<a class="color-blue" href="https://cupdf.com/category/documents" title="Category Documents">Documents

Research Article

Chemoprevention with Cyclooxygenase andEpidermal Growth Factor Receptor Inhibitors inFamilial Adenomatous Polyposis Patients: mRNASignatures of Duodenal NeoplasiaDon A. Delker1, Austin C.Wood2, Angela K. Snow2, N. Jewel Samadder1,2,Wade S. Samowitz2,3, Kajsa E. Affolter2,3, Kenneth M. Boucher1,2, Lisa M. Pappas2,Inge J. Stijleman2, Priyanka Kanth1, Kathryn R. Byrne1, Randall W. Burt1,2,Philip S. Bernard2,3, and Deborah W. Neklason1,2

Abstract

To identify gene expression biomarkers and pathways tar-geted by sulindac and erlotinib given in a chemopreventiontrial with a significant decrease in duodenal polyp burden at6 months (P < 0.001) in familial adenomatous polyposis(FAP) patients, we biopsied normal and polyp duodenaltissues from patients on drug versus placebo and analyzedthe RNA expression. RNA sequencing was performed onbiopsies from the duodenum of FAP patients obtained atbaseline and 6-month endpoint endoscopy. Ten FAP patientson placebo and 10 on sulindac and erlotinib were selected foranalysis. Purity of biopsied polyp tissue was calculated fromRNA expression data. RNAs differentially expressed betweenendpoint polyp and paired baseline normal were determinedfor each group and mapped to biological pathways. Key genesin candidate pathways were further validated by quantitativeRT-PCR. RNA expression analyses of endpoint polyp com-

pared with paired baseline normal for patients on placeboand drug show that pathways activated in polyp growth andproliferation are blocked by this drug combination. Directlycomparing polyp gene expression between patients on drugand placebo also identified innate immune response genes(IL12 and IFNg) preferentially expressed in patients on drug.Gene expression analyses from tissue obtained at endpoint ofthe trial demonstrated inhibition of the cancer pathwaysCOX2/PGE2, EGFR, and WNT. These findings provide molec-ular evidence that the drug combination of sulindac anderlotinib reached the intended tissue and was on target forthe predicted pathways. Furthermore, activation of innateimmune pathways from patients on drug may have contrib-uted to polyp regression. Cancer Prev Res; 11(1); 4–15. �2017AACR.

See related editorial by Shureiqi, p. 1

IntroductionFamilial adenomatous polyposis (FAP) is an autosomal dom-

inant inherited disorder due to germline mutations in the APC(adenomatous polyposis coli) gene (1, 2). FAP is characterized bythe formation of hundreds to thousands of adenomatous polypsin the colorectum and a nearly 100% lifetime risk of colorectalcancer if left untreated (3). Prophylactic colectomy has becomethe standard of care once the extent of colorectal polyposis isbeyond endoscopic control. FAP patients are also at greatlyincreased risk for duodenal neoplasia, with duodenal adenomaseventually forming in >50% of FAP patients and duodenal ade-

nocarcinoma occurring in up to 12% (3–8). As mutations in theAPC gene are central to the initiation and development of colo-rectal cancer with 80% of sporadic colorectal cancers having APCloss or inactivation, FAP is an excellent model to study themolecular events leading to development of colorectal cancerand other intestinal cancers.

Chemoprevention studies in FAP patients can provide clues asto how drugs modify tumor progression. Multiple studies haveshown that sulindac, a cyclooxygenase (COX) inhibitor andNSAID, significantly inhibits colorectal adenomatous polyps inFAP patients (9, 10). However, sulindac has failed to show a signi-ficant reduction in duodenal adenomas in FAP patients (11, 12).This is thought to be due to increased COX2 expression in theduodenal tissue compared with colonic tissue of FAP patients(13). Studies have suggested that APC inactivation and EGFRsignaling promote COX2 expression, leading to the developmentof intestinal neoplasms (14, 15). The convergence between theWNT and EGFR signaling pathways and COX2 activity was dem-onstrated in a mouse model of FAP in which a combination of aCOX and an EGFR inhibitor diminished small intestinal adenomadevelopment by 87% (16). These results led us to test the hypoth-esis that a combination of COX and EGFR inhibition wouldimpede adenoma formation in the duodenum of subjects withFAP. We recently reported on a positive clinical study where

1Department of Internal Medicine, University of Utah, Salt Lake City, Utah.2Huntsman Cancer Institute, University of Utah, Salt Lake City, Utah. 3Depart-ment of Pathology, University of Utah, Salt Lake City, Utah.

Note: Supplementary data for this article are available at Cancer PreventionResearch Online (http://cancerprevres.aacrjournals.org/).

Corresponding Author: Deborah W. Neklason, Huntsman Cancer Institute atUniversity of Utah, 1950 Circle of Hope, Salt Lake City, UT 84112. Phone: 801-587-9882; Fax: 801-585-5763; E-mail: [email protected]

doi: 10.1158/1940-6207.CAPR-17-0130

�2017 American Association for Cancer Research.

CancerPreventionResearch

Cancer Prev Res; 11(1) January 20184

Research. on July 25, 2020. © 2018 American Association for Cancercancerpreventionresearch.aacrjournals.org Downloaded from

Published OnlineFirst November 6, 2017; DOI: 10.1158/1940-6207.CAPR-17-0130

http://crossmark.crossref.org/dialog/?doi=10.1158/1940-6207.CAPR-17-0130&domain=pdf&date_stamp=2017-12-18

http://cancerpreventionresearch.aacrjournals.org/

patients with FAP were treated with either placebo or sulindac anderlotinib (sulindac–erlotinib). At 6 months, the median totalduodenal polyp burden had increased by 6 mm from baseline inthe placebo arm and decreased by 9 mm in the sulindac–erlotinibarm (P < 0.001; ref. 17).

Here, we report the gene expression analyses from duodenaltissue at endpoint compared with baseline for those subjectsenrolled in the trial. We show that the EGFR and COX2 pathwaysare activated in duodenal polyps and that the drug combinationof sulindac–erlotinib blocks this activation. In addition,we foundevidence for activation of IFNg and IL12 signaling pathways,suggesting that the recruitment of both Th1 and natural killer(NK) T cells may have contributed to the polyp regression (sizeand number) observed in the drug-treated arm (18, 19).

Materials and MethodsPatient cohort

This studywas approved by theUniversity of Utah InstitutionalReview Board (IRB#39278), conducted in accordance with rec-ognized ethical guidelines, and informed consent was obtainedfrom each subject. A randomized, two-arm chemoprevention trialwas conducted between July 2010 and June 2014 in which FAPpatientswere treated eitherwith placeboorwith sulindac (150mgtwice daily) plus erlotinib (75 mg/day) for 6 months (registeredwith ClinicalTrials.gov as NCT 01187901). Seventy-three indivi-duals with FAP completed the study. For patients with significantdiscomfort or evidence of toxicity, the dose was lowered duringthe course of the study as described previously (17). Endoscopicduodenal biopsies were taken for grossly uninvolved tissue atbaseline and endpoint, while polyps were obtained at endpointonly. Tissues were placed in RNAlater Stabilization Solution(Thermo Fisher Scientific).

