runx1 is a driver of renal cell ... - cancer research · kidney cancer is the seventh commonest...

CANCER RESEARCH | TUMOR BIOLOGYAND IMMUNOLOGY

RUNX1 Is a Driver of Renal Cell Carcinoma Correlatingwith Clinical Outcome A C

Nicholas Rooney1, Susan M. Mason1, Laura McDonald1, J. Henry M. D€abritz1, Kirsteen J. Campbell1,Ann Hedley1, Steven Howard1, Dimitris Athineos1, Colin Nixon1, William Clark1, Joshua D.G. Leach1,2,Owen J. Sansom1,2, Joanne Edwards2, Ewan R. Cameron3, and Karen Blyth1,2

ABSTRACT◥

The recurring association of specific genetic lesions with partic-ular types of cancer is a fascinating and largely unexplained area ofcancer biology. This is particularly true of clear cell renal cellcarcinoma (ccRCC) where, although key mutations such as loss ofVHL is an almost ubiquitous finding, there remains a conspicuouslack of targetable genetic drivers. In this study, we have identified apreviously unknown protumorigenic role for the RUNX genes inthis disease setting. Analysis of patient tumor biopsies together withloss-of-function studies in preclinicalmodels established the impor-tance of RUNX1 and RUNX2 in ccRCC. Patients with high RUNX1(and RUNX2) expression exhibited significantly poorer clinicalsurvival compared with patients with low expression. This wasfunctionally relevant, as deletion of RUNX1 in ccRCC cell lines

reduced tumor cell growth and viability in vitro and in vivo.Transcriptional profiling of RUNX1-CRISPR–deleted cells revealeda gene signature dominated by extracellular matrix remodeling,notably affecting STMN3, SERPINH1, and EPHRIN signaling.Finally, RUNX1 deletion in a genetic mousemodel of kidney cancerimproved overall survival and reduced tumor cell proliferation. Insummary, these data attest to the validity of targeting a RUNX1-transcriptional program in ccRCC.

Significance: These data reveal a novel unexplored oncogenicrole forRUNX genes in kidney cancer and indicate that targeting theeffects of RUNX transcriptional activity could be relevant forclinical intervention in ccRCC.

IntroductionKidney cancer is the seventh commonest cancer in the United

Kingdom with around 12,500 diagnoses and 4,500 deaths annually(Cancer Research UK; www.cancerresearchuk.org; accessed July2019). Around 85% of kidney cancers are classified as renal cellcarcinomas (RCC), of which, clear cell renal cell carcinoma (ccRCC)accounts for the vast majority (75%þ; refs. 1, 2). Since the mid 19700s,age-standardized kidney cancer mortality rates have increased by 74%while the incidence rate has increased by 85% in relative terms in thelast 25 years (Cancer Research UK). Various environmental riskfactors such as smoking, hypertension, and obesity contribute tokidney cancer development, but there is also a strong genetic contri-bution to the development of the disease (3). Many of these geneticalterations lead to changes in the transcriptional profile of the kidneycancer cells (4, 5).

Currently RCC represents a pressing clinical challenge due to itsincreasing incidence. Early-stage nonmetastatic RCC can be treated bypartial or radical nephrectomy, however, early-stage disease is oftenasymptomatic, resulting in patients more commonly presenting withadvanced disease that has a much poorer prognosis (6). Standard of

care for high-risk, advanced metastatic or recurrent RCC involvestargeted tyrosine kinase inhibitors (TKI) primarily against the VEGFandmTOR pathways, which have modest improvement over previouscytokine therapies (7, 8). Recently, combinatorial use of TKI withimmune checkpoint inhibitors against programmed cell death com-plex (PD-1 and PD-L1) have shown promising results in stage IIIclinical trials (9). However, the outlook for high-risk patients remainspoor and there is both a need for novel biomarkers of poor prognosisand identifying targetable genetic drivers.

Foremost among the genetic alterations that occur in kidney canceris the loss of the short arm of chromosome 3 that contains the tumorsuppressorsVHL and PBRM-1, BAP-1, and SETD2 and occurs in up to90% of cases of ccRCC (10, 11). VHL protein functions as an E3ubiquitin ligase targeting the hypoxia inducible factor (HIF) family oftranscription factors for proteasomal degradation. Loss of VHLtherefore causes a transcription factor–driven change in gene expres-sion, leading to the development of kidney cancer (1). The other tumorsuppressor genes commonly deleted (PBRM-1, BAP-1, SETD2) all actdirectly or indirectly, through epigenetic changes in methylationstatus, to cause alterations in gene expression of kidney cancercells (12). While much remains unknown about the transcriptionfactors important for kidney cancer, these genetic alterations highlightthe important role transcriptional misregulation plays in kidneycancer and the pressing need to identify the key factors involved.

RUNX1 is a member of an evolutionarily conserved family ofRUNXgenes that encode transcription factors. Together with its heterotypicbinding partner CBFb, RUNX1 forms a DNA-binding complexrequired for normal mammalian development (13). RUNX1 also hasestablished roles in various types of cancer (14) where classically,RUNX1 chromosomal translocations and mutations are key drivers ofhematopoieticmalignancies and leukemia (15). Increasingly, however,RUNX1 has been shown to play “context dependent” roles in solidtumors such as in the breast, where both RUNX1 gain and loss offunction has been associated with cancer (16–19). RUNX1 has alsobeen implicated in cancers of the ovary and uterus (20), prostate (21),

1CRUK Beatson Institute, Garscube Estate, Switchback Road, Bearsden,Glasgow, United Kingdom. 2Institute of Cancer Sciences, University of Glasgow,Bearsden, Glasgow, United Kingdom. 3School of Veterinary Medicine, Universityof Glasgow, Bearsden, Glasgow, United Kingdom.

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

Corresponding Author: Karen Blyth, CRUK Beatson Institute, Garscube Estate,Switchback Road, Bearsden, Glasgow G61 1BD, United Kingdom. Phone: 44-141-330-3686; Fax: 44-141-942-6521; E-mail: [email protected]

Cancer Res 2020;80:2325–39

doi: 10.1158/0008-5472.CAN-19-3870

�2020 American Association for Cancer Research.

AACRJournals.org | 2325Cancer Research.

by guest on August 31, 2020. Copyright 2020 American Association forhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from

http://crossmark.crossref.org/dialog/?doi=10.1158/0008-5472.CAN-19-3870&domain=pdf&date_stamp=2020-5-21

http://crossmark.crossref.org/dialog/?doi=10.1158/0008-5472.CAN-19-3870&domain=pdf&date_stamp=2020-5-21

http://www.cancerresearchuk.org

https://bloodcancerdiscov.aacrjournals.org

and skin. To date, very little is known about a functional role forRUNX1 in either normal kidney development or kidney cancer. Thereis some evidence of increased expression of a RUNX1 chromosomaltranslocation product in ccRCC patient samples (22) and RUNX1 hasbeen shown to be expressed in mouse models of kidney fibrosis (afeature of chronic kidney disease correlated with RCC), involvingRUNX regulation of TGFb-driven EMT (23).

Here we show that RUNX1 is expressed in human ccRCC and thathigh protein expression correlates with poorer survival. This is func-tionally relevant as deletion of RUNX1 in human ccRCC cell linesdisrupted tumor cell growth in vitro and in vivo, and enabled iden-tification of a novel set of RUNX1-dependent genes in ccRCC. Byutilizing a genetically engineered mouse (GEM) model of kidneycancer, we were able to interrogate the role of RUNX1 in tumorformation and genetically confirm that in vivo deletion of Runx1 slowskidney cancer development. Finally, we reveal that the related tran-scription factor RUNX2 is expressed in ccRCC, also associating withpoorer survival. Our results provide the first evidence that RUNXproteins are novel players in kidney cancer and functionally contributeto disease progression and clinical outcome.

Materials and MethodsAntibodies

The following antibodies were used: RUNX1 (8529), RUNX2(8486), GAPDH (3683), horseradish peroxidase-conjugated anti-rabbit secondary antibody (7074; Cell Signaling Technology); SER-PINH1 (10875-1-AP), and STATHMIN3 (11311-1-AP; ProteinTech).Primary antibodies used for immunoblotting at 1:1000 dilution: Ki67SP6 (RM-9106-S; Thermo Fisher Scientific).

ImmunoblottingCells were lysed in Pierce RIPA buffer (Thermo Fisher Scientific),

protein extracts resolved on 10% NuPAGE Novex Bis-Tris gels (LifeTechnologies) and transferred to Hybond-ECL nitrocellulose mem-branes (Amersham). All membranes were stripped and reprobed forGAPDH.

IHC analysisIHC staining for RUNX1/RUNX2 was performed on 4-mm forma-

lin-fixed paraffin-embedded (FFPE) sections previously dry heated at60�C for 2 hours. IHC performed on Agilent Autostainer link48.Sections manually dewaxed through xylene, graded alcohol, tap waterbefore heat-induced epitope retrieval (HIER) with sections heated to98�C (25 minutes); rinsed in Tris buffered saline with Tween (TBST),peroxidase blocked (Agilent, UK), washed in TBST before applicationof antibody at previously optimized dilution (RUNX1 1:75, RUNX21:300) for 40 minutes. Sections were washed in TBST before appli-cation of rabbit EnVision (Agilent) secondary antibody for 35minutesand rinsed in TBST before applying Liquid DAB (Agilent, UK) for 10minutes. Sections were washed in water, counterstained with hema-toxylin, and coverslipped using DPX. Ki67 (1:200 dilution) andSERPINH1 (1:80 dilution and high pH antigen retrieval). Digitalimages were captured on a Leica SCN400f slide-scanner (x20). Quan-tification of Ki67 performed manually using HALO image analysissoftware (Indica Labs).

