Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2016 Mar;64(3):661-73.
doi: 10.1016/j.jhep.2015.11.024. Epub 2015 Nov 26.

Hepatic Stellate Cell Transdifferentiation Involves Genome-Wide Remodeling of the DNA Methylation Landscape

Affiliations
Free PMC article

Hepatic Stellate Cell Transdifferentiation Involves Genome-Wide Remodeling of the DNA Methylation Landscape

Agata Page et al. J Hepatol. .
Free PMC article

Abstract

Background & aims: DNA methylation (5-mC) is an epigenetic mark that is an established regulator of transcriptional repression with an important role in liver fibrosis. Currently, there is very little knowledge available as to how DNA methylation controls the phenotype of hepatic stellate cell (HSC), the key cell type responsible for onset and progression of liver fibrosis. Moreover, recently discovered DNA hydroxymethylation (5-hmC) is involved in transcriptional activation and its patterns are often altered in human diseases. The aim of this study is to investigate the role of DNA methylation/hydroxymethylation in liver fibrosis.

Methods: Levels of 5-mC and 5-hmC were assessed by slot blot in a range of animal liver fibrosis models and human liver diseases. Expression levels of TET and DNMT enzymes were measured by qRT-PCR and Western blotting. Reduced representation bisulfite sequencing (RRBS) method was used to examine 5-mC and 5-hmC patterns in quiescent and in vivo activated rat HSC.

Results: We demonstrate global alteration in 5-mC and 5-hmC and their regulatory enzymes that accompany liver fibrosis and HSC transdifferentiation. Using RRBS, we show exact genomic positions of changed methylation patterns in quiescent and in vivo activated rat HSC. In addition, we demonstrate that reduction in DNMT3a expression leads to attenuation of pro-fibrogenic phenotype in activated HSC.

Conclusions: Our data suggest that DNA 5-mC/5-hmC is a crucial step in HSC activation and therefore fibrogenesis. Changes in DNA methylation during HSC activation may bring new insights into the molecular events underpinning fibrogenesis and may provide biomarkers for disease progression as well as potential new drug targets.

Keywords: DNMTs; Epigenetics; Hepatic myofibroblasts; Liver fibrosis; TETs.