Sample selection and RNA isolationA subset of tissue from 10 research participants who responded

to the drug combination (sulindac–erlotinib) and 10 researchparticipants who progressed on placebo was selected for molec-ular analyses (Table 1). Total RNA was prepared from biopsiesusing a Qiagen RNeasy Mini Kit (Qiagen #74106) following the

manufacturer's instructions and including the on-column RNase-free DNase treatment. RNA quantity and quality was determinedusing a Thermo Fisher Scientific NanoDrop Spectrophotometerand Agilent Bioanalyzer.

RNA sequencingFifty-two endpoint (20 uninvolved, 32 adenomas) and 17

baseline (all uninvolved) total RNA samples were treated withRiboZero Gold (Illumina) to remove ribosomal RNA prior tocDNA library preparation using the Illumina TruSeq StrandedTotal RNA Sample Prep Protocol. PCR-amplified libraries weresequenced on Illumina HiSeq 2500 instrument using 50-cyclesingle-read chemistry. Sequencing datasets are deposited to NCBIGEO submission #GSE94919. Sequencing reads were aligned totheRefSeqHg38human genomeusing theNovoalign application(Novocraft). Differentially expressed genes were calculated usingthe USeq application "DefinedRegionsDifferentialSeq" (DRDS)as described previously (20, 21). The DRDS application usesDESeq2 negative binomial statistics together with a BenjaminiandHochberg FDR to identify differentially expressed genes (22).For paired sample comparisons, the DESeq2 paired analysisapplication in R was used.

Total RNA from 12 endpoint polyps (6 from placebo and 6from drug endpoint) of similar 3 to 4 mm size, noted in Table 1,was further evaluated by targeted gene expression using theHuman Inflammation and Immunity Transcriptome 475 genepanel (Qiagen # RHS-005Z). This targeted approach improvessequencing depth and reduces biases in amplification. The panelincludes a diverse set of cytokines, growth factors, and transcrip-tion factors important in mediating general and specializedimmune responses. RNA is converted to cDNA and amplifiedwith a multiplex primer panel that labels each cDNA moleculewith a unique molecular tag. The quantified indexed library DNAwas pooled to an equimolar concentration alongside other sam-ples. The pooled libraryDNAwas amplified by emulsion PCR andenriched for positive ion sphere particles (ISP) using the IonTorrent One Touch System II (Life Technologies) and the Ion PIHi-QOT2Kit (Life Technologies).Templated ISPswere sequencedon a PIv3 micro-chip using the Ion Torrent Proton Machine (LifeTechnologies) and the Ion PI Hi-Q Sequencing 200 Kit (LifeTechnologies) for 130 cycles (520 flows). Sequence reads werealigned to the RefSeq Hg38 human genome using the STAR RNAreadmapper. Reads that aligned tomore than 3 sites in the humangenome or did not have at least 60 bp aligned were excluded.Differentially expressed genes between treatment groups weredetermined using the QIASeq secondary data analysis tool (Qia-gen). Student t test was applied after data normalization usingtotal molecular tag counts.

RNA quantification by qRT-PCRThe expression of MMP7, CD44, FOS, TM4SF5, EGFR, and

PTGS2 was confirmed by qRT-PCR. cDNA was synthesized usingthe SuperScript VILO cDNA Synthesis Kit (Invitrogen#11754250) following the manufacturer's instructions. Prede-signed IDT primer-probe sets were used together with PrimeTimeGene Expression Master Mix (IDT #1055771) to measure expres-sion of the following genes: CD44 (IDT #Hs.PT.58.2004193),EGFR (IDT #Hs.PT.58.20590781), FOS (IDT #Hs.PT.58.15540029), PTGS2 (#Hs.PT.58.77266), MMP7 (IDT #Hs.PT.58.40068681), and TM4SF5 (IDT #Hs.PT.58.39684117). Rel-ative gene expression was determined after normalization to

Translational Relevance

Familial adenomatous polyposis (FAP) patients have a100% lifetime risk of colorectal cancer and are also at increasedrisk for duodenal neoplasia, with duodenal adenomas even-tually forming in >50% of FAP patients. In most cases, pro-phylactic colectomy and frequent endoscopy is the standard ofcare for these patients. We recently completed a phase IIclinical trial of sulindac and erlotinib in a FAP patient cohortand found a significant decrease in duodenal polyp burden at6months for patients ondrug versus placebo.Here, we presentthe duodenal polyp RNA expression data from this cohort,which show almost complete inhibition of the tumor signal-ing pathways WNT, EGFR, and COX2/PGE2 and activation ofinnate immunity signaling pathways IL12 and IFNg in ade-nomas from patients on drug. The drug combination ofsulindac and erlotinib provides a promising alternative treat-ment of duodenal polyps in FAP patients.

Sulindac–Erlotinib Effect on Duodenal Polyp mRNA Expression

www.aacrjournals.org Cancer Prev Res; 11(1) January 2018 5




Table

1.Cha

racteristics

ofparticipan

tsan

dbiopsies

stud

ied

Duo

den

alpolypburden

Drugex

posure

End

oscopic

tissue

sformRNAan

alysis

Patient

number

Germlin

eAPC

mutation

Sample

number

Sex

Age

Arm

Base

sum

diam

End

point

sum

diam

Cha

nge

Avg

daily

dose

erlotinib(m

g)

Adve

rseev

ents

Baseline

uninvo

lved

End

point

uninvo

lved

End

point

polyp(s)

size

(inmm)

Tumorpurity

1c.39

27_39

31del5

008

Fem

ale

41

Drug

95

32�6

6%

47

Mucositis(1)

Yes

Yes

4a

0.67

2c.1660C>T

017

Male

45

Drug

5626

�54%

58Rash(1)

Yes

Yes

4a;5

;6;6

0.52;0.79;0

.80;0

.85

3c.426

_427

delAT

028

Male

55Drug

225

�77%

70Rash(1)

Yes

Yes

20.56

4c.1690C>T

027

Fem

ale

40

Drug

85

31�6

4%

46

Rash(1)

Yes

Yes

3a,b

0.46

5c.426

_427

delAT

055

Male

53Drug

134

�69%

27Mucositis(1);rash

(1)

Yes

Yes

2b0.47

6clinical

dx

073

Fem

ale

47

Drug

42

23�4

5%45

Rash(1)

Yes

Yes

3a0.51

7c.426

_427

delAT

155

Fem

ale

58Drug

63

23�6

3%49

Mucositis(1);rash

(1)

No

Yes

3a0.7

8del

promoter1B

173

Fem

ale

44

Drug

49

28�4

3%21

Rash(1)

Yes

Yes

3a0.54

9del

promoter1B

176

Male

28Drug

134

�69%

21Mucositis(2);rash

(2)

Yes

Yes

20.57

10c.4612_4613delGA

005

Male

56Drug

700

700

0%c

75Mucositis(2);rash

(1)