Tissue microarrayThe tissue microarray (TMA) contained cores from 184 patients

diagnosed with ccRCC within the Greater Glasgow NHS Trustbetween 1997 and 2008 and obtained from Greater Glasgow and

Clyde NHS Biorepository as described previously (24, 25). Briefly, toaddress tumor homogeneity, three cores measuring 0.6 mm2 fromthree different tumor-rich areas, as identified by a specialist pathol-ogist, were used to construct the three TMAs. After IHC for RUNX1 orRUNX2 (above) and hematoxylin costaining, the proportion of tumorcells with RUNX nuclear positivity was manually quantified using theweighted histoscore (H-Score) method. This involved calculating asemiquantitative score by multiplying the percentage of cells showingstaining by a score ranging from 0 to 3 representing increasingintensity of staining (score 0-no staining, score 1-weak staining, score2-moderate staining, and score 3, strong staining) providing a scorefrom 0 to 300 (25). Three TMA sections were stained at the same timeand average H-Scores obtained. H-Scores were stratified into quartiles(Q1–Q4), the upper quartile Q4 assigned as RUNX-High and remain-ing quartiles Q1–3 assigned as RUNX-Low. One third of the TMAwasindependently scored and agreement assessed by interclass correla-tion coefficient >0.8 (26). Klintrup–Makinen score is a pathologicallydefined measure of inflammatory infiltration described previously forthis TMA (24, 25, 27). Statistical analysis was performed using SPSSStatistics Version 21.0 (SPSS IBM). Associations between categorizedH-scores and available data on variables were analyzed using x2 tests.Kaplan–Meier curves were plotted with corresponding log-rank teststo assess the relationship between these markers and survival. Mul-tivariate analysis was performed using backwards Cox regressionconditional technique to test for independence (25).

Cell lines786-O cells (cultured in RPMI medium), Caki-2 cells (cultured in

McCoy 5a medium; Sigma), and HEK293 cells (cultured in DMEM)were provided by Professor Eyal Gottlieb (Beatson Institute,Glasgow, Scotland, 2014). All media were supplemented with 10%FCS, 2 mmol/L L-glutamine, penicillin/streptomycin, and 0.5 mg/mLamphotericin B (Sigma). All media reagents were from Gibco unlessotherwise stated. Cells were of low passage and cultured for approxi-mately 2 months after recovery from frozen vials. RCC cell lines (786-Oand Caki-2) were authenticated using Promega GenePrint 10 system.Short tandem repeat multiplex assay was performed that amplifies 9tetranucleotide repeat loci and the Amelogenin gender determiningmarker (December 2016). Cells were routinely tested for Mycoplasma.

shRNA and CRISPR/CAS9 RUNX1 gene silencingRUNX1MISSION shRNA lentivirus DNA constructs (Sigma) were

used for targeting human RUNX1 (sh1: TRCN0000338489, sh5:TRCN0000013660). Lentiviruses were produced by transfectingHEK293 cells with 10 mg of the relevant shRNA expression vector(pLKO) with 7.5 mg PsPax2 and 4 mg pVSVG packaging vectors(Tronolab) using the calcium chloride method; complete mediumreplacement was done 5 hours after transfection. Forty-eight hoursafter transfection, viral supernatant was removed, sterile filtered(0.45-mm pores), and used to infect adherent 786-O and Caki-2 cellsovernight in the presence of 8 mg/mL polybrene (Sigma). Live-cellvisualization of GFP confirmed successful transduction. Cells weremaintained in medium containing 2 mg/mL puromycin (Sigma). ForCRISPR/CAS9 deletion, guide RNAs (gRNA) targeting humanRUNX1 were designed using the Zhang Lab tool (MIT, Boston, MA).The gRNA sequence used was 50-ATGAGCGAGGCGTTGCCGCT-30. 786-O cells were transfected using lipofectamine (Thermo FisherScientific); 8mL of Lipofectamine in 250mL serum-freemedium (SFM)and 2 mg DNA in 250 mL of SFM (with GLN), were each left for10 minutes at room temperature then the Lipofectamine/DNA mixincubated at room temperature for 30 minutes before being added to

Rooney et al.

Cancer Res; 80(11) June 1, 2020 CANCER RESEARCH2326Cancer Research.



2 � 105 786-O cells (plated overnight) and incubated at 37�C for5 hours prior to amedium change. Forty-eight hours after transfection,cells were cultured in medium containing 2 mg/mL puromycin for48 hours. Transfected cells grew back as individual colonies, whichwere picked, expanded, and screened for RUNX1 deletion byimmunoblotting.

Cell growth assays (cell counting, xCELLigence, and MTS)A total of 2 � 104 786-O (pX Ctrl) and 786-O RUNX1 CRISPR

clones (CRISPR A1/CRISPR A3) were plated in triplicate in 12-wellplates, trypsinized, and counted using the Casyton cell counter96 hours later. Cell count for pX Ctrl cells was normalized to 1 andCRISPR clones expressed as a proportion; experiments were repeatedat least four times. A total of 7 � 103 Caki-2 cells were plated andcounted as above. For xCELLigence assay 3 � 103 786-O cells wereplated in quadruplicate into wells of an E Plate 16. The impedanceapplied to an electric field over time caused by cells growing in the plateis proportional to the number of cells in the plate and is represented as acell index when measured using the XCELLigence Real Time CellAnalysis System (Roche Diagnostics GmbH). Experiments performedin quadruplicate at least three times with separate batches of cells. ForMTS cell viability assays, 3 � 103 786-O cells or 1 � 103 Caki-2 cellswere plated in quadruplicate in 96-well plates; every 24 hours, a 20%volume of CellTiter96 MTS assay reagent (Promega) was added perwell and incubated for 1 hour prior to reading absorbance at A490.Experiments were repeated at least three times with separate batches ofcells.

EdU pulse chaseA total of 1 � 105 786-O cells were left to adhere overnight in

complete medium. Medium was removed and replaced with completemedium containing 10 mmol/L EdU and incubated at 37�C for 30minutes. Cells were washed twice in PBS and sampled immediately or6 hours later. Cells were costained with 50 mg/mL propidium iodide(Sigma) for 30 minutes with gentle rocking and then analyzed on anAttune NTX flow cytometer. All experimental conditions performedin triplicate three separate times. Flow cytometry data were analyzedusing FlowJo.

Sytox Green apoptosis assayA total of 3� 103 786-Ocells were plated (24-well plate) and allowed

to adhere overnight. The next day, medium was changed to completemedium containing 5 mmol/L of Sytox Green. The plate was imagedevery hour for 68 hours on an Incucyte FLR imaging system. Con-fluence and number of Sytox-positive cells per well were calculatedusing Incucyte software.

Scratch wound assay786-O cells were plated in a 96-well image lock plate and allowed

to adhere overnight in complete medium. At confluence, the platewas scratched using the wound maker (Essen Biosciences) andmedium changed. Closure of the wound was imaged every hourover 24 hours and analyzed using Incucyte ZOOM live cell imagingsystem. All experiments were performed in quadruplicate threetimes.

Animal studiesAll animal experiments performed under UK Home Office Project

Licences (60/4181 and 70/8645) with ethical approval from theBeatson Institute and the University of Glasgow under the Animal(Scientific Procedures) Act 1986 and EU directive 2010. Mice were

maintained in a purpose-built facility in a 12-hour light/dark cyclewithcontinual access to food and water.

Kidney capsule xenograftEight- to 10-week-old female CD1-Foxn1nu (nude) mice were

obtained from Charles River, UK. A total of 5 � 105 786-O� cellswere injected directly into the kidney capsule in 20 mL growth factor–reduced Matrigel. Mice were continually assessed for signs of kidneyimpairment, and kidney tumor developmentmonitored by ultrasoundImaging. Mice were humanely sacrificed at clinical endpoint or 18-week time-point. Parental 786-O cells were initially passaged oncein vivo through the kidney (as described above) and a secondary 786-Ocell line (referred to as 786-O�) was established in culture using anadapted version of the method described here (28). Briefly, the kidneywas excised and normal tissue removed. The tumorwas finely choppedinto a paste and incubated with 140 rpm rotation at 37�C for 10minutes in 10mL of 1mg/mL type 2 collagenase (Sigma). The tube wasvortexed vigorously for 30 seconds before a second 10-minute incu-bation. Cells were washed with RPMI and passed through sequential100-, 70-, and 40-mm filters. RUNX1 was deleted from the 786-O� lineby CRISPR/CAS9 as described above. 786-O� vector control andCRISPR cells were confirmed to not express CAS9 prior toengraftment.

