HuR/methyl-HuR and AUF1 regulate the MAT expressed during liver proliferation, differentiation, and carcinogenesis

Abstract

Background & aims: Hepatic de-differentiation, liver development, and malignant transformation are processes in which the levels of hepatic S-adenosylmethionine are tightly regulated by 2 genes: methionine adenosyltransferase 1A (MAT1A) and methionine adenosyltransferase 2A (MAT2A). MAT1A is expressed in the adult liver, whereas MAT2A expression primarily is extrahepatic and is associated strongly with liver proliferation. The mechanisms that regulate these expression patterns are not completely understood.

Methods: In silico analysis of the 3' untranslated region of MAT1A and MAT2A revealed putative binding sites for the RNA-binding proteins AU-rich RNA binding factor 1 (AUF1) and HuR, respectively. We investigated the posttranscriptional regulation of MAT1A and MAT2A by AUF1, HuR, and methyl-HuR in the aforementioned biological processes.

Results: During hepatic de-differentiation, the switch between MAT1A and MAT2A coincided with an increase in HuR and AUF1 expression. S-adenosylmethionine treatment altered this homeostasis by shifting the balance of AUF1 and methyl-HuR/HuR, which was identified as an inhibitor of MAT2A messenger RNA (mRNA) stability. We also observed a similar temporal distribution and a functional link between HuR, methyl-HuR, AUF1, and MAT1A and MAT2A during fetal liver development. Immunofluorescent analysis revealed increased levels of HuR and AUF1, and a decrease in methyl-HuR levels in human livers with hepatocellular carcinoma (HCC).

Conclusions: Our data strongly support a role for AUF1 and HuR/methyl-HuR in liver de-differentiation, development, and human HCC progression through the posttranslational regulation of MAT1A and MAT2A mRNAs.

Figure 1

HuR stabilizes MAT2A mRNA and AUF1 destabilizes MAT1A mRNA. A) Analysis of MAT2A mRNA from rat hepatocytes treated with AICAR (2 mM), AICAR+SAMe (4 mM), HGF (40 ng/ml) or HGF+SAMe for 4 h. Treatments were carried out in triplicate (p<0.05, *AICAR or HGF versus control, **SAMe+AICAR versus AICAR, ***SAMe+HGF versus HGF). B) After AICAR, SAMe or AICAR+SAMe treatments, rat hepatocytes were incubated with Actinomycin D (2 μg/ml) for 4 h. The levels of MAT2A and MAT1A mRNAs were normalized to GAPDH mRNA and represented on semi-logarithmic scale. Top graph *p<0.05, AICAR Act D or A+S Act D versus Act D, Bottom graph *p<0.05, SAMe versus control. C) Upper panel: Schematic representation of MAT2A and MAT1A mRNA, showing the biotinylated transcripts [5′UTR, Coding Region (CR), 3′UTR] and the predicted HuR and AUF1 motifs. Western blots demonstrate association between HuR or AUF1 with biotinylated MAT2A and MAT1A fragments. Biotin pull-down assays were carried out in triplicates using rat hepatocyte lysates. Bottom panel: RNP-IP analysis of MAT2A mRNA bound to HuR after SAMe, AICAR, AICAR+SAMe, HGF and HGF+SAMe treatments. The enrichment was calculated from triplicates (p<0.05, *AICAR or HGF versus control, **SAMe+AICAR versus AICAR, ***SAMe+HGF versus HGF). D) Three days after siRNA transfection, SAM-D and H4IIE cells were harvested to monitor the protein expression of HuR and MAT2A, or AUF1 and MAT1A, respectively. Western blots are representative of 3 independent experiments.

