Identification of Transcriptional Signatures of Colon Tumor Stroma by a Meta-Analysis

Hindawi Journal of Oncology Volume 2019, Article ID 8752862, 12 pages https://doi.org/10.1155/2019/8752862

Research Article Identification of Transcriptional Signatures of Colon Tumor Stroma by a Meta-Analysis

Md. Nazim Uddin ,1,2,3 Mengyuan Li,1,2,3 and Xiaosheng Wang 1,2,3

Biomedical Informatics Research Lab, School of Basic Medicine and Clinical Pharmacy, China Pharmaceutical University, Nanjing , China Cancer Genomics Research Center, School of Basic Medicine and Clinical Pharmacy, China Pharmaceutical University, Nanjing , China Big Data Research Institute, China Pharmaceutical University, Nanjing , China

Correspondence should be addressed to Xiaosheng Wang; [email protected]

Received 28 February 2019; Accepted 31 March 2019; Published 2 May 2019

Guest Editor: Nathaniel Weygant

Copyright © 2019 Md. Nazim Uddin et al. Tis is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Background. Te tumor stroma plays pivotal roles in infuencing tumor growth, invasion, and metastasis. Transcriptional signatures of colon tumor stroma (CTS) are signifcantly associated with prognosis of colon cancer. Tus, identifcation of the CTS transcriptional features could be useful for colon cancer diagnosis and therapy. Methods.Byameta-analysisofthree CTS gene expression profles datasets, we identifed diferentially expressed genes (DEGs) between CTS and colon normal stroma. Furthermore, we identifed the pathways, upstream regulators, and protein-protein interaction (PPI) network that were signifcantly associated with the DEGs. Moreover, we analyzed the enrichment levels of immune signatures in CTS. Finally, we identifed CTS-associated gene signatures whose expression was signifcantly associated with prognosis in colon cancer. Results.We identifed numerous signifcantly upregulated genes (such as CTHRC, NFEL, SULF, SOX, ENC,andCCND) and signifcantly downregulated genes (such as MYOT, ASPA, KIAA, ARHGEF, BCL-,andPPARGCA) in CTS versus colon normal stroma. Furthermore, we identifed signifcantly upregulated pathways in CTS that were mainly involved in cellular development, immune regulation, and metabolism, as well as signifcantly downregulated pathways in CTS that were mostly metabolism-related. Moreover, we identifed upstream TFs (such as SUZ12, NFE2L2, RUNX1, STAT3, and SOX2), kinases (such as MAPK14, CSNK2A1, CDK1, CDK2, and CDK4), and master metabolic transcriptional regulators (MMTRs) (such as HNF1A, NFKB1, ZBTB7A, GATA2, and GATA5) regulating the DEGs. We found that CD8+ T cells were more enriched in CTS than in colon normal stroma. Interestingly, we found that many of the DEGs and their regulators were prognostic markers for colon cancer, including CEBPB, PPARGC, STAT, MTOR, BCL, JAK,andCDK. Conclusions. Te identifcation of CTS-specifc transcriptional signatures may provide insights into the tumor microenvironment that mediates the development of colon cancer and has potential clinical implications for colon cancer diagnosis and treatment.

1. Background immunosurveillance evasion, and drug resistance of tumors [3]. Colorectal cancer (CRC) is the fourth most common Te tumor stroma is an important component of the tumor cancer and a leading cause of cancer mortality world- microenvironment (TME) and plays key roles in the tumor wide [4]. Transcriptional signatures of CRC stromal cells development [1]. Stromal cells are composed of many dif- have been associated with poor prognosis in CRC [5]. ferent types of cells, including vascular endothelial cells, Isella et al. demonstrated that the gene signatures of CRC pericytes, adipocytes, fbroblasts, osteoblasts, chondrocytes, stromal cells (cancer-associated fbroblasts, leukocytes, and extracellular matrix (ECM), and bone-marrow mesenchymal endothelial cells) were signifcantly upregulated in the stromal cells [2]. Te tumor stroma can promote ECM stem/serrated/mesenchymal transcriptional subtype of CRC remodeling, cellular migration, neoangiogenesis, invasion, which had a poor prognosis [6]. Calon et al. showed that 2 Journal of Oncology the CRC stromal transcriptional signatures correlated with . . Identification of PPI Network of the DEGs. We employed disease relapse [5]. Tese prior studies exhibited the signif- Network Analyst [11] to construct a PPI network of the icant roles of tumor stroma in CRC growth, invasion, and DEGs [11]. Two types of modules (function-frst modules and metastasis. connection-frst modules) of the PPI network were extracted. In this study, we performed a meta-analysis of three Te function-frst modules (FFMs) were constructed by path- colon tumor stromal transcriptome datasets using the bioin- way enrichment analysis and the connection-frst modules formatics approach. We identifed diferentially expressed (CFMs) were identifed by the random walk-based algorithm genes (DEGs) between colon tumor stroma (CTS) and [18]. normal stroma. On the basis of these DEGs, we identifed their associated pathways, upstream regulators, and protein- . .ComparisonoftheEnrichmentLevelsofCD+TCells protein interaction (PPI) network and certain prognostic between Two Classes of Samples. Te enrichment level of markers that were associated with survival of colon cancer CD8+Tcellsinasamplewasevaluatedbytheexpressionlevel patients. We also analyzed the enrichment levels of immune of CDA.WecomparedtheenrichmentlevelsofCD8+Tcells signatures in CTS. Tis study provides insights into CTS between two groups of samples using Student's t-test. molecular features that could have clinical implications for colon cancer diagnosis and treatment. .. Identification of DEGs between High-Stroma-Content and Low-Stroma-Content TCGA Colon Cancer Samples. We used 2. Methods ESTIMATE [19] to quantify the intratumoral stromal content (stroma score) of TCGA colon cancer samples. We identifed .. Datasets. WesearchedtheNCBIGeneExpression the DEGs between high-stroma-content (stroma score > Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/ median) and low-stroma-content (stroma score < median) geo/) using the keywords “colon cancer,” “stroma,” and tumors using Student's t-test. “tumor stroma” and identifed three CTS gene expression profles datasets (GSE31279, GSE35602, and GSE46824) [7–9]. In survival analyses, we used the TCGA colon cancer .. Survival Analyses. We compared the overall survival dataset (https://portal.gdc.cancer.gov/) and a SurvExpress (OS) and the disease-free survival (DFS) of colon cancer patients classifed based on gene expression levels (expression (http://bioinformatica.mty.itesm.mx/SurvExpress) built-in > < dataset (colon metabase) [10]. A summary of these datasets levels median versus expression levels median). Kaplan- is shown in Supplementary Table S1. Meier survival curves were used to show the survival dif- ferences, and the log-rank test was utilized to evaluate the signifcance of survival diferences. Te individual prognostic .. Identification of DEGs between CTS and Normal Stroma. genes were identifed and were ftted in a multivariate Cox We used the web tool Network Analyst [11] to identify the regression model. SurvExpress [10] was used for the multi- DEGs between CTS and normal stroma. Te ComBat method variate survival analysis. [12] in the tool was utilized to remove batch efects from the three CTS datasets (Supplementary Figure S1). Each individual dataset was normalized by base-2 log transformation 3. Results and quantile normalization, and the R package “limma” .. Identification of DEGs between CTS and Normal Stroma. was utilized to identify the DEGs between CTS and normal We identifed 694 DEGs between CTS and normal stroma by stroma. A meta-analysis of the three datasets was performed the meta-analysis. Tese DEGs included 295 downregulated using Cochran's combination test [13]. Te false discovery and 399 upregulated genes in CTS (Supplementary Tables rate (FDR), calculated by the Benjamini–Hochberg method S2 and S3). Figure 1 shows the top 25 upregulated and [14], was used to adjust for multiple tests. We determined the top 25 downregulated genes in CTS ranked on the basis of DEGs with a threshold of absolute combined efect size (ES) thecombinedES(thedetailedresultsofstatisticalanalysis >0.82 and FDR<0.05. for the top 10 upregulated and top 10 downregulated genes in CTS are shown in Supplementary Tables S4). CTHRC, .. Gene-Set Enrichment Analysis. We performed gene-set a gene involved in vascular remodeling, bone formation, enrichment analysis of the DEGs by GSEA [15]. Te KEGG and developmental morphogenesis, was upregulated in CTS pathways signifcantly associated with the upregulated and with the highest ES. It has been shown that CTHRC could < the downregulated DEGs were identifed (FDR 0.05), promote human CRC cell proliferation and invasion by respectively. activating Wnt/PCP signaling [20]. Tis gene also plays an important role in promoting ovarian cancer cell adhesion, . . Identification of Transcription Factors (TFs), Kinases, and migration, and metastasis through the activation of integrin Master Metabolic Transcriptional Regulators (MMTRs) at �3/FAK signaling [21]. NFEL, a gene regulating the cell Are Significantly Associated with the DEGs. To link gene cycle progression in colon cancer [22], was upregulated in expression signatures to upstream cell signaling networks, we CTS with the second highest ES. Interestingly, both CTHRC used eXpression2Kinases [16] to identify the upstream TFs and NFEL have been indicated as useful biomarker candi- and kinases that regulate the DEGs and utilized iRegulon [17] dates for CRC diagnosis because of their overexpression in to identify the MMTRs of the DEGs. adenomas and CRC relative to normal tissue [23]. SULF, Journal of Oncology 3