RNA sequencingA total of 5� 105 786-O cells (pXCtrl, CRISPRA1, andCRISPRA3)

were plated and sampled 48 hours later. Whole RNA was extractedusing RNeasy Mini Kit (Qiagen) according to the manufacturer'sprotocol. RNA was DNAse treated using RNAse-free DNAse set(Qiagen). RNA quality was tested on an Agilent 2200 Tapestationusing RNA screentape, all RNA integrity value ≥9.6. Libraries forcluster generation and DNA sequencing were prepared following anadapted method from Fisher and colleagues (29) using IlluminaTruSeq StrandedmRNALTKit. Quality and quantity ofDNA librariesassessed on an Agilent 2200 Tapestation (D1000 screentape) andQubit (Thermo Fisher Scientific) respectively. Libraries were run onIllumina Next Seq 500 using High Output 75 cycles kit (2� 36 cycles,paired end reads, single index). Quality checks on raw RNA-sequencing (RNA-seq) data files was done using fastqc version0.11.7 and fastq_screen version 0.11.4. RNA-seq paired-end readswere aligned to the GRCh38 (30) version of the mouse genome usingtophat2 version 2.1.0 with Bowtie version 2.2.6.0. Expression levelswere determined and statistically analyzed by a combination of HTSeqversion 0.6.1, the R environment, version 3.5.0, utilizing packages fromBioconductor data analysis suite and differential gene expressionanalysis is based on the negative binomial distribution using DESeq2.Full datasets produced from this study are publicly available on theSequence Read Archive database, accession number: PRJNA605312.

GEM model of kidney cancerCyp1aCre; Apcfl/fl; p21�/� mice (hereafter referred to as CAP) were

characterized in the Sansom lab as described previously (31). Thesemice were crossed with Runx1fl/fl mice (a kind gift from Prof. NancySpeck, University of Pennsylvania, Philadelphia, PA; Jax:010673,B6;129-Runx1tm3.1Spe/J; ref. 32) and/or Runx2fl/fl mice (33). Tumormice (equivalent numbers males/females in each cohort) were mon-itored for signs of tumor development and subsequently checked threetimes a week for signs of endpoint renal failure (blood in urine,hunching, swollen kidneys; ref. 31). At endpoint kidneys were fixedin 10% neutral buffered formalin and embedded in paraffin forsubsequent histologic analyses.

RUNX1 and Renal Cell Carcinoma

AACRJournals.org Cancer Res; 80(11) June 1, 2020 2327Cancer Research.



Figure 1.

Patients with ccRCC with high RUNX1 expression have poorer survival. A human tissue microarray containing 184 cores from patients with ccRCC was stained forRUNX1.A,Representative examples of RUNX1 staining in nontumor normal kidney and RCC cores. A range of RUNX1 expressionwas observed in patientswith ccRCCfrom negative to high. Magnified areas are in dashed boxes. Scale bar, 100 pixels. B,Quantification of RUNX1 expression as shown by RUNX1 histoscore (H-Score) forthe full TMA. Dashed red line, cut off for RUNX1 Low (quartile 1–3, H-Score: 0 to 26.7, n¼ 138) and RUNX1 High (upper quartile, H-Score: 30 to 225, n¼ 46). C,Kaplan–Meier curve showing reduced cancer-specific survival in RUNX1-High patients, log-rank P¼ 0.007 (survival data was available for 183 patients, RUNX1 High, n¼ 45).The cumulative percent survival for 5 years after diagnosis is shown below.Wilcoxon, P¼ 0.005. D, Average RUNX1 H-Score for patients divided into high KM score(KMHigh, averageRUNX1H-score¼ 34.1) versus lowKMscore (KMLow, averageRUNX1H-score¼ 15.2), t testP¼0.0027. E,Percent distribution of KM-LowandKM-High patients in RUNX1-Low and -High groups. RUNX1 Low, KM lo/hi % ¼ 55/45; RUNX1 High, KM lo/hi % ¼ 28/72; x2, P ¼ 0.001.

Rooney et al.




Statistical analysesThe TMA was analyzed using SPSS. All other statistical analyses

were performed using Graphpad Prism. The specific statistical testsused are indicated throughout. All error bars represent � SEM unlessotherwise indicated.

ResultsRUNX1 is expressed in human ccRCC and correlates with poorsurvival and increased inflammation

In silico analysis of The Cancer Genome Atlas (TCGA) ccRCCdataset (5) revealed that RUNX1 alterations occur in 6% of ccRCCcases and strikingly, the vast majority of these alterations are mRNAupregulations (96%; Supplementary Fig. S1A). A 6% alteration rate iscomparable with the rate at which other established genes involved inccRCC are altered (34) such as MTOR (11%), PI3KCA (8%), PTEN(8%), TP53 (7%), while interestingly, the pattern of alterations is muchmore varied for these genes (Supplementary Fig. S1B). Kaplan–Meiersurvival analysis of the patients with ccRCC with RUNX1 mRNAupregulation shows that they have a statistically significant decrease insurvival compared with the unaltered cohort (log-rank P ¼ 0.0008,RUNX1Unalteredmedian¼ 76.98months, RUNX1mRNA upregula-tion median¼ 36.21 months; Supplementary Fig. S1C). This is in linewith a recent report interrogating the TCGA dataset (35). Further-more, data obtained using the pan-cancer RNA-seq KM plotter (36)tool analyzing clinical survival data from 530 patients with ccRCC alsoshow that high RUNX1 expression correlates with poorer overallsurvival (P < 0.0001; Supplementary Fig. S1D).

Tissue microarrays (TMA) have previously been used to investigatethe protein expression level of RUNX1 in human epithelial tumors ofthe breast (17, 37) and ovary (38). Therefore, as an independentvalidation of in silico observations, we immunostained aTMAcontain-ing 184 tumor samples from patients with ccRCC. RUNX1 is clearlyexpressed in cell nuclei in a subset of ccRCC patient samples and is notexpressed in the nontumor kidney sample contained within the TMA(Fig. 1A). RUNX1 staining was scored by the weighted averagehistoscore (H-Score) method to quantify the range of RUNX1 expres-sion (see Materials and Methods). The TMA was stratified intoquartiles and the upper quartile (RUNX1-High, H-Score: 30–225,mean¼ 87.5, n¼ 46) was comparedwith the remaining lowest scoringcores (RUNX1-Low, Q1-Q3 H-Score: 0–26.7, mean ¼ 4.1, n ¼138; Fig. 1B). Patient survival information was available for 183patients. Kaplan–Meier survival analysis revealed that RUNX1-highpatients had a significantly poorer cumulative survival than RUNX1-low patients (log-rank P ¼ 0.007; Fig. 1C). The survival rate wasconsistently lower year on year and at 5 years from diagnosis was 68%for RUNX1-high patients compared with 88% for RUNX1-low, Wil-

coxon P ¼ 0.005 (Fig. 1C). Assessment of clinicopathologic char-acteristics showed that there was no significant association withRUNX1 H-Score and age, grade, necrosis, or recurrence (Table 1).However high RUNX1 expression was significantly associated with ahigh Klintrup–Makinen (KM) score, a pathologically defined measureof inflammation previously described for this TMA (see Materials andMethods). The average RUNX1 H-Score was significantly higher forpatients with a high KM score compared with low (34.1 vs. 15.2, t testP ¼ 0.0027; Fig. 1D). Accordingly, RUNX1-high patients weredistributed 28% KM low versus 72% KM high, compared with 55%KM low versus 45%KMhigh for RUNX1-low patients (Fig. 1E). Thesedata reveal for the first time that RUNX1 protein is aberrantlyexpressed in human ccRCC and that high RUNX1 expression is anindependent marker of poor prognosis [P ¼ 0.027HR ¼ 1.58; 95%confidence interval (CI),1.054–2.372] when combined with age, stage,grade, and tumor necrosis. These data also reveal that RUNX1-highpatients have an increase in inflammatory infiltration compared withtheir RUNX1-low counterparts.

RUNX1 is expressed in human ccRCC cell lines and deletionreduces cell growth

Having conclusively shown that RUNX1 expression correlates withpoorer survival in ccRCC, we wanted to ascertain a functional role forRUNX1 in this disease setting. To this end, RUNX1 expression wasmodulated in human ccRCC cell lines. Lentiviral delivery of differentshort hairpin RNAs (shRNA) was used to knockdown RUNX1expression in 786-O and Caki-2 cells (sh1 and sh5) compared witha scrambled control shRNA (Scr; Fig. 2A, inset; SupplementaryFig. S2A). shRNA-mediated knockdown of RUNX1 caused a decreasein cell index (proportional to the number of adherent viable cells) overa 125-hour period in culture in the 786-O cell line as assayed using thexCELLigence assay system (Fig. 2A and B). Cell number was alsosignificantly reduced in a second cell line (Caki-2) with RUNX1knockdown (Fig. 2C). In addition, cell viability after RUNX1 knock-down in 786-O and Caki-2 cells was reduced as assessed by the MTSassay (Supplementary Fig. S2B and S2C). To validate these findings,786-O cells were transfected with gRNA targeting RUNX1, and CAS9nuclease. Complete knockout of RUNX1 protein was confirmed in786-O RUNX1 CRISPR clones (CRISPR A1 and CRISPR A3) byimmunoblot (Fig. 2D, inset). CRISPRdeletion of RUNX1 also caused amore pronounced decrease in cell index (Fig. 2D andE) and a decreasein cell number (Fig. 2F) in both 786-O CRISPR clones.

To understand the nature of the growth defect observed in theRUNX1 knockout cells, the rate at which they were actively synthe-sizing DNA by incorporation of the thymidine analogue EdU wasassessed. 786-O control and RUNX1-deleted cells were pulsed withEdU by incubation for 30 minutes in medium containing EdU, then

Table 1. The relationship between RUNX1 and clinicopathologic characteristics of kidney cancer in the TMA study.

