Hindawi Computational and Mathematical Methods in Medicine Volume 2020, Article ID 6138039, 13 pages https://doi.org/10.1155/2020/6138039

Research Article Comprehensive Analysis of Immunoinhibitors Identifies LGALS9 and TGFBR1 as Potential Prognostic Biomarkers for Pancreatic Cancer

Yue Fan,1 Tianyu Li,1 Lili Xu,1 and Tiantao Kuang 2

1Department of Integrated TCM & Western Medicine, Zhongshan Hospital, Fudan University, Shanghai 200032, China 2Department of General Surgery, Zhongshan Hospital, Fudan University, Shanghai 200032, China

Correspondence should be addressed to Tiantao Kuang; [email protected]

Received 8 June 2020; Accepted 21 July 2020; Published 30 September 2020

Guest Editor: Lei Chen

Copyright © 2020 Yue Fan et al. This 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.

Pancreatic cancer (PC) is one of the most deadly cancers worldwide. To uncover the unknown novel biomarker used to indicate early diagnosis and prognosis in the molecular therapeutic field of PC is extremely of importance. Accumulative evidences indicated that aberrant expression or activation of immunoinhibitors is a common phenomenon in malignances, and significant associations have been noted between immunoinhibitors and tumorigenesis or progression in a wide range of cancers. However, the expression patterns and exact roles of immunoinhibitors contributing to tumorigenesis and progression of pancreatic cancer (PC) have not yet been elucidated clearly. In this study, we investigated the distinct expression and prognostic value of immunoinhibitors in patients with PC by analyzing a series of databases, including TISIDB, GEPIA, cBioPortal, and Kaplan- Meier plotter database. The mRNA expression levels of IDO1, CSF1R, VTCN1, KDR, LGALS9, TGFBR1, TGFB1, IL10RB, and PVRL2 were found to be significantly upregulated in patients with PC. Aberrant expression of TGFBR1, VTCN1, and LGALS9 was found to be associated with the worse outcomes of patients with PC. Bioinformatics analysis demonstrated that LGALS9 was involved in regulating the type I interferon signaling pathway, interferon-gamma-mediated signaling pathway, RIG-I-like receptor signaling pathway, NF-kappa B signaling pathway, cytosolic DNA-sensing pathway, and TNF signaling pathway. And TGFB1 was related to mesoderm formation, cell matrix adhesion, TGF-beta signaling pathway, and Hippo signaling pathway. These results suggested that LGALS9 and TGFBR1 might serve as potential prognostic biomarkers and targets for PC.

1. Introduction the molecular therapeutic field of PC is extremely of importance. The mortality of pancreatic cancer (PC), the fourth widely Previously, the application of the immune system to rec- occurred cancer with poor prognosis, has an overall five- ognize and eradicate tumors has made significant advance in year survival rate lower than 10% [1]. Due to the hidden the clinical use of cancer immunotherapy [7–9]. Notably, the symptoms at early stages, fewer than 15% of patients are emergence of immune checkpoints inhibitors typically inter- diagnosed with PC at a stage when they could be eligible fered negative regulators of T cell immunity including LAG3 for curative surgical resection [2]. To improve early detection [10–12], CTLA-4 [13, 14], PD-1 [15, 16], and TIM3 [17, 18]. and prognosis and to provide timely and effective treatment The advent of these “checkpoint inhibitors” has thoroughly for high-risk patients, predictive biomarkers for PC are altered and improved the former therapies for melanoma, required [3, 4]. To date, carbohydrate antigen 199 (CA199) lung cancer, and so on [19]. For instance, interference of and CA242, which are currently used in clinical settings as LAG3 relieved the exhaustion of T cells and heightened serum biomarkers for PC, are inadequate for early screening immunity against tumor due to the interaction among and prognosis [5, 6]. To uncover the unknown novel bio- LAG3 with MHC class II and galectin 3 [20]. Additionally, marker used to indicate early diagnosis and prognosis in tumor-infiltrating lymphocyte- (TIL-) produced TIM3 has 2 Computational and Mathematical Methods in Medicine been identified to display a key role in maintaining inactive the TCGA pipeline cancer [31]. Besides, the coexpression lymphocyte status or inducing lymphocyte apoptosis [21]. and interaction of selected proteins were probed referred to LGALS9 is a ligand of TIM-3 and expressed in a variety of cBioPortal guidelines [32]. cell types, especially in lymphoid organs and monocytes [22]. In addition, LGALS9 could impose unequal effects on 2.4. Gene Ontology and Pathway Enrichment Analysis. Gene immune cells in a tumor microenvironment [22]. TGF-β sig- ontology (GO) and Kyoto Encyclopedia of Genes and naling exhibited importance in biological signal regulation Genomes (KEGG) pathway enrichment analysis was per- including cell growth and death, differentiation, angiogene- formed using DAVID online tool. P <0:01 was set as the cut- sis, and inflammation [23]. Several recent studies demon- off criterion. strated that TGF-β signaling played a key role in immune response [24]. Understanding the potential functions and 2.5. Statistical Analysis. Student’s t-test was analyzed for sta- expression pattern of immune “checkpoint inhibitors” could tistical significance. Statistical analysis was performed by be helpful for the identification of novel prognosis and treat- SPSS 21.0 (SPSS Inc., Chicago, IL). ment biomarkers for PC. The occurrence and progression of newly produced strat- 3. Results egies, comprising microarray and RNA-sequencing, exerted a positive effect in molecular research and also gave impetus 3.1. Identification of Immunoinhibitor Expression Pattern to exploring accurate and safe treatment for PC [25–27]. in PC. The present study analyzed the expression pattern Here, we expanded PC-related knowledge in view of different of 23 immunoinhibitors in PC using TCGA database, includ- databases, thus generating a conclusive analysis of the link ing CD160, CD244, KIR2DL1, KIR2DL3, BTLA, CSF1R, between the function of immune checkpoint inhibitors and HAVCR2, TIGIT, LAG3, PDCD1, VTCN1, PDCD1LG2, the diagnosis along with the development of PC. LGALS9, CD96, TGFBR1, TGFB1, CTLA4, ADORA2A, PVRL2, IL10, IDO1, IL10RB, and KDR. As shown in 2. Materials and Methods Figure 1, we found that IDO1, CSF1R, VTCN1, KDR, LGALS9, TGFBR1, TGFB1, IL10RB, and PVRL2 were highly 2.1. Survival Analysis. Kaplan-Meier plotter (http://www expressed in PC tissues. .kmplot.com/) is an online database containing microarray gene expression data, and survival information derived from 3.2. Increasing Expression of Immunoinhibitors Was Observed Gene Expression Omnibus, TCGA, and the Cancer Biomed- in PC Samples. The GEPIA database was used to compare the ical informatics Grid. Kaplan-Meier (K-M) Plotter database difference of expression of 9 overexpressed immunoinhibitors was used to analyze prognostic parameter of expected candi- in transcription level between cancers and normal tissues dates [28]. K-M survival curves and logrank test were per- (Figure 1). The data demonstrated TGFB1 (Figure 2(a)), formed to disintegrate correlation, such as gene expression PVRL2 (Figure 2(b)), CSF1R (Figure 2(c)), TGFBR1 with overall survival (OS) or first progression (FP) or post (Figure 2(d)), VTCN1 (Figure 2(e)), LGALS9 (Figure 2(f)), progression survival (PPS), respectively. Significant differ- IL10RB (Figure 2(g)), KDR (Figure 2(h)), and IDO1 ence was indicated as P <0:05. (Figure 2(i)) mRNA levels were significantly upregulated in patients with PC compared to normal tissues. 2.2. Construction of Protein Interaction Network. A func- tional protein interaction network was constructed as indi- 3.3. Immunoinhibitors Were Positively Correlated to the cated in website (http://string-db.org/) [29]. Among them, Advanced Stage and Grade in PC. Furthermore, the TISIDB 50 selected proteins indeed associating with Homo sapiens database analysis showed TGFB1 was positively correlated were selected, followed by calculating confidence score as to the advanced grades of PC samples (Figure 3). However, more than 0.9. we did not observe a significant upregulation of PVRL2 (Figure 3(b)), CSF1R (Figure 3(c)), TGFBR1 (Figure 3(d)), 2.3. TISIDB, GEPIA, TCGA, and CBioPortal Analysis. VTCN1 (Figure 3(e)), LGALS9 (Figure 3(f)), IL10RB TISIDB is an integrated repository portal for tumor- (Figure 3(g)), KDR (Figure 3(h)), and IDO1 (Figure 3(i)) in immune system interactions. The present study used TISIDB advanced grades of PC samples. (http://cis.hku.hk/TISIDB) database to detect the relation- Interestingly, our data also revealed the correlation ship between centromere protein expression and clinical between Immunoinhibitors level and stages of PC samples. stages, lymphocytes, immunomodulators, and chemokines The results displayed that expression of VTCN1 was raised in PC. Gene Expression Profiling Interactive Analysis in grade 2 and grade 3 samples compared to grade 1 PC sam- (GEPIA) [30] was a powerful tool to determine key interac- ples, but expression of VTCN1 was decreased in grade 4 sam- tive and customizable functions including differential expres- ple after being normalized to that in grade 1/2 PC samples sion analysis, profiling plotting, correlation analysis, patient (Figure 4(b)). Meanwhile, our data showed IL10RB was survival analysis, similar gene detection, and dimensionality enhanced in grade 2, grade 3, and grade 4 PC samples com- reduction analysis, which was used to determine mRNA pared to grade 1 PC samples (Figure 4(g)). However, no expression in 9,736 tumors and 8,587 normal tissues. The obvious difference between the expression of TGFB1 cBioPortal system was used to investigate cancer genomic (Figure 4(a)), PVRL2 (Figure 4(b)), CSF1R (Figure 4(c)), and clinical-related characters within 105 cancer subjects in TGFBR1 (Figure 4(d)), LGALS9 (Figure 4(f)), KDR Computational and Mathematical Methods in Medicine 3