Yes

Yes

6;6

;6;6

0.76;0

.78;0

.67;0.70

11c.53

1þ2_

531þ

3insT

021

Male

52Placebo

61

163

167%

NA

Yes

Yes

8;6

0.68;0.71

12c.20

93T

>G024

Male

46

Placebo

1225

108%

NA

Rash(1)

Yes

Yes

20.67

13c.20

93T

>G031

Fem

ale

56Placebo

2251

132%

NA

Mucositis(1)

Yes

Yes

4b;4

a0.49;0

.65

14del

exon11-18

072

Fem

ale

50Placebo

5583

51%

NA

Yes

Yes

3a0.54

15c.694C>T

080

Fem

ale

37Placebo

2061

205%

NA

Yes

Yes

2b0.44

16c.904C>T

095

Male

59Placebo

107

185

73%

NA

Yes

Yes

4b;4

a0.49;0

.64

17del

promoter1B

150

Male

54Placebo

112

181

62%

NA

No

Yes

4a;4

0.63;

0.58

18c.4612_4613delGA

151

Fem

ale

38Placebo

3898

158%

NA

No

Yes

3a0.55

19c.53

1þ2_

531þ

3insT

175

Male

19Placebo

1221

75%

NA

Yes

Yes

4a

0.8

20c.39

27_39

31del5

013

Fem

ale

58Placebo

98

135

38%

NA

Yes

Yes

8,6

,40.66;0

.72;

0.75

NOTE:T

he"sum

diam"refers

tothesum

ofthediameter

ofallp

olypsen

doscopically

iden

tified

asdescribed

previously

(17).

aIndicates

3to

4mm

polypused

inHum

anInflam

mationan

dIm

mun

ityTranscriptomepan

el(6

drug;6

placebo).

bIndicates

polypexclud

edfordifferentialexpressionan

alysis(<50

%tumorpurity;

2drug;3placebo).

c Patient

10was

ondrugan

dha

dno

chan

gein

sum

diameter

but

areductionin

volumedue

toachan

gein

polyphe

ight.

Delker et al.

Cancer Prev Res; 11(1) January 2018 Cancer Prevention Research6




the geometric mean expression of the internal control genesUBC (IDT #Hs.PT.39a.22214853), GAPDH (IDT #Hs.PT.58.40035104), and KRR1 (IDT #Hs.PT.58.4223891). Eachgene was run in triplicate on a Bio-Rad CFX96 Real-Time PCRSystem. Data analysis was performed with the Bio-Rad CFXManager software. GraphPad Prism 7 was used for plottingqRT-PCR results and for statistical analysis.

Bioinformatic analysesTumor purity was assessed using the R application Estimation

of Stromal and Immune cells in Malignant Tumors using Expres-sion data (ESTIMATE) as described previously (23, 24). Normal-ized RPKMvalues from all expressed genes (�10 reads) were usedto determine RNA expression from immune and stromal cells ineach of the 32 duodenal polyp RNA sequencing (RNA-Seq)datasets. Percent tumor purity was then calculated by combiningimmune and stromal scores.

Principal component analysis (PCA) and hierarchical cluster-ing was used to identify sets of differentially expressed genes thatseparate samples based on placebo versus drug treatment(25, 26). Ingenuity Pathway Analysis (IPA) was used to predictsignaling pathways changing between the groups. Signaling path-ways regulated by known transcription factors, cytokines, growthfactors, and kinases (upstream regulators) were identified using aFisher exact test.

We determined a z-score using IPA that infers the activationstate of an upstream regulator based on the direction of foldchange of its target genes (27). A z-score �2 was consideredstatistically significant.

IHCFormalin-fixed paraffin-embedded (FFPE) polyp tissue from

10 patients on placebo (19 duodenum, 3 colon) and 10 patientson drug (7 duodenum and 3 colon) resected for clinical patho-logic evaluation were stained for CD56 (a marker for NK cells)according to ARUP clinical laboratory test number 2003589.Briefly, sections were cut at 4mm, melted at 60�C for 30 minutes,stained on the BenchMark Ultra (VentanaMedical Systems) usingthe CD56 mouse mAb, clone 123C3.D5 (Abgent, catalog #AH10009) at a dilution of 1:40, and detected using the UltraViewUniversal DAB Detection Kit (Roche). The sections were counter-stained with hematoxylin. CD56þ cells were counted in up to 10high-power fields (HPF, 40�). Some polyp tissue had less than 10HPF of dysplastic tissue. Associations of CD56 counts withreceiving drug treatment were estimated through Poisson regres-sion, offset by the log of HPFs present. An alpha level of 0.05 wasused to determine statistical significance. SAS 9.4 was used forstatistical analysis. Using a likelihood ratio test comparing Pois-son and negative binomial models, we found that our data areoverdispersed using the Poisson model, so a negative binomialmodel was used.

ResultsThe primary objective of this study was to determine whether

the sulindac–erlotinib drug combination was affecting theintended molecular targets/pathways in duodenal polyps. Duo-denal tissue from 20 individuals, 10 from each arm of the trial,was selected on the basis of change in duodenal polyp burdenover the 6-month investigation period. Some patients underwentdose reduction, so the average consumed daily dose is reported

(Table 1). In addition, we report the reported adverse events, theindividual polyps used from the patients, reported as size in mm,and the corresponding calculated tumor purity.

Gene expression comparisons from RNA-SeqThe average number of unique aligned reads per sample was

14.8 million. Approximately 74% of human RefSeq genes(18,698/25,199) had a minimum of 10 exonic reads in one ormore samples and were therefore considered expressed in thehumanduodenum. Figure 1 is a schematic showing thenumber ofsubjects and samples used to identify differences in gene expres-sion between: (i) baseline uninvolved and endpoint uninvolved;(ii) baseline uninvolved and endpoint polyp; and (iii) endpointuninvolved and endpoint polyp. Overall, the dynamic range ofgene expression observed across these relatively small adenomas(<10 mm) was less than typically seen in invasive cancers (28).Thus, a 2.0-fold change cutoff was used to select genes thatdistinguish between subjects treated with placebo versus drug.A 1.5-fold change was used for analyses involving pathwaydiscovery.

Endpoint versus baseline comparisons. PCA of differentiallyexpressed genes from endpoint versus baseline comparisonsshowed clear separation of polyps from patients on drug versuspolyps from patients on placebo (Fig. 2A). Principal compo-nent 1 (PC1) accounted for 20% of the variability observed inthe data and separated polyps from patients on drug frompatients on placebo. Principal component 2 (PC2) accountedfor 14% of the variation in the data and separated 2 polyps(6mmA_017 and 6mmB_017), bottom center, from patient 2on drug, from the other 11 polyps from patients on drug. Thesetwo polyps may represent acquired resistance to the drug.Principal component 3 accounted for 8% of the variation inthe gene expression data.