Clinicopathological RUNX1 High RUNX1 LowCharacteristic n (%) n (%) P

Age (� 61/> 61) 21/24 (47/53) 68/70 (49/51) 0.211Grade (I/II/III/IV) 2/16/17/10 (4/36/38/22) 11/38/65/19 (8/29/49/14) 0.332T-Stage (I/II/III/IV) 22/6/15/2 (49/14/33/4) 57/24/48/4 (43/18/36/3) 0.802Necrosis (not necrotic/necrotic) 3/19 (14/86) 8/38 (17/83) 0.861Recurrence (no/yes) 30/16 (65/35) 106/32 (77/23) 0.121Klintrup Makinen (Low/High) 13/33 (28/72) 76/62 (55/45) 0.002

Note: All statistics were analyzed by Pearson x2. Clinicopathologic scoring as published previously (24, 25).





Figure 2.

Deletion of RUNX1 causes a growth defect in human ccRCC cell lines.A, Representative immunoblot for RUNX1 (inset) shows partial knockdown in 786-O cells stablytransduced with two different shRUNX1 lentiviruses (786-O sh1 and 786-O sh5) compared with scrambled control (786-O Scr), reprobed for GAPDH loading control.shRUNX1 cells grow slower compared with Scr control in xCELLigence assay. n ¼ 4 independent experiments performed in quadruplicate. B, Cell index of 786-OshRUNX1 cells at 125 hours. � , ANOVA P¼0.0339; Scr versus sh5.C,Caki-2 ccRCC cells stably transfected with shRUNX1 (Caki-2 sh1 and Caki-2 sh5); average numberof cells 96 hours after plating is shown; n¼ 3 independent experiments, performed in triplicate, ANOVA P values: � , sh1¼0.0226; �� , sh5¼0.0052.D,Representativeimmunoblot for RUNX1 (inset) on 786-O vector control (pX Ctrl) and RUNX1 CRISPR cells (CRISPR A1 and CRISPR A3), reprobed for GAPDH loading control. 786-ORUNX1 CRISPR–deleted cells had a growth defect in xCELLigence assay. n¼ 3 independent experiments, performed in quadruplicate. E, Cell index of RUNX1 CRISPRcells at 125 hours. ANOVA P values: �� , A1¼0.0058; ��, A3¼ 0.0025. F,Normalized cell counts of 786-O RUNX1 CRISPR cells comparedwith control at 96 hours afterplating.N¼4 independent experiments, performed in triplicate. ANOVAP values: �� , A1,P¼0.0021; �� , A3,P¼0.0069.G,Representativeflowcytometry plots for pXcontrol and RUNX1 CRISPR clones (A1 andA3) time-point T6 (6 hours after 30-minute EdU pulse). y-axis is Log EdU-647 fluorescence; x-axis is PI staining. Quadrantsapplied whereby Q1 and Q2 are EdUþ and Q3 and Q4 are EdU�. Box represents G1� population. Numbers in quadrants are percent of total single cells analyzed. H,Average percent of EdUþ cells at T6. I, Average G1� % population of EdUþ cells revealed G1

� population is higher in pX control cells. ANOVA P values: �� , A1,P¼0.0001; � , A3, P¼0.01. All flow cytometry are n¼ 3 independent experiments performed in triplicate. J,Average number of SYTOX green dead cells per well as aproportion of % confluence. ANOVA, P values: A1 ¼ 0.0006, A3 ¼ 0.0015. N ¼ 4 independent experiments, performed in quadruplicate, analyzed using Incucytesoftware. All error bars � SEM.

Rooney et al.




sampled by fixation in 4%PFA immediately after EdU incubation (T0)or 6 hours later (T6). The cells were costained for EdU and PI(propidium iodide) and analyzed by flow cytometry as shown for theT6 time-point (Fig. 2G). There was no difference in total EdUincorporation between control and RUNX1-deleted cells at T0 (Sup-

plementary Fig. S2D) and at T6 (Fig. 2H). However, there was a clearreduction in the G1� population representing EdUþ cells that havetransitioned through S-phase and returned to G1 in both the RUNX1–deleted cell lines (Fig. 2I and population highlighted in box inFig. 2G).This suggests that theRUNX1-deleted cells face a delay in transitioning

Figure 3.

RUNX1 deleted cells have reduced in vitro cell migration and in vivo tumor formation. A, Representative images of scratch wound closure (yellow) at 0, 12, and24 hours after wounding for control and RUNX1-deleted 786-O cells. Scale bar, 300 mm. B, Quantification of wound closure as shown by relative wound density,ANOVA P values: A1, P¼ 0.0182; A3, P¼ 0.039. C, Relative wound density at 12-hour time-point is significantly reduced in RUNX1 deleted cells (A1 and A3). ANOVAP values: A1, P ¼ 0.0164; A3, P ¼ 0.0212. All scratch wound assays were performed in quadruplicate in three independent experiments. Error bars, � SEM.D, Representative ultrasound images of orthotopic recipient kidneys with 786-O� control and RUNX1-deleted cells over 18 weeks from surgery at indicated time-points. In control injected kidney, the purple dashed line represents normal kidney outline; blue dashed line, tumor outline. The proportion of tumor-bearing kidneysas identifiedbyultrasound is presentedbelow thepanels for each time-point;P¼0.011 (Fisher exact test) at 18weeks.E,Hematoxylin andeosin (H&E) and IHC imagesof kidney tumors from control (n ¼ 4) and RUNX1-deleted (n ¼ 1) injected kidneys stained for RUNX1 and RUNX2. Arrows, tumor areas in H&E. The RUNX1-deletedinjected kidney was negative for RUNX1 staining in the epithelium of the small tumor region, whereas RUNX1 was still present in stromal cells. Scale, 100 mm.





Figure 4.

RNA-seq reveals a RUNX1-dependent gene set in RCC cells. A, Heatmap of genes in 786-O ccRCC cells that were identified from analysis of RNA-seq data ashaving significant differential expression (P < 0.05; >2-fold change, FC) in the same direction between CRISPR clone A1 (n ¼ 3) and control pX (n ¼ 3), andbetween CRISPR clone A3 (n ¼ 3) and control pX. Blue, downregulation of gene expression (row Z-Score <0); red, upregulation of gene expression (rowZ-Score > 0). Venn diagram of significantly differentially expressed (FC > 2 and Padj < 0.05) genes in CRISPR clone A1 (yellow) and clone A3 (purple) versuscontrol and the overlap. The 724 genes were significantly differentially expressed in both clones, with 710 changed in the same direction in both, shown inbrackets. (Continued on the following page.)

Rooney et al.




through the S/G2 stages of the cell cycle. Finally, the number of deadcells was assessed by time-lapse imaging of the control and RUNX1-deleted cells in the presence of SYTOX Green nucleic acid stain. Thisrevealed that the number of SYTOX-positive dead cells per well, as aproportion of percent confluence, was higher in the RUNX1-deletedcells compared with control, especially at earlier time-points (Fig. 2J).Confluence and the number of SYTOX-positive dead cells per well areshown individually in Supplementary Fig. S2E and S2F. Together,these data indicate that knockout of RUNX1 causes a reduction in cellgrowth in ccRCC cell lines and that RUNX1CRISPR cells have a subtledelay in progression through the cell cycle and an increase in cell death.

Knockout of RUNX1 in 786-O ccRCC cells reduces in vitro cellmigration and in vivo tumor formation

To further investigate the effect of RUNX1 knockout in physio-logically relevant assay systems, the effect of deletion on cell migrationusing in vitro scratch wound assays was assessed. This revealed thatRUNX1-deleted cells exhibited decreased wound closure and reducedrelative wounddensity over a 24-hour period (Fig. 3A–C). To establishwhether RUNX1 deletion effects ccRCC development in vivo, andto circumvent the low tumorigenicity of the 786-O cell line, wegenerated a secondary cell line (hereafter referred to as 786-O�) bypassaging 786-O cells through the kidney in vivo (see Materials andMethods). RUNX1was deleted in these 786-O� cells by CRISPR/CAS9as performed above (Supplementary Fig. S3A) and these 786-O�

RUNX1-deleted cells showed a similar growth defect to the parentalcells (Supplementary Fig. S3B and S3C). RUNX1-deleted and control786-O� cells were injected directly into the kidney capsule of CD1-Nu/Nu recipient mice and their tumor growth was monitored by ultra-sound over an 18-week period. This revealed that at 10 weeks post-surgery, 7 of 13mice injected with the control cells had formed tumorscompared with 0 of 13 for the RUNX1-deleted 786-O� cells (Fig. 3D).When sacrificed at 18 weeks, 8 of 13 mice injected with RUNX1-proficient cells had grossly observable kidney tumors, whereas just 1 of13 of the recipients with RUNX1-deleted cells had a small tumorgrowth (P¼ 0.011; Fisher exact test;Fig. 3D; Supplementary Fig. S3D).Four of 13 control mice exhibited gross lung metastases while none ofthe RUNX1-deleted group did. Kidney tumors from the control groupand the single tumor arising in the RUNX1-deleted cohort werestained for RUNX1 and its closely related family member RUNX2.RUNX1was highly expressed in all control tumors tested (n¼ 4) whileit was absent from the RUNX1-deleted tumor as expected (Fig. 3E).However, it was notable that RUNX2 was present in both control andRUNX1-deleted tumors. These data support our findings that RUNX1is important for growth and survival of human ccRCC cells and thatdeletion of RUNX1 hampers tumor growth and development in vivo.

