Hepatocellular carcinoma (HCC) is a complex disease characterized by genomic and epigenomic changes. Each tumor represents a specific mutation profile in specific gene families that can determine the composition of the tumor and dictate its nature in terms of the tumor cell’s ability to remain undifferentiated, migrate, and metastasize (Waarta et al. 2022). Epi/genome-wide studies have been useful in deciphering the association of drug response with the expression profiles of tumor cells with specific mutations (Chiu et al. 2019; Li et al. 2020). Thus, both genetic and epigenetic tumor profiles not only shape the nature of tumor cells but also predict the response and outcomes of clinical intervention.
In our study, we used cell line models comprising three HCC cell lines, namely, HepG2, Huh7, and QGY7703, to understand the miRNA landscape present in these three cancer cell lines. With the use of specific probes targeting approximately 800 miRNAs, we quantified mature miRNAs that were differentially upregulated or downregulated in each of the cell lines. The differentially expressed miRNAs in HepG2 vs Huh7 cells were associated with the upregulation of the RAS, EGFR, and VEGFA signalling pathways. Furthermore, pathway enrichment analysis revealed downregulation of genes in the RUNX and NOTCH families and upregulation of kinases involved in G1/S transition in QGY7703 cells compared with Huh7 cells. Interestingly, RUNX and NOTCH are well-known tumor suppressors in HCC (Zhu et al. 2021; Krajinović et al. 2023). Downregulation of these two genes in QGY7703 cells may be critical for determining the rigor and growth of QGY7703 cells, which are faster-growing cell lines than Huh7 cells. Furthermore, our bioinformatic analysis revealed that, compared with HepG2 cells, QGY7703 cells exhibited enhanced Wnt signalling, which can explain why the mesenchymal properties of QGY7703 cells (Grant et al. 2012; Santhekadur et al. 2014) are more pronounced than those of HepG2 cells, which have a more or less epithelial morphology (Donato et al. 2015), making the former cell line more aggressive in terms of migratory and invasive ability than the latter.
It has been observed that nearly 13–44% of HCC patients overexpress Alk (4Liu et al. 2016). Increased cytokine expression (Song et al. 2021), ALK expression (Liu et al. 2016), and IGF1R (Guan et al. 2021) expression are known to fuel cell proliferation and survival. Our pathway interaction analysis identified several downregulated pathways in the HepG2 vs Huh7 comparison as well. These included a reduction in cytokine, ALK, and IGF1R signalling cascades. This can explain why HepG2 cells have a morphology similar to that of differentiated hepatocytes compared with that of Huh7 cells, which have a greater growth rate and shorter doubling time.
Many studies have shown that mutations in the TP53 and CTNNB1 genes are independent drivers of HCC suggesting that mutations in any of these two genes are sufficient to initiate tumors (Friemel et al. 2016). However, the mechanisms that contribute to tumor formation due to such mutations are poorly understood, and the complex epigenome and its target interactome are yet to be dissected completely, making it difficult to study the downstream targets, pathway overlaps, and regulators involved. An integrated bioinformatic approach is a powerful tool for understanding and obtaining a larger picture of complex cellular networks involved in diseases such as cancer. We employed an integrated bioinformatic approach to identify the upregulated and downregulated miRNAs in HCC clinical samples with CTNNB1 and TP53 mutations based on the small-RNA sequencing reads available from the TCGA-LIHC public cohort. We then compared our in vitro miRNA expression data with those of a public dataset (HepG2 vs Huh7: CTNNB1 mutant vs TP53 mutant) to identify differentially expressed miRNAs that showed similar trends in expression patterns. At least 9 different miRNAs (miR-885-5p, miR-424-5p, miR-130-5p, miR-296-5p, miR-382-5p, miR-181b5p, miR-181d-5p, miR-520d-3p, and miR-328-5p) were differentially expressed and exhibited a similar trend in their expression patterns based on their mutational profiles in both our in vitro and TCGA samples. Hub-gene prediction revealed MYC, BCL2, CASP3, PTEN, STAT3, ESR1, HIF1A, MTOR, CCND1, and H3C1 as the top interacting gene partners of these 9 miRNAs. Interestingly, we detected HCC and HCV infection as one of the pathways enriched in these hub genes according to the panel of 9 miRNAs mentioned above.
Identifying and predicting drug response is a critical step in cancer management. We performed a drug-gene interaction analysis using hub genes to identify drugs that could serve as possible targets and identified at least 878 drug candidates that were available as targets for nine gene targets (ESR1, PTEN, HIF1A, MTOR, BCL2, MYC, CASP3, CCND1, and STAT3), of which several drugs were for antineoplastic purposes. Finally, a pancancer drug analysis identified at least three drugs, namely, sorafenib (a HIF1A/PTEN target), pembrolizumab, and nivolumab (a PTEN target), that are already FDA-approved for HCC treatment and management. Interestingly, PTEN expression is lost or reduced in most HCCs and it is reported that restoring its expression can improve sorafenib resistance and mitigate sorafenib’s activity such as metabolic reprogramming (Zhoa et al. 2020; Miao et al. 2021) and increased apoptosis in HCC cells (Ruan et al. 2012). Likewise, PTEN-loss is also associated with PD-L1-mediated reduction of INF-γ and CD8 + T cells that can further facilitate tumor progression and metastasis (Zhoa et al. 2020; Vidotto et al. 2020). Tumors exhibiting PTEN loss can be potentially treated by pembrolizumab and nivolumab which can potentially target PD-L1 (Finn et al. 2020; Fessas et al. 2023). Further targeted pre-clinical and clinical studies can confirm the use of these drugs in the clinical management of HCC patients presenting specific mutations in either TP53 or CTNNB1 genes.
In the present study, employing a systems biology approach, using human HCC cell lines and clinical datasets of HCC patients with different mutational profiles from public datasets, we identified miRNA expression landscapes and predicted their target genes, interactomes, and associated pathways, thereby providing an overall insight into the epigenetic landscape of different HCC cellular subtypes based on mutational statuses. In addition, our study identified druggable genes that can be repurposed for the management of HCC. Although the identified gene targets are experimentally validated, the inclusion of functional studies can add more power to the study.