Hierarchical clustering analysis of the top genes that composeprincipal components 1 and 2 and are in the WNT signalingpathway (14 genes), PGE2 pathway (10 genes), or EGFR pathway(10 genes) are shown in Fig. 2B. Endpoint versus baseline com-parisons showa separation of polyps frompatients ondrug versuspolyps from patients on placebo. The gene expression patterns ofthese genes suggest that the drug combination blocks adenomaprogression through WNT signaling and specifically blocks thetargeted PGE2 and EGFR signaling processes. Two outlier polypsidentified from patient 2 (D6mmB_017 and D6mmA_017) inPCA showed high expression of a subset of genes (MMP7, AXIN2,CXCL5, EGR1, FOS; Fig. 2B), suggesting a lack of drug effect, andpossible drug resistance in these polyps.

When comparing endpoint samples with paired baseline unin-volved duodenal tissue, 3 patients (one on drug and two onplacebo) were excluded due to a lack of baseline tissue (Table 1).Five polyps were also excluded from the final analysis becausetumor purity was estimated at <50% (Table 1; Supplementary Fig.S1). Differential gene expression with the <50% tumor purityincluded and excluded is presented in Supplementary Table S1.When comparing endpoint polyp samples with paired baselineuninvolved duodenum from patients randomized to placebo, weidentified 977 differentially expressed genes (fold change � 2.0;FDR < 0.05) by DESeq2 analysis (Supplementary Table S1). Thesame comparison from patients on drug yielded only 51 differ-entially expressed genes, suggesting that the drug combinationrestores the normal duodenal biology (Supplementary Table S1).






Comparing endpoint uninvolved duodenum with paired base-line uninvolved duodenum, we identified 1 differentiallyexpressed gene in patients on sulindac–erlotinib and 11 differ-entially expressed genes in patients on placebo (SupplementaryTable S1).

Endpoint only comparison. To evaluate whether there were drug-specific effects on normal tissue, we compared endpoint unin-volved duodenum between patients on sulindac–erlotinib andpatients on placebo (Supplementary Table S2). No differentiallyexpressed genes (fold change�2.0; FDR<0.05)were found.Onlyone differentially expressed gene, NANOS3, was found compar-ing endpoint adenomas with paired endpoint uninvolved duo-denum from drug-treated patients. In contrast, 493 differentiallyexpressed genes were found comparing endpoint adenomas withpaired endpoint uninvolved duodenum frompatients onplacebo(Supplementary Table S2). Again, this set of differentiallyexpressed genes includes multiple known genes involved inadenoma polyp progression, includingCD44,MMP7, andCEMIP(also known as KIAA1199; ref. 29).

qRT-PCR validation. CD44, MMP7, FOS, TM4SF5, EGFR, andPTGS2 gene expression, representing the three target signalingpathways WNT, EGFR, and COX2, were evaluated by qRT-PCRusing PrimeTime gene expression assays (IDT; Fig. 3). Relative

gene expression was determined by the 2�DDCt method usingGAPDH, KRR1, andUBC as the internal control genes (30).CD44andMMP7 show a significant increase in gene expression in polypas compared with normal from the placebo group (P < 0.05),consistent with previous reports (Fig. 3; refs. 29, 31). In polypsfrom subjects on placebo, we observe upregulation of EGFRmRNA, a major effector after APC transformation (32, 33). Wealsoobserve downregulation ofTM4SF5, whichhas been reportedto be associated with elevated TM4SF5 protein expression, sug-gesting activation of a previously observed feedback mechanisminvolving proteasome inhibition in response to elevated TM4SF5protein levels (34). Notably for all six genes, there is no significantchange in gene expression of normal versus polyp from the druggroup. When comparing gene expression between polyps fromsubjects on placebo versus subjects on drug, the FOS gene is theonly one to show a significant difference (P < 0.05). The twopolyps from patient 2 may represent acquired resistance to thedrug, as bothhadhigher FOS expression andwere outliers (Fig. 2).Certain genes, such as PTGS2, had wide variation in gene expres-sion; thus, additional confirmation was performed using qRT-PCR. The different platforms had high correlation in measure-ment, suggesting biologic (and not experimental) variability(Supplementary Fig. S2). Previous studies evaluating the levelsof PTGS2mRNA in intestinal polyps have beenmixed, with somestudies showing elevated RNA levels (35, 36) and others showing

20 patients, 10 on sulindac–erlotinib and 10 on placeboevaluated for gene expression changes in duodenal

polyps and uninvolved tissue (Table 1)

5 polyps excluded withpolyps <50% purity

Supplementary Fig. 1

Paired statistics with 17patients with both baseline

and endpoint samples

Baseline uninvolved• vs. endpoint uninvolved (17 patients, 17 tissues)• vs. endpoint polyp (12 patients, 23 tissues)Fig. 2, Fig. 3, SupplementaryTable 1

Endpoint uninvolved• vs. endpoint polyp (18 patients, 27 tissues)

Identification and validation of key genes and pathwaysaffected on sulindac–erlotinib therapyTable 2, Table 3, Supplementary Table 3

Supplementary Table 2

Paired statistics with 18 patientsendpoint samples

Figure 1.

Flow diagram of patient samplesused for gene expression analysis.Duodenal polyp and uninvolvedtissue samples from 10 patients onsulindac–erlotinib (drug) and10 patients on placebo were used forgene expression analysis.

Delker et al.





no change compared with uninvolved tissue (37), which isconsistent with the major regulation of COX2 being posttransla-tional (38). Even in studies showing statistical differences in thelevel of PTGS2, induction is significantly varied across subjects,which is similar to our findings in duodenal polyps from FAPpatients.

Pathway discoveryThe differentially expressed list of 2,637 genes representing

1.5-fold differential expression and FDR <0.05 (SupplementaryTable S1) were uploaded into IPA software (2,591 of the genesannotated in IPA) to identify the pathways affected in duodenaltissue when normal epithelia becomes a polyp. The software alsoidentified which of these pathway changes were repressed bysulindac–erlotinib treatment. Multiple known signaling path-ways important in adenoma development were predicted to beactivated in adenomas from patients on placebo, includingCTNNB1 (WNT), EGFR, TNF, and PGE2 pathways (Table 2;Supplementary Table S3). In contrast, adenomas from patientsondrug showed almost complete loss of cancer pathway signalingin agreement with the reduction in polyp burden observed inthese patients. The z-score predicts the activation state of theupstream regulator based on the direction of fold change of itsgene targets. A z-score �2 suggests the upstream regulator issignificantly activated. The Fisher exact P value is based on theoverlap of differentially expressed genes and known regulatortargets with no information regarding activation state. We alsoobserved inhibition of multiple signaling pathways in polypsfrom patients on placebo that include pathways related to intes-tinal development and tumor suppression (CDX2) (39, 40),heterochromatin assembly (CBX5), cell adhesion (CTNNA) andIFN signaling (IFNa and IFNg).

IPA analysis also revealed immune signaling pathways that aredownregulated in placebo polyps and not downregulated in drugpolyps, most notably IFNa and IFNg . Because PGE2, a product ofCOX2 enzyme, is potent inducer of IL10, which in turn suppressesIFNg signaling (41), this observation would be consistent withincreased COX2 expression in placebo polyps. Similarly, polypsfrom patients treated with sulindac–erlotinib would not haveCOX2 overexpression and would not have suppression of IFNg .To further explore this finding, an inflammation and immunitytranscriptome panel was run on 6 placebo polyps and 6 drugpolyps to enhance coverage. These results confirmed that IFNa,IFNg , and IL12 are more active in polyps from patients on drug,whereas PGE2 is less active in polyps from patients on drug(Table 3; Supplementary Tables S4 and S5).