14

12

10

8

6

4

2

0 IL10 KDR IDO1 CD96 LAG3 BTLA TIGIT CSF1R CD244 CD160 PVRL2 TGFB1 CTLA4 IL10RB PDCD1 VTCN1 LGALS9 TGFBR1 HAVCR2 KIR2DL1 KIR2DL3 ADORA2A PDCD1LG2

Figure 1: Identification of immunoinhibitor expression pattern in PC. The present study analyzed the expression pattern of 23 immunoinhibitors in PC using TCGA database, including CD160, CD244, KIR2DL1, KIR2DL3, BTLA, CSF1R, HAVCR2, TIGIT, LAG3, PDCD1, VTCN1, PDCD1LG2, LGALS9, CD96, TGFBR1, TGFB1, CTLA4, ADORA2A, PVRL2, IL10, IDO1, IL10RB, and KDR.

(Figure 4(h)), and IDO1 (Figure 4(i)) and the stage in the 3.5. To Assess the Coexpression and Interaction Gene with PC patients was taken on. Immunoinhibitors in PC Patients. We evaluated the associa- tion of candidate gene expression with immunoinhibitors 3.4. Analysis of Immunoinhibitor Feature in Prognostic PC by Pearson’s correlation analysis. The immunoinhibitor- Patients. We deeply explored the profile of immunoinhibi- target pair with absolute Pearson’s correlation coefficient tors implicated in prognostic PC patients. Our data revealed value > 0:5 was considered significant. The networks were that the increasing level of TGFBR1 (Figure 5(d)), VTCN1 constructed using Cytoscape software. As presented in (Figure 5(e)), LGALS9 (Figure 5(f)), and IDO1 (Figure 5(i)) Figure 6, the coexpression network included 9 immunoinhi- mRNA was closely pertained to poor OS. However, the dys- bitors, 1250 targets, and 1304 edges. From the analysis, we regulation of TGFB1 (Figure 5(a)), PVRL2 (Figure 5(b)), observed LGALS9 may have a primary role in this network CSF1R (Figure 5(c)), IL10RB (Figure 5(g)), and KDR and possessed approximately 30% coexpressing targets with (Figure 5(h)) was not related with OS in PC. IL10RB, PVRL2, and IDO1. 4 Computational and Mathematical Methods in Medicine

TGFB1 PVRL2 CSF1R ⁎ ⁎ 7 ⁎

6 8 6 5

6 4 4

4 3

2 Relative expression levels Relative expression levels Relative expression levels 2 2 1

0 0 0

PAAD PAAD PAAD (num(T) = 179; num(N) = 171) (num(T) = 179; num(N) = 171) (num(T) = 179; num(N) = 171) (a) (b) (c) TGFBR1 VTCN1 LGALS9 6 ⁎ 6 ⁎

5 5 8

4 4 6

3 3 4 2 2 Relative expression levels Relative expression levels Relative expression levels 2 1 1

0 0 0

PAAD PAAD PAAD (num(T) = 179; num(N) = 171) (num(T) = 179; num(N) = 171) (num(T) = 179; num(N) = 171) (d) (e) (f) IL110RB KDR IDO1 ⁎ ⁎ ⁎ 5 8

6 4 6

3 4 4 2 Relative expression levels

Relative expression levels 2 Relative expression levels 2 1

0 0 0

PAAD PAAD PAAD (num(T) = 179; num(N) = 171) (num(T) = 179; num(N) = 171) (num(T) = 179; num(N) = 171) (g) (h) (i)

Figure 2: Increasing expression of immunoinhibitors was observed in PC samples. TGFB1 (a), PVRL2 (b), CSF1R (c), TGFBR1 (d), VTCN1 (e), LGALS9 (f), IL10RB (g), KDR (h), and IDO1 (i) mRNA levels were significantly upregulated in patients with PC compared to normal tissues. Computational and Mathematical Methods in Medicine 5

15 15 15

14

10 10 13

12 5 5

11 Relative expression of CSF1R (log2 CPM) Relative expression of TGFB1 (log2 CPM) Relative expression of PVRL2 (log2 CPM) 0 10 0 G1 G2 G3 G4 G1 G2 G3 G4 G1 G2 G3 G4 (a) (b) (c) 14 15 15

13 10 12 10

11 5

10 5 0 9 Relative expression of VTCN1 (log2 CPM) Relative expression of LGALS9 (log2 CPM) Relative expression of TGFBR1 (log2 CPM) 8 –5 0 G1 G2 G3 G4 G1 G2 G3 G4 G1 G2 G3 G4 (d) (e) (f) 13 15 15

12

10 10 11

10 5 5

9 Relative expression of KDR (log2 CPM) Relative expression of IDO1 (log2 CPM) Relative expression of IL10RB (log2 CPM) 8 0 0 G1 G2 G3 G4 G1 G2 G3 G4 G1 G2 G3 G4 (g) (h) (i)

Figure 3: Immunoinhibitors were positively correlated to the advanced grade in PC. TISIDB database analysis revealed the expression levels of TGFB1 (a), PVRL2 (b), CSF1R (c), TGFBR1 (d), VTCN1 (e), LGALS9 (f), IL10RB (g), KDR (h), and IDO1 (i) in grade 1, 2, 3, and 4 PC samples.

3.6. Assessment of the Function of LGALS9 and TGFB1 in PC LGALS9 was related to the RIG-I-like receptor signaling Patients. We finally validated the role of LGALS9 and TGFB1 pathway, NF-kappa B signaling pathway, cytosolic DNA- after the analysis of GO and KEGG in the DAVID system sensing pathway, and TNF signaling pathway (Figure 7(b)). using their genes. After bioinformatics analyzing, LGALS9 Also, we found that TGFB1 was related to mesoderm for- was involved in regulating type I interferon signaling path- mation, cell matrix adhesion, substrate adhesion-dependent way, defense response to virus, interferon-gamma-mediated cell spreading, in utero embryonic development, extracellu- signaling pathway, response to virus, innate immune lar matrix organization, outflow tract septum morphoge- response, and negative regulation of viral genome replication nesis, mitral valve morphogenesis, and Hippo signaling (Figure 7(a)). KEGG pathway analysis demonstrated that (Figure 7(c)). And KEGG pathway analysis showed TGFB1 6 Computational and Mathematical Methods in Medicine

15 15 15

14

10 10 13

12 5 5

11 Relative expression of TGFB1(log2 CPM) Relative expression of PVRL2 (log2 CPM) Relative expression of CSF1R (log2 CPM) 0 10 0 Stage I Stage II Stage III Stage IV Stage I Stage II Stage III Stage IV Stage I Stage II Stage III Stage IV (a) (b) (c) 14 15 15