Identification of a RUNX1-regulated gene signature in ccRCCAsRUNX1deletion causes a defect in ccRCC cell growth, wewanted

to understand the significant downstream players by assessing howdeletion of RUNX1 affects the global transcriptional profile in human

786-O ccRCC cells. RNA sequencing was performed on whole RNAextracts from control and RUNX1 CRISPR cells (as used in Fig. 2).Several hundred genes were significantly differentially expressed (P <0.05, >2-fold up or downregulation) in either RUNX1-deleted cell line(A1 ¼ 1185, A3 ¼ 1296). This revealed a novel RUNX1 regulatedsignature of 724 genes common to both clones that were significantlydifferentially expressed, with 710 altered in the same direction in bothRUNX1 CRISPR clones compared with the control cells (Fig. 4A).Excluding uncharacterized genes, pseudogenes, and novel transcripts,661 genes are significantly differentially expressed with 394 upregu-lated and 267 downregulated on RUNX1 deletion in both clones.Principal component analysis revealed exceptionally high agreementbetween the datasets with 97% of the variance explained by RUNX1deletion (Supplementary Fig. S4). Full lists of the regulated genes areavailable in Supplementary Data File S1 where they are ranked by foldchange and significance. Gene ontological analysis using Metacorerevealed the main biological pathways that were altered on RUNX1deletion. This encompasses a range of pathways such as cell adhesionand ECM remodeling, Eph, and Ephrin signaling, angiogenesis andglutathione metabolism (Fig. 4A). The most altered pathway was celladhesion and ECM remodeling, which included changes in expres-sion of genes such as MMP1, MMP16, SERPINE2, Fibronectin, andSyndecan 2 (Fig. 4B). The average fold change on the x-axis [x ¼Average log2(fold change)] was plotted against significance on the y-axis [y ¼ �log10(Max(Padj)] in a volcano plot to visually depict themost significantly differentially expressed genes (Fig. 4C). Two suchgenes, STMN3, which encodes a protein that plays a role in micro-tubule dynamics in the cell cycle (upregulated þ46.3x, redcircle Fig. 4C) and SERPINH1 (HSP-47), increased expression ofwhich has been shown to be a marker of poor prognosis in ccRCC(downregulated�4.1�, blue circle Fig. 4C) were validated byWesternblot analysis, which supported the findings of the RNA-seq data(Fig. 4D). Interestingly, the second most altered gene ontology wasEph and Ephrin signaling, which are downstream targets of the WNTsignaling pathway that is itselfmodulated byRUNX1 activity (Fig. 4E).Finally,CPT1A,which has been shown to be suppressed in ccRCC, wasincreased on RUNX1 deletion (Fig. 4F). These data have, for the firsttime, identified a group of genes whose expression is significantlyaltered as a consequence of the level of RUNX1 in human ccRCC. Thisshows that deregulation of RUNX1 expression affects a wide range ofkey pathways, many of which are related to kidney cancer and cancerprogression.

RUNX1 deletion improves survival in a genetic mouse model ofkidney cancer

To further explore the functional role of RUNX1 in a physiologicsetting, we turned to a GEM model where we could intrinsicallymodulate RUNX1 levels. First, we ascertained the levels of RUNX1in a GEM model of kidney cancer available in our lab in which Crerecombinase expressed in the kidney epithelium drives deletion of thetumor suppressorApc on a p21-null background (31). Normal kidneys

(Continued.)Table of statistically significant pathwaysmodulated in RUNX1-deleted 786-O cells compared with control, produced by Metacore (Clarivate Analytics)GeneGOanalysis of RNA-seqdata.P value andFDRare shown.B,RNA-seq read counts of selected targets from the toppathway “Cell adhesion andECM remodeling.”C,Volcanoplot of average log2-fold change (x-axis) versus –log10maxPadj values (y-axis). The averageandmax relate to values for bothCRISPRclonesA1 andA3. Thepoints highlighted in red are the 724 differentially altered genes with average log2(fold change) >1 and Max(Padj values) < 0.05. Red circle, STMN3; blue circle,SERPINH1. D, Left, average read counts from RNA-seq data for STMN3 (upregulatedþ46.3�; P < 0.0001) and SERPINH1 (downregulated�4.1�; P < 0.0001). Right,corresponding validation at the protein level; representative immunoblot probed for STATHMIN3 and sequentially reprobed for SERPINH1 and GAPDH (loadingcontrol).E,RNA-seq read counts of selected targets from the second top pathway “Eph and Ephrin signaling.” F,RNA-seq read count forCPT1A. Mean and SDof threebiological replicates shown for all RNA-seq read counts (P values < 0.0003).





and kidney tumors from thismodel (AH-Cre;Apcfl/fl;p21�/� referred toas CAP) were stained for RUNX1 to reveal that while RUNX1 is notexpressed in normal kidney, it is significantly upregulated in kidneytumors (Fig. 5A). We proceeded to cross this CAP model with aconditional knockout of Runx1 (Runx1fl/fl) (32). RUNX1 deletion inthe tumors ofCAP;Runx1fl/flmicewas confirmed at the protein level byIHC, which showed absence of RUNX1 (Fig. 5B). Cohorts of CAP;Runx1þ/þ and CAP;Runx1fl/fl mice were aged until clinical endpoint.Kaplan–Meier analysis shows that survival of CAP;Runx1fl/flmice wassignificantly extended (log-rank P ¼ 0.0365) compared with theirCAP;Runx1þ/þ counterparts, with a mean survival of 104.6 versus78.6 days, t test P ¼ 0.0415 (Fig. 5C and D). Tumors were immu-nostained for the proliferation marker Ki67, which exhibited lowerpositive staining in tumors from the CAP;Runx1fl/fl mice comparedwith CAP;Runx1þ/þ (Fig. 5E). This was confirmed by quantification

using the HALO imaging platform, which revealed that tumors fromCAP;Runx1þ/þmice had a higher proportion of Ki67þ cells than fromCAP;Runx1fl/flmice (34% vs. 24.4%, t test, P¼ 0.0154; Fig. 5F). Finally,tumors immunostained for SERPINH1 (downregulated in RNA-seq,Fig. 4E) revealed SERPINH1 is highly expressed inCAP;Runx1þ/þ

tumors compared with normal kidney (Fig. 5G). Deletion of RUNX1causes a significant decrease in SERPINH1 levels in kidney tumors inline with our RNA-seq data. Taken together, these data from our GEMmodel of kidney cancer confirm in vivo that deletion of RUNX1 leadsto improved survival and less tumor proliferation.

High RUNX2 expression also correlates with poorer survival inhuman ccRCC

While deletion of Runx1 significantly delayed tumorigenesis inthe GEM model, these animals still succumbed to disease. We

Figure 5.

RUNX1 deletion delays kidney cancer in agenetic mouse model. A, Representativeimages of two normal kidneys and kidneytumors from the AH-Cre;Apcfl/fl;p21�/�

(referred to as CAP) genetic mouse modelof kidney cancer stained for RUNX1. Scalebar, 100 mm. B, IHC for RUNX1 in two rep-resentative kidney tumors from CAP;Runx1fl/fl mice confirmed deletion of RUNX1in the tumors. C, Kaplan–Meier survivalcurve for CAP;Runx1þ/þ versus CAP;Runx1fl/fl mice showing improved survivalon Runx1 deletion, log-rank P ¼ 0.0365. D,Mean average lifespan for CAP;Runx1þ/þ

(78.6 days,n¼ 16) versusCAP;Runx1fl/flmice(104.6 days, n ¼ 19); t test, P ¼ 0.0415. E,Representative imagesof twodifferentCAP;Runx1þ/þ and CAP;Runx1fl/fl tumors (n1 andn2) stained for the proliferationmarker Ki67.Scale bar, 100 mm. F, Quantification byHALO analysis of the % of Ki67þ cells;Runx1þ/þ ¼ 34% (n ¼ 12); Runx1fl/fl ¼24.4% (n ¼ 9); t test, P ¼ 0.0154. G, Rep-resentative IHC images of tumors (4 vs. 4)from CAP;Runx1þ/þ versus CAP;Runx1fl/fl

mice stained for SERPINH1.