The RNA expression data suggest activation of the specific T-celland NK-mediated immune response through IL12 and IFNg . Wethus stained FFPE tissues from the clinical trial for the presence ofNK cells using IHC for CD56 (neural cell adhesion molecule;NCAM). We observe an increase in NK cells in polyps frompatients on drug versus placebo, but the observation is notstatistically significant. Polyps from the group treated with drughad an average CD56 count of 1.18 per HPF ranging from 0 to3.57 per HPF. The placebo group had an average CD56 count of0.846 per HPF, ranging from 0 to 1.83. Upon fitting the data to anegative binomialmodel,wefinda1.43 increase in count perHPFin the polyps from patients on drug (95% CI, 0.77–2.66; P ¼0.2641). Although there is a slight increase inNK cells, whichmayrepresent activation of T-cell–mediated immune response, itis not statistically different from the polyps from patientson placebo.

DiscussionWe recently completed and reported the clinical findings of a

chemoprevention trial aimed at preventing progression of duo-denal adenomas in patients with FAP (17). Combined treatment

Figure 2.

Differential gene expression in duodenal polyps from subjects treated withsulindac–erlotinib versus placebo and focused on intended targeted pathwaysofWNT, EGFR, and PGE2. A, PCA using the 977 genes differentially regulated at2-fold or higher in placebo polyps as compared with baseline uninvolved.Figure includes endpoint polyps with >50% purity by ESTIMATE from subjectson drug (red) or on placebo (blue). B, Hierarchical clustering values arerepresented as log2 ratios of endpoint polyps with >50% purity compared withpaired baseline uninvolvedduodenum fromeach patient. Samples are labeled asdrug (D) or placebo (P), the size of polyp in mm, and the sample number,corresponding toTable 1. Red, increased expression; blue, decreased expression;and black, no change compared with baseline uninvolved duodenum.






Figure 3.

Quantitative RT-PCR confirmation of differentially regulated genes representing WNT, EGFR, and PGE2 signaling pathways. A and B, Genes in WNT signalingpathway. C and D, Representative of EGFR signaling pathway. (Continued on the following page.)

Delker et al.





Figure 3.

(Continued. ) E and F, PGE2 signaling pathway. Each individual sample is graphed along with the median and 95% confidence interval. P values are reportedfor Mann–Whitney tests. Nonsignificant P values are noted as ns. The open circles are the five <50% tumor purity polyps. qRT-PCR failed for theendpoint uninvolved duodenum sample from patient 176, so this data point is absent from all graphs.






with sulindac–erlotinib resulted in a 56% reduction in duodenalpolyp burden after 6months, whereas the placebo armhad a 31%increase in duodenal polyp burden. The goal of this study was tocharacterize the molecular changes associated with adenomatouspolyp regression in FAP patients treated with sulindac–erlotinib.RNA-Seq technology was used to define the human duodenumtranscriptome in FAP patients treated with sulindac–erlotinib orplacebo.

One of the challenges resulting from the success of the clinicaltrial was that there was very limited polyp tissue available frompatients on drug. Consequently, the power to discover molecularchanges was limited, ranging from 51% to 80%depending on thenumber of paired comparisons (Supplementary Data). Even withthis limitation, the robustness of these gene expression–basedsignatures provided convincing evidence that the sulindac–erlotinib drug combination was inhibiting the intended WNT,EGFR, and PGE2 pathways. This fits with our current biologic andgenetic understanding of the progression of colon cancer, expandsour understanding of how these drugs drive the regression ofduodenal polyps in FAP patients, and identifies new biomarkersfor diagnostics and therapeutics.

Adenomas from patients on placebo displayed a wide range ofgene expression changes, including changes in RNA transcriptsassociated with increased WNT, EGFR, and PGE2 signaling.Increased expression of many WNT signaling targets, includingAXIN2, LGR5, MMP7, and MYC observed in our study are inagreement with previously published gene expression studiesfrom sporadic and/or FAP patient cohorts (29, 31, 42). Thesegenes play an important role in regulating cell proliferation andtumor progression in colon adenomas. AXIN2 and LGR5 are bothconsidered negative regulators of WNT signaling, while MMP7and MYC are positive regulators. AXIN2 or conductin protein ispart of the multiprotein APC complex that regulates the stabilityof b-catenin, and LGR5 is an established cell surface proteinmarker of intestinal stem cells (43, 44). MMP7 protein plays arole in the breakdown of extracellular matrix, T-cell migration,

and tumor metastasis (45). MYC protein plays multiple roles inthe development of cancer, including the regulation of cell pro-liferation, apoptosis, and epithelial-to-mesenchymal transition(46). The upregulation of these genes provides strong support ofWNT activation in FAP duodenal adenomas in our study similarto previous gene expression studies.

The RNA levels of many target genes of both EGFR and PGE2signaling were also significantly increased in duodenal adenomasfrom patients on placebo but not on drug. The increased expres-sion of immediate early genes FOS and EGR1 in duodenaladenomas is consistent with increased EGFR signaling and cellproliferation. Increased PGE2 signaling was associated withincreased mRNA levels of the stem cell marker CD44. PGE2increases the number and survival of CD44þ stem cells in humancolorectal cancer and animal models of colon tumor metastasis(47). The increased activation of EGFR and PGE2 pathways inadenomatous polyps from FAP patients in our study furthersupports the use of the combined sulindac–erlotinib therapy inour phase II clinical trial for the treatment of duodenal adenomasin FAP patients.

Pathways selectively downregulated in placebo polyps but notdrug polyps compared with paired uninvolved duodenal tissueincluded a-catenin (CTNNA) and chromobox 5 (CBX5). a-Cate-nin is an actin-binding protein whose cellular distribution isregulated by b-catenin (48). Reciprocally a-catenin inhibitsWnt/b-catenin–mediated transcription through dual binding ofb-catenin and actin. The observed decrease in a-catenin signalingin duodenal polyps observed in our study may be the result ofaltered a-catenin localization and/or reduced transcriptionalregulation through a-catenin. Chromobox 5, also known asheterochromatin protein 1 alpha (HP1a), is a nuclear proteinthat mediates transcriptional silencing through interactions withH3K9 methyltransferase and DNA methyltransferases 1 and 3(49). HP1a is also important in accurate chromosomal segrega-tion during mitosis (50). Decreased HP1a signaling observed inplacebo polyps may reflect an increase in transcriptional activity

Table 2. Ingenuity (IPA) upstream regulator analysis of differentially expressed genes in endpoint polyp compared with uninvolved baseline

Placebo polyp(n ¼ 10, 2,591 genes)

Drug polyp(n ¼ 13, 236 genes)

Upstream regulator Symbol Molecule type z-score P z-score P

Tumor necrosis factor TNF Cytokine 2.06 4.95E�16 1.62 3.08E�02Beta-catenin CTNNB1 Transcription factor 3.04 2.29E�11 1.89 4.61E�05Epidermal growth factor EGF Growth factor 2.38 3.31E�07 NA NSEpidermal growth factor receptor EGFR Kinase 3.38 3.66E�05 1.56 1.80E�02Prostaglandin E2 PGE2 Endogenous chemical 1.95 1.83E�03 1.10 3.52E�02Chromobox 5 CBX5 Transcription factor �3.09 1.90E�11 �1.89 1.28E�04Interferon gamma IFNG Cytokine �1.17 3.18E�10 0.393 1.76E�03Caudal type homeobox 2 CDX2 Transcription factor �2.29 1.23E�09 0.00 3.85E�02Interferon alpha IFNA Group �3.74 3.78E�04 �1.27 3.33E�02Alpha-catenin CTNNA Group �2.14 2.86E�03 NA NS

Abbreviations: NA, not applicable because upstream regulator was not significantly represented in differentially expressed gene list; NS, not significant.

