Hepatocyte-specific, PPARγ-regulated mechanisms to promote steatosis in adult mice

Abstract

Peroxisome proliferator-activated receptor γ (PPARγ) is the target for thiazolidinones (TZDs), drugs that improve insulin sensitivity and fatty liver in humans and rodent models, related to a reduction in hepatic de novo lipogenesis (DNL). The systemic effects of TZDs are in contrast to reports suggesting hepatocyte-specific activation of PPARγ promotes DNL, triacylglycerol (TAG) uptake and fatty acid (FA) esterification. As these hepatocyte-specific effects of PPARγ could counterbalance the positive therapeutic actions of systemic delivery of TZDs, the current study used a mouse model of adult-onset, liver (hepatocyte)-specific PPARγ knockdown (aLivPPARγkd). This model has advantages over existing congenital knockout models, by avoiding compensatory changes related to embryonic knockdown, thus better modeling the impact of altering PPARγ on adult physiology, where metabolic diseases most frequently develop. The impact of aLivPPARγkd on hepatic gene expression and endpoints in lipid metabolism was examined after 1 or 18 weeks (Chow-fed) or after 14 weeks of low- or high-fat (HF) diet. aLivPPARγkd reduced hepatic TAG content but did not impact endpoints in DNL or TAG uptake. However, aLivPPARγkd reduced the expression of the FA translocase (Cd36), in 18-week Chow- and HF-fed mice, associated with increased NEFA after HF feeding. Also, aLivPPARγkd dramatically reduced Mogat1 expression, that was reflected by an increase in hepatic monoacylglycerol (MAG) levels, indicative of reduced MOGAT activity. These results, coupled with previous reports, suggest that Cd36-mediated FA uptake and MAG pathway-mediated FA esterification are major targets of hepatocyte PPARγ, where loss of this control explains in part the protection against steatosis observed after aLivPPARγkd.

Keywords: Cd36; LC/MS; Mogat1; adult-onset hepatocyte-specific knockdown; diet-induced steatosis.

Figure 1. Hepatocyte-specificity expression of AAV8-TBGp driven transgene

Hepatocyte-specific expression of GFP in AAV8-TBGp-EGFP injected wild-type mice (green, A,B). GFP expression was absent in non-hepatocyte cells (yellow arrows) in sinusoids (Sn), central vein (CV), portal vein (PV), bile duct (BD) or artery (HA). TBGp-GFP was not expressed in hepatic stellate cells (HSC, desmin +, red, C) or macrophages (Mac, F4/80+, red, D). Sections were counterstained with DAPI (blue nuclei, A–D). Hepatic GFP (E) and Cre (F) expression was detected only in liver extracts of AAV8-TBGp-EGFP and AAV8-TBGp-Cre injected mice, respectively. In order to confirm the hepatocyte-specific activity of Cre recombinase, PPARγ^fl/fl mice were injected with AAV8-TBGp-Cre and expression of PPARγ was reduced in hepatic extracts but not adipose tissue (G). eWAT, epididymal fat; iWAT, inguinal fat; Ctx, cortex; Int, intestine. Asterisks indicate difference between AAV8-TBGp-Cre injected mice as compared to AAV8-TBGp-Null mice. *, p<0.05; ***, p<0.0001. n=3–6 mice/group.

Figure 2. Adult-onset hepatocyte-specific PPARγ knock-down (aLivPPARγkd) in chow-fed mice (A–C) and diet-induced obese mice (D–F)

A) PPARγ^fl/fl mice were injected at 10 weeks of age with 1.5*10¹¹GC AAV8-TBGp-Null (C, open columns) or 1.5*10¹¹GC AAV8-TBGp-Cre (Kd, shaded columns) via lateral tail vein and killed 1 or 18 weeks after. B) Hepatic PPARγ mRNA and (C) protein expression of 1 wk and 18 wks aLivPPARγkd and their littermate control mice. Representative image of the western-blot for hepatic PPARγ and Histone [as loading control]). Of note, nuclear protein variability within total hepatic extracts (assessed by Histone) alters the amount of PPARγ protein detected. n=4–5 mice/group. D) LF-fed PPARγ^fl/fl mice were injected at 10 weeks as described above and immediately half of the mice were fed at high-fat diet (HF, 60% Kcal from fat) to induce liver steatosis while the rest were maintained on a low fat diet (LF, 10% Kcal from fat). Mice were killed 14 weeks after. E). Hepatic PPARγ mRNA and (F) protein expression of 14 wks LF- and HF-fed aLivPPARγkd and their littermate controls. Representative image of the western-blot for hepatic PPARγ and Histone [as loading control]. Asterisks indicate differences between C and Kd. *, p<0.05; ***, p<0.0001. Letters indicate differences between 1wk and 18 wks or LF and HF-fed mice within group. b, p<0.01; c, p<0.0001. n=5–6 mice/group