class dataset

CTHRC1 SULF1 Tumor stroma NFE2L3 CCND1 ENC1 Normal stroma CLDN1 SOX9 SEMA5A PCDH17 BCL6B GSE46824 VCAN NUAK1 KDELR3 COL5A1 GSE35602 KIAA1217 CKAP4 DCBLD1 PROX1 GSE31279 PLAUR PTBP3 ESM1 MLXIPL SLC16A3 SEMA3G GZMB MYOT C7 GNG7 ARHGEF37 ASPA BCL2 RFX2 SLC6A16 KIAA2022 SRSF5 CBX7 METTL7A EFCAB1 PELI2 ANKRD13A C1QTNF9 SEPP1 FXYD1 GATS PDK4 TMOD1 C16orf89 PCF11 FGF13 CCDC7 GSM775185 GSM775225 GSM775250 GSM775191 GSM775259 GSM775265 GSM775271 GSM775275 GSM775279 GSM775287 GSM871410 GSM871412 GSM871414 GSM871416 GSM871418 GSM871420 GSM871422 GSM871423 GSM871426 GSM871428 GSM871429 GSM871432 GSM871433 GSM1138839 GSM1138840 GSM1138842 GSM1138850 GSM1138851 GSM1138852 GSM1138853 GSM1138854 GSM1138855 GSM1138856 GSM1138857 GSM1138858 GSM1138859 GSM1138860 GSM1138861 GSM1138862 GSM1138843 GSM1138863 GSM1138841 GSM1138844 GSM1138849 GSM1138845 GSM1138848 GSM1138847 GSM1138846 GSM775277 GSM775284 GSM775256 GSM775247 GSM775182 GSM775188 GSM775222 GSM775262 GSM775268 GSM775273 GSM871436 GSM871437 GSM871441 GSM871440 GSM1138872 GSM1138864 GSM1138865 GSM1138866 GSM1138868 GSM1138869 GSM1138871 GSM1138870 GSM1138867

Low High

Figure 1: Gene expression pattern of the top 25 upregulated and top 25 downregulated genes in colon tumor stroma (CTS) relative to colon normal stroma ranked on the basis of the combined efect size (ES) identifed by Network Analyst [11]. whose expression in tumor stroma is a prognostic marker .. Identiﬁcation of Pathways Signiﬁcantly Associated with the in advanced pancreatic cancer [24], was upregulated in CTS DEGs. GSEA [15] identifed 44 KEGG pathways that were with the third highest ES. Te overexpression of this gene signifcantly associated with the upregulated genes in CTS. has been associated with a poor prognosis in urothelial Tese pathways were mainly involved in cellular development carcinoma [25]. SOX,thegeneupregulatedinCTSwith (p53 signaling, Wnt signaling, apoptosis, Notch signaling, the fourth highest ES, has been shown to be overexpressed focal adhesion, endocytosis, ECM-receptor interaction, cell in CRC and its overexpression was an independent adverse adhesion molecules, adherens junction, tight junction, gap prognosticator in CRC [26]. Some other genes upregulated junction, and regulation of actin cytoskeleton), immune in CTS have been demonstrated to be overexpressed in CRC regulation (leukocyte transendothelial migration, comple- and their expression was negatively associated with CRC ment and coagulation cascades, natural killer cell medi- prognosis, such as ENC, CCND,VCAN, SEMA A,and ated cytotoxicity, Toll-like receptor, chemokine signaling, NOS [27–31]. Interestingly, both PCDH and BCL B were and cytokine-cytokine receptor interaction), and metabolism upregulated in CTS, while they had reduced expression in (purine metabolism and pyrimidine metabolism) (Figure 2, CRC[32,33].ItindicatesthatPCDH and BCL B could be Supplementary Table S5). Previous studies have shown that specifcally expressed in CTS cells but not in colon cancer some of these pathways were signifcantly associated with cells. colon cancer [38–41]. For example, the Wnt and Notch Many of the signifcantly downregulated genes in CTS pathways were associated with colon cancer development have been associated with CRC [34–37]. For example, MYOT, [38, 39]. Te cytokine-cytokine receptor interaction pathway ASPA,andKIAA were downregulated in CRC [34], the wassignifcantlyenrichedinCRC[34].TeECMandECM- downregulation of ARHGEF was associated with a poor associated proteins [39], the glycosaminoglycan metabolism, prognosis in CRC [35], higher expression levels of BCL- were and chondroitin sulfate/dermatan sulfate metabolism path- correlated with a better survival prognosis in CRC [36], and ways played key roles in mediating tumor microenvironment PPARGCA was a negative predictor for CRC prognosis [37]. [40, 41]. Altogether, a number of the abnormally expressed genes In addition, GSEA identifed six KEGG pathways that in CTS compared to colon normal stroma identifed by the were signifcantly associated with the downregulated genes in meta-analysishavebeenassociatedwithCRCpathologyand CTS (Supplementary Figure S2). Most of these pathways were prognosis. metabolism-related, including purine metabolism, histidine 4 Journal of Oncology