Table 3. Top signaling pathways in Inflammation and Immunity Transcriptome panel that are activated or inhibited in polyps from patients treated with sulindac–erlotinib compared with patients on placebo

Upstream regulator Symbol Molecule type z-score P

Interferon alpha IFNA Group 3.063 4.52E�22Interferon gamma IFNG Cytokine 2.965 5.29E�21Interleukin 12 IL12 Complex 2.675 1.55E�15Signal transducer and activator of transcription 4 STAT4 Transcription factor 2.403 3.00E�06Prostaglandin E2 PGE2 Chemical �2.225 1.52E�05

NOTE: Positive (þ) z-score or negative (�) z-score indicate that these pathways are more or less active in patients treated with sulindac–erlotinib versus placebo.

Delker et al.





and/or decrease in chromosomal stability necessary for duodenalpolyp development and progression. Together, the reduced activ-ity of a-catenin and HP1a are both suggestive of increasedtranscriptional activity associated with tumor development andgrowth.

Following sulindac–erlotinib therapy, a pronounced reductionin the number of genes differentially expressed in adenomas wasobserved in the patients on drug. The average size of adenomascollected from patients on drug was smaller than those collectedfrompatients on placebo, yet retained their adenomatous appear-anceuponpathologic examination. Thepronounced regressionofduodenal adenomas in our phase II clinical trial was surprisingbased on previous trials using sulindac therapy alone (11, 12). Atbest, we anticipated an inhibition of further duodenal polypgrowth in our FAP patient cohort. Although this is clearly spec-ulative, and further work will be required, our results may suggestthat the combination therapy of sulindac–erlotinib is restoringsome level of innate immunosurveillance and adenoma cellkilling in the duodenum of FAP patients on drug. The derepres-sion of IFNa signaling and presence of CD56þ NK cells supportthis idea. It has been shown that increased levels of PGE2 cause adecrease IL12 signaling, which is important in the recruitment ofNK cells and cytotoxic T lymphocytes (51). Both mouse andhuman cancer cells deficient in COX or PGE2 show increasedimmune signaling and T-cell–dependent growth control com-pared with cancer cells expressing COX (52). These findings areconsistent with our studies that show sulindac–erlotinib restoresthe expression of genes important in the innate immune responseand NK-cell surveillance and function.

On the basis of RNA analysis, we observed potential resistanceto the drug combination in multiple polyps from patient 2. Thischemoprevention trial, however, was limited in its ability toevaluate resistance by the minimal polyp tissue that was frompatients on drug and available at the end of the trial. It will beimportant for future work to examine the molecular changes inpolyps that are persistent, do not regress, and likely representresistance to EGFR and COX inhibition. Future chemopreventiontrials with this drug combination will benefit by extending thetreatment period beyond 6 months by which time-resistantpolyps would be more established and evident. In addition,sampling for future studies should consider use of highly sensitivenew technologies, such as single-cell RNA and DNA sequencing,liquid biopsy for circulating tumor DNA and cytokines in blood,single-cell mass cytometry for protein expression, and metabo-lomics. These other endpoints would enable one to overcomelimitations of sample number as well as detecting distinct respon-sive and resistant cellswithin a single polyp. In a follow-up clinicaltrial (NCT02961374), efficacy and tolerability of erlotinib alone,given once weekly, in FAP patients with Spigelman stage II to IIIduodenal polyposis is being examined. Secondary aims willsimilarly evaluate the gene expression profiles of duodenal tis-sues. These analyses will be important to examine whether inhi-bition of the COX2 pathway is required to derepress IFNasignaling.

In summary, our analysis of gene expression in duodenal tissuefrom patients on drug compared with patients on placebodescribes key changes in cancer, inflammation, and innate immu-nity signaling pathways. Many of the genes and pathwaysdescribed in our study support previous findings of molecularchanges observed in colon adenomas from FAP patients. Theobserved reduction in duodenal polyp number and size togetherwith the inhibition of WNT, EGFR, and PGE2 signaling andincrease in IFNg signaling provide important insights into themechanisms of duodenal polyp regression in FAP patients treatedwith sulindac–erlotinib.

Disclosure of Potential Conflicts of InterestN.J. Samadder has received speakers bureau honoraria from Cook Medical,

Inc., is a consultant/advisory board member for Janssen, and has providedexpert testimony for Medico-Legal Consulting. R. Burt is a consultant/advisoryboard member for Thetis Pharma. No potential conflicts of interest weredisclosed by the other authors.

Authors' ContributionsConception and design:D.A. Delker, N.J. Samadder, R.W. Burt, D.W. NeklasonDevelopment of methodology: D.A. Delker, A.C. Wood, N.J. Samadder,D.W. NeklasonAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): A.C. Wood, A.K. Snow, N.J. Samadder, K.E. Affolter,I.J. Stijleman, P. Kanth, K.R. Byrne, R.W. Burt, D.W. NeklasonAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): D.A. Delker, A.C. Wood, A.K. Snow, K.M. Boucher,L.M. Pappas, P.S. Bernard, D.W. NeklasonWriting, review, and/or revision of the manuscript: D.A. Delker, A.K. Snow,N.J. Samadder,W.S. Samowitz, K.E. Affolter, K.M. Boucher, P. Kanth, K.R. Byrne,R.W. Burt, P.S. Bernard, D.W. NeklasonAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): D.A. Delker, R.W. Burt, D.W. NeklasonStudy supervision: N.J. Samadder, R.W. Burt, D.W. NeklasonOther (evaluated pathology): W.S. Samowitz

AcknowledgmentsWe thank Dr. Matt Topham for helpful discussions of signaling pathways,

Michelle W. Done, Megan Keener, Therese Berry, and Danielle Sample for studycoordination effort, and the research participants who are committed to findingsolutions for managing their condition.

Grant SupportThis workwas supported byNIHHHSN2612012000131 (to P.S. Bernard,D.

W. Neklason, D.A. Delker, A.C. Wood, and I.J. Stijleman), NCIPO1-CA073992(to R.W. Burt, D.W. Neklason, A.K. Snow, N.J. Samadder, W.S. Samowitz, K. M.Boucher, L.M. Pappas, P. Kanth, and K.R. Byrne), National Cancer InstituteCancer Center Support Grant P30-CA042014 (to K.M. Boucher and L.M.Pappas), National Center for Advancing Translational Sciences of the NIHunder Award Number UL1TR00106, and Huntsman Cancer Foundation.

