The genomic landscape of Mongolian hepatocellular carcinoma

Julián Candia,Xin Wei Wang,Enkhjargal Bayarsaikhan,Mayank Tandon,Undarmaa Tudev,Anuradha Budhu,Ann Chao,Justin Lack,Jigjidsuren Chinburen,Lkhagva-Ochir Tovuu,Marshonna Forgues

doi:10.1038/s41467-020-18186-1

Abstract

Mongolia has the highest incidence of hepatocellular carcinoma (HCC) in the world, but its causative factors and underlying tumor biology remain unknown. Here, we describe molecular characteristics of HCC from 76 Mongolian patients by whole-exome and transcriptome sequencing. We present a comprehensive analysis of mutational signatures, driver genes, and molecular subtypes of Mongolian HCC compared to 373 HCC patients of different races and ethnicities and diverse etiologies. Mongolian HCC consists of prognostic molecular subtypes similar to those found in patients from other areas of Asia, Europe, and North America, as well as other unique subtypes, suggesting the presence of distinct etiologies linked to Mongolian patients. In addition to common driver mutations (TP53, CTNNB1) frequently found in pan-cancer analysis, Mongolian HCC exhibits unique drivers (most notably GTF2IRD2B, PNRC2, and SPTA1), the latter of which is associated with hepatitis D viral infection. These results suggest the existence of new molecular mechanisms at play in Mongolian hepatocarcinogenesis.

Highlights

Mongolia has the highest incidence of hepatocellular carcinoma (HCC) in the world, but its causative factors and underlying tumor biology remain unknown
In order to uncover molecular subtypes of Mongolian HCC, we implemented an unsupervised clustering approach coupled with survival analysis (Supplementary Fig. 1) and found four molecular subclasses, labeled MO1-4 (Fig. 1a)
It is interesting to notice, that MO2 appears to have an order of magnitude fewer tumor-vs-nontumor differentially expressed genes compared to the other subclasses; correspondingly, MO2 will be shown to carry fewer copy number variations (CNVs) and structural variants (SVs)

Summary

Results

In order to uncover molecular subtypes of Mongolian HCC, we implemented an unsupervised clustering approach coupled with survival analysis (Supplementary Fig. 1) and found four molecular subclasses, labeled MO1-4 (Fig. 1a) Associated to these four molecular subclasses, we found 575 signature genes, each of them significantly up- or down-regulated in one subclass relative to the other subclasses (Supplementary Data 3). AFP is a well-characterized biomarker for diagnosis, pathological grade, progression, and survival of HCC patients[14], whereas cirrhosis, on the other hand, is a well-known intermediate stage in the progression from chronic liver disease and fibrosis to liver tumorigenesis[15] In agreement with these findings, Kaplan–Meier plots of overall survival showed that subclasses MO1-2 correspond to statistically significant better prognosis compared with MO3-4 (Fig. 1b). We determined the mutational landscape of Mongolian HCC (Fig. 2) compared to previous studies of driver mutations in 373 HCC patients from a

40 Log-rank test p values

Methods

Code availability