MYC determines lineage commitment in KRAS-driven primary liver cancer development

Abstract

Background & aims: Primary liver cancer (PLC) comprises hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (iCCA), two frequent and lethal tumour types that differ regarding their tumour biology and responses to cancer therapies. Liver cells harbour a high degree of cellular plasticity and can give rise to either HCC or iCCA. However, little is known about the cell-intrinsic mechanisms directing an oncogenically transformed liver cell to either HCC or iCCA. The scope of this study was to identify cell-intrinsic factors determining lineage commitment in PLC.

Methods: Cross-species transcriptomic and epigenetic profiling was applied to murine HCCs and iCCAs and to two human PLC cohorts. Integrative data analysis comprised epigenetic Landscape In Silico deletion Analysis (LISA) of transcriptomic data and Hypergeometric Optimization of Motif EnRichment (HOMER) analysis of chromatin accessibility data. Identified candidate genes were subjected to functional genetic testing in non-germline genetically engineered PLC mouse models (shRNAmir knockdown or overexpression of full-length cDNAs).

Results: Integrative bioinformatic analyses of transcriptomic and epigenetic data pinpointed the Forkhead-family transcription factors FOXA1 and FOXA2 as MYC-dependent determination factors of the HCC lineage. Conversely, the ETS family transcription factor ETS1 was identified as a determinant of the iCCA lineage, which was found to be suppressed by MYC during HCC development. Strikingly, shRNA-mediated suppression of FOXA1 and FOXA2 with concomitant ETS1 expression fully switched HCC to iCCA development in PLC mouse models.

Conclusions: The herein reported data establish MYC as a key determinant of lineage commitment in PLC and provide a molecular explanation why common liver-damaging risk factors such as alcoholic or non-alcoholic steatohepatitis can lead to either HCC or iCCA.

Impact and implications: Liver cancer is a major health problem and comprises hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (iCCA), two frequent and lethal tumour types that differ regarding their morphology, tumour biology, and responses to cancer therapies. We identified the transcription factor and oncogenic master regulator MYC as a switch between HCC and iCCA development. When MYC levels are high at the time point when a hepatocyte becomes a tumour cell, an HCC is growing out. Conversely, if MYC levels are low at this time point, the result is the outgrowth of an iCCA. Our study provides a molecular explanation why common liver-damaging risk factors such as alcoholic or non-alcoholic steatohepatitis can lead to either HCC or iCCA. Furthermore, our data harbour potential for the development of better PLC therapies.

Keywords: Cancer; Cell of origin; Hepatocellular carcinoma (HCC); Hepatocyte; Intrahepatic cholangiocarcinoma (iCCA); Liver; MYC.

Fig. 1.. MYC regulates lineage commitment of hepatocytes during liver tumorigenesis.

(A, B) Experimental outline. Intrahepatic delivery of transposable vectors encoding for GFP and Kras^G12D (pT-CaGIK) (A) or Myc and Kras^G12D (pT-CaMIK) (C) and intrahepatic tumor burden (B, D). (E, F) Representative micrographs of HE, Masson Trichrome (M. Trichrome) staining and immunohistochemistry staining against K19 and HNF4alpha (scale bar, 100 μm; n = 6). (G) Hepatocyte lineage tracing. Representative images of native fluorescence of Kras^G12D (H) and Myc^OE; Kras^G12D (I) tumors in ROSA26mT-mG; p19^ARF−/− mice (left image scale bar, 200 μm; magnified inset scale bar, 100 μm; n = 3).

Fig. 2.. Co-expression of oncogenic beta-catenin does not alter lineage commitment of Kras^G12D driven liver carcinomas.

(A) Intrahepatic delivery of pT-CaKIG (Kras^G12D; GFP) and pT-CaN90BC (N90-CTNNB1; RFP) in p19^ARF−/− mice. (B) Representative micrographs of native fluorescence of tumors developed as in A (scale bar, 200 μm; magnified inset scale bar, 100 μm; n = 3). (C) Representative micrographs of immunohistochemistry stainings against K19 and M. Trichrome (scale bar, 100 μm). (D) Western blot analysis to detect the expression of beta-catenin from pT-CaN90BC vector. Notice the lower molecular weight of N90beta-catenin compared to wild-type protein.

Fig. 3.. MYC dependent transcriptional and chromatin changes in transformed hepatocytes.

(A) Experimental outline. pT-CaKIR, (Kras^G12D; RFP). (B) Heatmap of differentially expressed genes (DEGs). (C) Top enriched hallmark gene-sets in DEGs. (D) Histograms of distance to nearest TSS (“Transcription Start Site”) for the indicated groups. (E) Transcription factor binding sites (“TFBS”) enriched in either promoter (“P”) or enhancer (“E”) regions, more accessible in MK (left) along with predicted regulators of DEGs upregulated by MYC (right) (*p<=0.00001). (F) Same as (E) but considering regions more accessible in K, and DEGs downregulated by MYC. C, control; MK, Kras^G12D; pT3-EF1alpha-Myc; K, Kras^G12D; pT3-EF1alpha-empty.

Fig. 4.. Transcriptomic Analyses in PLC of the TIGER-LC cohort.

(A, B) Hierarchical clustering of PLC samples in the TIGER-LC cohort based on the gene expression of MYC target genes: in (A) the Dang_Bound_by_Myc gene set (Pearson’s Chi-squared test with Yates’ continuity correction, X-squared = 97.681, df = 1, p < 2.2e-16) and in (B) MYC responsive genes in HepG2 cells (X-squared = 100.89, df = 1, p < 2.2e-16) were used. (C, D) FOXA1 and FOXA2 target enrichment in human HCC and iCCA tumor samples and corresponding adjacent tissue from the TIGER-LC cohort (****p < 0.0001).

Fig. 5.. ETS1 and FOXA TFs impact lineage commitment in PLC.

(A-D) Representative micrographs of immunohistochemistry stainings against K19, HNF4alpha, M. Trichrome and HE staining of Myc^OE; Kras^G12D; p19^ARF−/− tumors expressing the indicated shRNAs or cDNA (scale bar, 100 μm; n = 3). (E) Schematic representation of the proposed model. Transformed hepatocytes lose gene expression of lineage determining genes driven by FOXA1/2 and activate ETS1 dependent iCCA/cholangiocytic genes. Oncogenic levels of MYC cooperate with FOXA1/2 to maintain and enforce the activity of hepatocyte/HCC specific enhancers and promoters, against ETS1 activity on iCCA/cholangiocytic cis regulatory regions, determining lineage commitment towards HCC.