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received May 1, 2017; revised August 31, 2017; accepted October 2, 2017;published OnlineFirst November 6, 2017.

References1. Groden J, Thliveris A, SamowitzW, CarlsonM,Gelbert L, AlbertsenH, et al.

Identification and characterization of the familial adenomatous polyposiscoli gene. Cell 1991;66:589–600.

2. Kinzler KW, Nilbert MC, Su LK, Vogelstein B, Bryan TM, Levy DB, et al.Identification of FAP locus genes from chromosome 5q21. Science1991;253:661–5.






3. Jasperson KW, Tuohy TM, Neklason DW, Burt RW. Hereditary and familialcolon cancer. Gastroenterology 2010;138:2044–58.

4. Biasco G, Pantaleo MA, Di Febo G, Calabrese C, Brandi G, Bulow S. Risk ofduodenal cancer in patients with familial adenomatous polyposis. Gut2004;53:1547.

5. Bulow S, Bjork J, Christensen IJ, Fausa O, J€arvinen H, Moesgaard F, et al.Duodenal adenomatosis in familial adenomatous polyposis. Gut2004;53:381–6.

6. OfferhausGJ,Giardiello FM,KrushAJ, Booker SV, Tersmette AC, KelleyNC,et al. The risk of upper gastrointestinal cancer in familial adenomatouspolyposis. Gastroenterology 1992;102:1980–2.

7. Jagelman DG, DeCosse JJ, Bussey HJ. Upper gastrointestinal cancer infamilial adenomatous polyposis. Lancet 1988;1:1149–51.

8. Vasen HF, Bulow S, Myrhoj T, Mathus-Vliegen L, Griffioen G, Buskens E,et al. Decision analysis in the management of duodenal adenomatosis infamilial adenomatous polyposis. Gut 1997;40:716–9.

9. Giardiello FM, Hamilton SR, Krush AJ, Piantadosi S, Hylind LM, Celano P,et al. Treatment of colonic and rectal adenomas with sulindac in familialadenomatous polyposis. N Engl J Med 1993;328:1313–6.

10. Giardiello FM, Yang VW, Hylind LM, Krush AJ, Petersen GM, Trimbath JD,et al. Primary chemoprevention of familial adenomatous polyposis withsulindac. N Engl J Med 2002;346:1054–9.

11. Debinski HS, Trojan J, Nugent KP, Spigelman AD, Phillips RK. Effect ofsulindac on small polyps in familial adenomatous polyposis. Lancet1995;345:855–6.

12. Nugent KP, Farmer KC, Spigelman AD, Williams CB, Phillips RK. Ran-domized controlled trial of the effect of sulindac on duodenal and rectalpolyposis and cell proliferation in patients with familial adenomatouspolyposis. Br J Surg 1993;80:1618–9.

13. Brosens LA, Iacobuzio-Donahue CA, Keller JJ, Hustinx SR, Carvalho R,Morsink FH, et al. Increased cyclooxygenase-2 expression in duodenalcompared with colonic tissues in familial adenomatous polyposis andrelationship to the -765G ->C COX-2 polymorphism. Clin Cancer Res2005;11:4090–6.

14. Coffey RJ, Hawkey CJ, Damstrup L, Graves-Deal R, Daniel VC,Dempsey PJ,et al. Epidermal growth factor receptor activation induces nuclear targetingof cyclooxygenase-2, basolateral release of prostaglandins, and mitogen-esis in polarizing colon cancer cells. Proc Natl Acad Sci U S A 1997;94:657–62.

15. Eisinger AL, Nadauld LD, Shelton DN, Peterson PW, Phelps RA, ChidesterS, et al. The adenomatous polyposis coli tumor suppressor gene regulatesexpression of cyclooxygenase-2 by amechanism that involves retinoic acid.J Biol Chem 2006;281:20474–82.

16. Roberts RB, Min L, Washington MK, Olsen SJ, Settle SH, Coffey RJ, et al.Importance of epidermal growth factor receptor signaling in establishmentof adenomas and maintenance of carcinomas during intestinal tumori-genesis. Proc Natl Acad Sci U S A 2002;99:1521–6.

17. SamadderNJ,NeklasonDW,BoucherKM,ByrneKR,KanthP, SamowitzW,et al. Effect of sulindac and erlotinib vs. placebo on duodenal neoplasia infamilial adenomatous polyposis: a randomized clinical trial. JAMA2016;315:1266–75.

18. Nastala CL, Edington HD, McKinney TG, Tahara H, Nalesnik MA, BrundaMJ, et al. Recombinant IL-12 administration induces tumor regression inassociationwith IFN-gammaproduction. J Immunol 1994;153:1697–706.

19. Galon J, Angell HK, Bedognetti D,Marincola FM. The continuum of cancerimmunosurveillance: prognostic, predictive, and mechanistic signatures.Immunity 2013;39:11–26.

20. Delker DA, McGettigan BM, Kanth P, Pop S, Neklason DW, Bronner MP,et al. RNA sequencing of sessile serrated colon polyps identifies differen-tially expressed genes and immunohistochemical markers. PLoS One2014;9:e88367.

21. Nix DA, Courdy SJ, Boucher KM. Empirical methods for controlling falsepositives and estimating confidence in ChIP-Seq peaks. BMC Bioinformat-ics 2008;9:523.

22. Love MI, Huber W, Anders S. Moderated estimation of fold change anddispersion for RNA-seq data with DESeq2. Genome Biol 2014;15:550.

23. Yoshihara K, Shahmoradgoli M, Martinez E, Vegesna R, Kim H, Torres-Garcia W, et al. Inferring tumour purity and stromal and immune celladmixture from expression data. Nat Commun 2013;4:2612.

24. AranD, SirotaM,Butte AJ. Systematic pan-cancer analysis of tumour purity.Nat Commun 2015;6:8971.

25. de HoonMJ, Imoto S, Nolan J, Miyano S. Open source clustering software.Bioinformatics 2004;20:1453–4.

26. Saldanha AJ. Java Treeview–extensible visualization of microarray data.Bioinformatics 2004;20:3246–8.

27. Kramer A, Green J, Pollard J Jr, Tugendreich S. Causal analysis approachesin Ingenuity Pathway Analysis. Bioinformatics 2014;30:523–30.

28. di Pietro M, Sabates Bellver J, Menigatti M, Bannwart F, Schnider A, RussellA, et al. Defective DNA mismatch repair determines a characteristictranscriptional profile in proximal colon cancers. Gastroenterology2005;129:1047–59.

29. Sabates-Bellver J, Van der Flier LG, de Palo M, Cattaneo E, Maake C,Rehrauer H, et al. Transcriptome profile of human colorectal adenomas.Mol Cancer Res 2007;5:1263–75.

30. Livak KJ, Schmittgen TD. Analysis of relative gene expression data usingreal-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods2001;25:402–8.