Figures

Figure 1
Figure 1. Levels of 5mC, 5hmC and their regulatory enzymes in animal liver fibrosis models
(A) mRNA and (B) protein levels of TET2, TET3, DNMT1, DNMT3a, DNMT3b were assessed in liver samples obtained from controls or rats on MCD diet. (C) Genomic DNA samples were used to evaluate global levels of 5-hmC and 5-mC. Methylene blue staining was used as loading control. The results were analysed using GelAnalyzer software and plotted on the graph. (D) mRNA and (E) protein levels of TET2, TET3, DNMT1, DNMT3a, DNMT3b and genomic DNA (F) were assessed in liver samples obtained from livers of sham operated or bile duct ligated rats. (G) ) mRNA and (H) protein levels of TET2, TET3, DNMT1, DNMT3a, DNMT3b and genomic DNA (I) were assessed in liver samples obtained from rats treated with olive oil as a control or chronic CCl4 injured rats. Error bars represent mean values ± standard error of the mean (SEM) *p<0.05; **p<0.005; ***p<0.001.
Figure 2
Figure 2. Levels of 5mC, 5hmC and their regulatory enzymes in human liver diseases
(A) mRNA and (B) protein levels of TET2, TET3, DNMT1, DNMT3a, DNMT3b were analysed in normal human liver samples or from freshly explanted primary sclerosing cholangitis liver (C) Genomic DNA samples were used to evaluate global levels of 5-hmC and 5-mC. Methylene blue staining was used as loading control. The results were analysed using GelAnalyzer software and plotted on the graph. (D-F) analysis of normal human livers or explanted alcoholic liver disease livers as already described in A to C (n=5). Error bars represent mean values ± standard error of the mean (SEM) *p<0.05; ***p<0.001.
Figure 3
Figure 3. Levels of 5mC, 5hmC and their regulatory enzymes in in vivo isolated HSCs from olive oil (control) or chronic CCl4 treated rats
(A) cytospins were carried out and stained for αSMA, cells were counted and number of αSMA positive cells expressed as percentage of total cell number. SDS-PAGE was carried out on the protein samples and immunoblotted for TGFβ1, GAPDH was used as loading control. (B) mRNA and (C) protein levels of TET2, TET3, DNMT1, DNMT3a, DNMT3b were analysed from in vivo activated or control HSCs (D) Genomic DNA samples were used to assess global levels of 5-hmC and 5-mC. Error bars represent mean values ± standard error of the mean (SEM) *p<0.05; ***p<0.001.
Figure 4
Figure 4. Levels of 5mC, 5hmC and their regulatory enzymes in isolated rat HSCs, which were allowed to spontaneously activate on plastic, in vitro
(A) protein levels of TET2, TET3, DNMT1, DNMT3a, DNMT3b were assessed in quiescent (day 1) and activated HSC (day 10) samples. (B) mRNA was isolated from quiescent HSCs (day 1) or HSCs cultured for 3, 5, 7 and 10 days and level of DNMT1 measured, (C) DNMT3a, (D) DNMT3b, (E) TET2 and (F) TET3 (G-H) gDNA was extracted from different stages of in vitro activation (days 1, 3, 5, 7 and 10) and used to measure the amount of 5-hmC (G) or (H) 5-mC by slot-blot. The results were analysed using GelAnalyzer software. Error bars represent mean values ± standard error of the mean (SEM) *p<0.05.
Figure 5
Figure 5. Genome-aligned 5-mC sites, identified by high-throughput sequencing (RRBS-Seq), were plotted by chromosomal location, displaying their respective methylation percentage
(A) Overlapping CpG methylated sites were identified by overlaying the 5-mC sites from the control and CCl4 groups. (B) The identified overlapping CpG methylated sites were allocated into genomic region with a defined annotated position within the genome. Genomic features corresponding to 5-mC methylation sites identified in (A) were plotted as a percentage of total. (C) Unique 5-mC sites found in the control HSCs only, were plotted onto the genome map (D) Genomic features corresponding to 5-mC methylation sites identified in (D) were plotted as a percentage of total. (E) Unique 5-mC sites found in the in vivo activated HSC only, were identified by removing the overlapping sites from the remainder of the 5m-CpG sites from each sample (F) Genomic features corresponding to 5-mC methylation sites identified in (E) were plotted as a percentage of total.
Figure 6
Figure 6. 5-hmC sites, identified by high-throughput sequencing (oxBS-Seq), were plotted by chromosomal location
(A) The amount of hydroxymethylation for each site was obtained by subtracting the methylation obtained from RRBS-Seq from the oxBS-Seq. Overlaying the control and CCl4 groups identified overlapping sites between the two groups which were plotted onto the genome map. Sites coloured in red show a significant (p < 0.05) log2fold-change in hydroxymethylation. (B) Genomic features corresponding to 5-hmC methylation sites identified in (A) were plotted as a percentage of total. (C) Unique 5-hmC sites found in the control HSCs were plotted onto the genome map (D) Genomic features corresponding to 5-hmC methylation sites identified in (D) were plotted as a percentage of total. (E) Unique 5-hmC sites found in the in vivo activated HSC CCl4 were identified by removing the overlapping sites from the remainder of the 5-hmC sites from each sample. (F) Genomic features corresponding to 5-hmC methylation sites identified in (E) were plotted as a percentage of total.
Figure 7
Figure 7. Rat HSC were transfected with control and DNMT3a siRNA
(A) mRNA levels of DNMT3a were quantified by qRT-PCR in cells treated with control or DNMT3a siRNA. (B) Twenty μg of whole cell protein extract from two preparations of rat HSCs treated with DNMT3a siRNA were separated by SDS-PAGE and immunoblotted for DNMT3a and GAPDH as loading control. qRT-PCR analysis of (C) TGFβ1, (D) Collagen 1A1 and (E) α-SMA expression in cells treated with control or DNMT3a siRNA was performed. Error bars represent mean values ± standard error of the mean (SEM) *p<0.05; **p<0.01; ***p<0.001. (F) DNA methylation pattern at Collagen 1A1 promoter as determined by RRBS was plotted in qHSC and aHSC, as shown by red blocks/bars. DNA methylation becomes erased in aHSCs. (G) DNA methylation pattern at α-SMA as determined by RRBS was plotted in qHSC and aHSC, as shown by red blocks/bars. DNA methylation largely becomes erased in aHSCs.

Comment in

Similar articles

See all similar articles

Cited by 23 articles

See all "Cited by" articles

Publication types

Substances

Feedback