Rooney et al.




hypothesized that the related RUNX family protein RUNX2 might beexpressed and contribute to disease progression. Indeed, RUNX2 wasexpressed both in the CAP;Runx1þ/þ and CAP;Runx1fl/fl tumors(Fig. 6A). Attempts to model deletion of RUNX2 in this model ofkidney cancer were hampered by the nonviability of AH-Cre;Runx2fl/fl

mice (suggesting a possible limiting requirement for RUNX2 inembryonic development). Heterozygous deletion of Runx2 did notaffect survival either on a Runx1þ/þ or Runx1fl/fl background in theCAP model. Although, it is important to note that RUNX2 proteinexpression was still observed in these tumors (SupplementaryFig. S5A–S5D). Interestingly, however, in silico analysis of RUNX2expression in the TCGA human ccRCC dataset (5) revealed thatRUNX2 is altered in 8% of patients with ccRCC (93.5% are mRNAupregulations, Supplementary Fig. S5E). The pan-cancer RNA-seqKM plotter tool (described in Supplementary Fig. S1C) revealedhuman patients with ccRCC with high RUNX2 expression had poorersurvival (log-rank P ≤ 0.0001; Supplementary Fig. S5F). It is note-worthy that RUNX3 is also upregulated in ccRCC; however, unlikeRUNX1 and RUNX2, it does not significantly correlate with disease-free survival (Supplementary Fig. S5E and S5F). To directly assess theexpression of RUNX2 in human patients with ccRCC, the same TMAin Fig. 1 was used to show that RUNX2 protein is also expressed inhuman ccRCC (Fig. 6B). While the RUNX2 H-Score was on averagelower than RUNX1 (Supplementary Fig. S5G), when stratified intoquartiles based on RUNX2H-Score (Supplementary Fig. S5H), patientswith high RUNX2 also had a statistically significant decrease in survivalcompared with patients with low or no RUNX2 expression (log-rankP ¼ 0.0478; Fig. 6C). At five years post diagnosis, survival for theRUNX2-low quartile was 87% compared with 73% for RUNX2 high(Fig. 6C, inset table). A positive correlation between RUNX1-high andRUNX2-high expression was also observed (Supplementary Fig. S5I).Assessment of clinicopathologic characteristics for RUNX2 expressionalso showed no correlation with age, grade, necrosis, and recurrence(Supplementary Table S1). However, similar to RUNX1, RUNX2 alsocorrelated with a high KM score (Fig. 6D and E). These data reveal thatthe related transcription factor RUNX2 is also important in humanccRCC with high expression being indicative of a poorer prognosis.

Taken together, we have identified a novel role for the RUNX familyof transcription factors in kidney cancer where both RUNX1 andRUNX2 are expressed and act in an oncogenic fashion that aids theprogression of the disease.

DiscussionThis study underscores the importance and functional relevance of

the developmentally important transcription factor RUNX1 in kidneycancer. Interrogation of The Cancer Genome Atlas shows RUNX1 tobe upregulated in ccRCC (this study and refs. 35, 39), which we cannow corroborate using histochemical analysis of an independentcohort. Patients with ccRCC with poor prognosis have high RUNX1expression, while deletion of RUNX1 reduced kidney cancer cellgrowth and prevented or delayed tumor development. In addition,we have shown for the first time that RUNX2 is also expressed inpatients with poor prognosis. This work opens up a new and unex-plored avenue of research into theRUNX genes’ enigmatic functions inneoplastic disease and identifies theRUNX genes as novel players in thegenetic landscape of kidney cancer.

Initial gene expression observations in silico revealed that altera-tions in RUNX1 occur at a similar frequency as perturbation of otherkidney cancer drivers. Strikingly, almost all RUNX1 alterations weremRNA upregulations, suggesting increased expression is the mode by

which RUNX1 contributes to ccRCC. RUNX1 activity appears to beassociatedwith the neoplastic state in renal cells as normal tissues showlittle evidence of expression. This may reflect the specific ccRCCtranscriptome, defined by the recurring molecular changes that typifythis disease. Given the importance of hypoxia-induced factor (HIF)activation in the pathogenesis of this disease, it is worth noting that anumber of studies have pointed to the interplay between HIFs andRUNX transcription factors including: physical interaction; coregula-tion of target genes; and in the case of RUNX2, stabilization of the HIFprotein (40–42). Moreover, the RUNX genes are themselves regulatedby HIFs, raising the possibility that they are both downstream targets,and act to potentiate the oncogenic signal (43). Molecular character-ization of RCC subtypes revealed that increased immune cell infiltra-tion gene expression signatures associatedwith the poorest performingpatients specifically in ccRCC (4, 44, 45) and immune checkpointinhibitors are currently in clinical trial for advanced disease (8, 9). Inthis regard, it is interesting that our study reveals a positive correlationbetween RUNX and local inflammatory cell infiltration. Increasedsystemic inflammation is a known feature of renal cancer (46, 47) andintegration of KM score with systemic biomarkers was able to predictpoor prognosis in ccRCC (27). Gene ontology profiles suggest thatimmune and inflammatory processes dominate the expression land-scape of ccRCC (10, 11).

RUNX1 deletion in ccRCC cell lines perturbed the cell cycle andreduced cell viability while Runx1 deletion in our GEM modeldecreased tumor growth and tumor cell proliferation. This geneticallyconfirmed an oncogenic role for RUNX1, which was also highlyupregulated in another murine model of ccRCC (48). There is con-siderable evidence that RUNX1 has an important role in proliferationin organisms as diverse as nematodes (49), sea urchins (50), andmammals; although whether it promotes or restricts cell divisiondepends on the cellular context (51, 52). Similarly RUNX1 is knownto regulate cell survival differently in different cell types (53, 54), anddownstream mediators of survival have been identified in some tissuesystems (55).Our cell line data is given greater physiologic relevance bythe observation that RUNX1-null ccRCC cells almost entirely failed togrow in a kidney xenograft model. Furthermore, our data showing apro-proliferative effect in cell lines and tumors together with enhancedcell survival suggests an exclusively oncogenic role for RUNX1 in thecontext of renal cancer cells.

Using RNA-seq, we revealed that deletion of RUNX1 inducesprofound gene expression changes. KEGG analysis of RCC expressionprofile studies has emphasized that upregulated genes are associatedwith significant cell adhesion changes and interactions between cyto-kines and their receptors (56). In this regard, it is worth noting that ourgene analysis showed enriched expression of genes involved in cell–ECM interactions and cell–cell interactions such as Eph–Ephrinsignaling, suggesting RUNX1 may be contributing to a commononcogenic pathway in renal cancer. One of the most significantlydownregulated genes in our human RUNX1-deleted RCC cells andmurine tumors was SERPINH1. Importantly, SERPINH1 is a potentialnegative prognostic marker in ccRCC (57) and is required for collagenfolding and secretion (58), therefore its expression may contribute tochanges in tissue architecture that promote tumor development.SERPINH1 also associates with enhanced TGFb signaling and bothRUNX1 and RUNX2 have been shown to be involved in TGFb-induced kidney fibrosis (23, 59). The role of collagen in ccRCC iscurrently unclear; however, collagen density and alignment haverecently been shown to be significantly higher in patients withhigh-grade tumors compared with low grade (60). The fatty acidmetabolism enzyme CPT1A increased markedly on RUNX1 deletion,





Figure 6.

RUNX2 is expressed in humanRCCand correlateswith poorer survival.A, Serial sections fromCAP;Runx1þ/þ andCAP;Runx1fl/flmurine kidney tumors immunostainedfor both RUNX1 and RUNX2 showed high RUNX2 expression in both cohorts. Magnified sections are in dashed boxes. Scale bars, 100 mm. CAPmice are AH-Cre;Apcfl/fl;p21�/�. B, Representative examples of RUNX2 staining in nontumor normal kidney and RCC cores from human TMA as used in Fig. 1. A range of RUNX2expression was observed in patients with RCC from negative to high. Magnified areas are in dashed boxes. Scale bars, 100 mm. C, Kaplan–Meier curve showingreduced cancer-specific survival in RUNX2-High patients; log-rank P ¼ 0.0478 and inset life table showing cumulative % survival for 5 years after diagnosis.Wilcoxon, P ¼ 0.045. D, RUNX2 H-Score is significantly higher in patients with a high KM score (KM Low ¼ 5.3; KM High ¼ 16.1; t test, P ¼ 0.0005).E, Proportion of KM-Low and KM-High patients in RUNX2-Low and -High groups, RUNX2-Low KM lo/hi % ¼ 57/43; RUNX2-High KM lo/hi % ¼ 28/72.

Rooney et al.




suggesting a negative correlation. CPT1A is reduced in ccRCC wheresuppression causes lipid droplet accumulation (a prominent feature ofccRCC) and tumor development (61). Intriguingly, suppression ofCPT1A in ccRCC is mediated by the HIF family and is therefore anexample of potential RUNX1 interplay with the VHL–HIF signalingaxis. We also observed a pronounced increase in STATHMIN3, amicrotubule-binding protein important for the formation of mitoticspindles. Overexpression of STATHMIN3 has been associated withdelayed cell cycle in leukemia (62). Future studies using chromatinimmunoprecipitation analysis would provide valuable insights intowhich genes are directlymodulated by RUNX activity and functionallycontribute to the RUNX1-related phenotype in renal cells.

The long-term trend for kidney cancer is one of growing globalincidence, and improved treatments for advanced disease remains anunmet clinical need. Human patients with the highest RUNX1 expres-sion in our study had the poorest prognosis and a 20% reduction insurvival rate at 5 years post diagnosis (68% vs. 88%). Indeed, RUNX1associated with poorer survival independent of age, grade, and stage.These data identify RUNX1 as a novel prognostic biomarker and as apotential therapeutic target in human ccRCC. This is encouraginggiven the active pursuit of therapeutic agents that can block thetranscriptional function of the RUNX proteins (63). However, thewider consequences of directly targeting RUNX in kidney cancerwould need to be established in the context of the sustained require-ment for RUNX function in other tissues.