- LoＡ10 (FDR) 051 2 3 4 FNTB, and EP300) and 12 (PKNOX2, GATA2, MAPK10,

Pathways in cancer 5.21 TEAD1, TOX, MEF2A, GATA5, ELK1, MAZ, NHLH1, ATF1, Small cell lung cancer 2.69 ECM-receptor interaction 2.69 andRAD21)MMTRsfortheupregulatedandthedown- Leukocyte transendothelial migration 2.69 Focal adhesion 2.69 Wnt signaling pathway 2.69 regulated genes in CTS, respectively (Supplementary Table Glycosaminoglycan biosynthesis-chondroitin sulfate 2.69 p53 signaling pathway 2.69 S7), and built the regulatory networks associated with these Complement and coagulation cascades 2.31 Purine metabolism 2.26 Natural killer cell mediated cytotoxicity 1.98 MMTRs (Figure 4). In the networks, ATF6 (activating Cell adhesion molecules (CAMs) 1.98 Endocytosis 1.98 transcription factor 6), a TF regulating unfolded protein Toll-like receptor signaling pathway 1.88 Adherens junction 1.61 Chemokine signaling pathway 1.48 response during endoplasmic reticulum (ER) stress, targeted Tight junction 1.48 Apoptosis 1.44 163upregulatedgenes,andPKNOX2(PBX/knotted1home- Notch signaling pathway 1.44 Gap junction 1.44 obox 2), which plays key roles in regulating cell proliferation, Cytokine-cytokine receptor interaction 1.41 Pyrimidine metabolism 1.35 Regulation of actin cytoskeleton 1.32 diferentiation, and death, targeted 131 downregulated genes. Interestingly, two members of the GATA family of TFs Figure 2: Signiﬁcantly upregulated KEGG pathways in CTS relative (GATA2 and GATA5) were the MMTRs that regulated the to colon normal stroma identiﬁed by GSEA [ ]. FDR: false discovery downregulated genes in CTS (Figure 4(b)). rate. Altogether, the identifcation of upstream TFs, kinases, and MMTRs signifcantly associated with the DEGs may provide insights into the TME that mediates the development metabolism, glycine, serine, and threonine metabolism, and of colon cancer. drug metabolism-cytochrome p450. Tese pathways have been associated with colon and other cancers [42–44]. For . . CD+ T Cells Are More Enriched in CTS than in Normal example, impaired purine metabolism was associated with Stroma. We compared the enrichment levels of CD8+ T cells the progression of cancer [42]. Histidine metabolism could between CTS and normal stroma and found that CD8+ T cells boostcancertherapy[43].CytochromeP450enzymeswere showed signifcantly higher enrichment levels in CTS than associated with the metabolism of anticancer drugs and their in normal stroma (Student's t-test, p=0.016) (Figure 5). Tis expression was associated with a poor prognosis in CRC suggests an antitumor immune response activity in the TME patents [44]. of colon cancer.

.. Identification of Upstream TFs, Kinases, and MMTRs . . Identification of Prognostic Factors in Colon Cancer Based Significantly Associated with the DEGs. We identifed 11 sig- on the DEGs and eir Upstream Regulators. We investigated nifcant upstream TFs regulating the DEGs, including SUZ12, the association between the transcriptional signatures of CTS NFE2L2, RUNX1, ESR1, STAT3, TCF3, FOSL2, SALL4, AR, and survival prognosis (overall survival (OS) and disease- SMC3, and SOX2, of which the genes encoding RUNX1 free survival (DFS)) in the TCGA colon cancer dataset. Te and SALL4 were upregulated in CTS (Figure 3(a)). Most transcriptional signatures included the top 10 upregulated of these TFs have been associated with colon cancer [45– and top 10 downregulated genes in CTS on the basis of ES, 45 49]. For example, SUZ12 was the most signifcant upstream hub genes (≥3 degrees) from the zero-order PPI network of TF which could contribute to the CRC development [45]. the DEGs (Supplementary Table S8), and the genes encoding RUNX1 mutations were associated with the CRC risk [46]. 11 TFs, 124 kinases, and 21 MMTRs regulating the DEGs. We TCF3 and FOSL2 were associated with the tumorigenesis of foundthattheexpressionofmanyofthesetranscriptional CRC [47, 48]. Te overexpression of SOX2 was associated signatures was signifcantly associated with the survival of withtheprogressionandapoorprognosisincoloncancer colon cancer patients. For example, the expression of CEBPB, [49]. a gene signifcantly upregulated in CTS and a hub node in the Moreover, we identifed 124 signifcant protein kinases PPI network, had a signifcant negative correlation with OS in that regulate the DEGs (Figure 3(b), Supplementary Table colon cancer (Figure 6(a)). Te negative correlation between S6). Tese kinases mainly included cell cycle regula- CEBPB expression and survival has also been demonstrated tion kinases (CDKs), signaling MAP kinases (MAPKs, in other cancer types, such as high-grade serous ovarian MAP2Ks, and MAP3Ks), and ribosomal kinases (RPS6KA1, cancer [52]. PPARGC was signifcantly downregulated in RPS6KA3, and RPS6KA5). MAPK14 was the most sig- CTSandwasahubnodeinthePPInetwork,whileits nifcant upstream kinase negatively regulating the forma- expression had a signifcant positive correlation with OS in tion of colitis-associated colon tumors [50]. Furthermore, colon cancer (Figure 6(a)). PPARGCA was indicated as a weconstructedaTF-kinaseinteractionnetworkofthese tumor suppressor in colon cancer [53] and ovarian cancer TFs and kinases (Figure 3(c)). In the network, the most [54],aswellasanegativeprognosticbiomarkerforCRC[37]. connected TFs included SUZ12, NFE2L2, RUNX1, STAT3, Our data indicate that the deregulation of these genes in CTS FOSL2, AR, SMC3, ESR1, and TCF3, and the most con- is prognostic for colon cancer patients. nected kinases included MAPK14, CDK1, CSNK2A1, CDK2, Among the upstream regulators (TFs, kinases, and MAPK3, HIPK2, ERK1, and CDK4. It indicates that the cell MMTRs) of the DEGs, the expression of STAT, RPS KA , cycle regulation may play a pivotal role in CTS. IKBKE, ERBB, MTOR,andNFKB had a positive correlation MMTRs are interesting biomarkers and targets for with OS in colon cancer, while the expression of CDK, metabolism-targeted cancer therapy [51]. We identifed 9 CDK ,andBRD hadanegativecorrelationwithOSincolon (HNF1A, NFKB1, ZBTB7A, ATF6, TEAD4, TFAP2B, JAZF1, cancer (Figure 6(a)). Te deregulation of these genes has been Journal of Oncology 5

- LoＡ10 (p value) - LoＡ10 (p value) 04321 0215105 0

SUZ12 4.71 MAPK14 23.48 CDK1 23.03 NFE2L2 3.47 CSNK2A1 22.27 MAPK1 20.33 RUNX1 2.61 CDK4 18.06 MAPK3 17.1 2.51 ESR1 CDK2 16.52 GSK3B 15.96 STAT3 2.38 HIPK2 15.61 TCF3 2.28 ERK1 15.47 AKT1 13.81 FOSL2 2.02 MAPK8 13.7 ERK2 13.62 SALL4 1.99 JNK1 11.5 CK2ALPHA 11.47 AR 1.88 PKBALPHA 10.32 RPS6KA3 10.06 SMC3 1.62 ABL1 9.99 GSK3BETA 9.77 SOX2 1.37 DNAPK 9.28 Transcription Factor Protein Kinase (a) (b)

