PNPLA3 downregulation exacerbates the fibrotic response in human hepatic stellate cells

Abstract

Non-alcoholic steatohepatitis (NASH) results, in part, from the interaction of metabolic derangements with predisposing genetic variants, leading to liver-related complications and mortality. The strongest genetic determinant is a highly prevalent missense variant in patatin-like phospholipase domain-containing protein 3 (PNPLA3 p.I148M). In human liver hepatocytes PNPLA3 localizes to the surface of lipid droplets where the mutant form is believed to enhance lipid accumulation and release of pro-inflammatory cytokines. Less is known about the role of PNPLA3 in hepatic stellate cells (HSCs). Here we characterized HSC obtained from patients carrying the wild type (n = 8 C/C) and the heterozygous (n = 6, C/G) or homozygous (n = 6, G/G) PNPLA3 I148M and investigated the effect of genotype and PNPLA3 downregulation on baseline and TGF-β-stimulated fibrotic gene expression. HSCs from all genotypes showed comparable baseline levels of PNPLA3 and expression of the fibrotic genes α-SMA, COL1A1, TIMP1 and SMAD7. Treatment with TGF-β increased PNPLA3 expression in all 3 genotypes (~2-fold) and resulted in similar stimulation of the expression of several fibrogenic genes. In primary human HSCs carrying wild-type (WT) PNPLA3, siRNA treatment reduced PNPLA3 mRNA by 79% resulting in increased expression of α-SMA, Col1a1, TIMP1, and SMAD7 in cells stimulated with TGF-β. Similarly, knock-down of PNPLA3 in HSCs carrying either C/G or G/G genotypes resulted in potentiation of TGF-β induced expression of fibrotic genes. Knockdown of PNPLA3 did not impact fibrotic gene expression in the absence of TGF-β treatment. Together, these data indicate that the presence of the I148M PNPLA3 mutation in HSC has no effect on baseline activation and that downregulation of PNPLA3 exacerbates the fibrotic response irrespective of the genotype.

Fig 1. PNPLA3 genotype does not impact activation of HSCs in basal or TGF-β stimulated states.

Primary HSCs (n = 20) were genotyped as wild-type, heterozygous, or homozygous for the I148M PNPLA3 mutation. After culture for 24h the cells were imaged and shown to have robust expression of GFAP, an identifying marker of hepatic stellate cells (A). These cells were then treated with or without 5ng/ml TGF-β and mRNA was isolated and qPCR was performed for markers of HSC activation. Baseline activation of HSC was assessed in control cells without TGF-β treatment and activated HSCs were assessed with TGF-β treatment for wildtype (black bar), heterozygous (light grey bar), and homozygous (dark grey bar) donors by expression of α-SMA (B), collagen 1 (C), TIMP1 (D), and SMAD7 (E). Levels of PNPLA3 (F) were measured across genotypes in baseline and TGF-β stimulated conditions. A small molecule ALK5 inhibitor (1mM) was added simultaneously with the TGF-β to serve as a positive control for reversal of activation. Values are mean +/- SD, p = ns within all treatments vs WT genotype by 1-way ANOVA.C/C = wild type PNPLA3; C/G = heterozygous for rs738409; G/G = homozygous for rs738409. n = 6.

Fig 2. PNPLA3 knockdown activates primary human HSCs regardless of rs738409 genotype.

Primary HSCs were treated with control 5nM scrambled or PNPLA3 siRNA (Ambion) followed by treatment with or without 5ng/ml TGF-β and cultured for 24 hours. Expression of α-SMA, collagen 1, TIMP1, and SMAD7 in wildtype (A), heterozygous (B), and homozygous (C) donors was determined by qPCR. Values are mean +/- SD; *p<0.05 **p<0.01 ***p < .001 by 2-way ANOVA vs scramble siRNA. n = 5–6.

Fig 3. Upregulation of PNPLA3 in primary human HSCs by ACC inhibition.

WT PNPLA3 HSCs donors (n = 3) were treated with a small molecule ACC inhibitor at 10uM for 16 hours in the presence or absence of 5 ng/ml TGF-β. A small molecule ALK5 inhibitor (1mM) was added simultaneously with the TGF-β to serve as a positive control for reversal of activation. Transcriptional regulation of SREBP1C (A) and PNPLA3 (B) was demonstrated by qPCR. mRNA expression of HSC activation markers αSMA (C), COL1A1 (D), COL1A4 (E) were also determined by qPCR. Values are mean +/- SEM; *p<0.05 **p<0.01 ***p < .001 by t-test vs untreated condition. n = 2–6.