Figure 2

MAT2A expression during de-differentiation of cultured hepatocytes. A) Expression of MAT2A mRNA over a time course in rat hepatocytes in presence/absence of SAMe (4 mM) *p<0.05, time of treatment versus previous time. B) Representative western blots of HuR and methyl-HuR in rat hepatocytes during de-differentiation. HuR versus time 0 h and methyl-HuR versus time 0 ratios from densitometric analysis are presented; each assay was carried out in triplicate. C) The association of HuR with MAT2A mRNA was assayed by RNP-IP analysis using cytoplasmic fractions of rat hepatocytes incubated as described above. MAT2A mRNA was normalized to GAPDH mRNA in HuR-IPs and represented relative to the levels of MAT2A mRNA in IgG-IP samples. p<0.05, *time of treatment versus time 0 h, **SAMe versus control. D) RNP-IP analysis of MAT2A mRNA bound to methyl-HuR; the enrichment was calculated from triplicate samples, *p<0.05, SAMe versus control. E) Polysome gradient analysis in rat hepatocytes cultured for 24 h in presence/absence of SAMe. MAT2A mRNA levels were plotted as a percentage of the total MAT2A mRNA levels. The translational activity of each fraction is as follows: NB, not bound polysomes; NT, not translated, moderately translated (LMW, low-molecular-weight polysomes), and actively translated (HMW, high-molecular-weight polysomes) (upper panel). HuR and methyl-HuR protein in each fraction were analyzed by Western blot analysis. Each assay was carried out in triplicate.

Figure 3

MAT1A expression during de-differentiation of cultured hepatocytes. A) MAT1A mRNA expression in rat hepatocytes at the indicated times, *p<0.05, time of treatment versus time 0 h. B) Representative Western blot analyses of AUF1 and MATI/III levels in rat hepatocytes undergoing de-differentiation; data are representative of 3 independent experiments. C) RNP-IP analysis of the association of AUF1 with MAT1A mRNA in cytoplasmic fractions of rat hepatocytes incubated as indicated. The enrichment of MAT1A mRNA in AUF1-IPs was calculated as described in Figure 2C. p<0.05, *time of treatment versus time 0 h, **SAMe versus control.

Figure 4

Role of HuR, methyl-HuR and AUF1 during liver development. A, B) mRNA expression of MAT2A, HuR, MAT1A and AUF1 in fetal livers (E16, E18), and livers from pups (P1 and P5) and adult rats (3 months), normalized to GAPDH mRNA, p<0.05, *ages of development versus E16. C) Levels of HuR, methyl-HuR and AUF1 as evaluated by Western blot analysis. Ponceau S staining was used as loading control (Supplemental Figure 3). The ratio of methyl-HuR/HuR was calculated, *p<0.05, ages of development versus E16. D, E) Binding of HuR or AUF1 to target mRNAs during liver development, as assessed by RNP-IP and real-time PCR analysis, *p<0.05, ages of development versus E16. F) RNP-IP analysis of MAT2A mRNA bound to methyl-HuR, enrichment represents the average from triplicate experiments, *p<0.05, ages of development versus E16.

Figure 5

The levels of MAT2A mRNA are regulated by HuR in GNMT KO mice. A) MAT2A mRNA expression in GNMT KO mice expressed as fold change versus WT (*p<0.05). B) Levels of HuR, methyl-HuR, and loading control GAPDH in total extracts. C) RNP-IP analysis of HuR binding to MAT2A mRNA. D) RNP-IP analysis of MAT2A mRNA bound to methyl-HuR, *p<0.05, GNMT KO versus WT.

Figure 6

HuR, methyl-HuR and AUF1 detection in human HCC. A) Representative immunofluorescence analysis of HuR, methyl-HuR and AUF1 protein in normal and human HCC samples. B) The relative immunofluorescence intensity in cancer tissues was calculated using Image J software and expressed as fold change of the relative immunofluorescence intensity in normal tissues. Data are representative of experiments realized in 22 HCC patients and 4 normal biopsies, and fold changes are significantly different (*p<0.05).

Figure 7

Diagram of MAT2A, MAT1A and their post-translational regulators HuR, methyl-HuR, and AUF1 in hepatocyte de-differentiation, development and malignant transformation. Mature hepatocytes express high levels of MAT1A mRNA and low levels of AUF1, while MAT2A mRNA is in low abundance due to its negative regulator, methyl-HuR. During de-differentiation, the levels of AUF1 mRNA in hepatocytes increases and the ratio of methyl-HuR/HuR decreases. This leads to a switch from MAT1A to MAT2A mRNA expression. SAMe treatment of hepatocytes prevents these changes. Malignant transformation of hepatocytes is accompanied by similar expression patterns for AUF1, HuR and methyl-HuR, as well as for MAT1A and MAT2A mRNAs. During liver development, the opposite is observed, with decreased AUF1 levels and increased methyl-HuR/HuR ratios.