Kinase Intermediate protein Transcription factor Phosphorylation PPI

(c)

Figure 3: e significant upstream transcriptional factors (TFs) and kinases that regulate the differentially expressed genes (DEGs) between CTS and colon normal stroma identified by eXpressionKinases [ ]. (a) Signifcant upstream TFs regulating the DEGs. (b) Signifcant upstream kinases regulating the DEGs. (c) A TF-kinase interaction network of the signifcant upstream TFs and kinases regulating the DEGs. 6 Journal of Oncology

LZTS1S11 ENPEP TM7SF3 ZMYND19 SMYD2 POSTN FAM89A VAMP5 PLAUR GOLGA2 CST1 FADD GUK1 ARHGAP18A

FBXO28 KRT17

CLEC11A ASAP1 TFAP2B TMED3 SERPINB5INBINB5 FOXQ1 SERP1 PSMA5 GPRC5A EIF4E2 TMEM204 APLN TMEM200A IL1RAP TNFRSF11B DOPEY2 HYAL2 GPBAR1 DLC1 TM4SF1 CEACAM6 NCEH1 FAM122B FTSJ1 NT5DC2 LCN2AMMECR1 TIMP1 AGFG1 LEF1 EIF2AK4IF2AK4F22 K4K4 AFAP1L1 HS6ST2 CEACAM3 PECAM1 S100A11 MET LRRC15 JAZF1 CARS GRP NOS3 SPSB4 PTPRB LBX2BXBX2AP2S1 RCN3 MPZL2 MEOX1 HSPA12B OTUB2 KDELR2 RCN1 BAMBI SMAGP CDR2L SLC39A4 LRP88 FARSB TWF2 EFHD1 SEMA3G GPR4 CEP72 SEC61BFKBP144 PODXL BCL6B SULT2B1 TUBA1C VEZT ATF6 SEMA5A EMID1 TRIB3 KDELR3 RAB2A KCNQ1 FUT1 PCDH17 SEC24B KDELR1 PLVAP PRDX4 FGFRL1 SLC24A3 PLXDC2 CKAP4 GPC1PC SULF1 SALL4 HEATR3 CHPF

NME2 GNPNAT1 LRRC8E EFNB2 BGN SLC29A2 TMEM158 TBC1D4 FNTB TSPAN17 HS3ST3A1 NDFIP2 CYP4F11 NRAS VDAC1 FGR SLC12A8 ARL5B ARRDC1 OXTR FZD2 SLC35E1 FAM43AF GATA2 INHA GZMB TCEA1 S100A12 RCC2 NCOR2 AGRN UACA FOXC1 TMEM105 ETV5 CYP51A1 VSNL1 LOX USP13 LYZ HAS2 ICA1 NDUFA4L2 SERPINA1 C4orf32 PVR ST6GALNAC5 UBE2I TFPT ADAMTS6 FAM64A LAMC2 TESC SOX9 NOTCH3 YRDC TWIST1 BDNF CEBPB CRIP2 PHPT1PT1 NCOA7 LACTBB CDK6 SPTBN5 E2F7 CNN2 FOXA2 CHSY1 NUAK1 EHD4 ZNF668 P4HB TPM4 CTHRC1 ZBTB7A PYCR1 BACE2 TNIP3 LMNB2 PLCB4 SFXN1 SFTA2 IER3 SNX4 PLOD2 CLDN4 RPS6KA4 PPP1R14C PIP4K2B EVI5 CECR6 CLK2C ATP8B18B18BB11 UNC5B BAZ1ABAZBBAA SIX4 HDGF GSTO1 DAP PABPC1 SERPINE1 IRAK1 SLC25A22 EGR2 TMEM44 NFKB1 FRMD5 INSIG1 TMEM184B SLC1A7 CRB3 SCO2 CHMP4C TRAF2 TPBG ADAM12 GADD45AGADD4GADDDD44 CARD10 F12 HIST1H1B

ASPH RHOV CLDN1 CHTF18C EP300 FZD6 CXADR SLC1A4S RALA ITGA2 RAB3IP

SLC5A1 HNF1A YY1 CBX3 ICAM1 TEAD4 PLEK2NCF2NCCFC NFE2L3 NXT1 NOD2 VIL1

SPINK1SPS CDS1 FTL BICD1 DCBLD1 CCNB1IP1AURKBAAUURRB1IPIPP1P ESM1 REPIN1 UPF3A

(a)

DCAF5 ZNF304 MST1

ZNF5400 BHMT PAN2 USP44 CD5L

ANKRD13A ZPBP2

TRIM3C10orf76C10o0o CLIP4CLC AICDA LOC401052 GNG7G RASGRP22 MYCBP2 PASD1PAS

USP8 MTMR10 ATF1 KLK2LK2K2 KCNB1 FAM9C ZNF704 IGSF10 FLJ25758 MEF2A CXCL12CX ELK1 ENDOU RAD21 GATS MYOT

LSM14A

OMA1MMAMA1A1A TOX TMEM220 SLC22A7 SIK3 SLC17A3 FCRL4 ACHE C16orf89 ICA1L FAM120AOS CNR2 GATA2 SHBG GKAP1GKK ZNF772 IL9 EPM2A ZFP36L1 TXNIP CHST9 FAM131B FNIP1 CDC42EP4 RASSF5 CLCN6 NDUFB8 SMAP2 DNAH5DDN CDH12 METTL7A SSBP2 DISP1

TUBB STEAP4 SNAP91 CASP8 SCN4B MAPK10 OSBPL11 LYST C7 TEAD1 FAM102A RABGAP1L BCL6 SEMA6A TMEM100 TTBK2 PXK MAZ CBFA2T2 RGS9 ARHGEF37 HVCN1 SV2A DNMT3A STAT6 GPR162 ANK2 MAP2K6 FGF13 ABCB7 PTBP2 PDE4D CTNND2 ERICH1 MAOBM KIAA0513 ARHGAP15 ZNF536 ZNF407 GABRQ SLC9A3SLC9ASLCSLC9 ANKS1BA SCUBE1 GATA5 MFN2 SMARCA2 PKNOX2 PTH1R IGSF11 VAT1L MT1B FHL1 PDE9A KCNIP3 GPLD1 NPVF TRERF1TRRRERF1RERFREERF1EERFPRKG2RF1F11 SEMA3D ZNF268 FHIT TGM4M4 RSPO2 ATRNL1 KCNIP4KCKKCNICNNIP4PP4 RFX2 C1QTNF9 CRADD COLEC12 CTNNA2 ZNF780B APOBEC2APOBECAAPPO ECC2C2 SLITRK2 NEUROG3 BCL2 EFCAB1 PIK3C2B NHLH1 NPTX1 XK YTHDC1 PDE6B FAM124A P2RY1 TMEM116 PPARGC1A ADAMTS19 ARHGAP20 CR2 EPHB6B66 UST PCF11 LRRTM33 RCAN2 RIT2 SLC9A9 FSIP2 MTX3 PELI2 RSPO1 CA14 BAALC ANTXR2 AKAP7 GAS8 CCDC136 FBXO15 CRTC1 CTF1 CRY2 ESYT3

BRS3 ATP5L H3F3B

(b)