13 10 12 10

11 5

10 5 0 9 Relative expression of VTCN1 (log2 CPM) Relative expression of LGALS9 (log2 CPM) Relative expression of TGFBR1 (log2 CPM) 8 –5 0 Stage I Stage II Stage III Stage IV Stage I Stage II Stage III Stage IV Stage I Stage II Stage III Stage IV (d) (e) (f) 13 15 15

12

10 10 11

10 5 5

9 Relative expression of KDR (log2 CPM) Relative expression of IDO1 (log2 CPM) Relative expression of IL10RB (log2 CPM) 8 0 0 Stage I Stage II Stage III Stage IV Stage I Stage II Stage III Stage IV Stage I Stage II Stage III Stage IV (g) (h) (i)

Figure 4: Immunoinhibitors were positively correlated to the advanced stages in PC. TISIDB database analysis revealed the expression levels of TGFB1 (a), PVRL2 (b), CSF1R (c), TGFBR1 (d), VTCN1 (e), LGALS9 (f), IL10RB (g), KDR (h), and IDO1 (i) in stage I, II, III, and IV PC samples. was related to pathways in cancer, TGF-beta signaling path- immune checkpoint pathways responsible for curbing immu- way, focal adhesion, signaling pathways regulating pluripo- nity [35]. Monoclonal antibodies targeting immune check- tency of stem cells, regulation of actin cytoskeleton, Hippo points exhibited huge advance in cancer therapy. Currently, signaling pathway, and shigellosis (Figure 7(d)). some researches have revealed that patients with different can- cer recovered better after treatment of immunoinhibitors. 4. Discussion Developing prospective methods based on immunoinhibitors could be of significance to explore novel biomarkers in the PC with a poor prognosis was regarded as one of the most PC diagnosis and prognosis. deadly carcinomas [33]. The growth and development of Can- In this study, we analyzed the expression pattern of 23 cer were reported to be involved in the process of immune immunoinhibitors in PC using TCGA database and found that suppression [34]. Cancer cells could stimulate various IDO1,CSF1R,VTCN1,KDR,LGALS9,TGFBR1,TGFB1, Computational and Mathematical Methods in Medicine 7

TGFB1 PVRL2 CSF1R 1.0 1.0 1.0 HR = 0.71 (0.42 – 1.21) HR = 0.68 (0.45 – 1.03) HR = 1.25 (0.79 – 1.97) 0.8 P 0.8 P 0.8 logrank = 0.21 logrank = 0.064 logrank P = 0.34 0.6 0.6 0.6

0.4 0.4 0.4 Probability Probability Probability 0.2 0.2 0.2

0.0 0.0 0.0 0 20406080 020406080 0 20406080 Time (months) Time (months) Time (months) Expression Expression Expression Low Low Low High High High (a) (b) (c) TGFBR1 VTCN1 LGALS9 1.0 1.0 1.0 HR = 1.53 (1.02 – 2.31) HR = 1.55 (1.01 – 2.39) HR = 1.78 (1.15 – 2.74) 0.8 0.8 logrank P = 0.045 0.8 logrank P = 0.04 logrank P = 0.0081 0.6 0.6 0.6

0.4 0.4 0.4 Probability Probability Probability 0.2 0.2 0.2

0.0 0.0 0.0 0 20406080 020406080 0 20406080 Time (months) Time (months) Time (months) Expression Expression Expression Low Low Low High High High (d) (e) (f) IL10RB KDR IDO1 1.0 1.0 1.0 HR = 1.29 (0.79 – 2.12) HR = 0.56 (0.33 – 0.95) HR = 1.74 (1.12 – 2.7) 0.8 logrank P = 0.31 0.8 0.8 logrank P = 0.013 logrank P = 0.03 0.6 0.6 0.6

0.4 0.4 0.4 Probability Probability Probability 0.2 0.2 0.2

0.0 0.0 0.0 0 20406080 020406080 0 20406080 Time (months) Time (months) Time (months) Expression Expression Expression Low Low Low High High High (g) (h) (i)

Figure 5: Analysis of the correlation between immunoinhibitor expression and survival time in PC patients. Kaplan-Meier plotter database analysis revealed the correlation between the levels of TGFB1 (a), PVRL2 (b), CSF1R (c), TGFBR1 (d), VTCN1 (e), LGALS9 (f), IL10RB (g), KDR (h), and IDO1 (i) and overall survival time in PC patients. 8 Computational and Mathematical Methods in Medicine

