Gallbladder Cancer (GBC) is among the most common malignant tumor of the gastrointestinal system. The only effective treatment is complete surgical resection for this cancer which can be given to only about 10% of patients because this cancer is usually diagnosed at advanced stages, however, recurrence rates after surgical resections are high (Zhu et al., 2010). GBC patients have an overall median survival of 19 months and a 5-year survival rate of 28.8% (Zhu et al., 2020) despite standard treatment, including chemotherapy, radiotherapy and targeted therapy. Therefore, it is necessary to understand the molecular mechanism associated with GBC and identify novel targets for therapeutic applications in GBC.

MicroRNAs (miRNAs) are a class of small non-coding RNAs (ncRNAs) of approximately ∼22 nucleotides long. Their mRNA targets are based on limited sequence complementarity between the miRNAs seed region (first 2–7 nucleotides from the 5′ end) and regions in the 3′ untranslated region (3’ -UTR) of the mRNA. miRNAs play a key role in the negative regulation of target genes and thus change the cellular environment. Nearly 60% of the total known mRNAs are regulated by miRNAs (Friedman et al., 2008). Another group of ncRNA—long non-coding RNAs (lncRNAs), which are >200 nucleotides in length, also affect mRNA stability (Sebastian-delaCruz et al., 2021). LncRNAs containing a ‘miRNA response element’ can compete with other RNAs and are believed to act as competing endogenous RNAs (ceRNAs). LncRNA-mediated ceRNA regulatory network, namely, lncRNA/miRNA/mRNA axis, is important in promoting tumorigenesis and can potentially serve as a handle to identify key therapeutic targets (Xu et al., 2022).

There are several high-throughput studies available in GBC to identify differentially expressed mRNAs (DEGs) and proteins (DEPs) (Kim et al., 2008; Miller et al., 2009; Huang et al., 2014; Wang et al., 2020a). Knowledge of the non-coding RNAs regulating the expression of these cancer-associated genes/proteins and the corresponding regulatory networks would be important to understand the pathogenesis of GBC and for therapeutic applications. Various groups have analyzed the differential expression of miRNAs (DEMs) in tissue (Letelier et al., 2014; Zhou et al., 2014; Goeppert et al., 2019), serum/plasma extracellular vesicles (Ueta et al., 2021; Yang et al., 2022) using high-throughput studies in GBC. Similarly, other researchers have analyzed differentially expressed lncRNA (DELs) using high-throughput studies (Ma et al., 2016; Wang et al., 2017a; Wu et al., 2017), however, the majority of the studies on lncRNA in GBC are targeted studies. Few of these studies revealed the “lncRNA-miRNA-mRNA network” in GBC in a targeted manner (Wang et al., 2016a; Hu et al., 2019; Yang et al., 2020; Zhang et al., 2020). Another study has constructed a ceRNA network using DELs, DEGs data and predicted miRNAs based on the DEGs from a single GBC dataset (GSE76633) (Kong et al., 2019a). In view of the availability of various targeted and high-throughput studies at lncRNA, miRNA, mRNA, and protein levels in GBC, the construction of a “ceRNA regulatory network” (lncRNA/miRNA/mRNA axis) based on multiple experimental datasets would be highly relevant.

We aimed to uncover the potential ceRNA regulatory networks regulating cancer-associated processes involved in the development of GBC. The present study applied multi-omics datasets available in GBC, including high-throughput GBC-proteomics data from our lab (Akhtar et al., 2023), and RNA interaction databases (predicted/experimentally verified) to identify novel lncRNA-miRNA-mRNA regulatory networks in GBC which may be explored for their therapeutic applications in GBC.

The study design and workflow of the study is shown in Figure 1. ‘Tissue-based’ datasets (DELs, DEMs, DEGs, and DEPs) were used from high-throughput and/or targeted studies and performed IPA analysis for miRNA-mRNA interactions (predicted and experimentally observed) using DEMs and DEGs or DEPs. The “experimentally observed targets” were further used for associated canonical pathways and PPI analysis to identify hub molecules. In addition, miRNA targets DE at both mRNA and protein levels with a positive correlation in expression were also analyzed. Finally, lncRNA-miRNA-mRNA regulatory networks were constructed for the selected targets possibly functional in GBC.

The molecular mechanism associated with the development and progression of GBC is not clear. An understanding of the ceRNA regulatory networks targeting the “tumor-associated proteins” in GBC would be highly important. In the present study, we integrated the “tissue-based” datasets (lncRNAs, miRNAs, mRNAs and proteins) from high-throughput and/or targeted studies to identify novel ceRNA regulatory networks in GBC. IPA analysis was performed for miRNA-mRNA interactions (predicted and experimentally observed) using DEMs and DEGs or DEPs. We focused on “experimentally observed targets” for associated canonical pathways and PPI analysis to identify hub molecules. In this study, we included the protein dataset to explore the interactions in which the miRNA target is DE at both mRNA and protein levels with a positive correlation in expression in GBC. Then, lncRNA-miRNA-mRNA regulatory networks were constructed for the selected targets, and the following strategy was used for screening the potential ceRNA regulatory network. First, the miRNA-mRNA interactions with strong evidence were screened as per miRTarBase and a literature survey was done for any report on these interactions in cancer conditions. Then, lncRNA-miRNA pairs were also screened for any report of cancer. Further, using the miRNA-mRNA interactions and lncRNA-miRNA interactions, we propose ceRNA networks (lncRNA-miRNA-mRNA) possibly functional in GBC and may have a potential for therapeutic applications in GBC.