Figure 4: Regulatory networks of the master metabolic transcriptional regulators (MMTRs) and their targeted diﬀerentially expressed genes (DEGs) between CTS and normal stroma identiﬁed by iRegulon []. (a) Regulatory network of the MMTRs and their targeted upregulated genes in CTS. (b) Regulatory network of the MMTRs and their targeted downregulated genes in CTS. Te green color octagon indicates MMTRs and purple color oval indicates DEGs. Journal of Oncology 7

0.6

0.4

0.2

0.0

CD8+ T cell enrichment level level CD8+ T cell enrichment −0.2 p = 0.016

−0.4

Tumor stroma Normal stroma Figure 5: CD+ T cells have signiﬁcantly higher enrichment levels in CTS than in colon normal stroma. Student's t-test p value is shown. associated with tumor progression in a wide variety of cancer downregulated pathways were mostly metabolism-related. types [55–60]. Tese results revealed the abnormal alterations of cellular In addition, we identifed 18 transcriptional signatures development, immune regulation, and metabolism pathways of CTS whose expression was signifcantly associated with in CTS. We found that CD8+ T cells were more enriched in DFS in colon cancer individually (Supplementary Figure CTS than in colon normal stroma, suggesting an immune S3). Tese genes included CEBPB, BCL, PAN, NOS, infltration microenvironment in CTS. Furthermore, we FTL, ARHGEF, SMC, EP, JAK, RPS KA, RPS KA, identifed numerous CTS transcriptional signatures whose PRKACA, HIPK, HIPK, MAPK, GSKA, CLK,andCDK. expression was signifcantly associated with prognosis in It indicates that these CTS transcriptional signatures could be colon cancer, such as CEBPB, PPARGC, STAT, MTOR, biomarkers for colon cancer relapse. BCL, JAK,andCDK. Tese transcriptional signatures Furthermore, we used the multivariate analysis to validate are mainly involved in immune regulation (CEBPB, STAT, the association between the prognostic CTS transcriptional and JAK), metabolism (PPARGC and MTOR), cell cycle signatures and survival using the colon metabase data [10]. (CDK), and apoptosis (BCL), suggesting that the deregu- ForOSanalysis,atotalof482patientsweresplitinto lation of these pathways in CTS may contribute to the altered two groups: high-risk group (N=241) versus low-risk group prognosis in colon cancer. (N=241) based on the prognostic index (Supplementary To verify the association of the identifed transcriptional Figure S4A). As expected, the high-risk group had worse signatures with CTS, we analyzed the TCGA colon cancer OS than the low-risk group (Figure 6(b)). Similarly, for dataset. We divided these cancers into high-stroma-content DFS analysis, we divided patients into the high-risk group and low-stroma-content groups on the basis of their intra- (N=272) and the low-risk group (N=273) based on the tumoral stromal content evaluated by ESTIMATE [19] and prognostic index (Supplementary Figure S4B) and found that foundthat153upregulatedgenesinCTShadsignifcantly the high-risk group had worse DFS compared to the low-risk higher expression levels in the high-stroma-content group group (Figure 6(c)). Tese results proved the prognostic value than in the low-stroma-content group. Tese genes included of these CTS transcriptional signatures in colon cancer. 18 hub genes in the PPI network of DEGs and 6 TFs, 40 kinases, and 12 MMTRs encoding genes that regulated the 4. Discussion DEGs (Supplementary Figure S5, Table S9). We also found 27 downregulated genes in CTS which had signifcantly Te tumor stroma constitutes an important component of lower expression levels in the high-stroma-content group, the TME that mediates tumor growth, immune evasion, and including 14 hub genes, and genes encoding 2 TFs, 18 metastasis [1]. Tus, it is important to identify molecular kinases, and 3 MMTRs (Supplementary Figure S5, Table S9). features in the tumor stroma. To this end, we performed Interestingly, most of the downregulated hub genes in CTS a meta-analysis of three CTS transcriptome datasets for were also downregulated in the high-stroma-content colon identifying CTS-associated transcriptional signatures. We cancers (Supplementary Figure S5). Tese results indicate identifed a number of upregulated and downregulated genes that many transcriptional signatures of CTS identifed by in CTS compared to colon normal stroma. Furthermore, the meta-analysis are tumor stroma-specifc. In addition, we we identifed upregulated and downregulated pathways sig- found that CD8+ T cells had signifcantly higher enrichment nifcantly associated with these deregulated genes in CTS. levels in CTS versus colon normal stroma (Student's t-test, Teupregulatedpathwaysweremainlyinvolvedincellular p=0.016), as well as in the high-stroma-content colon can- development, immune regulation, and metabolism, and the cers versus the low-stroma-content colon cancers (Student's 8 Journal of Oncology

CEBPB PPARGC1A STAT3 NFKB1

p = 0.036 p = 0.003 p = 0.009 p = 0.03 0.8 0.8 0.8 0.8 0.4 0.4 0.4 0.4 0.0 0.0 0.0 0.0 Probability of overall survival overall of Probability 0 1000 3000 0 1000 3000 0 1000 3000 0 1000 3000

CDK1 CDK5 BRD2 RPS6KA5

p = 0.008 p = 0.04 p = 0.047 p = 0.045 0.8 0.8 0.8 0.8 0.4 0.4 0.4 0.4 0.0 0.0 0.0 0.0 Probability of overall survival overall of Probability 0 1000 3000 0 1000 3000 0 1000 3000 0 1000 3000

ERBB2 IKBKE MTOR

p = 0.018 p = 0.043 p = 0.024 0.8 0.8 0.8 0.4 0.4 0.4 0.0 0.0 0.0 Probability of overall survival overall of Probability 0 1000 3000 0 1000 3000 0 1000 3000 Overall survival time (days) Higher expression level Lower expression level (a) Concordance Index=63.4,Log-Rank Equal Curves p=1.105e-06,R^2=0.074/0.988 Concordance Index=61.64,Log-Rank Equal Curves p=0.0001785,R^2=0.044/0.935 Risk Groups Hazard Ratio=2.07 (conf.int.1.53~2.79),p=1.959e-06 Risk Groups Hazard Ratio=2(conf.int.1.38~2.89),p=0.0002425 1.0 1.0 Low-risk group 0.8 Low-risk group 0.8

0.6 0.6 High-risk group 0.4 High-risk group 0.4 0.2 0.2

Probability of overall survival overall of Probability 0.0 Probability of disease of survival free Probability 0.0 0 50 100 150 200 0 50 100 150 200 241 137224679134202 273 124153464132211 134172860100154241 145153363118190272 Overall survival time (months) Disease-free survival time (months) 241,+:173,CI=55.8 273,+:230,CI=57.6 241,+:125,CI=57 272,+:190,CI=56.9 (b) (c)

Figure 6: e CTS gene signatures whose expression is associated with prognosis in colon cancer. (a) Kaplan-Meier survival curves show the gene signatures whose expression is signifcantly associated with overall survival (OS) in colon cancer in the TCGA colon cancer dataset (log- rank test, p<0.05). (b) Multivariate Cox regression analysis shows that the OS-associated CTS gene signatures are prognostic for OS in colon cancer in a SurvExpress built-in dataset (colon metabase) [10]. (c) Multivariate Cox regression analysis shows that the DFS-associated CTS gene signatures are prognostic for DFS in colon cancer in a SurvExpress built-in dataset (colon metabase) [10]. DFS: disease-free survival. Journal of Oncology 9

High-stroma-content colon cancers Low-stroma-content colon cancers

1.0 p= 0.499 p= 0.031 Probability of disease-free of survival Probability 0.0 0.2 0.4 0.6 0.8 0.0 0.2 0.4 0.6 0.8 1.0 0 1000 2000 3000 4000 0 1000 2000 3000 4000

Disease-free survival time (days)

Higher CD8+ T cells enrichment levels Lower CD8+ T cells enrichment levels

Figure 7: e higher enrichment levels of CD+ T cells were associated with better disease-free survival in the low-stroma-content colon cancers, but not in the high-stroma-content colon cancers. ESTIMATE [19] was used to quantify the intratumoral stromal content (stroma score) of TCGA colon cancer samples. High-stroma-content: stroma score > median; low-stroma-content: stroma score < median.

