Fig 1.
Transcriptional alterations accompanying Mdr2-/- liver disease progression.
(A) Volcano plots reporting differentially expressed genes (up-regulated brown, down-regulated green) identified by comparing WT and inflamed Mdr2-/- livers (left panel) and Mdr2-/- inflamed livers and tumors (right panel). The y-axis shows the -Log10 of the P-values determined by Cuffdiff analysis. (B) Gene ontology analyses on the sets of differentially expressed genes (up- and down-regulated) in each comparison. (C) Representative RNA-seq snapshots of differentially expressed genes. (D) Representative genes belonging to cell adhesion and ECM/cytoskeleton organization (Col4a5, Chd1, Mmp2), to glutathione metabolism (Gclc, Gstt2, Gstm1,) and to metabolism of xenobiotics by CYP450 enzymes (Cyp2c39, Cyp4a32, Nr1i3). Differences were assessed using two-sided Mann-Whitney test (p < 0.05).
Fig 2.
Stage-specific gene expression profile of Mdr2-/- livers.
Clusters have been generated by considering differentially expressed genes with FC ≥2, q-value ≤0.05, FPKM ≥2. The number of genes per cluster is indicated in each plot. Bottom panels report selected GO terms enriched for each cluster. Right panels show PWMs overrepresented on the promoters (+/-1000 bp from TSS) of genes of each cluster, whose cognate TF expression trend is consistent with the cluster.
Fig 3.
Mdr2-/- tumor development is characterized by global overexpression of endobiotics detoxification pathway members.
Ingenuity pathway analysis was used to generate the network overlaid with relative gene expression levels of inflammation to tumor transition. Node colors indicate the upregulated (red) and downregulated (green) genes relative to the comparison of tumor samples over inflamed livers. Items depicted by a dashed circle represent several members of the Cyp2c (specifically, Cyp2c29, Cyp2c37, Cyp2c38, Cyp2c39, Cyp2c40, Cyp2c44, Cyp2c50, Cyp2c55, Cyp2c67, Cyp2c68, Cyp2c69, Cyp2c70) and Cyp4a (Cyp4a10, Cyp4a12a, Cyp4a12b, Cyp4a14, Cyp4a32) families. Network edges (the relationship between nodes represented by lines and arrows) represent direct (solid lines) and indirect (dashed lines) interactions between molecules as supported by information in the Ingenuity knowledge base. Each functional class of molecules is represented by a different node shape.
Fig 4.
H3K27Ac profiling of Mdr2-/- samples.
H3K27Ac ChIP-seq were performed on samples from normal (WT), inflamed (INF) and HCC (nodules) livers (TUM) from mice treated with clodronate liposomes. (A) PCA analysis. 2 out of 15 samples were removed as outliers. (B) GO categories associated with differentially acetylated genomic regions of WT, inflamed and liver samples, as inferred from a GREAT analysis. (C) Snapshots of representative WT, inflamed and tumor samples showing H3K27Ac and RNA-seq data of differentially expressed genes. (D) Motif finding analysis of the genomic regions analyzed in (B). The PWMs indicated have been filtered based on the expression of TFs on each group.
Fig 5.
Toxic bile acids are reduced to normal levels in HCC stage.
Bile acid quantification of Mdr2-/- livers by LC-MS/MS analysis. Bile acids were extracted from macrophage-depleted livers from 15/17-month-old Mdr2-/- mice (HCC stage), 8-month-old Mdr2-/- mice (inflamed stage) and age matched FVB/NJ mice (WT control) and quantified by MS. Each dot of the beeswarm plots (panel B, C and D) represents the sum of all bile acids quantified in one liver sample (5 animals per groups). Values of each quantified bile acid are listed in Table 1. The central black bars indicate the median with the 1st and 3rd quartile. In all analyses, differences were assessed using Welch Two Sample t-test (p < 0.05). Data are shown as nanogram of bile acids on mg of proteins extracted. (A) PCA of total bile acids plotted in B (B) Sum of total bile acids concentration, (C) Sum of hydrophobic bile acids, (D) LCA.
Table 1.
Bile acids content in Mdr2-/- livers.
Fig 6.
CAR inhibition blocks cancer progression in Mdr2-KO livers.
Nodules from HCC livers of Mdr2-/- mice vehicle (UT) or CAR inhibitor treated (T) (50 mg/kg) were measured by caliper, collected for histological analysis and for RT-QPCR expression (see also S10 Table). (A) Q-PCR analysis of representative CAR target genes (UT, nodule number = 22; T, nodule number = 12). (B) Total number of lesions per mouse. (C) Total number of carcinomas per mouse. The central black bars indicate the median. (D) Tumor content measured as a percentage of HCC in each nodule. The number of nodules in the two groups are reported in parentheses. (E) Size differences in nodules from treated and untreated Mdr2-/- mouse groups. In all analyses, differences were assessed using one-sided Mann-Whitney test (p < 0.05).