/HIF path-way, the newly emerging chromatin remodeling/histonemethylation pathway, and the phosphoinositide 3-kinase/AKTpathway in ccRCC (28). Because more than 60% of ccRCCtumors overexpress the HAVCR/KIM-1 gene and this results inaugmented HIF-1A expression atmRNA and protein levels, it isapparent that the HAVCR/KIM-1 gene is involved in relevantpathways driving ccRCC development and progression.

This study demonstrates that HAVCR/KIM-1 acts above theIL-6/gp130/STAT-3/HIF-1 axis, mediating, in part, the effectsinduced by hypoxia in this pathway. The hypothesized mech-anism of action of HAVCR/KIM-1 is depicted in Fig. 7. Uponshedding by MMPs, HAVCR/KIM-1 ectodomain binds to puta-tive HAVCR/KIM-1 receptors, triggers IL-6 production, andactivates the gp130/STAT-3/HIF-1A pathway in ccRCC cells.Although MMP1 and MMP3 have been involved in HAVCR/KIM-1 shedding, mainly in situations inducing kidney injury(29), little is known on HAVCR/KIM-1–dependent MMPs inrenal cancer. Besides this paracrine/autocrine effect on tumorcells, it shall be considered the possible impact of HAVCR/KIM-1 ectodomain on tumor-associated stromal cells, includ-ing endothelia, tumor-associated fibroblasts, T lymphocytes,and macrophages (30, 31). Soluble or surface HAVCR/KIM-1expressed by renal epithelial cells upon injury has been pro-posed as the endogenous ligand for CD300lb/LMIR-5/CLM-7/mIREM-3 receptors expressed in myeloid cells in an ischemia–reperfusion injury model (32).

In conclusion, by inducing IL-6 expression, HAVCR/KIM-1activates STAT-3 and promotes expression of growth andangiogenic factors on tumor and likely in nontumor-associatedcells that would help tumor growth and metastasis. Theelucidation of HAVCR/KIM-1 downstream pathways has

Figure 5. HAVCR/KIM-1 control of HIF-1A expression depends on IL-6 levels. A, IPA pathways showing up and downregulation of the HIF-1A pathwayin cells with HAVCR/KIM-1 overexpression (right) or silencing (left). Red and green colors indicate up- and downregulation, respectively. B, HIF-1AmRNA (left) and protein (right) levels in 769-P cells with HAVCR/KIM-1 overexpression (HK16R) or silencing (sh851) and their corresponding controlcells (CR4 and shCV) were measured by RT-qPCR and Western blot analysis, respectively. Actin was included as a loading control. C, RT-qPCR todetermine VEGFA (left) and GLUT1 (right) levels in 769-P cells with HAVCR/KIM-1 overexpression (HK16R) or silencing (sh851) and their control cell lines(CR4 and shCV), respectively (left). D, RT-qPCR to determine IL-6 levels in IL-6–silenced 769-P cells with HAVCR/KIM-1 overexpression (HK16RshIL-6) orcontrol (HK16RshCV; left) and RT-qPCR to determine HIF-1A levels in HK16RshIL-6 and control HK16RshCV 769-P cells (right; �, P < 0.05).






Figure 6. pSTAT-3 S727 and pSTAT-3 Y705 expression in tumor samples from patients with ccRCC and correlation with clinical outcome. A, representativeexamples of expression patterns in ccRCC. pSTAT-3 Y705 (top) and pSTAT-3 S727 (bottom) expression in ccRCC tumors (T) and normal counterparts (N).Original magnification, �200; scale bar, 200 mm. B, Kaplan–Meier estimates of 100 months overall F€urhman grade 1 and 2 (left) and 3 and 4 (right)according to pSTAT-3 S727 levels. C, Kaplan–Meier estimates of 100 months overall pT1-2 N0 M0 (left) and pT3-4 N0-1 M0-1 (right) according topSTAT-3 S727 levels. D, Kaplan–Meier estimates of 100 months overall low risk (left) and high risk (right) according to pSTAT-3 S727 levels. Metastaticpatients have been excluded. The intensity score was defined as follows: 0, no appreciable staining in cells; 1, weak-intensity cells; 2, intermediateintensity of staining; and 3, strong intensity of staining. [H-score ¼ 1 � (% weak) þ 2 � (% moderate) þ 3 � (% intense) ranging from 0 to 300]. Correlationbetween pSTAT-3 S727 HistoScore expression level (blue line, HS above 100; green line, HS below 100) andmean survival of patients with ccRCC accordingto F€urhman grade (B), clinical stage (C), and the UICC risk group (D) are indicated.