Figure 3. Impact of aLivPPARγkd in hepatic gene expression of molecular mechanisms controlling hepatic TAG levels

A) 1 wk (top) and 18 wks (bottom) chow-fed and B) 14 wks LF-fed (top) and HF-fed (bottom) aLivPPARγkd-induced regulation of hepatic gene expression. Graphs represent the natural logarithm of the relative change in the gene expression of aLivPPARγkd mice as compared to their littermates controls (set at 0, x-axis), within age (A) or diet (B). Absolute values are included in Supplemental Table 1A,B. Asterisks show significant changes between C and aLivPPARγkd within age (A) or diet (B). *, p<0.05; **, p<0.01, ***, p<0.0001. Livers were collected at 1100h (4h after food removal). FA_ox: fatty acid oxidation, TAG_hydr: TAG hydrolysis, VLDL_syn: VLDL synthesis, DNL: de novo lipogenesis and TAG_syn: TAG synthesis. Selected genes of these metabolic pathways represented in figure 3: peroxisome proliferator-activated receptor α (PPARα), acyl-CoA synthetase long-chain family member 1 (Acsl1), carnitine palmitoyltransferase 1α (Cpt1α), hepatic nuclear factor 4 α (Hnf4α), PPARγ co-activator 1 α; (Pgc1α), Cyp4a10, adipose triglyceride lipase (Atgl), hormone-sensitive lipase (Hsl), monoacylglycerol lipase (Mgll), apolipoprotein B (ApoB), microsomal triglyceride transfer protein (Mttp), sterol response element binding protein 1c (Srebp1c), acetyl-CoA carboxilase 1 (Acc1), fatty acid synthase (Fasn), fatty acid elongase (Elovl6), stearoyl –CoA desaturase 1 (Scd1), hepatic lipase (Hl), low density lipoprotein lipase receptor (Ldlr), very-low density lipoprotein receptor (Vldlr), lipoprotein related protein 1 (Lrp1), fatty acid translocase (Cd36), glycerol phosphate acyltransferase (Gpat1), monoacylglycerol acyltransferase 1 or 2 (Mogat1/2), diacylglycerolacyltransferase 1/2 (Dgat1/2).

Figure 4. aLivPPARγkd reduces HF diet-induced FA levels, but has little impact on FA indices of DNL

A) Hepatic FA levels of 16:0, palmitate; 16:1, palmitoleate; and 18:2, linoleate, levels and B) hepatic FA ratios indicative of DNL include the SCD-index (16:1/16:0) and the DNL-index (16:0/18:2), where control (open columns, C) and aLivPPARγkd (closed columns, Kd) mice were fed a LF- or a HF-diet for 14 weeks. Asterisks indicate differences between C and Kd. ***, p<0.0001. Letters indicate differences between LF and HF-fed mice within group. a, p<0.05; c, p<0.0001. n=4–6 mice/group.

Figure 5. aLivPPARγkd reduces HF diet-induced hepatic TAG and DAG levels while increasing MAG levels, independent of diet, indicative of impaired hepatic MAG pathway activity in aLivPPARγkd mice

A) Schematic representation of acylglycerol synthesis by glycerol-3-phosphate (G3P) pathway that produces DAG by subsequent re-esterification of FA in G3P and lysophosphatidic acid (LPA) or by monoacylglycerol (MAG) pathway that produces DAG after re-esterification of FA in MAG. B) Relative hepatic TAG, DAG and MAG levels assessed by liquid chromatography/mass spectrometry (LC/MS) in control (open columns) and aLivPPARγkd (close columns) mice. TAG, DAG and MAG are shown as relative values of LF-fed controls. Asterisks indicate differences between control and aLivPPARγkd. ***, p<0.0001. Letters indicate differences between LF and HF-fed mice within group. b, p<0.01; c, p<0.0001 n=5–6 mice/group.