Abstract:
Hepatocellular carcinoma (HCC) is one of the most prevalent worldwide cancer. Various studies reported that chronic infections with hepatitis B virus (HBV) or hepatitis C virus (HCV), reactivation of telomerase expressions, liver diseases and alcohol usage represent major contributors of HCC. Moreover, the reactivation of telomerase expression is one of the most important factors that contribute about 90% in HCC. Telomerase reactivation enables the uncontrolled cell proliferation resulting in malignant transformation. Telomerase complex is composed of the core enzyme, telomerase reverse transcriptase (hTERT), the telomerase RNA component (TERC) that acts as an RNA template, and dyskerin that stabilizes the telomerase complex. In telomerase complex, hTERT has been identified as crucial factor in reactivation of telomerase. Moreover, in reactivation of telomerase, hTERT promoter mutations have major contribution i.e., ~60%. It has been reported recently that NCOA3 binds with hTERT promoter and activates hTERT expression to promote HCC growth [1]. However, NCOA3 is phosphorylated by up regulated Dual-Specificity Tyrosine Phosphorylation–Regulated Kinase 3(Dyrk3) in normal conditions. Dyrk3 is down regulated via activating transcription factor 4(ATF4) which also form a negative feedback loop with Dyrk3. This results in no phosphorylation of NCAO3 and hTERT activation. Consequently, un-phosphorylated NCOA3 binds and activates ATF4, significantly promoting the purine synthesis pathway and HCC progression. Therefore, in this study we target Dyrk3 and hTERT via structural drug design and machine learning approaches. We identified hydrophobicity and hydrogen bonding as important features contributing positively towards the biological activity of both targets. Our machine learning model also predict hydrophobicity as important feature for distinguishing actives from in actives. Hence, targeting HCC via using such chemical compounds having these features must be fruitful towards its cure.