LINC00661RP11-1C1.6 C6 CXCL6 GAS6-AS2 FLRT2 LINC00379 TEAD2 UNC5CL SLC23A1 AMBP AC087762.1 LRRC75B GGT1LL21NC02-21A1.1 RP11-680F8.1 ATP13A4 GRB10 FRAS1 SOX9-AS1 GATM ERBB4 AC007349.5 RP11-206P5.2 AC079779.4 VTCN1 CCDC160AC010136.2 MTMR12 LINC00671 BAAT SERPINA5 PLEKHS1 MUC15CFI PNLDC1PDGFD VEPH1TMEM51 ITIH5 ZNF704PKNOX2 FLT3 AC147651.5 RP11-678G14.3VAV3 BMF GS1-24F4.2 ZCCHC16 CTD-2600O9.1 ACADL NFIA PDE5ACTD-2626G11.2 RP4-541C22.5 SLC4A4 AC002076.10 AJAP1 CNDP2 RP11-277P12.20 SLC34A2 RERG AC005152.3CTA-992D9.8ACSM3 RP11-404P21.3 HOMER2 STOX2 AL162759.1 RP11-659P15.1 CTD-2337I7.1 RP5-1121A15.3 ZNF209P CLDN10 TMEM56-RWDD3 SERPINA4RP11-680E19.1 TLR5 WWC1 ADAMTS10 RP11-420K14.2 ERICH5 LINC00864 FAM171A1 CCL2 TRPV6 KCNJ15RP11-680E19.2 AQP1 NR1H4 RP11-519G16.5 AC007349.4 CYS1 AMHR2 GGTLC1 PIK3AP1 SPP1 SOD2 GRIK1-AS1 SORBS2 LINC00987 MUM1L1 RP11-1036E20.7 NCOA7 SLC30A2 MTNR1A PSG8 CYP39A1SSC4D CACHD1 SERAC1 PADI2 TGFBR1 FAM149A CTSLP1 RP5-1148A21.3 RP11-347E10.1 RP13-714J12.1 C3orf58 FAIM PICALM VNN2 HABP2ONECUT1 SERPINA3 TSPAN6 GRB14 MMP10 HOGA1 TOR1AIP1 GPRC5B CTC-513N18.6 43718 GLIS3 SLC3A1 NFIA-AS2 RNF13 KCNJ16 ACVR1 NRIP1 ATXN1 AC003973.3RP11-429K17.1 TRABD2B ARL13B TMEM245 CTC-236F12.4 DCDC2CCDC141 NID1 RP11-43F13.3RP3-335N17.2 SBK1 HNF1B TMPRSS13 LSM11 CRIM1 SOX6 PYGO1 C21orf62 C8orf34-AS1 RP11-553L6.5 CFAP221 NEURL3 RBSN STAT3 SCNN1G TRIM50 NR5A2 RNF11 RAB31 LTBP1 HIATL1 TEAD1 RP11-543D5.2 ZMYM6NB BICC1 F11 SLC25A30 SCD5 CX3CR1 CCL28 ZFAND3 ARL2BPLRRC58 LEPROT EOGT YTHDF3 PKD2 SETD7PXYLP1 FAM150B TTPAHS6ST1 RP11-363G10.3 ATP1A1 TMEM217AC005307.3 PSG9 NCMAP ITGAV SLC30A7 KPNA3 GMNN ITGA1 TWSG1 FKBP14 C10orf11 CREB5 MIER1 WDR82 APOBEC2PQLC3 SLAIN2PRKG1 RASSF8 MAPK1 RAB14 LAMC1 KRT75 SBSPON ALDH3A2 PTPN14 PCSK5 MMP7 AC013275.2AP000354.4SLC16A7 CH507-396I9.6 LYPD1 MTMR2 ANXA9 SLC17A4 SNTB2 ZEB1 OGFOD1 FOXP1 PTPN9 TFPI2 IDO1 PDCL ROCK1 ZFHX3 RELL1 WWTR1 CLIC4 ZBTB6 F11-AS1 ZC3H13 AC104777.1 RAB18 MTR SERPING1 RP11-569G13.3RP11-420K14.1 GBP1 KIF20B RP11-485F13.1 GNL3L CTSO ZBTB8A HUNK AKAP7 ZNF676 ANO6 TMEM167B ZFP91 SFRP5 LAP3 ARNT RBM12 ACSL4 PTPRG PITX2 SUSD6 IGHV1OR16-3 FAM168A RIC1 FERMT2 MLK4 RP11-352D13.6 BIRC3GBP1P1 PPP4R2 RP11-10J5.1 NUB1 SSH1 SRGAP2C GNAQ MITFCAPZA2 IFFO2 RP11-497E19.1 KRR1 RP11-465I4.3 KDSR TMTC3 LRRN4 RP11-77K12.1 PSMA4 SOCS5 LUZP1 RP11-570J4.1 RP11-401E9.3 TRAFD1 MSANTD3-TMEFF1HIAT1 C16orf72 DACT1 ERCC4 DYNC1LI2 KCTD10 GNG12 ABL1 TCHP UNG IFNG RP11-158D2.1 43719 BMPR2 RP11-815J21.4 LATS2 GNLY BTN3A3 JAM3 WWC2 CCDC59 CTD-2384A14.1 ZMAT3 PBRM1 GTF2H3CXCL10RNU6ATAC39P 43710 BMPR1A MAP3K2 CTA-384D8.35 AC073325.1 MPP7RP4-784A16.4 KIRREL OSBPL11 ZNF385D-AS1 IL2RA ITGB1 RAB23 PEA15 PALLD HIP1 AL109763.2 GZMB AIDA ANKRD40 AKAP14 CSGALNACT2SGCD EXT2 RP11-33N16.2 WARS RP11-202G18.1 FSTL1 TMPO STAT1 UBD RP11-476H20.1 RABGAP1CDC37L1 RP11-121A13.3 ERI1RP3-347M6.1 LRRTM1 RP11-797M17.1 KCTD20 ENAH PLEKHA3 PHKBP1 HLA-DOALILRP2 AC012501.3 RBFOX2 TRIM32 MIB1 C5orf24 YIPF5 UBTD2 NUDCD1 DDR2 PDLIM5METTL24 AGL APOL3 LGALS17A RP11-384J4.2 LPP FAM26F AC012501.2 ZFP1 DTHD1 KLHL20 SNX18 MRPL42 RP11-328J2.1 AC084193.1 AIM2 EDNRA TNS1 NEDD4 C5orf51 CTC-480C2.1RP11-351A20.1 PADI3 IL18BP LAG3 FAM98A DCAF10 ZNF281 TUSC7 TCF12 CDR2 UBE2Q2 TNFSF13B FBXW2 ATXN1L CXCL11 MAP4 FAM181A-AS1 TAP2HLA-DRB5 EIF4A1P1 KLHDC7B CAMSAP2 MORF4L1P1 TRIM53CP RP11-1072N2.2 RP11-268P4.2RP4-665J23.4 STON1 KANK2 ITGB1P1 PSMA6 AFF4 PEAK1 PHACTR2 GYPB AC069363.1 CSNK1A1 VCL GINS4 RP11-519M16.1 CSTF2 PSMA3SLC20A2 TROVE2 ZBTB38 SPIN1 KRT8P2 IL21 MSRB3 CALD1 RACGAP1 FAM133A RAPGEF2 PARP9 RSU1 PARVA CDKL5 GAPVD1 IL12RB1 IGKV2-24 DCAF4L2TRMT112P5 HIPK1 NUP37 CR1L CFAP97 AC008067.2 ZNF35 SMAD2 IRG1 SEC23A CRTC3 ATL3 PVRL2 LGALS9 RP13-147D17.3 CXCL9 PPTC7 GPN3 SNX9 SNAP23 FAM198B RP11-291B21.2RP11-145M4.2 CTA-384D8.31 COL4A1 CCSER2 RFC5 SH3BGRL RP11-62E14.2 IGHV2OR16-5 EVI5 ADAMTS20 CD274 RP11-503G7.1 GPR25 RP1-229K20.8 FAM102B AMOTL1 SLC52A3 FBXL12 RP11-44K6.4 IFIT2 IFIH1 GBP4 APOL6 RNF185 NFIC CD38 NEDD1 DR1 DIXDC1 BAD NOP10 SYK RN7SKP266 TMEM219 RAN APOBEC3GRIPK2 SCYL1 SLC15A3 SLC35C1 UBE2L6 RARRES3 FBXO6 RP11-6N13.1 DBNL HLA-B IRF1 PYDC2 UBE2N PARPBP TMSB10 OAS3 RP11-692P14.1 RHBDF1 SYMPK MOV10 NLRC5 TRIM21 B2M OR4P1P RP11-231P20.6 APEX2 GZMH LINC01608 RHOF RP11-456O19.5 RP3-395M20.8 CMPK2 BATF2 RP11-3J1.1 TDGP1 SEPW1 DRAM1 DYNLRB1 IRF3 GJB3 XAF1 PSME2P2 NBN SLC22A6 QPCTL CCDC88B DDX60L PSME1 PSME2 RP11-74C13.4 XRCC1 BAZ1A PAQR4 HLA-E RPL9P15 FCGR3B RP11-145G20.1CHST11 HS6ST2-AS1 DYNLT1 SNRPD2 PARP10 DPYD-IT1 CLASRP DAB2IP PSMB10 EPHA2 USP30-AS1 SNX18P27 C11orf98 TACSTD2 RRAS RAB20 HKDC1 LSM8 PSMB8 DTX3L MFAP3 CLPTM1 DNTTIP1 RP11-777B9.5 OASL PSMB9 ETV7 KIR2DL4 AC092421.1 STXBP2 NRIR ZNFX1 OR2G3CTD-2374C24.1 HLA-DRB1 RP11-380D23.1 ATP5J2 PSMB3 SH2D4A TRAPPC6A AP5B1 ARHGAP11A RHOC PPP1R35 ETHE1 RN7SKP54 HSH2D NMI GBP5 ZYG11B ADAM15PPP1R13LFBXO46 NACC1 BLOC1S3 IFI44 TSPAN15 PARP14 TYMP TAP1 NFE2L3 RP11-419K12.1 CLDN4 CCDC61 SP100 CAPN5 RPL7P53 HEG1 NUDC BCAR1 IFI27 HCP5 TM4SF18 RP11-774D14.1 IFI44L SQRDL IRF9 TBL3 DCAF15 HNRNPABEXOC3L4 OAS2 EPSTI1 LINC01419 SERTAD1 NFKB2 RMDN3 KRT8 TOMM40 SHKBP1 OR7E14P SFXN3 GMIP LAMP3 IGLV3-24 PPP4R1L PRSS22 MYPOP POMP UBR4 EHD1 CTHRC1P1 HLA-DQA1 SFN PML DPP3 HLA-F MTMR14 SDCBP2 RNF213 FUT3 TRIM10 SEPHS2C1orf116 IFIT3 EPHB6 OTUB1 GNB2 TMEM265OCEL1 EPS8L1 DMWD RP11-161H23.5 ANXA11 LGALS3BP TRIM69TMEM106A MAP3K11 SERF2 ST14 GIPC1GSTP1 IFIT1 USP18 SH3RF2 CDK18 KDR NR2F6 TICAM1 CLTB CACNA1C-IT1 ZNF267 CDH5 C19orf66 RHOQ PHKG2 SPDYE13P SHISA5 RTP4 HLA-C OSTM1 PPAP2A CTB-191K22.5 PSMB2 PPP2R1A LASP1 PPP1CA CTTNBP2NL KIAA1586 BUD31 PNPLA2 PLCB3 IFI6 DGUOK GLTSCR1L CNKSR1 NQO1 HERC1 S100A16 DOK4 FBLIM1 WASF3 ZNF366 AIM1L VASP EFNB1 PPA1 UNC93B1 ACOT9 NDUFC2 SCNM1 CBLC CAPN1 SLC12A9 PPP4CCTD-2231E14.8 LRRC1 MPZL1 ANGPT1 MPDZ PIH1D1 CFL1 CTNNB1 KLHL15 ROBO4 INF2 BAK1 CLDN23 UTRN GPR137 PAGR1 CEP68 SMARCA4RP11-818C3.1 ERCC2 DDX60 MLKL LRCH2 AC104651.3 STX4 FUT4 TTC9C TTC28 RNF152 KCNN4 FAT1P1 DDA1 