Pancreatic cancer (PC) is the 12th most prevalent cancer in men and the 11th most common cancer in women, with about 450,000 new cases diagnosed worldwide every year (1). The high mortality and poor prognosis are primarily due to lack of noticeable and distinctive clinical signs or biomarkers for early detection. Aggressive metastatic spread of PC contributes to the difficulty of treatment (2). Generally, PC is classified into two types: the most frequent pancreatic adenocarcinoma, which arises in the pancreas’ exocrine glands, and the less common pancreatic neuroendocrine tumor, which occurs in the pancreas’ endocrine tissue (1). Another forms of categorization is based on whether the tumors present in the entire (solid or cystic) or on the prevailing cell differentiation structure (ductal, acinar, or endocrine). Solid types include pancreatic ductal adenocarcinoma, neuroendocrine neoplasms, acinar cell carcinomas, and pancreatoblastomas. Mucinous cystic neoplasms, intraductal papillary mucinous neoplasms, and strong pseudopapillary neoplasms are some of the less harmful cystic forms (3).

The options of PC treatment are very restricted and highly dependent on the stage of the disease. Chemotherapy treatment remains the primary choice for patients with advanced and metastatic tumors. Radiation is another treatment for unresectable, metastatic cancer when combined with chemotherapy (4). The National Comprehensive Cancer Network (NCCN) and the American Society of Clinical Oncology (ASCO) recommend modified FOLFIRINOX, gemcitabine and capecitabine, and single-agent gemcitabine or fluorouracil. S-1, an oral 5-fluorouracil prodrug as a standard treatment for PC. However, those treatments only marginally prolong life expectancy by approximately 3% (5). Furthermore, patients with PC are typically diagnosed at advanced stage with remote metastases. Therapeutic resistance is a persistent challenge, dumping therapeutic efficacy and prognosis of PC (6). Therefore, it is urgent to develop more non-surgical therapeutic approaches to effectively treat PC. One of the strategies to tackle those issues is by utilizing the available drugs for new indications.

Drug repurposing is a method to identify new indications of a drug (7). This approach has several advantages in the drug discovery for specific indications. First, the risk of failure is lower because the repurposed drugs are relatively safe if early-stage studies have been completed in animal models and humans. Second, this approach is able to reduce the time for pharmacokinetic and toxicological studies. Third, less capital is required (8). Genomic information has been widely utilized for drug repurposing. One of the databases that provide genomic information is cBioPortal. cBioPortal is an open-access resource for the interactive exploration of multidimensional cancer genomics datasets (9). Herein, we employed cBioPortal as a main resource and integrated different annotations to prioritize the PC risk genes. The detailed flowchart of genomic-driven drug repurposing is illustrated in Figure 1 .

This study utilized a genomic database to retrieve somatic mutation genes of PC. Five functional annotations were applied to build a scoring system and prioritized PC risk drug target identifications. We found 318 drug target genes.12 genes of them had the highest score (score = 5) after prioritization. Molecular mechanism of the 12 genes were listed in Table 3 . After prioritization of candidate drugs, we found 13 potential drugs (tamoxifen, sunitinib, bosutinib, everolimus, afatinib, palbociclib, vorinostat, axitinib, clomifene, raloxifene fulvestrant, midostaurin and toremifene) for PC.

Out of 12 genes, some potential gene targets were highlighted by our pipelines, including BCR and MTOR (55, 56), which were targeted by bosutinib and everolimus, respectively. Genes including KRAS, TP53, and SMAD4 had also been reported from previous studies (57, 58). These results indicated that five functional annotations are useful tools to priortize the important genes.

CMap Touchstone analysis showed that ESR1 as a dominated target protein for PC. ESR1 is a ligand-activated transcription factor that belongs to the steroid/nuclear receptor superfamily (59). The estrogen-responsive element (ERE) on the promoters of ESR1 target genes dimerizes and binds to coactivators in response to estrogen binding (60). ESR1 is known to involve in various cancers, such as endometrial and ovarian cancers (61). A previous study mentioned that ESR1 is also expressed in a subset of pancreatic adenocarcinoma, most notably in mucinous tumors (55, 56). Tamoxifen is the ESR1-targeted drug with both preclinical trial and clinical results for PC (62, 63). Furthermore, several clinical trials have reported using estrogen therapy on PC treatment (53, 63–65). However, the results are still controversial. Another drug that targeted ESR1 is fulvestrant, originally used for metastatic breast cancer. Fulvestrant works by inhibiting the dimerization of the ESR1 receptor and exerts no estrogen agonist effect (66). Therefore, if compared with tamoxifen, side effects of fulvestrant seem to be more favorable.

Midostaurin is a candidate drug from our prioritization process. The target for midostaurin is PRKCA that is a phospholipid-dependent, cytoplasmic serine/threonine kinase acting as an intracellular signal transducer (67). PRKCA activates different proteins such as cellular proliferation, differentiation, and gene expression (68). In addition, the expression of PRKCA has been associated with an elevated expression of the multidrug resistance phenotype (MDR) (69). Clinically, PRKCA is a target for types of cancer including acute myeloid leukemia (AML), breast cancer, and ovarian cancer (70). However, the reference of PRKCA as a treatment target for PC is still limited. Therefore, we propose PRKCA as a potential novel target. In this study, midostaurin targeted PRKCA with a score of 89, indicating the similar profiles between midostaurin andgemcitabine.

This study focused on bioinformatic approaches to prioritize drug candidates for PC. However, we should emphasize that further experimental and clinical studies are necessary to confirm the findings. The advantage of these approaches is narrowing down the drug targets and improving the success rate of drug development by prioritizing the best candidates. It is worth noting that this study has some limitations. First, not all genes retrieved from our analysis were druggable. Second, we used only one drug, gemcitabine, as the standard treatment for PC; this approach omits other potential candidate drugs to be found. Finally, CMap did not provide specific cell lines for PC.

Using our pipeline, we reported that there are 13 compounds that are the most potential PC candidate drugs. Two drugs are with preclinical results, eight are under clinical trials, and three have no previous reports yet. Two drugs (midostaurin and fulvestrant) were proposed as novel drugs repurposing for PC treatment. Here, we highlighted that our research demonstrated the viability of using public genomic information as a potential drug discovery method.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding authors.

Conceptualization: EM and W-CC. Data curation: EM, WA, LMI, C-NK and W-CC. Formal analyses: EM, WA, LMI and W-CC. Data interpretation and discussion: EM, WA, LMI, W-CH, C-NK and W-CC. Writing—review and editing: EM, WA, LMI, W-CH, C-NK and W-CC. Manuscript revision: EM, C-NK and W-CC. Supervision: W-CH, C-NK and W-CC. All authors contributed to the article and approved the submitted version.

This work was supported by grants from the Ministry of Science and Technology, Taiwan (no. MOST109-2320-B-038-013) and Taipei Medical University, Taiwan (12310-106079;12310-10739) for Yusuke Nakamura Chair Professorship.

We thank Min-Rou Lin, Wan-Hsuan Chou, Dr. Che-Mai Chang (Taipei Medical University, Taiwan) for helpful comments during the revision and Nien-Yu Yang (Taipei Medical University, Taiwan; The University of Manchester, UK) for the assistance with Figure 1 draft.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.