We found “p53 signaling pathway” to be among the top pathway associated with 183 miRNA targets. PPI network analysis using these miRNA targets showed 5 hub molecules, 3 of them (p53, CCND1 and CTNNB1) were associated with p53 signaling pathway. TP53, a tumor suppressor gene, affects the cell cycle mechanism and programmed cell death (apoptosis) (Lowe and Lin, 2000; Yildirim et al., 2015). It is reported to be overexpressed in ∼56% of GBC cases (Ghosh et al., 2013) and 70% of GBC cases (Misra et al., 2000). Cyclin D1 (CCND1) participates in the cell cycle phase transition (G1/S phase) (Luo et al., 2017). Accumulation of β-catenin promotes the transcription of many oncogenes such as c-Myc and CyclinD-1 (Shang et al., 2017). Both CCND1 and CTNNB1 are negatively regulated by the p53 genes (Liu et al., 2001; Swaminathan et al., 2012).

Further, lncRNA-miRNA-mRNA regulatory networks for p53, CCND1 and CTNNB1 revealed novel ceRNAs possibly regulating the expression of p53 in GBC. We found three p53-miRNAs interactions (miR-125b-5p, miR-34a-5p, and miR-30a-5p), with strong evidence (miRTarBase), regulating the expression of p53 (Figure 6A). Le et al. (2009) highlighted the importance of miR-125b, a brain-enriched miRNA, in the negative regulation of p53 and p53-induced apoptosis during development and stress response. miR-125b has been reported as an oncogene that inhibits cell apoptosis by negatively regulating p53 expression (Wu et al., 2013). There is no report on the regulation of p53 through miR-34a-5p and miR-30a-5p in cancer. In one of our networks (Figure 6B), miR-193a-3p and miR-195-5p are regulating the expression of CCND1, and they are also reported to participate in the pathogenesis of hepatocellular carcinoma and pancreatic ductal adenocarcinoma, respectively, by targeting CCND1 (Chen et al., 2019; Wang et al., 2020b). miRNA-195 inhibits cell proliferation, migration and invasion in epithelial ovarian carcinoma (EOC). In addition, a negative correlation of miR 195 expression with that of CDC42 and CCND1 expression levels was also observed in EOC (Hao et al., 2020). We found miR-200a to be regulating the expression of CTNNB1 in our network (Figure 6C). miR-200a is demonstrated to target CTNNB1 in nasopharyngeal carcinoma (Xia et al., 2010).

Literature search on “lncRNA-miRNA pairs” in cancer revealed one lncRNA (MALAT1) -miR125b interaction in laryngocarcinoma (Zong et al., 2021), three lncRNA (TUG1, NEAT1, MALAT1) -miR-34a interactions reported in endometrial cancer, Nasopharyngeal Cancer, Melanoma (Liu et al., 2017; Ji et al., 2019; Li et al., 2019) and three lncRNA (PVT1, NEAT1, and MALAT1) -miR-30a-5p interactions in papillary thyroid carcinoma, gastric cancer, Hepatocellular Carcinoma (Feng et al., 2018; Pan et al., 2018; Rao et al., 2021). Previous studies revealed that the lncRNAs H19 and NEAT1 were found to directly target miR-193a-3p in Hepatocellular Carcinoma and Lung Adenocarcinoma, respectively (Ma et al., 2018; Xiong et al., 2018). PVT1 is reported to target miR-195 in osteosarcoma (Zhou et al., 2016). H19 has been reported to competitively bind to miR-200a and indirectly regulate β-catenin in colorectal cancer (Yang et al., 2017). MALAT1 is found to competitively bind to miR-200a-3p in non-small cell lung cancer (Wei et al., 2019).

Here, six ceRNA regulatory networks (MALAT1-miR125b-p53; PVT1/MALAT1-miR-195-CCND1, H19/NEAT1-miR-193a-3p-CCND1, H19/MALAT1-miR-200a-CTNNB1) were found for which lncRNA-miRNA and/or miRNA-mRNA interactions have been reported in other cancers. In view of the correlation in expression of these lncRNA, miRNA and mRNA, the ceRNA networks appear to play key regulatory roles and may be explored in GBC.