DOK1 FAM111A EXOC3L2 IL6ST ZNF576 ARPC1B RELB FAM13C ZNF382 ANKFY1 TRIP6 RPS6KA1 HERC6 ARPC5 SGK223 MMRN2 RBM42 ADGRE5 PRKCE ERCC6L2 PPM1D AC093732.1 COG4 NFKBIE CORO1B IRF7 DBI PCDH12 PJA2 MARK4 GNG5 LYPLA2 S100A13 GUK1 CDC37 POLD4 DDX58 DSTYK EFHD2 RN7SL575P OAZ2 EMCN DIS3L TSC22D4 FXYD3 BABAM1 PARP12 BST2 ZBP1 RELA ALDH2 MX1 SLAMF8 TCAF2 LPPR4 IRGQ ZNF438 IFI35 MXD1 TLR8 BHLHB9 PPP1R37 CAPG IFITM1 CNTN4 EPC2 RNF181 CHCHD1 BCL3 FAM32A SAMD9L SASH1 TEK NEGR1 SMG9 SLC10A3 AC108056.1 IL15RA TOR4A GABRG1 ZIK1 MKL2 ZNHIT1 C19orf33 TBC1D10B GSDMB SSX2IP SERINC1 DNAJC17 EBP VAMP8 CASP1 CARD16 FADD PDCD1LG2 APPBP2 ZFP82 TMEM189 FERMT3 RPS6KA4 PREX2 ZKSCAN2 BCL7C SMIM22 PRKD2 OAS1 FCGR3A NEUROD1 SPECC1L LRRC8C BBC3 HSD3B7 FBXL19 MCU CXorf36TMEM170B ACVRL1RNF180 RPS6KB2 SNRPA HPS1 LRRC25 PTPRB STX10 PLOD3 RTKN SLC4A11 TMEM62 RIPK3 HLA-DRA ZDHHC15 ARL15 CDC42EP2 SH3BGRL3 CDK2AP2 IL32 PLVAP AC006126.4 EPAS1 CEACAMP11 S100A6 HELZ2 PHF11 ARHGEF15 NRIP2 ZKSCAN7PCDH17 LAMTOR5 MX2 HLA-A MAN1C1 ZNF568 ADAM8 PNKP SLC7A7 PDLIM1 C11orf95 TMOD2 ERG EPHA3 MOSPD3 ADRM1 GPR75 GNPDA2 CYP2U1 ZSWIM4 SLC9A1 NDST1 MPG TMEM179B EML2 POTEB2 FLT1 CLEC14A SETBP1 CYBA HRH2 AIF1 LIN7A CHMP2A RP11-44M6.7PHLDA2 FHL5 PPAP2B BRI3 UBA7 DHX58 LILRB4 IPO11 FAM114A2 ZBED3 CHMP4B SPTAN1 PLSCR1 GIMAP1-GIMAP5 PSAP EBF3 WDFY3-AS2 PLAUR SPI1 PREX1 43898 ZNF271P EIF6 MYO5A FCGR1B EDNRB COL4A3BP POP7 RTN2 ADAP2 CORO2B COX6B1 ITPKC RAB11A LIFR LSR CD14 P2RX7 CCR2 CD302 SLC25A53 PCDH18 FECH CD200 EIF4E2 RP11-399B17.1 ZNF727 GET4 FOXA3 PSENEN DNM2 PILRA ZNF542P AP2S1 DIAPH1 CNOT11 MS4A7 PTPRO ARHGAP31 NDNL2 NRP1 ADPRHL2 CCRL2 GAS7 HGF VASH1 ZNF454 MDFI LCP2 MNDA TRIM23 ZNF614 ERCC1 TRIM5 CD33 MYCT1 RECK LCMT2 FAM172A CTNNA1 MYD88 MS4A6ASLC31A2 EVC2 CD93 GIMAP8 AKT3 RHOJ GEMIN7 CTDSP1 PRR13 A2M FCER1G CAPNS1 RASGRP4 WIPF1 CADM1GLP2R SHROOM4 CASP10 LST1 TCF4 NCOA1 JAM2 SHE P2RY13DOCK4 ADGRL4 FZD4 TNIP1 MVP NPL FRMD4A ZNF518B PLSCR4 REV3L BLOC1S6 ST8SIA4 GALM TM6SF1 EPB41L2 FOLR2 MAP3K3 ZNF569 S1PR1 DUSP19 STAT5B TRIM34 RGS18 ADAMTS3 GPR4 PSMA5 TFEC C1orf162 CYBB TAL1 TNRC6C TRIM14 LILRB2 NLRC4 DAB2 CCR5 RGL1 FLT4 TMEM92LGALS3 NHSL2 ZNF25 ZNF135 MAP2K4 RPL7AP64 CXorf21 PLEK PECAM1 CYYR1 ABCC9 SLC16A2 TMX3 GCSAML ATP8B2 NFAM1 PRCP RP11-389C8.2 TNFAIP8L2 GPR65 SLC1A3 GLIPR2 CNN3 RP11-456K23.1 CD34 QKI RBMS3 MMP16 SNRK FRYL P2RY12 ZEB2 GABPA ZNF829 EID1 HECW2 KLF3 DNMBP LAPTM5 ADORA3 ATP8B4 APOLD1 CALCRL PIK3R5 GIMAP4 MERTK NXPE3 ELOVL1 USP51 GOLPH3L FGL2 BMP2K SIRPB2 FAM49A LRRK2 SAMHD1 HCLS1 GIMAP6 CCDC102B RAMP2-AS1PABPC5 F11R TMEM248 PIK3CG C1S CSF2RA SIGLEC9C3AR1 SH2B3 SECISBP2L PLCL1 ADGRF5 PIAS2 IST1 MAPK3 KIAA0355 VDR MYO1C GPR141 AC109826.1 STARD8 RNF146 MAPK6 AMBRA1 LINC01140 NCKAP1L CDC40 RP11-121A8.1 HACD4 ADPRH AOAH EVI2A KIDINS220 IKZF4 RP11-73M7.1 STYK1 ZNF660 THSD1 PIK3R3 CARD6ATG16L1 CEACAM6 AC007246.3 C1QB HCK STK38 Previous12Next SP1 MSR1 ITGB2 C10orf128 CLEC4A DLC1 RNASEL VSIG4 LINC01094 TRPC6 FBXL3 PRKCA STK24 CD4ZNF804A USP39 PPM1M FCGR1C TTBK2 BFAR CORO2A MAT2B CELF2 MYO1F ARSB LDB2 DSCR3 MS4A4A NR3C1 TIE1 HIPK3 FAM118B ARHGAP26 CLDN12 PPT1 GAB3 OGFRL1 SVEP1 LINC01057 ECD GGTA1P PLB1 MAGI2-AS3 GLCE MARK2 PRRC2C RNF19B AMICA1 PIK3R6 FAT4 ARHGAP12 ZNF207 IL10RA FPR3FCGR2B CLEC7A TBCEL LACTB PSEN1 APH1A LIMA1 PALD1 FPR1 IFNGR2 TNFSF12 MPEG1 ACBD3 ZDHHC9 WDR55 TMEM183A MLX PLEKHO2 SLA SAMD12-AS1 CHP1 CASS4 SLC43A3 SLC44A1 SCP2 CDS1 SAMSN1 TGFB1 ZFP64 TLR7 HLA-DPA1 TLR4 MFSD1 ARHGEF6 CTNND1 IGSF6 STAU1 DAZAP2 AP000275.65 RNASE6 LILRA1 NRROS SIRPA KDM5B MTM1 ISG20L2 WWP2 ARRB2 ZDHHC5SLC2A10 SLC35D1 CMKLR1 CRTAM ESRP2 ATG4C CD86 GPNMB AXL RP11-803D5.1 MICU1 CD163 CNOT1 SPATS2L DSECD300A FCGR2AGLIPR1 GAL3ST4 TMEM127VPS37C IFNAR1 HAVCR2 TSPAN4 MYO9B NUP98 SRI TET3 GRHL2 RAB9A IRF2 HLA-DMB RASSF2 BPNT1 EFCAB14 C1QA GPR34 FAM101B SLC35A3 CASP6 C1QC SLCO2B1 RP11-750H9.5 TMC7 KCNK1 CCR1 ITGAM LAIR1 SRM LRRC32 CMTM6 GOLGB1 TLR1 COL5A1 LIMK2HNRNPF CASP7 SLC12A4 PLEKHO1 NIPAL2 IFNAR2 C21orf59 CD226 TNFRSF1B SCARF2 BAP1 ERLIN1 NUMB STAT6 AP1S2 CD84 LY96 CYFIP1 AMMECR1 F13A1 SIGLEC7DOK2 C8orf58 PLXDC2 MRC2 CMTM3 FAP GALNT12 XPNPEP1TMEM2TCF7L2 TJP2 PBDC1 ZBTB7A KIAA1217 AC112721.1 SH3GL1 PTRF CTSZ CMPK1 SLC25A13 SLC9A9 PPP1R18 ETV3 PGM2 C1orf131 ERGIC1 LIPH SRGN TTYH2 SRGAP2 ZYX GNB4 RP11-497H16.7 GNAI2 RGS19 DIAPH2 SCAMP2 GNGT2 SRPX2 RP11-420L9.5 NLGN2 AEBP1 RETSATGART CALML4 ACSM5 C1R ARL4C ZSWIM6 ALPK1 LY75 SRSF3 ZDHHC3 HLA-DPB1 OLFML2B CLEC11A EHD2 MARVELD3 MBOAT1 RASSF6 EIF4EBP2 C3orf52 TBCCD1 CSF1 HDAC7 NT5DC2 C2orf27A TSSC2 ARPC2 CD68 SLC39A13 NRM CASP3 RP11-394I13.1 NAGK SPRED2NFKB1 AIMP1 JOSD1 LRRC57 CD276C18orf54 CTC-205M6.5 RCN3 MYL12A EFTUD1 ETV6 COL18A1 LINC00152 C14orf119 FMNL3 EFS HTRA2 CNNM4 RBM47 TEP1 MMP14 AC133644.2 MAD2L2 NTAN1FAM65A FBXO34 RP11-671J11.5 THOC5 VDAC1 BCAS1 EMP3 KIAA0930 MCTP2 MED14 PARP4 KIAA1755 MAP7D1 TMEM144 GALK2 Showing 1 to 100 of 200FARP2 entries P4HA3 INAFM1 MFAP2 H2AFY SLC35B3 SRC CNN2 FBXL7 LOXL1 COLGALT1 HECTD3 FAM120A SACS TSG101 CSF1R FTH1P8 SULF2 FAM225A NRIP3 MIS18A SLC30A6 LINC01561 MAF CD248SPHK1 METRNL CASP8 PDXDC1 GTPBP1 KIF7 TSPAN3 HMGN1 CPNE2 ZBTB17 RP11-680A11.5 FSCN1 EML4 NFE2L2 C1QTNF6 RENBP COL5A3 SCAF4 CGN CALHM2 TNFRSF4 LOXL3 ADAM19 MAN2B1 NRBP1 STX6 ELF4 TMEM256-PLSCR3 DDX23 BCL10 ARAP1 WTIP PTK7 MPV17 ADAMTS7 SMAGP CTSB LAMA4 HSPA12B COL6A2 HOMER3 SLC37A1 CLINT1 PFN1 TAX1BP3 NOTCH3 FKBP1A PCOLCE PHLDB1 CAD MSC LAMB1 LINC01116TMEM200BRUNX1 RHOG FTL AC108463.2 EML3 POLR2G FKBP10 MXRA8 HLXBATF3 RILPL1 ARSILEF1-AS1 CAPZB KDELC1 COL4A2LGALS1 RAB34 UBTD1 CASC15 PPM1F TIMP1 TSHZ3 DRAP1 IL10RB B4GALNT1 SKI HIC1 LAMB2 HSPG2 TRPV2 ACTB TMSB4X AP2M1P3H1RP11-732A19.5TOM1 PACS1 GPR68 BGN VIM RAB32 LRRC17 RARRES2 CERCAM LIMK1 EMILIN1 SPNS1 PLXND1 PLOD1 EVA1B BMP1 FAM89BGGT5 IGDCC4TMEM109 ARID5A TWIST1 RAB3IL1 BASP1 COL6A1 HTRA1 RP11-626H12.2 DFNA5 WDR81 NID2 EFEMP2 FTSJ1 TUBB ITGA5 PLA2G15 C11orf68 FHL3 ARHGAP22 SYDE1GTDC1 ISLR LRP1 GAS2L1 CCM2 PLEKHM2 PTGIR BCKDK TUBB6 RP5-1172A22.1 FTH1