−8 t-test, p=3.3∗10 ). It indicates that CD8+ T cells tend to have GEO: Gene Expression Omnibus elevated infltration in the TME of colon cancer. Interestingly, TCGA: Te Cancer Genome Atlas we found that the higher enrichment levels of CD8+ T cells FDR: False discovery rate were associated with better DFS in the low-stroma-content GSEA: Gene-set enrichment analysis colon cancers, but not in the high-stroma-content colon FFM: Function-frst module cancers (Figure 7). It suggests that the immune cells exert an CFM: Connection-frst module. antitumor efect only when they have infltrated into tumor cells and that the immune cells in the tumor stroma may not Data Availability have such a direct antitumor efect. Tis study has identifed a number of CTS-associated Te datasets (GSE31279, GSE35602, and GSE46824) were transcriptional signatures that could be biomarkers for colon downloaded from the NCBI GEO database (https://www cancer diagnosis and prognosis and may provide therapeutic .ncbi.nlm.nih.gov/geo/), the TCGA colon cancer dataset was targets for colon cancer. However, to translate these fndings downloaded from the website https://portal.gdc.cancer.gov/, into clinical application, further experimental and clinical and the colon metabase dataset was from SurvExpress validation would be necessary. (http://bioinformatica.mty.itesm.mx/SurvExpress).

5. Conclusions Conflicts of Interest Te identifcation of CTS-specifc transcriptional features Te authors declare that they have no conficts of interest may provide insights into the mechanism that mediates the regarding the publication of this paper. development of colon cancer and thus has potential clinical implications for colon cancer diagnosis and treatment. Authors’ Contributions Abbreviations Md. Nazim Uddin performed data analyses and helped CTS: Colon tumor stroma write and prepare the manuscript. Mengyuan Li performed DEGs: Diferentially expressed genes data analyses and helped prepare the manuscript. Xiaosheng PPI: Protein-protein interaction Wang conceived the research, designed analysis strategies, ES: Efect size and wrote the manuscript. All the authors read and approved TME: Tumor microenvironment the fnal manuscript. ECM: Extracellular matrix TFs: Transcription factors Acknowledgments MMTRs: Master metabolic transcription factors OS: Overall survival Tis work was supported by China Pharmaceutical Uni- DFS: Disease-free survival versity (grant numbers 3150120001 and 2632018YX01 to CRC: Colorectal cancer Xiaosheng Wang). 10 Journal of Oncology

Supplementary Materials [2]K.M.Bussard,L.Mutkus,K.Stumpf,C.Gomez-Manzano, and F. C. Marini, “Tumor-associated stromal cells as key Supplementary Figure S1: illustration of PCA and density contributors to the tumor microenvironment,” Breast Cancer plotsasvalidationtoolsforbatchefectremoval.Plotof Research,vol.18,no.1,p.84,2016. principal components: (A) before batch efect removal and [3] D. Hanahan and L. M. Coussens, “Accessories to the crime: (B) afer batch efect removal. Plot of density: (C) before functions of cells recruited to the tumor microenvironment,” batch efect removal and (D) afer batch efect removal. Te Cancer Cell,vol.21,no.3,pp.309–322,2012. multidimensional scaling of the datasets demonstrates that, [4] F. Bray, J. Ferlay, I. Soerjomataram, R. L. Siegel, L. A. Torre, and before the batch efect adjustment, each dataset obviously A. Jemal, “Global cancer statistics 2018: GLOBOCAN estimates separated from all the others, whereas, afer batch efect of incidence and mortality worldwide for 36 cancers in 185 countries,” CA: A Cancer Journal for Clinicians,vol.68,no.6, adjustment, samples from all the datasets are incorporated pp. 394–424, 2018. clearly. Supplementary Figure S2: six KEEG pathways that [5]A.Calon,E.Lonardo,andA.Berenguer-Llergo,“Stromal are signifcantly associated with the downregulated genes gene expression defnes poor-prognosis subtypes in colorectal in colon tumor stroma relative to normal stroma. Supple- cancer,” Nature Genetics,vol.47,pp.320–329,2015. mentary Figure S3: the genes (DEGs and their upstream [6] C. Isella, A. Terrasi, S. E. Bellomo et al., “Stromal contribution regulators) whose expression is signifcantly associated with to the colorectal cancer transcriptome,” Nature Genetics,vol.47, disease-free survival in colon cancer (log-rank test, p<0.05). no. 4, pp. 312–319, 2015. Supplementary Figure S4: patients divided into the high- [7] M. Abba, S. Laufs, M. Aghajany, B. Korn, A. Benner, and riskgroupandthelow-riskgroupbasedontheprognostic H. Allgayer, “Look who’s talking: Deregulated signaling in gene signatures identifed. A. Overall survival. B. Disease-free colorectal cancer,” Cancer Genomics & Proteomics,vol.9,no.1, survival. Supplementary Figure S5: numbers of overlapping pp.15–25,2012. genes between the DEGs between colon tumor stroma and [8]N.Nishida,M.Nagahara,T.Satoetal.,“Microarrayanalysis normal stroma and their upstream regulators and the DEGs of colorectal cancer stromal tissue reveals upregulation of two between high-stroma-content and low-stroma-content colon oncogenic miRNA clusters,” Clinical Cancer Research,vol.18, no. 11, pp. 3054–3070, 2012. cancers. UP TCGA: upregulated diferentially expressed genes between high-stroma-content and low-stroma-content [9] “GEO Accession viewer,” https://www.ncbi.nlm.nih.gov/geo/query/ acc.cgi?acc=GSE46824. TCGA colon cancer samples and DOWN TCGA: downreg- [10] R. Aguirre-Gamboa, H. Gomez-Rueda, E. Mart´ınez-Ledesma ulated diferentially expressed genes between high-stroma- et al., “SurvExpress: an online biomarker validation tool and content and low-stroma-content TCGA colon cancer sam- database for cancer gene expression data using survival anal- ples. Supplementary Table S1: a summary of the datasets ysis,” PLoS ONE,vol.8,no.9,p.e74250,2013. used in this study. Supplementary Table S2: upregulated [11] J. Xia, E. E. Gill, and R. E. W. Hancock, “NetworkAnalyst genes in colon tumor stroma versus colon normal stroma. for statistical, visual and network-based meta-analysis of gene Supplementary Table S3: downregulated genes in colon expression data,” Nature Protocols,vol.10,no.6,pp.823–844, tumor stroma versus colon normal stroma. Supplementary 2015. TableS4:thetop10upregulatedandtop10downregulated [12] W. E. Johnson, C. Li, and A. Rabinovic, “Adjusting batch efects genes in colon tumor stroma. Supplementary Table S5: 44 in microarray expression data using empirical Bayes methods,” KEGG pathways that were signifcantly associated with the Biostatistics,vol.8,no.1,pp.118–127,2007. upregulated genes in colon tumor stroma (CTS). Supple- [13] W. G. Cochran, “Te combination of estimates from diferent mentary Table S6: upstream transcription factors and kinases experiments,” Biometrics,vol.10,no.1,pp.101–129,1954. regulating the diferentially expressed genes between colon [14] Y. Benjamini and Y. Hochberg, “Controlling the false discovery tumor stroma and normal stroma. Supplementary Table rate: a practical and powerful approach to multiple testing,” JournaloftheRoyalStatisticalSocietyB:Methodological,vol.57, S7: master metabolic transcriptional regulators (MMTRs) no. 1, pp. 289–300, 1995. (iRegulon normalized enrichment score NES > 3.0) regulat- [15] A. Subramanian, P. Tamayo, V. K. Mootha et al., “Gene set ing the diferentially expressed genes between colon tumor enrichment analysis: a knowledge-based approach for inter- stroma and normal stroma. Supplementary Table S8: hub preting genome-wide expression profles,” Proceedings of the genes in the protein-protein interaction network of the NationalAcadamyofSciencesoftheUnitedStatesofAmerica, diferentially expressed genes between colon tumor stroma vol. 102, no. 43, pp. 15545–15550, 2005. and normal stroma. Supplementary Table S9: overlapping [16]D.J.B.Clarke,M.V.Kuleshov,B.M.Schilderetal.,“EXpres- genes between the DEGs between colon tumor stroma and sion2Kinases (X2K) Web: Linking expression signatures to normal stroma and their upstream regulators and the DEGs upstream cell signaling networks,” Nucleic Acids Research,vol. between high-stroma-content and low-stroma-content colon 46,no.W1,pp.W171–W179,2018. cancers. (Supplementary Materials) [17] R. Janky, A. Verfaillie, H. Imrichovaetal.,“iRegulon:froma´ gene list to a gene regulatory network using large motif and References track collections,” PLoS Computational Biology,vol.10,no.7, Article ID e1003731, 2014. [1]K.C.Valkenburg,A.E.deGroot,andK.J.Pienta,“Targeting [18]D.Petrochilos,A.Shojaie,J.Gennari,andN.Abernethy, the tumour stroma to improve cancer therapy,” Nature Reviews “Using random walks to identify cancer-associated modules in Clinical Oncology,vol.15,no.6,pp.366–381,2018. expression data,” BioData Mining,vol.6,p.17,2013. Journal of Oncology 11