Out of 242, a total of 11 interactions include 9 targets reported to be DE at both mRNA and protein levels in GBC. Out of 9, we found lncRNA-miRNA-mRNA regulatory networks for 8 of them (CDX2, MTDH, TOP2A, TSPAN8, EZH2, TAGLN2, LMNB1, and PTMA) (Figure 7). The same strategy as explained earlier was used for screening the potential ceRNA regulatory networks possibly functional in GBC. We found miRNA-mRNA interaction for a total of 4 genes (EZH2, CDX2, TAGLN2, and PTMA). EZH2 has different roles in cancer, such as oncogenic, tumor suppressor, cancer cell metastasis, cancer immunity, and metabolism. Previous studies have shown its overexpression in different cancers, including prostate cancer, breast cancer, etc. (Duan et al., 2020). PTMA is upregulated and associated with the development of various cancers, including esophageal squamous cell carcinoma, colorectal, bladder, lung, and liver cancer (Zhu et al., 2019). The Caudal-type homeobox transcription factor 2 (CDX2) gene is a specific intestinal transcription factor that is involved in the embryonic development and differentiation of the intestine. Its overexpression in gastric carcinoma cells significantly inhibits cell growth and proliferation (Xie et al., 2010). Transgelin 2 (TAGLN2) is known to bind to actin to facilitate the formation of cytoskeletal structures. Downregulation of transgelin 2 is reported to promote breast cancer metastasis (Yang et al., 2019).

We found miRNA-mRNA interactions (miR-26a-5p/miR-101-3p-EZH2; miR-181c-5p-CDX2; miR1-3p-TAGLN2/PTMA) involving four genes. Interestingly, miR-26a-5p-EZH2 interaction was found to be involved in cell proliferation, cell invasion and apoptosis in GBC cells (Wang et al., 2016b). EZH2 is reported to be a direct target of miR-26a in Uveal Melanoma (UM) cells. Further, the knockout of EZH2 mimicked the tumor inhibition of miR-26a in UM cells (Li et al., 2021). A recent study demonstrated that miR-101-3p prevented retinoblastoma cell proliferation by targeting EZH2 and HDAC9 (Jin et al., 2018), prevented autophagy in endometrial cancer cells by targeting EZH2 (Wang and Liu, 2018) and inhibits invasion and metastasis in renal cell carcinoma by Targeting EZH2 (Dong et al., 2021). PTMA was identified as a target gene regulated by the miR-1 in bladder cancer (Yamasaki et al., 2012).

A literature search on “lncRNA-miRNA pairs” in cancer revealed two lncRNA (NEAT1, MALAT1) -miR-101-3p interactions in lung cancer (Wang et al., 2017b; Kong et al., 2019b), two lncRNA (TUG1, MALAT1) -miR-26a-5p interactions observed in colon cancer and colorectal cancer (Tian et al., 2019; Zhou et al., 2021) and two lncRNA (TUG1, MALAT1) -miR-1-3p interactions in hepatic carcinoma, esophagus cancer (Li et al., 2020; Tang et al., 2022). We found six ceRNA regulatory networks (TUG1/MALAT1-miR-26a-5p-EZH2; NEAT1/MALAT1-miR-101-3p- EZH2; TUG1/MALAT1-miR-1-3p-PTMA) targeting two genes, EZH2 and PTMA. In view of the functional role of EZH2 and PTMA in cancer, as discussed earlier, the ceRNA regulatory network targeting these two genes may be investigated in GBC.

Overall, we identified twelve ceRNA regulatory networks which might be functional in GBC, however, the experimental validation of these networks (expression analysis) in the clinical samples is the limitation of the present study. In future, in vitro and in vivo studies would be planned that might establish the functional role of these networks in GBC.

The present study used the experimental data from high-throughput/targeted studies on miRNAs, mRNAs and proteins in GBC and identified 183 miRNA targets. IPA analysis showed “p53 signaling pathway” to be the top pathway associated with them. Three of the targets were among the top 5 hub molecules in PPI network analysis. A total of 9 targets were reported to be DE at both mRNA and protein levels. Further, lncRNA-miRNA-mRNA regulatory networks were constructed for 3 targets (TP53, CCND1, and CTNNB1) associated with p53 signaling pathway and 9 targets (CDX2, MTDH, TOP2A, TSPAN8, EZH2, TAGLN2, LMNB1, and PTMA) DE at both mRNA and protein level. Overall, twelve ceRNA regulatory networks (MALAT1-miR125b-p53; PVT1/MALAT1-miR-195-CCND1, H19/NEAT1-miR-193a-3p-CCND1, H19/MALAT1-miR-200a-CTNNB1, TUG1/MALAT1-miR-26a-5p-EZH2; NEAT1/MALAT1-miR-101-3p- EZH2; TUG1/MALAT1-miR-1-3p-PTMA) were identified for which lncRNA-miRNA and/or miRNA-mRNA interactions have been reported in other cancers. These and other lncRNA-miRNA-mRNA regulatory networks may be experimentally validated and explored for their therapeutic applications in GBC.