31. Segditsas S, Sieber O, Deheragoda M, East P, Rowan A, Jeffery R, et al.Putative direct and indirect Wnt targets identified through consistent geneexpression changes in APC-mutant intestinal adenomas from humans andmice. Hum Mol Genet 2008;17:3864–75.

32. Pai R, Soreghan B, Szabo IL, Pavelka M, Baatar D, Tarnawski AS. Prosta-glandin E2 transactivates EGF receptor: a novel mechanism for promotingcolon cancer growth and gastrointestinal hypertrophy. Nat Med2002;8:289–93.

33. Al-Salihi MA, Ulmer SC, Doan T, Nelson CD, Crotty T, Prescott SM, et al.Cyclooxygenase-2 transactivates the epidermal growth factor receptorthrough specific E-prostanoid receptors and tumor necrosis factor-alphaconverting enzyme. Cell Signal 2007;19:1956–63.

34. Kim JY, Nam JK, Lee SA, Lee MS, Cho SK, Park ZY, et al. Proteasomeinhibition causes epithelial-mesenchymal transition upon TM4SF5 expres-sion. J Cell Biochem 2011;112:782–92.

35. Eberhart CE, Coffey RJ, Radhika A, Giardiello FM, Ferrenbach S, DuBoisRN. Up-regulation of cyclooxygenase 2 gene expression in human colo-rectal adenomas and adenocarcinomas. Gastroenterology 1994;107:1183–8.

36. Nosho K, Yoshida M, Yamamoto H, Taniguchi H, Adachi Y, Mikami M,et al. Association of Ets-related transcriptional factor E1AF expression withoverexpression ofmatrixmetalloproteinases, COX-2 and iNOS in the earlystage of colorectal carcinogenesis. Carcinogenesis 2005;26:892–9.

37. Gustafsson A, Hansson E, Kressner U, Nordgren S, Andersson M, Wang W,et al. EP1-4 subtype, COX and PPAR gamma receptor expression incolorectal cancer in prediction of disease-specific mortality. Int J Cancer2007;121:232–40.

38. Parfenova H, Balabanova L, Leffler CW. Posttranslational regulation ofcyclooxygenase by tyrosine phosphorylation in cerebral endothelial cells.Am J Physiol 1998;274:C72–81.

39. Grainger S, Savory JG, LohnesD. Cdx2 regulates patterning of the intestinalepithelium. Dev Biol 2010;339:155–65.

40. Hryniuk A, Grainger S, Savory JG, Lohnes D. Cdx1 and Cdx2 function astumor suppressors. J Biol Chem 2014;289:33343–54.

41. Kalinski P. Regulation of immune responses by prostaglandin E2.J Immunol 2012;188:21–8.

42. Obrador-Hevia A, Chin SF, Gonzalez S, Rees J, Vilardell F, Greenson JK,et al.Oncogenic KRAS is not necessary forWnt signalling activation inAPC-associated FAP adenomas. J Pathol 2010;221:57–67.

43. Lustig B, Jerchow B, Sachs M, Weiler S, Pietsch T, Karsten U, et al. Negativefeedback loop of Wnt signaling through upregulation of conductin/axin2in colorectal and liver tumors. Mol Cell Biol 2002;22:1184–93.

44. Barker N, van Es JH, Kuipers J, Kujala P, van den BornM, CozijnsenM, et al.Identification of stem cells in small intestine and colon by marker geneLgr5. Nature 2007;449:1003–7.

45. Yokoyama Y, Grunebach F, Schmidt SM,Heine A,H€antschelM, StevanovicS, et al. Matrilysin (MMP-7) is a novel broadly expressed tumor antigenrecognized by antigen-specific T cells. Clin Cancer Res 2008;14:5503–11.

46. Dang CV. MYC on the path to cancer. Cell 2012;149:22–35.47. Wang D, Fu L, Sun H, Guo L, DuBois RN. Prostaglandin E2 promotes

colorectal cancer stem cell expansion and metastasis in mice. Gastroen-terology 2015;149:1884–95.

48. Daugherty RL, Serebryannyy L, Yemelyanov A, Flozak AS, Yu HJ, Kosak ST,et al. alpha-Catenin is an inhibitor of transcription. ProcNatl Acad SciU SA2014;111:5260–5.


Delker et al.




49. Maison C, Almouzni G. HP1 and the dynamics of heterochromatinmaintenance. Nat Rev Mol Cell Biol 2004;5:296–304.

50. Chu L, Huo Y, Liu X, Yao P, Thomas K, Jiang H, et al. The spatiotemporaldynamics of chromatin protein HP1alpha is essential for accurate chro-mosome segregation during cell division. J Biol Chem 2014;289:26249–62.

51. Walker W, Rotondo D. Prostaglandin E2 is a potent regulator of interleu-kin-12- and interleukin-18-induced natural killer cell interferon-gammasynthesis. Immunology 2004;111:298–305.

52. Zelenay S, van der Veen AG, Bottcher JP, Snelgrove KJ, Rogers N, Acton SE,et al. Cyclooxygenase-dependent tumor growth through evasion of immu-nity. Cell 2015;162:1257–70.






2018;11:4-15. Published OnlineFirst November 6, 2017.Cancer Prev Res Don A. Delker, Austin C. Wood, Angela K. Snow, et al. Patients: mRNA Signatures of Duodenal NeoplasiaFactor Receptor Inhibitors in Familial Adenomatous Polyposis Chemoprevention with Cyclooxygenase and Epidermal Growth

Updated version

10.1158/1940-6207.CAPR-17-0130doi:

Access the most recent version of this article at:

Material

Supplementary

1

http://cancerpreventionresearch.aacrjournals.org/content/suppl/2017/11/28/1940-6207.CAPR-17-0130.DCAccess the most recent supplemental material at:

Cited articles

http://cancerpreventionresearch.aacrjournals.org/content/11/1/4.full#ref-list-1

This article cites 52 articles, 15 of which you can access for free at:

Citing articles

http://cancerpreventionresearch.aacrjournals.org/content/11/1/4.full#related-urls

This article has been cited by 2 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://cancerpreventionresearch.aacrjournals.org/content/11/1/4To request permission to re-use all or part of this article, use this link



http://cancerpreventionresearch.aacrjournals.org/lookup/doi/10.1158/1940-6207.CAPR-17-0130

http://cancerpreventionresearch.aacrjournals.org/content/suppl/2017/11/28/1940-6207.CAPR-17-0130.DC1

http://cancerpreventionresearch.aacrjournals.org/content/suppl/2017/11/28/1940-6207.CAPR-17-0130.DC1

http://cancerpreventionresearch.aacrjournals.org/content/11/1/4.full#ref-list-1

http://cancerpreventionresearch.aacrjournals.org/content/11/1/4.full#related-urls

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

mailto:[email protected]

http://cancerpreventionresearch.aacrjournals.org/content/11/1/4


Home | Cancer Prevention Research - Chemoprevention with … · 2Huntsman Cancer Institute, University of Utah, Salt Lake City, Utah. 3Depart-ment of Pathology, University of Utah,

Documents