Cuadros et al.





providedwith novel biomarkers for tumor prognosis, as well aswith putative novel therapeutic targets to be further developed.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: T. Cuadros, M.R. Vil�a, J. Morote, A. MeseguerDevelopment of methodology: T. Cuadros, I. de Torres, A. MeseguerAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): T. Cuadros, E. Trilla, E. Sarr�o, J. Vilardell, I. de Torres,S. Ram�on y CajalAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): T. Cuadros, E. Trilla, E. Sarr�o, M.R. Vil�a, J. Vilardell,M. Salcedo, J. L�opez-Hellin, A. S�anchez, S. Ram�on y Cajal, E. Itarte, J. Morote,A. MeseguerWriting, review, and/or revision of the manuscript: T. Cuadros, E. Trilla,E. Sarr�o, M.R. Vil�a, I. de Torres, M. Salcedo, E. Itarte, J. Morote, A. Meseguer

Administrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): E. Trilla, I. de Torres, J. MoroteStudy supervision: E. Trilla, I. de Torres, S. Ram�on y Cajal, A. Meseguer

Grant SupportThis work was supported in part by the Fundaci�o La Marat�o de TV3 (Ref:

052410 to A. Meseguer); the Ministerio de Ciencia e Innovaci�on (BFU2009-10189,to E. Itarte; SAF2011-2950 and SAF2005-05167 to A. Meseguer); the Instituto deSalud Carlos III (FIS PI081351 and REDINREN 2.0 REF: RD12/0021/0013 to A.Meseguer); the Fundaci�on Renal�I~nigo �Alvarez de Toledo (FRIAT to A.Meseguer),and the Fundaci�on Senefro (SEN to A. Meseguer). A. Meseguer's research groupholds the Quality Mention from the Generalitat de Catalunya (2009SGR).

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received June 11, 2013; revised November 15, 2013; acceptedDecember 4, 2013;published OnlineFirst January 3, 2014.

References1. Hock LM, Lynch J, Balaji KC. Increasing incidence of all stages of

kidney cancer in the last 2 decades in the United States: an analysis ofsurveillance, epidemiology and end results program data. J Urol2002;167:57–60.

2. Jemal A, Siegel R, Ward E, Hao Y, Xu J, Murray T, et al. Cancerstatistics. Cancer J Clin 2008;58:71–96.

3. Chow W-H, Dong LM, Devesa SS. Epidemiology and risk factors forkidney cancer. Nat Rev Urol 2010;7:245–57.

4. Escudier B, Kataja V ESMO Guidelines Working Group. Renal cellcarcinoma: ESMOClinical PracticeGuidelines for diagnosis, treatmentand follow-up. Ann Oncol 2010;21:137–9.

5. Coppin C, Kollmannsberger C, Le L, Porzsolt F, Wilt TJ. Targetedtherapy for advanced renal cell cancer (RCC): a Cochrane sys-tematic review of published randomised trials. BJU Int 2011;108:1556–63.

6. Vil�a MR, Kaplan GG, Feigelstock D, Nadal M, Morote J, Porta R, et al.Hepatitis A virus receptor blocks cell differentiation and is overex-pressed in clear cell renal cell carcinoma. Kidney Int 2004;65:1761–73.

7. Kaplan G, Totsuka A, Thompson P, Akatsuka T, Moritsugu Y, Fein-stone SM. Identification of a surface glycoprotein on African greenmonkey kidney cells as a receptor for hepatitis A virus. EMBO J 1996;15:4282–96.