[19] K. Yoshihara, M. Shahmoradgoli, E. Mart´ınez et al., “Inferring bioinformatical analysis,” International Journal of Molecular tumour purity and stromal and immune cell admixture from Sciences,vol.18,no.4,2017. expression data,” Nature Communications,vol.4,p.2612,2013. [35] Q. Sun, M. Li, and X. Wang, “Te Cancer Omics Atlas: [20] X.-M. Yang, H.-Y. You, Q. Li et al., “CTHRC1 promotes An integrative resource for cancer omics annotations,” BMC human colorectal cancer cell proliferation and invasiveness by Medical Genomics,vol.11,no.1,p.63,2018. activating Wnt/PCP signaling,” International Journal of Clinical [36] Q. Huang, S. Li, P. Cheng et al., “High expression of anti- and Experimental Pathology,vol.8,no.10,pp.12793–12801,2015. apoptotic protein Bcl-2 is a good prognostic factor in colorectal [21] B. Guo, H. Yan, L. Li, K. Yin, F. Ji, and S. Zhang, “Collagen triple cancer: result of a meta-analysis,” World Journal of Gastroen- helix repeat containing 1 (CTHRC1) activates Integrin �3/FAK terology,vol.23,no.27,pp.5018–5033,2017. signaling and promotes metastasis in ovarian cancer,” Journal of [37] S.-H. Yun, M.-S. Roh, J.-S. Jeong, and J.-I. Park, “Peroxisome Ovarian Research,vol.10,no.1,p.69,2017. proliferator-activated receptor � coactivator-1� is a predictor of [22] A. M. Chowdhury, H. Katoh, A. Hatanaka et al., “Multiple lymph node metastasis and poor prognosis in human colorectal regulatory mechanisms of the biological function of NRF3 cancer,” Annals of Diagnostic Pathology,vol.33,pp.11–16,2018. (NFE2L3) control cancer cell proliferation,” Scientiﬁc Reports, [38]F.E.Bertrand,C.W.Angus,W.J.Partis,andG.Sigounas, vol.7,no.1,p.12494,2017. “Developmental pathways in colon cancer,” Cell Cycle,vol.11, [23] M. Palma, L. Lopez, M. Garc´ıa et al., “Detection of collagen no. 23, pp. 4344–4351, 2012. triple helix repeat containing-1 and nuclear factor (erythroid- [39] A. M. Socovich and A. Naba, “Te cancer matrisome: From derived 2)-like 3 in colorectal cancer,” BMC Clinical Pathology, comprehensive characterization to biomarker discovery,” Semi- vol.12,p.2,2012. nars in Cell & Developmental Biology,2018. [24]R.M.S.Gill,A.Michael,L.Westley,H.M.Kocher,J.I.Murphy, [40] N. Afratis, C. Gialeli, D. Nikitovic et al., “Glycosaminoglycans: and G. K. Dhoot, “SULF1/SULF2 splice variants diferentially key players in cancer cell biology and treatment,” FEBS Journal, regulate pancreatic tumour growth progression,” Experimental vol.279,pp.1177–1197,2012. Cell Research,vol.324,no.2,pp.157–171,2014. [41] Basappa, K. S. Rangappa, and K. Sugahara, “Roles of gly- [25] H. Lee, B. Yeh, T. Chan et al., “Sulfatase-1 overexpression cosaminoglycans and glycanmimetics in tumor progression and indicates poor prognosis in urothelial carcinoma of the urinary metastasis,” Glycoconjugate Journal,vol.31,no.6-7,pp.461–467, bladderanduppertract,”Oncotarget,vol.8,no.29,pp.47216– 2014. 47229, 2017. [42] J. Yin, W. Ren, X. Huang, J. Deng, T. Li, and Y. Yin, “Potential [26] B. Lu,Y.Fang,J.Xuetal.,“AnalysisofSOX9expressionin¨ mechanisms connecting purine metabolism and cancer ther- colorectal cancer,” American Journal of Clinical Pathology,vol. apy,” Frontiers in Immunology,vol.9,2018. 130, no. 6, pp. 897–904, 2008. [43] C. Frezza, “Histidine metabolism boosts cancer therapy,” Nature,vol.559,no.7715,pp.484-485,2018. [27]S.Arikan,C.Cacina,E.Guler,S.C¸ ulcu, G. Tuna, and I. Yaylim- Eraltan, “Te efects of NOS3 Glu298Asp variant on colorectal [44]M.Kumarakulasingham,P.H.Rooney,S.R.Dundasetal., cancer risk and progression in Turkish population,” Molecular “Cytochrome P450 profle of colorectal cancer: Identifcation of Biology Reports,vol.39,no.3,pp.3245–3249,2012. markers of prognosis,” Clinical Cancer Research, vol. 11, no. 10, pp. 3758–3765, 2005. [28] S. Chida, H. Okayama, M. Noda et al., “Stromal VCAN expression as a potential prognostic biomarker for disease recurrence [45] Y.-L. Liu, X. Gao, Y. Jiang et al., “Expression and clinico- in stage II-III colon cancer,” Carcinogenesis,vol.37,no.9,pp. pathological signifcance of EED, SUZ12 and EZH2 mRNA 878–887, 2016. in colorectal cancer,” Journal of Cancer Research and Clinical Oncology, vol. 141, no. 4, pp. 661–669, 2015. [29] S. Demirkol, I. Gomceli, M. Isbilen et al., “A combined ULBP2 [46] M. L. Slattery, A. Lundgreen, J. S. Herrick, B. J. Caan, J. D. Potter, and SEMA5A expression signature as a prognostic and predic- and R. K. Wolf, “Associations between genetic variation in tive biomarker for colon cancer,” Journal of Cancer,vol.8,no.7, RUNX1, RUNX2, RUNX3, MAPK1 and eIF4E and risk of colon pp. 1113–1122, 2017. and rectal cancer: additional support for a TGF-�-signaling [30] M. Fujita, Y. Furukawa, T. Tsunoda, T. Tanaka, M. Ogawa, and pathway,” Carcinogenesis,vol.32,no.3,pp.318–326,2011. Y. Nakamura, “Up-regulation of the ectodermal-neural cortex 1 [47] K. Taniue, A. Kurimoto, Y. Takeda et al., “ASBEL –TCF3 (ENC1) gene, a downstream target of the �-catenin/T-cell factor complex is required for the tumorigenicity of colorectal cancer complex, in colorectal carcinomas,” Cancer Research,vol.61,no. cells ,” Proceedings of the National Acadamy of Sciences of the 21, pp. 7722–7726, 2001. United States of America, vol. 113, no. 45, pp. 12739–12744, 2016. [31] Y.Li, J. Wei, C. Xu, Z. Zhao, and T. You, “Prognostic signifcance [48] T. Bae, K. Rho, J. W. Choi, K. Horimoto, W. Kim, and S. Kim, of cyclin D1 expression in colorectal cancer: a meta-analysis “Identifcation of upstream regulators for prognostic expression of observational studies,” PLoS ONE,vol.9,no.4,ArticleID signature genes in colorectal cancer,” BMC Systems Biology,vol. e94508, 2014. 7,no.1,p.86,2013. [32]X.Hu,X.Sui,L.Lietal.,“Protocadherin17actsasatumour [49] T. J. Miller, M. J. McCoy, C. Hemmings, M. K. Bulsara, B. suppressor inducing tumour cell apoptosis and autophagy, and Iacopetta, and C. F. Platell, “Te prognostic value of cancer is frequently methylated in gastric and colorectal cancers,” e stem-like cell markers SOX2 and CD133 in stage III colon cancer Journal of Pathology,vol.229,no.1,pp.62–73,2013. is modifed by expression of the immune-related markers [33]Y.Gu,A.Li,H.Sunetal.,“BCL6Bsuppressesproliferation FoxP3, PD-L1 and CD3,” Pathology,vol.49,no.7,pp.721–730, and migration of colorectal carcinoma cells through inhibition 2017. of the PI3K/AKT signaling pathway,” International Journal of [50] J. Gupta, I. delBarcoBarrantes, A. Igea et al., “Dual function of Molecular Medicine,vol.41,no.5,pp.2660–2668,2018. p38� MAPK in colon cancer: suppression of colitis-associated [34] Y. Guo, Y. Bao, M. Ma, and W. Yang, “Identifcation of key tumor initiation but requirement for cancer cell survival,” candidate genes and pathways in colorectal cancer by integrated Cancer Cell,vol.25,no.4,pp.484–500,2014. 12 Journal of Oncology