Figure 6: Construction of the coexpression network of immunoinhibitors in PC patients.

IL10RB, and PVRL2 were highly expressed in PC tissues. acted as a candidate for the treatment of cancer [39]. The Moreover, the analysis revealed that IDO1, CSF1R, VTCN1, level of B7-H4 on tumor cells with adverse clinical and path- KDR, LGALS9, TGFBR1, TGFB1, IL10RB, and PVRL2 mRNA ologic features endowed B7-H4 with clinical significance level was significantly upregulated in patients with PC com- [40]. Moreover, the expression of B7-H4 in tumor- pared to normal tissues. Kaplan-Meier plotter results demon- associated macrophages was correlated with Foxp3+ regula- strated that the increasing level of TGFBR1, VTCN1, and tory T cells (Tregs) [41]. Because the expression of B7-H4 LGALS9 mRNA was closely pertained to poor OS. was on a variety of tumor cells and tumor-related macro- TGF-β was a primary executor of the stability and toler- phages, blocking of B7-H4 could improve the tumor micro- ance of the internal environment of immune system, includ- environment, thus enabling antigen-specific clearance of ing controlling many component functions [36]. Thus, tumor cells [41]. Our study suggested the enhanced level of disrupting TGF-β signal could result in inflammatory dis- VTCN1 was related to OS time and PFS (progression-free eases and tumorigenesis. In addition, TGF-β is also a prelim- survival) time of PC. Nevertheless, no increasing level of inary immunosuppressor in the tumor microenvironment VTCN1 was shown in neither PC nor normal tissues. [37]. Current researches have reported TGF-β was engaged Transmembrane receptor TIM-3 was encoded by in tumor immune evasion and adverse reactions to tumor HAVCR2 and expressed on a variety of cells [42]. The expres- immunotherapy [37]. Nevertheless, our study confirmed that sion of TIM-3 is closely related to exhaustion and impaired TGFBR1 and TGFB1 are upregulated in PC samples. Kaplan- function of T cells. The interaction between TIM-3 and Meier analysis showed that TGFBR1 was associated with galectin 9 has been demonstrated to induce Th1 cell apopto- reduced OS and PFS time in PC patients. sis, leading to reduced responses from autoimmunity and VTCN1 exists on the surface of antigen-presenting cells antitumor immunity [22] and also making TIM-3 as a poten- (APC) and interacts with ligands that bind to T-cell surface tial target for ICB. Of note, our study firstly exposed the receptors [38]. The activity of B7-H4 was illustrated to be upregulated level of LGALS9 in PC patients was associated related to the reduced inflammatory CD4+ T cell response with shorter OS and PFS time. as previously described. Studies have indicated VTCN1 IDO1 was responsible for converting tryptophan (Trp) expression was positively linked to tumor progression and into downstream catabolic product, called caninuria. Computational and Mathematical Methods in Medicine 9

Q value Innate immune response 0.00009 Defense response to virus

Type I interferon signaling pathway

Response to virus

Interferon−gamma−mediated signaling pathway

Negative regulation of viral genome replication

NIK/NF−kappaB signaling

Positive regulation of type I interferon production 0.00006

Cellular response to mechanical stimulus

Defense response Antigen processing and presentation of peptide antigen via MHC class I Negative regulation of type I interferon production

Response to interferon−gamma

Positive regulation of T cell mediated cytotoxicity 0.00003 Response to interferon−beta

0.05 0.10 0.15 Count 10 20 (a) Q value 0.003 Herpes simplex infection

Infuenza A

Viral carcinogenesis

Measles

Hepatitis C

Epstein−Barr virus infection 0.002 Antigen processing and presentation

TNF signaling pathway

NF−kappa B signaling pathway

RIG−I−like receptor signaling pathway

Cytosolic DNA−sensing pathway

Proteasome 0.001 Type I diabetes mellitus Allograf rejection

Graf−versus−host disease

0.04 0.06 0.08 0.10 Count 5 15

10 20 (b)