In general, the relationship between the RUNX genes and otherhematopoietic and solid tumors is complex with both a tumorsuppressor and a pro-oncogenic role described in leukemia (14, 15),breast (18), and prostate cancer (21). As such ccRCC may be theoptimal choice for exploring novel therapeutic agents that blockRUNX function. This is further given credence considering that therelated family member RUNX2 is also associated with poorersurvival and increased inflammation in patients with ccRCC, andwas also highly expressed in our GEM model. Indeed RUNX1 andRUNX2 co-occurred in a selection of patient samples as well as inGEM tumors. Although our staining was carried out on serialsections, dual IHC for both RUNX1 and RUNX2 could give valuableinsights into the spatial localization and consequence of co-occurrence of the RUNX transcription factors. Although beyondthe scope of this study, it will be important to dissect this interplaybetween the RUNX proteins in ccRCC and how they each contrib-ute to the disease phenotype. Furthermore, while we have notspecifically investigated RUNX3 in our system, in silico analysisrevealed it is also upregulated in kidney tumors at the mRNA level.Intriguingly, akin to that observed in pancreatic adenocarcino-ma (64), transcriptomic upregulation of RUNX3 did not relate to

patient survival. Nonetheless, Whittle and colleagues elegantlydemonstrated that high RUNX3 in their pancreatic cancer TMAdid correlate with poor prognosis and conveyed a prometastaticphenotype. Therefore, it will be interesting to study RUNX3 furtherin the context of ccRCC to ascertain whether its role recapitulatesthat seen in pancreatic cancer. Future studies will use compoundgenetic models and anti-RUNX drugs to investigate the conse-quence of total ablation of RUNX function in kidney cancer.

Disclosure of Potential Conflicts of InterestO.J. Sansom reports receiving a commercial research grant from Novartis,

AstraZeneca, and Cancer Research Technology/BMS. No potential conflicts ofinterest were disclosed by the other authors.

Authors’ ContributionsConception and design: N. Rooney, O.J. Sansom, E.R. Cameron, K. BlythDevelopment of methodology: N. Rooney, S.M. Mason, C. Nixon, O.J. Sansom,E.R. Cameron, K. BlythAcquisition of data (provided animals, acquired and managed patients, providedfacilities, etc.):N. Rooney, S.M. Mason, L. McDonald, J.H.M. D€abritz, K.J. Campbell,S. Howard, D. Athineos, J EdwardsAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): N. Rooney, S.M. Mason, K.J. Campbell, A. Hedley,D. Athineos, J.D.G. Leach, J. Edwards, E.R. Cameron, K. BlythWriting, review, and/or revision of the manuscript: N. Rooney, S.M. Mason,K.J. Campbell, W. Clark, J.D.G. Leach, O.J. Sansom, J. Edwards, E.R. Cameron,K. BlythAdministrative, technical, or material support (i.e., reporting or organizing data,constructing databases): S.M. Mason, A. Hedley, C. Nixon, W. Clark, J.D.G. LeachStudy supervision: K. BlythOther (histopathology interpretation): J.D.G. Leach

AcknowledgmentsWe would like to acknowledge the Core Services and Advanced Technologies at

the Cancer Research UK Beatson Institute (C596/A17196), with particular thanks tothe Biological Services Unit, Histology and Molecular Technologies. We thank theNHS Greater Glasgow and Clyde Biorepository for supplying the TMA; TomHamilton and Sandeep Dhayade for assistance with xenograft surgery; ProfessorEyal Gottlieb for providing 786-O and Caki-2 cell lines and David Stevenson foradvice on CRISPR strategy.We thank CatherineWinchester for critical reading of themanuscript. This work was funded by CRUK core funding C596/A17196 (KBand OJS labs - CRUK Beatson Institute) and Renal Cancer Research Fund Scotland(to J. Edwards) who supported the TMA production.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.

Received December 10, 2019; revised February 17, 2020; accepted March 6, 2020;published first March 10, 2020.

References1. Nabi S, Kessler ER, Bernard B, Flaig TW, Lam ET. Renal cell carcinoma: a review

of biology and pathophysiology. F1000Research 2018;7:307.2. Hsieh JJ, Purdue MP, Signoretti S, Swanton C, Albiges L, Schmidinger M, et al.

Renal cell carcinoma. Nat Rev Dis Primers 2017;3:17009.3. Tahbaz R, Schmid M, Merseburger AS. Prevention of kidney cancer inci-

dence and recurrence: lifestyle, medication and nutrition. Curr Opin Urol2018;28:62–79.

4. Ricketts CJ, De Cubas AA, Fan H, Smith CC, LangM, Reznik E, et al. The cancergenome atlas comprehensivemolecular characterization of renal cell carcinoma.Cell Rep 2018;23:313–26.

5. Cancer Genome Atlas Research Network. Comprehensive molecular character-ization of clear cell renal cell carcinoma. Nature 2013;499:43–9.

6. Escudier B, Porta C, Schmidinger M, Rioux-Leclercq N, Bex A, Khoo V, et al.Renal cell carcinoma: ESMO clinical practice guidelines for diagnosis, treatmentand follow-updagger. Ann Oncol 2019;30:706–20.

7. Motzer RJ, Hutson TE, Tomczak P,MichaelsonMD, Bukowski RM, RixeO, et al.Sunitinib versus interferon alfa in metastatic renal-cell carcinoma. N Engl J Med2007;356:115–24.

8. Ghali F, Patel SH, Derweesh IH. Current Status of Immunotherapy forlocalized and locally advanced renal cell carcinoma. J Oncol 2019;2019:7309205. DOI: 10.1155/2019/7309205.

9. Motzer RJ, PenkovK,Haanen J, Rini B, Albiges L, Campbell MT, et al. Avelumabplus axitinib versus sunitinib for advanced renal-cell carcinoma. N Engl J Med2019;380:1103–15.





10. Mitchell TJ, Turajlic S, Rowan A, Nicol D, Farmery JHR, O'Brien T, et al. Timingthe landmark events in the evolution of clear cell renal cell cancer: TRACERxrenal. Cell 2018;173:611–23.

11. Turajlic S, Xu H, Litchfield K, Rowan A, Horswell S, Chambers T, et al.Deterministic evolutionary trajectories influence primary tumor growth: TRA-CERx renal. Cell 2018;173:595–610.

12. Varela I, Tarpey P, Raine K, Huang D, Ong CK, Stephens P, et al. Exomesequencing identifies frequent mutation of the SWI/SNF complex gene PBRM1in renal carcinoma. Nature 2011;469:539–42.

13. Okuda T, vanDeursen J, Hiebert SW,GrosveldG, Downing JR. AML1, the targetof multiple chromosomal translocations in human leukemia, is essential fornormal fetal liver hematopoiesis. Cell 1996;84:321–30.

14. Ito Y, Bae SC,Chuang LS. TheRUNX family: developmental regulators in cancer.Nat Rev Cancer 2015;15:81–95.

15. Sood R, Kamikubo Y, Liu P. Role of RUNX1 in hematological malignancies.Blood 2017;129:2070–82.

16. Ellis MJ, Ding L, Shen D, Luo J, Suman VJ, Wallis JW, et al. Whole-genomeanalysis informs breast cancer response to aromatase inhibition. Nature 2012;486:353–60.

17. Ferrari N, Mohammed ZM, Nixon C, Mason SM, Mallon E, McMillan DC, et al.Expression of RUNX1 correlates with poor patient prognosis in triple negativebreast cancer. PLoS One 2014;9:e100759.

18. Rooney N, Riggio AI, Mendoza-Villanueva D, Shore P, Cameron ER, BlythK. Runx genes in breast cancer and the mammary lineage. In:RUNX proteins in development and cancer. Singapore: Springer Singa-pore; 2017. p. 353–68.

19. Banerji S, Cibulskis K, Rangel-Escareno C, Brown KK, Carter SL, Frederick AM,et al. Sequence analysis of mutations and translocations across breast cancersubtypes. Nature 2012;486:405–9.

20. Riggio AI, Blyth K. The enigmatic role of RUNX1 in female-related cancers -current knowledge & future perspectives. FEBS J 2017;284:2345–62.

21. Takayama K, Suzuki T, Tsutsumi S, Fujimura T, Urano T, Takahashi S,et al. RUNX1, an androgen- and EZH2-regulated gene, has differentialroles in AR-dependent and -independent prostate cancer. Oncotarget2015;6:2263–76.

22. Xiong Z, Yu H, Ding Y, Feng C, Wei H, Tao S, et al. RNA sequencing revealsupregulation of RUNX1-RUNX1T1 gene signatures in clear cell renal cellcarcinoma. Biomed Res Int 2014;2014:450621.

23. Zhou T, Luo M, Cai W, Zhou S, Feng D, Xu C, et al. Runt-related transcriptionfactor 1 (RUNX1) promotes TGF-beta-induced renal tubular epithelial-to-mesenchymal transition (EMT) and renal fibrosis through the PI3K subunitp110delta. EBioMedicine 2018;31:217–25.

24. Roseweir AK, Qayyum T, Lim Z, Hammond R, MacDonald AI, FraserS, et al. Nuclear expression of Lyn, a Src family kinase member, isassociated with poor prognosis in renal cancer patients. BMC Cancer2016;16:229.

25. Lua J, Qayyum T, Edwards J, Roseweir AK. The prognostic role of the non-canonical nuclear factor-kappa B pathway in renal cell carcinoma patients.Urol Int 2018;101:190–6.