Figure 7. Pathway diagram forHAVCR/KIM-1 action in ccRCCcells. 1, upon shedding by MMPs,HAVCR/KIM-1 ectodomain bindsto putative HAVCR/KIM-1receptors thatmight include TIM-4,CD300 family members, or others;2, soluble ectodomain binding toreceptors triggers IL-6 production;3, secreted IL-6 binds to its ligand-binding receptor gp80 andthereupon to the signal transducergp130; 4, phosphorylatedtyrosines in gp130 form dockingsites for activator of transcriptionSTAT-3 that, upon binding,become activated throughphosphorylation at Y705 and atS727; 5, activated STAT-3 bindsto specific sites in target genepromoters inducing transcriptionof HIF-1A, among others.






8. Ichimura T, Bonventre JV, Bailly V, Wei H, Hession CA, Cate RL, et al.Kidney injury molecule-1 (HKIM-1), a putative epithelial cell adhesionmolecule containing a novel immunoglobulin domain, is up-regulatedin renal cells after injury. J Biol Chem 1998;273:4135–42.

9. McIntire JJ, Umetsu SE, Akbari O, Potter M, Kuchroo VK, Barsh GS,et al. Identification of Tapr (an airway hyperreactivity regulatory locus)and the linked Tim gene family. Nat Immunol 2001;2:1109–16.

10. Bailly V, Zhang Z, Meier W, Cate R, Sanicola M, Bonventre JV.Shedding of kidney injury molecule-1, a putative adhesion proteininvolved in renal regeneration. J Biol Chem 2002;277:39739–48.

11. Han WK, Bailly V, Abichandani R, Thadhani R, Bonventre JV. Kidneyinjury molecule-1 (HKIM-1): a novel biomarker for human renal prox-imal tubule injury. Kidney Int 2002;62:237–44.

12. HanWK, Alinani A, Wu CL, Michaelson D, LodaM,McGovern FJ, et al.Human kidney injurymolecule-1 is a tissue and urinary tumormarker ofrenal cell carcinoma. J Am Soc Nephrol 2005;16:1126–34.

13. Lin F, Zhang PL, Yang XJ, Shi J, Blasick T, Han WK, et al. HumanKidney Injury Molecule-1 (HKIM-1): a useful immunohistochemicalmarker for diagnosing renal cell carcinoma and ovarian clear cellcarcinoma. Am J Surg Pathol 2007;31:371–81.

14. Cuadros T, Trilla E, Vil�aMR, de Torres I, Vilardell J,MessaoudNB, et al.hHAVRc-1/KIM-1 represents a susceptibility gene for development ofclear cell renal cell carcinoma (ccRCC) and shedding of hHAVRc-1/KIM-1 ectodomain a predictive biomarker of tumor progression. Eur JCancer 2013;49:2034–47.

15. Gogusev J, Augusti M, Chr�etien Y, Droz D. Interleukin-6 and TNF a

production in human renal cell carcinoma. Kidney Int 1993;44:585–92.16. Miki S, Iwano M, Miki Y, Yamamoto M, Tang B, Yokokawa K, et al.

Interleukin-6 (IL-6) functions as an in vitro autocrine growth factor inrenal cell carcinomas. FEBS Lett 1989;250:607–10.

17. Blay JY, Negrier S, Combaret V, Attali S, Goillot E, Merrouche Y, et al.Serum level of interleukin 6 as a prognostic factor in metastatic renalcell carcinoma. Cancer Res 1992;52:3317–22.

18. Takenawa J, Kaneko Y, Fukumoto M, Fukatsu A, Hirano T, FukuyamaH, et al. Enhanced expression of interleukin-6 in primary human renalcell carcinomas. J Natl Cancer Inst 1991;83:1668–72.

19. Kishimoto T. The biology of interleukin-6. Blood 1989;74:1–10.20. Heinrich PC, Behrmann I, Muller-Newen G, Schaper F, Graeve L.

Interleukin-6-type cytokine signalling through the gp130/Jak/STATpathway. Biochem J 1998;334:297–314.