[51]S.R.Rosario,M.D.Long,H.C.Afronti,A.M.Rowsam,K.H. Eng, and D. J. Smiraglia, “Pan-cancer analysis of transcriptional metabolic dysregulation using the cancer genome atlas,” Nature Communications,vol.9,2018. [52] D. Liu, X.-X. Zhang, M.-C. Li et al., “C/EBP� enhances platinum resistance of ovarian cancer cells by reprogramming H3K79 methylation,” Nature Communications,vol.9,no.1,p.1739,2018. [53] J. Feilchenfeldt, M.-A. Brundler,C.Soravia,M.T¨ otsch,¨ and C. A. Meier, “Peroxisome proliferator-activated receptors (PPARs) and associated transcription factors in colon cancer: reduced expression of PPARgamma-coactivator 1 (PGC-1),” Cancer Let- ters,vol.203,no.1,pp.25–33,2014. [54] Y. Zhang, Y. Ba, C. Liu et al., “PGC-1alpha induces apoptosis in human epithelial ovarian cancer cells through a PPARgamma- dependent pathway,” Cell Research,vol.17,no.4,pp.363–373, 2007. [55] J. Concetti and C. L. Wilson, “NFKB1 and cancer: friend or foe?” Cells,vol.7,no.9,2018. [56]U.Saini,S.Naidu,A.C.Elnaggaretal.,“ElevatedSTAT3 expression in ovarian cancer ascites promotes invasion and metastasis: A potential therapeutic target,” Oncogene,vol.36, no. 2, pp. 168–181, 2017. [57] W. J. Yang, Y. Qu, B. X. Tan et al., “Prognostic signifcance of preoperative IKBKE expression in esophageal squamous cell carcinoma,” OncoTargets and erapy,vol.11,pp.1305–1314, 2018. [58] W. Sung, Y. Lin, P. Wu et al., “High nuclear/cytoplasmic ratio of Cdk1 expression predicts poor prognosis in colorectal cancer patients,” BMC Cancer,vol.14,p.951,2014. [59]K.Zhuang,J.Zhang,M.Xiongetal.,“CDK5functionsasa tumor promoter in human colorectal cancer via modulating the ERK5–AP-1 axis,” Cell Death & Disease,vol.7,no.10,p.e2415, 2016. [60]Y.Hu,J.Zhou,F.Yeetal.,“BRD4inhibitorinhibitscolorectal cancer growth and metastasis,” International Journal of Molecu- lar Sciences,vol.16,no.1,pp.1928–1948,2015. M EDIATORSof INFLAMMATION

The Scientifc Gastroenterology Journal of Research and Practice Diabetes Research Disease Markers World Journal Hindawi Hindawi Publishing Corporation Hindawi www.hindawi.com Volume 2018 Hindawi Hindawi http://www.hindawi.comwww.hindawi.com Volume 20182013 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Journal of International Journal of Immunology Research Endocrinology Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Submit your manuscripts at www.hindawi.com

BioMed PPAR Research Research International Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Journal of Obesity

Evidence-Based Journal of Stem Cells Complementary and Journal of Ophthalmology International Alternative Medicine Oncology Hindawi Hindawi Hindawi Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2013

Parkinson’s Disease

Computational and Mathematical Methods Behavioural AIDS Oxidative Medicine and in Medicine Neurology Research and Treatment Cellular Longevity Hindawi Hindawi Hindawi Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018