26. Kirkegaard T, Edwards J, Tovey S, McGlynn LM, Krishna SN,Mukherjee R, et al.Observer variation in immunohistochemical analysis of protein expression, timefor a change? Histopathology 2006;48:787–94.

27. Qayyum T, McArdle PA, Lamb GW, Going JJ, Orange C, Seywright M, et al.Prospective study of the role of inflammation in renal cancer. Urol Int 2012;88:277–81.

28. Valente MJ, Henrique R, Costa VL, Jer�onimo C, Carvalho F, Bastos ML, et al.A rapid and simple procedure for the establishment of human normal andcancer renal primary cell cultures from surgical specimens. PLoS One 2011;6:e19337.

29. Fisher S, Barry A, Abreu J, Minie B, Nolan J, Delorey TM, et al. A scalable, fullyautomated process for construction of sequence-ready human exome targetedcapture libraries. Genome Biol 2011;12:R1.

30. Church DM, Schneider VA, Graves T, Auger K, Cunningham F, Bouk N, et al.Modernizing reference genome assemblies. PLoS Biol 2011;9:e1001091.

31. Cole AM, Ridgway RA, Derkits SE, Parry L, Barker N, Clevers H, et al. p21 lossblocks senescence following Apc loss and provokes tumourigenesis in the renalbut not the intestinal epithelium. EMBO Mol Med 2010;2:472–86.

32. Growney JD, Shigematsu H, Li Z, Lee BH, Adelsperger J, Rowan R, et al. Loss ofRunx1 perturbs adult hematopoiesis and is associated with a myeloproliferativephenotype. Blood 2005;106:494–504.

33. Ferrari N, Riggio AI, Mason S, McDonald L, King A, Higgins T, et al. Runx2contributes to the regenerative potential of the mammary epithelium. Sci Rep2015;5:15658.

34. Brugarolas J. Molecular genetics of clear-cell renal cell carcinoma. J Clin Oncol2014;32:1968–76.

35. Fu Y, Sun S, Man X, Kong C. Increased expression of RUNX1 in clear cell renalcell carcinoma predicts poor prognosis. PeerJ 2019;7:e7854.

36. Nagy A, Lanczky A, Menyhart O, Gyorffy B. Validation of miRNA prognosticpower in hepatocellular carcinoma using expression data of independentdatasets. Sci Rep 2018;8:9227.

37. Hong D, Messier TL, Tye CE, Dobson JR, Fritz AJ, Sikora KR, et al. Runx1stabilizes the mammary epithelial cell phenotype and prevents epithelial tomesenchymal transition. Oncotarget 2017;8:17610–27.

38. Keita M, Bachvarova M, Morin C, Plante M, Gregoire J, Renaud MC, et al. TheRUNX1 transcription factor is expressed in serous epithelial ovarian carcinomaand contributes to cell proliferation, migration and invasion. Cell Cycle 2013;12:972–86.

39. Qin S, Shi X, Wang C, Jin P, Ma F. Transcription factor and miRNAinterplays can manifest the survival of ccRCC patients. Cancers 2019;11.DOI: 10.3390/cancers11111668.

40. Peng ZG, Zhou MY, Huang Y, Qiu JH, Wang LS, Liao SH, et al. Physical andfunctional interaction of Runt-related protein 1 with hypoxia-inducible factor-1alpha. Oncogene 2008;27:839–47.

41. Kwon TG, Zhao X, Yang Q, Li Y, Ge C, Zhao G, et al. Physical and functionalinteractions between Runx2 and HIF-1alpha induce vascular endothelial growthfactor gene expression. J Cell Biochem 2011;112:3582–93.

42. Lee SH, Che X, Jeong JH, Choi JY, Lee YJ, Lee YH, et al. Runx2 protein stabilizeshypoxia-inducible factor-1alpha through competition with von Hippel-Lindauprotein (pVHL) and stimulates angiogenesis in growth plate hypertrophicchondrocytes. J Biol Chem 2012;287:14760–71.

43. Tamiya H, Ikeda T, Jeong JH, Saito T, Yano F, Jung YK, et al. Analysis of theRunx2 promoter in osseous and non-osseous cells and identification ofHIF2A asa potent transcription activator. Gene 2008;416:53–60.

44. Chen F, Zhang Y, Senbabaoglu Y, Ciriello G, Yang L, Reznik E, et al.Multilevel genomics-based taxonomy of renal cell carcinoma. Cell Rep2016;14:2476–89.

45. Geissler K, Fornara P, Lautenschlager C, Holzhausen HJ, Seliger B, Riemann D.Immune signature of tumor infiltrating immune cells in renal cancer. Oncoim-munology 2015;4:e985082.

46. Chang Y, An H, Xu L, Zhu Y, Yang Y, Lin Z, et al. Systemic inflammation scorepredicts postoperative prognosis of patients with clear-cell renal cell carcinoma.Br J Cancer 2015;113:626.

47. de Vivar Chevez AR, Finke J, Bukowski R. The role of inflammation in kidneycancer. Adv Exp Med Biol 2014;816:197–234.

48. Harlander S, Schonenberger D, Toussaint NC, Prummer M, Catalano A, BrandtL, et al. Combined mutation in Vhl, Trp53 and Rb1 causes clear cell renal cellcarcinoma in mice. Nat Med 2017;23:869–77.

49. Nimmo R, Antebi A, Woollard A. mab-2 encodes RNT-1, a C. elegans Runxhomologue essential for controlling cell proliferation in a stem cell-like devel-opmental lineage. Development 2005;132:5043–54.

50. Robertson AJ, Coluccio A, Knowlton P, Dickey-Sims C, Coffman JA. Runxexpression is mitogenic andmutually linked toWnt activity in blastula-stage seaurchin embryos. PLoS One 2008;3:e3770.

51. Friedman AD. Cell cycle and developmental control of hematopoiesis by Runx1.J Cell Physiol 2009;219:520–4.

52. Hoi CS, Lee SE, Lu SY, McDermitt DJ, Osorio KM, Piskun CM, et al. Runx1directly promotes proliferation of hair follicle stem cells and epithelial tumorformation in mouse skin. Mol Cell Biol 2010;30:2518–36.

53. Blyth K, Slater N, Hanlon L, Bell M, Mackay N, Stewart M, et al. Runx1promotes B-cell survival and lymphoma development. Blood Cells Mol Dis2009;43:12–9.

54. Cai X, Gao L, Teng L, Ge J, Oo ZM, Kumar AR, et al. Runx1 deficiency decreasesribosome biogenesis and confers stress resistance to hematopoietic stem andprogenitor cells. Cell Stem Cell 2015;17:165–77.

55. Kilbey A, Terry A, Wotton S, Borland G, Zhang Q, Mackay N, et al. Runx1orchestrates sphingolipid metabolism and glucocorticoid resistance in lympho-magenesis. J Cell Biochem 2017;118:1432–41.

56. Xing Q, Huang Y, Wu Y, Ma L, Cai B. Integrated analysis of differentiallyexpressed profiles and construction of a competing endogenous long non-codingRNA network in renal cell carcinoma. PeerJ 2018;6:e5124.

Cancer Res; 80(11) June 1, 2020 CANCER RESEARCH2338

Rooney et al.

Cancer Research. by guest on August 31, 2020. Copyright 2020 American Association forhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from


57. Qi Y, Zhang Y, Peng Z, Wang L, Wang K, Feng D, et al. SERPINH1 over-expression in clear cell renal cell carcinoma: association with poor clinicaloutcome and its potential as a novel prognostic marker. J Cell Mol Med2018;22:1224–35.

58. Ito S, Nagata K.Biology of Hsp47 (Serpin H1), a collagen-specific molecularchaperone. Semin Cell Dev Biol 2017;62:142–51.

59. Kim JI, Jang HS, Jeong JH, Noh MR, Choi JY, Park KM. Defect in Runx2 geneaccelerates ureteral obstruction-induced kidney fibrosis via increased TGF-betasignaling pathway. Biochim Biophys Acta 2013;1832:1520–7.

60. Best S, Liu Y, Keikhosravi A, Drifka C, Woo K, Mehta G, et al. Collagenorganization of renal cell carcinoma differs between low and high grade tumors.BMC Cancer 2019;19:490.

61. DuW, Zhang L, Brett-Morris A, Aguila B, Kerner J, Hoppel CL, et al. HIF driveslipid deposition and cancer in ccRCC via repression of fatty acid metabolism.Nat Commun 2017;8:1769.

62. Rubin CI, AtwehGF. The role of stathmin in the regulation of the cell cycle. J CellBiochem 2004;93:242–50.

63. Illendula A, Gilmour J, Grembecka J, Tirumala VSS, Boulton A,Kuntimaddi A, et al. Small molecule inhibitor of CBFbeta-RUNXbinding for RUNX transcription factor driven cancers. EBioMedicine2016;8:117–31.

64. Whittle MC, Izeradjene K, Rani PG, Feng L, Carlson MA, DelGiorno KE, et al.RUNX3 controls a metastatic switch in pancreatic ductal adenocarcinoma. Cell2015;161:1345–60.

AACRJournals.org Cancer Res; 80(11) June 1, 2020 2339


Cancer Research. by guest on August 31, 2020. Copyright 2020 American Association forhttps://bloodcancerdiscov.aacrjournals.orgDownloaded from


runx1 is a driver of renal cell ... - cancer research · kidney cancer is the seventh commonest...

Documents