Figure 7: Continued. 10 Computational and Mathematical Methods in Medicine

Q value

Extracellular matrix organization

In utero embryonic development 0.005

Wnt signaling pathway

Cell−matrix adhesion

Endocytosis 0.004

Substrate adhesion−dependent cell spreading

Mesoderm formation 0.003 Positive regulation of osteoblast diferentiation

Positive regulation of bone mineralization

Regulation of endocytosis 0.002 Patterning of blood vessels

Hippo signaling

Outfow tract septum morphogenesis 0.001 Adherens junction assembly

Mitral valve morphogenesis

0.02 0.03 0.04 Count 3 7 4 8 5 6 9 (c) Q value 0.004 Pathways in cancer

Focal adhesion

Regulation of actin cytoskeleton

Signaling pathways regulating pluripotency of stem cells

Hippo signaling pathway 0.003

TGF−beta signaling pathway

Platelet activation

Vascular smooth muscle contraction

Leukocyte transendothelial migration 0.002 ECM−receptor interaction

Arrhythmogenic right ventricular cardiomyopathy (ARVC)

Shigellosis

Hypertrophic cardiomyopathy (HCM) 0.001 Adherens junction

Pathogenic escherichia coli infection

0.02 0.04 0.06 Count 5.0 12.5 7.5 15.0 10.0 (d)

Figure 7: Assessment of the function of LGALS9 and TGFB1in PC patients. GO (a) analysis and KEGG pathway (b) analysis of LGALS9 in PC patients. GO analysis (c) and KEGG pathway analysis (d) of LGALS9 in PC patients. Computational and Mathematical Methods in Medicine 11

Emerging studies have shown that IDO1 was expressed in efforts should be contributed to our findings, followed by a large number of human cancers. In the transcription providing a much more promising clinical strategy for an level, IDO1 displayed powerful relevance with T cell infil- early diagnosis and prognostic marker in PC therapy. tration [43]. Even though the expression of CSFR1, KDR, IL10RB, and Data Availability PVRL2 was upregulated in PC samples, which was not asso- ciated with the prognosis of PC. CSFR1 was mostly found in The datasets used during the present study are available from aggressive cell models and participated in the invasion and the corresponding author upon reasonable request. migration of tumor cells, and its expression is related to the poor prognosis of cancer patients. Positive feedback existed Conflicts of Interest between the expressions of CSF1 and EGF in tumors [44]. By blocking the signal transduction mediated by the EGF The authors declare that they have no competing interests. receptor or CSF-1 receptor, incomplete feedback loop would inhibit the migration and invasion of macrophages and ’ tumor cells. Activated VEGF-VEGFR2 could accumulate Authors Contributions Treg cells and control the migration of T lymphocytes [45]. Yue Fan is the first author. The IL-10R signal on effector T cells and Treg cells is essen- tial to keep immune tolerance [46]. Emerging studies have identified PVR2 as a new immune checkpoint [47]. Acknowledgments Of note, this study revealed that LGALS9 and TGFBR1 This work has been supported by the Development Project of were upregulated in PC compared to normal tissues. More- Shanghai Peak Disciplines-Integrative Medicine (20180101). over, we showed LGALS9 and TGFBR1 were significantly associated with the prognosis in PC. Despite the fact that LGALS9 and TGFBR1 were not significantly correlated to References the grades, we indeed observed LGALS9 had an upregulated “ trend and TGFBR1 had a downregulated trend. We thought [1] X. I. Zhao, D. C. Li, X. G. Zhu et al., B7-H3 overexpression in pancreatic cancer promotes tumor progression,” International the limited sample size may contribute to this result. Also, Journal of Molecular Medicine, vol. 31, no. 2, pp. 283–291, the coexpression plus bioinformatics analysis revealed that 2013. immunoinhibitors were involved in regulating multiple “ fl [2] A. Lambert, C. Gavoille, and T. Conroy, Current status on the in ammatory and immune response-related pathways as place of FOLFIRINOX in metastatic pancreatic cancer and previously described. Very interestingly, we found LGALS9 future directions,” Therapeutic Advances in Gastroenterology, was involved in regulating type I interferon signaling path- vol. 10, no. 8, pp. 631–645, 2017. way, interferon-gamma-mediated signaling pathway, RIG- [3] B. Wu, K. Wang, J. Fei et al., “Novel three‑lncRNA signature I-like receptor signaling pathway, NF-kappa B signaling predicts survival in patients with pancreatic cancer,” Oncology pathway, cytosolic DNA-sensing pathway, and TNF signal- Reports, vol. 40, no. 6, pp. 3427–3437, 2018. ing pathway. We also found that TGFB1 was related to meso- [4] T. E. Newhook, E. M. Blais, J. M. Lindberg et al., “A thirteen- derm formation, cell matrix adhesion, TGF-beta signaling gene expression signature predicts survival of patients with pathway, and Hippo signaling pathway. These pathways pancreatic cancer and identifies new genes of interest,” PLoS had been demonstrated as key regulators of tumorigenesis One, vol. 9, no. 9, article e105631, 2014. and immune therapy. [5] X. Du, X. Zheng, Z. Zhang et al., “A label-free electrochemical Several limitations should also be noted in this study. immunosensor for detection of the tumor marker CA242 First, we showed LGALS9 and TGFB1 had a crucial role in based on reduced graphene oxide-gold-palladium nanocom- ” PC with a series of bioinformatics analysis. The further posite, Nanomaterials, vol. 9, no. 9, p. 1335, 2019. experimental validations of their functions in PC could [6] W. Yan, L. Xu, Q. Wu et al., “A case report of spontaneous strengthen our conclusion. Second, more clinical samples rupture of intracranial epidermoid cyst with dramatic increase should be collected to detect the expression of these immu- of serum carbohydrate antigen 199: a three-year follow-up study,” BMC Neurology, vol. 15, no. 1, p. 198, 2015. noinhibitors in PC, which could provide more evidences to “ confirm their prognostic value. [7] C. S. Hinrichs, A. Kaiser, C. M. Paulos et al., Type 17 CD8+ T cells display enhanced antitumor immunity,” Blood, vol. 114, no. 3, pp. 596–599, 2009. 5. Conclusion [8] B. Wang, C. Sun, S. Wang et al., “Image-guided dendritic cell- based vaccine immunotherapy in murine carcinoma models,” Conclusively, our data suggested that immunoinhibitor American Journal of Translational Research, vol. 9, no. 10, mRNA level was dramatically upregulated, but negatively pp. 4564–4573, 2017. correlated with OS for PC. All the data suggested these genes [9] H. L. Kaufman, T. Amatruda, T. Reid et al., “Systemic versus could be used as an emerging prognostic indicator and tar- local responses in melanoma patients treated with talimogene gets in PC patients. Our findings would give a hint to have laherparepvec from a multi-institutional phase II study,” Jour- a better understanding of the mechanism implicated in PC nal for Immunotherapy of Cancer, vol. 4, no. 1, p. 12, 2016. and stretched out more precise immunotherapeutic treat- [10] Y. Imai, K. Hasegawa, H. Matsushita et al., “Expression of mul- ments for PC prognosis. Nevertheless, more researches and tiple immune checkpoint molecules on T cells in malignant 12 Computational and Mathematical Methods in Medicine