21. Heinrich PC, Behrmann I, Haan S, Hermanns HM, M€uller-Newen G,Schaper F. Principles of interleukin (IL)-6-type cytokine signalling andits regulation. Biochem J 2003;374:1–20.

22. Jung JE, LeeHG,Cho IH,ChungDH, YoonSH, YangYM, et al. STAT-3is a potential modulator of HIF-1-mediated VEGF expression in humanrenal carcinoma cells. FASEB J 2005;19:1296–8.

23. Horiguchi A, Oya M, Marumo K, Murai M. STAT3, but not ERKs,mediates the IL-6-induced proliferation of renal cancer cells, ACHNand 769-P. Kidney Int 2002;61:926–38.

24. Horiguchi A, Oya M, Shimada T, Uchida A, Marumo K, Murai M.Activation of signal transducer and activator of transcription 3in renal cell carcinoma: a study of incidence and its association withpathological features and clinical outcome. J Urol 2002;168:762–5.

25. Guo C, Yang G, Khun K, Kong X, Levy D, Lee P, et al. Activation ofSTAT-3 in renal tumors. Am J Transl Res 2009;1:283–90.

26. Pantuck AJ, Zeng G, Belldegrun AS, Figlin RA. Pathobiology, prog-nosis and targeted therapy for renal cell carcinoma: exploiting thehypoxia-induced pathway. Clin Cancer Res 2003;9:4641–52.

27. Dorevi�c G, Matusan-Ilijas K, Babarovi�c E, Hadzisejdi�c I, Grahovac M,Grahovac B, et al. Hypoxia inducible factor-1a correlates with vascularendothelial growth factor A and C indicating worse prognosis in clearcell renal cell carcinoma. J Exp Clin Cancer Res 2009;28:40.

28. Cancer Genome Atlas Research Network. Comprehensive molec-ular characterization of clear cell renal cell carcinoma. Nature 2013;499:43–9.

29. Lim AI, Chan LY, Lai KN, Tang SC, Chow CW, Lam MF, et al. Distinctrole of matrix metalloproteinase-3 in kidney injury molecule-1 shed-ding by kidney proximal tubular epithelial cells. Int J Biochem Cell Biol2012;44:1040–50.

30. Kujawski M, Kortylewski M, Lee H, Herrmann A, Kay H, Yu H. STAT-3mediates myeloid cell dependent tumor angiogenesis in mice. J ClinInvest 2008;118:3367–77.

31. Calon A, Espinet E, Palomo-Ponce S, Tauriello DV, Iglesias M,C�espedes MV, et al. Dependency of colorectal cancer on a TGF-b-driven program in stromal cells for metastasis initiation. Cancer Cell2012;22:571–84.

32. Yamanishi Y, Kitaura J, Izawa K, Kaitani A, Komeno Y, Nakamura M,et al. TIM1 is an endogenous ligand for LMIR5/CD300b: LMIR5deficiency ameliorates mouse kidney ischemia/reperfusion injury.J Exp Med 2010;207:1501–11.

Cuadros et al.





2014;74:1416-1428. Published OnlineFirst January 3, 2014.Cancer Res Thaïs Cuadros, Enric Trilla, Eduard Sarró, et al. Patient OutcomeRenal Cell Carcinoma and Determines Tumor Progression and HAVCR/KIM-1 Activates the IL-6/STAT-3 Pathway in Clear Cell

Updated version

10.1158/0008-5472.CAN-13-1671doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2014/01/06/0008-5472.CAN-13-1671.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/74/5/1416.full#ref-list-1

This article cites 32 articles, 8 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/74/5/1416.full#related-urls

This article has been cited by 3 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/74/5/1416To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-13-1671

http://cancerres.aacrjournals.org/content/suppl/2014/01/06/0008-5472.CAN-13-1671.DC1

http://cancerres.aacrjournals.org/content/74/5/1416.full#ref-list-1

http://cancerres.aacrjournals.org/content/74/5/1416.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerres.aacrjournals.org/content/74/5/1416


havcr/kim-1 activates the il-6/stat-3 pathway in clear ......molecular and cellular pathobiology...

Documents