Aim. Identification of candidate secreted biomarkers for hepatocellular carcinoma (HCC) diagnosis. Methods. Genes upregulated in HCC tissue and encoding secreted proteins were identified by RNA-seq. Gene expression changes in HCC were evaluated by RT-qPCR and meta-analysis of public databases. Biomarker properties were studied using ROC-curves, correlation and survival analysis. Results. PDGFA was identified by RNA-seq as an overexpressed gene encoding for a secreted protein in 5 HCC cases. PDGFA and GPC3 upregulation was revealed in 17 of 19 HCC samples and in most cases from the public databases. Combination of PDGFA and GPC3 discerned HCC and non-tumor tissue better than PDGFA or GPC3 alone. PDGFA overexpression was associated with better overall survival of the patients at early HCC stage and with weaker tumor invasion into blood vessels. Conclusion. PDGFA is a valuable secreted biomarker for HCC that might be used in combination with GPC3 to increase its sensitivity. K e y w o r d s: hepatocellular carcinoma, PDGFA, tumor biomarker, NGS ISSN 1993-6842 (on-line); ISSN 0233-7657 (print)

Background: Hepatocellular carcinoma (HCC) is the most common form of malignant liver tumors, characterized by unfavorable prognosis and low sensitivity to chemotherapy. HCC diagnosis is complicated by late manifestation of symptoms and lack of effective biomarkers. The existing MRI approach does not cover 20% of HCC cases (hypovascular variants). Another problem is differential diagnosis between G1-stage HCC and subclasses of hepatocellular adenoma (HCA). Present study examines the ability of potential biomarkers, previously identified by our group, to differentiate HCC from benign liver tumors represented by HCA and focal nodular hyperplasia (FNH). We also analyzed gene expression changes associated with the development of FNH and HCA in order to identify molecular markers capable of distinguishing between these two neoplasm types. Methods: 61 pairs of surgical biopsies of tumor and non-tumorous liver tissue of patients with HCA (5 cases), FNH (6 cases) and HCC (50 cases) were used in the study. Expression levels of RAB3B, IQGAP3, GPC3, HKDC1, TOP2A, GNAZ, PDGFA and CENPF genes were evaluated using RT-qPCR. Data on gene expression changes were statistically processed and sorted using hierarchical cluster analysis. Results: Significant (p < 0.05) increase in expression level of IQGAP3 (p ¼ 8.8x10 À8 ), GPC3 (p ¼ 4.2x10 À5 ), CENPF (p ¼ 5.1x10 À4 ), and TOP2A (p ¼ 0.042) was detected in HCC tissue but not in benign tumor samples when compared to respective non-tumor samples. HKDC1 and RAB3B overexpression was observed in both HCC and benign tumors. HCA and FNH cases differ considerably by pattern of gene expression changes. HKDC1 expression level is higher in FNH than in HCA (p ¼ 0.017). Conclusions: Overexpression of IQGAP3, GPC3, CENPF and TOP2A genes is specific for HCC, but not HCA and FNH, so these genes are promising candidate biomarkers for differential diagnosis of benign and malignant liver tumors. Activation of RAB3B and HKDC1 genes in FNH and HCA tissue suggests their possible role in the development of these neoplasms. Specific patterns of gene expression changes described for HCA and FNH indicate the difference in molecular mechanisms underlying their pathogenesis and provide a tool for distinguishing these neoplasms from each other.