ascites from epithelial ovarian carcinoma,” Oncology Letters, [27] Y. Mao, J. Shen, Y. Lu et al., “RNA sequencing analyses reveal vol. 15, no. 5, pp. 6457–6468, 2018. novel differentially expressed genes and pathways in pancre- [11] H. C. Puhr and A. Ilhan-Mutlu, “New emerging targets in can- atic cancer,” Oncotarget, vol. 8, no. 26, pp. 42537–42547, 2017. cer immunotherapy: the role of LAG3,” ESMO Open, vol. 4, [28] A. Nagy, A. Lanczky, O. Menyhart, and B. Gyorffy, “Validation no. 2, article e000482, 2019. of miRNA prognostic power in hepatocellular carcinoma [12] L. P. Andrews, A. E. Marciscano, C. G. Drake, and D. A. Vig- using expression data of independent datasets,” Scientific nali, “LAG3 (CD223) as a cancer immunotherapy target,” Reports, vol. 8, no. 1, p. 9227, 2018. Immunological Reviews, vol. 276, no. 1, pp. 80–96, 2017. [29] D. Szklarczyk, A. Franceschini, M. Kuhn et al., “The STRING [13] X. Du, F. Tang, M. Liu et al., “A reappraisal of CTLA-4 check- database in 2011: functional interaction networks of proteins, point blockade in cancer immunotherapy,” Cell Research, globally integrated and scored,” Nucleic Acids Research, vol. 28, no. 4, pp. 416–432, 2018. vol. 39, pp. D561–D568, 2010. [14] E. D. Kwon, B. A. Foster, A. A. Hurwitz et al., “Elimination of [30] Z. Tang, C. Li, B. Kang, G. Gao, C. Li, and Z. Zhang, “GEPIA: a residual metastatic prostate cancer after surgery and adjunc- web server for cancer and normal gene expression profiling tive cytotoxic T lymphocyte-associated antigen 4 (CTLA-4) and interactive analyses,” Nucleic Acids Research, vol. 45, blockade immunotherapy,” Proceedings of the National Acad- no. W1, pp. W98–W102, 2017. emy of Sciences of the United States of America, vol. 96, [31] P. Sidaway, “Pancreatic cancer: TCGA data reveal a highly het- no. 26, pp. 15074–15079, 1999. erogeneous disease,” Nature Reviews Clinical Oncology, vol. 14, [15] D. E. Dolan and S. Gupta, “PD-1 pathway inhibitors: changing p. 648, 2017. the landscape of cancer immunotherapy,” Cancer Control, [32] J. Gao, B. A. Aksoy, U. Dogrusoz et al., “Integrative analysis of vol. 21, no. 3, pp. 231–237, 2014. complex cancer genomics and clinical profiles using the cBio- [16] L. Zitvogel and G. Kroemer, “Targeting PD-1/PD-L1 interac- Portal,” Science signaling, vol. 6, p. pl1, 2013. tions for cancer immunotherapy,” Oncoimmunology, vol. 1, [33] B. Y. Li, L. J. He, X. L. Zhang, H. Liu, and B. Liu, “High expres- pp. 1223–1225, 2014. sion of RAB38 promotes malignant progression of pancreatic [17] A. Friedlaender, A. Addeo, and G. Banna, “New emerging tar- cancer,” Molecular Medicine Reports, vol. 19, no. 2, pp. 909– gets in cancer immunotherapy: the role of TIM3,” ESMO 918, 2019. Open, vol. 4, article e000497, Suppl 3, 2019. [34] R. J. McKallip, M. Nagarkatti, and P. S. Nagarkatti, “Delta-9- [18] I. Herrera-Camacho, M. Anaya-Ruiz, M. Perez-Santos, tetrahydrocannabinol enhances breast cancer growth and L. Millan-Perez Pena, C. Bandala, and G. Landeta, “Cancer metastasis by suppression of the antitumor immune response,” immunotherapy using anti-TIM3/PD-1 bispecific antibody: a Journal of Immunology, vol. 174, no. 6, pp. 3281–3289, 2005. patent evaluation of EP3356411A1,” Expert Opinion on Ther- [35] V. Chandramohan, X. Bao, X. Yu et al., “Improved efficacy apeutic Patents, vol. 29, no. 8, pp. 587–593, 2019. against malignant brain tumors with EGFRwt/EGFRvIII tar- [19] A. H. Sharpe, “Introduction to checkpoint inhibitors and can- geting immunotoxin and checkpoint inhibitor combinations,” cer immunotherapy,” Immunological Reviews, vol. 276, no. 1, Journal for Immunotherapy of Cancer, vol. 7, no. 1, p. 142, pp. 5–8, 2017. 2019. [20] S. Buisson and F. Triebel, “MHC class II engagement by its [36] S. H. Wrzesinski, Y. Y. Wan, and R. A. Flavell, “Transforming ligand LAG-3 (CD223) leads to a distinct pattern of chemo- growth factor- and the immune response: implications for kine and chemokine receptor expression by human dendritic anticancer therapy,” Clinical Cancer Research, vol. 13, no. 18, cells,” Vaccine, vol. 21, no. 9-10, pp. 862–868, 2003. pp. 5262–5270, 2007. [21] C. Kyi and M. A. Postow, “Immune checkpoint inhibitor com- [37] M. Pickup, S. Novitskiy, and H. L. Moses, “The roles of TGFβ binations in solid tumors: opportunities and challenges,” in the tumour microenvironment,” Nature Reviews Cancer, Immunotherapy, vol. 8, no. 7, pp. 821–837, 2016. vol. 13, no. 11, pp. 788–799, 2013. [22] M. Nishino, N. H. Ramaiya, H. Hatabu, and F. S. Hodi, “Mon- [38] J. R. Podojil and S. D. Miller, “Potential targeting of B7-H4 for itoring immune-checkpoint blockade: response evaluation and the treatment of cancer,” Immunological Reviews, vol. 276, biomarker development,” Nature Reviews Clinical Oncology, no. 1, pp. 40–51, 2017. vol. 14, no. 11, pp. 655–668, 2017. [39] C. Chen, Q. X. Qu, Y. Shen et al., “Induced expression of B7- [23] I. Fabregat, J. Fernando, J. Mainez, and P. Sancho, “TGF-beta H4 on the surface of lung cancer cell by the tumor-associated signaling in cancer treatment,” Current Pharmaceutical macrophages: a potential mechanism of immune escape,” Design, vol. 20, no. 17, pp. 2934–2947, 2014. Cancer Letters, vol. 317, no. 1, pp. 99–105, 2012. [24] S. Sanjabi, S. A. Oh, and M. O. Li, “Regulation of the immune [40] N. Xie, J. B. Cai, L. Zhang et al., “Upregulation of B7-H4 pro- response by TGF-β: from conception to autoimmunity and motes tumor progression of intrahepatic cholangiocarci- infection,” Cold Spring Harbor perspectives in biology, vol. 9, noma,” Cell Death & Disease, vol. 8, no. 12, p. 3205, 2017. no. 6, 2017. [41] I. Kryczek, S. Wei, G. Zhu et al., “Relationship between B7-H4, [25] Y. Pan, F. Lu, Q. Fei et al., “Single-cell RNA sequencing reveals regulatory T cells, and patient outcome in human ovarian car- compartmental remodeling of tumor-infiltrating immune cells cinoma,” Cancer Research, vol. 67, no. 18, pp. 8900–8905, induced by anti-CD47 targeting in pancreatic cancer,” Journal 2007. of Hematology & Oncology, vol. 12, no. 1, p. 124, 2019. [42] T. A. W. Holderried, L. de Vos, E. G. Bawden et al., “Molecular [26] L. Zhang, S. Yu, C. Wang, C. Jia, Z. Lu, and J. Chen, “Establish- and immune correlates of TIM-3 (HAVCR2) and galectin 9 ment of a non‑coding RNAomics screening platform for the (LGALS9) mRNA expression and DNA methylation in mela- regulation of KRAS in pancreatic cancer by RNA sequencing,” noma,” Clinical Epigenetics, vol. 11, no. 1, p. 161, 2019. International Journal of Oncology, vol. 53, no. 6, pp. 2659– [43] L. Zhai, E. Ladomersky, K. L. Lauing et al., “Infiltrating T cells 2670, 2018. increase IDO1 expression in glioblastoma and contribute to Computational and Mathematical Methods in Medicine 13

decreased patient survival,” Clinical Cancer Research, vol. 23, no. 21, pp. 6650–6660, 2017. [44] M. Yang, D. McKay, J. W. Pollard, and C. E. Lewis, “Diverse functions of macrophages in different tumor microenviron- ments,” Cancer Research, vol. 78, no. 19, pp. 5492–5503, 2018. [45] Y. Tada, Y. Togashi, D. Kotani et al., “Targeting VEGFR2 with Ramucirumab strongly impacts effector/ activated regulatory T cells and CD8(+) T cells in the tumor microenvironment,” Journal for Immunotherapy of Cancer, vol. 6, no. 1, p. 106, 2018. [46] G. Cheng, A. Yu, and T. R. Malek, “T-cell tolerance and the multi-functional role of IL-2R signaling in T-regulatory cells,” Immunological Reviews, vol. 241, no. 1, pp. 63–76, 2011. [47] H. Stamm, F. Klingler, E. M. Grossjohann et al., “Immune checkpoints PVR and PVRL2 are prognostic markers in AML and their blockade represents a new therapeutic option,” Oncogene, vol. 37, no. 39, pp. 5269–5280, 2018.