Hepatic p63 regulates steatosis via IKKβ/ER stress

Abstract

p53 family members control several metabolic and cellular functions. The p53 ortholog p63 modulates cellular adaptations to stress and has a major role in cell maintenance and proliferation. Here we show that p63 regulates hepatic lipid metabolism. Mice with liver-specific p53 deletion develop steatosis and show increased levels of p63. Down-regulation of p63 attenuates liver steatosis in p53 knockout mice and in diet-induced obese mice, whereas the activation of p63 induces lipid accumulation. Hepatic overexpression of N-terminal transactivation domain TAp63 induces liver steatosis through IKKβ activation and the induction of ER stress, the inhibition of which rescues the liver functions. Expression of TAp63, IKKβ and XBP1s is also increased in livers of obese patients with NAFLD. In cultured human hepatocytes, TAp63 inhibition protects against oleic acid-induced lipid accumulation, whereas TAp63 overexpression promotes lipid storage, an effect reversible by IKKβ silencing. Our findings indicate an unexpected role of the p63/IKKβ/ER stress pathway in lipid metabolism and liver disease.

Figure 1. Effect of hepatic p53 down-regulation on liver steatosis.

(a) p53 protein levels in the liver after a tail vein injection of an associate adenovirus serotype 8 (AAV8) expressing either GFP or Cre in p53floxed mice fed a standard diet (STD) (n=7 per group). Effects of the liver-specific silencing of p53 on (b) in hematoxylin-eosin (upper panel) and oil red O staining (lower panel) of mice liver sections (n=3 per group); (c) total liver TG content, serum AST, ALT and TG levels (n=8 AAV8-GFP and 10 AAV8-Cre mice); (d) liver protein levels of FAS, pJNK/JNK, pIRE/IRE, XBP1s, pPERK, peIF2α/eIF2α, cleaved caspase 3, cleaved caspase 7, ApoB100 and ApoB48; (e) mRNA expression of CPT1, ACADM, ACADL and FATP2; (f) serum ketone bodies; (g) glucose and (h) insulin tolerance test (n=8 AAV8-GFP and 10 AAV8-Cre mice). The values of AAV8 GFP mice were always normalized to 100% (n=7 per group). Protein GAPDH or transferrin levels were used to normalize protein levels. Dividing lines indicate splicings in the same gel. Uncropped blots of this Figure accompanied by the location of molecular weight markers are shown in Supplementary Fig. 10. Data are presented as mean±standard error mean (s.e.m.). Statistical significance, *P<0.05 and **P<0.01, was tested using Student t-test comparing AAV8-GFP and AAV8-Cre mice.

Figure 2. Hepatic rescue of p53 in mice fed a high fat diet ameliorates steatosis of global p53 null mice.

(a) GFP protein levels in the liver and brown adipose tissue (BAT) of WT and p53 null mice after a tail vein injection of adenoviruses encoding either GFP or p53 (n=5 per group). (b) p53 gene expression in the liver of WT and p53 null mice after the injection of adenoviruses encoding either GFP or p53. (c) Representative photomicrographs of hematoxylin-eosin (upper panel) and oil red O staining (lower panel) of mice liver sections (n=3 per group); (d) total liver TG content, serum AST and TG levels (n=6 GFP and 5 p53 in WT mice; n=5 GFP and 4 p53 in KO mice); (e) mRNA expression of CPT1, ACADM, ACADL and FATP2; and (f) protein levels of FAS, pJNK/JNK, pIRE/IRE, XBP1s, pPERK, peIF2α/eIF2α, cleaved caspase 3, cleaved caspase 7, ApoB100 and ApoB48 in the liver of WT and p53 null mice fed a HFD after the over-expression of hepatic p53. Western blots were performed separately in WT and p53 null mice, and the values of Ad GFP mice were always set to 100% (n=5 GFP and p53 in both WT and p53 null mice). GAPDH or transferrin were used to normalize protein levels. Dividing lines indicate splicings in the same gel. Uncropped blots of this Figure accompanied by the location of molecular weight markers are shown in Supplementary Fig. 11. Data are presented as mean±standard error mean (s.e.m.). Statistical significance, *P<0.05 and **P<0.01, was tested using Student t-test comparing WT and KO mice. ND: non detected.

Figure 3. p53 modulates p63 in the liver.

(a) p63 protein levels in WT and p53 null mice fed a standard diet (n=7 per group). (b) p63 liver protein levels in WT and p53 null mice fed a high fat diet injected with p53 dominant positive adenovirus and GFP (n=5 GFP and 4 p53 in both WT and p53 null mice). (c) p63 liver protein levels in p53 floxed mice injected with either AAV8-GFP or AAV8-CRE (n=7 per group). (d) Protein levels of TAp63 and ΔNp63 isoforms in the liver of WT and p53 null mice fed a standard diet. (e) Protein levels of TAp63 and ΔNp63 isoforms in the liver of WT and p53 null mice fed a high fat diet injected with p53 dominant positive adenovirus and GFP. GAPDH was used to normalize protein levels. Dividing lines indicate splicings in the same gel. Uncropped blots of this Figure accompanied by the location of molecular weight markers are shown in Supplementary Fig. 12. Data are presented as mean±standard error mean (s.e.m.). Statistical significance, *P<0.05 and **P<0.01, was tested using two-tailed Student t-test.

Figure 4. High fat diet increases hepatic TAp63α.

(a) Body weight; (b) cumulative food intake (n=10 per group); (c) protein levels of TAp63 and ΔNp63 isoforms; and (d) protein levels of FAS, pIRE, XBP1s and pPERK in the liver of mice fed a high fat diet during 2 and 4 weeks (n=6 per group). GAPDH was used to normalize protein levels. Dividing lines indicate splicings in the same gel. Uncropped blots of this Figure accompanied by the location of molecular weight markers are shown in Supplementary Fig. 13. Data are presented as mean±standard error mean (s.e.m.). Statistical significance, *P<0.05, **P<0.01 and ***P<0.001. For multiple comparison (a–c) a one way ANOVA followed by Bonferroni or Kruskal-Wallis test was performed. Student t-test was used in western blot (d) comparing STD and VHFD 2 weeks.

Figure 5. Down-regulation of p63 ameliorates p53 knockdown- and high fat diet-induced steatosis.

(a) p63 protein levels in the liver of mice with down-regulated p53 fed a standard diet (STD) after a tail vein injection of a lentiviral particle that encodes either GFP or a shRNA p63 (n=7 per group). Effect of the simultaneous liver silence of p53 and p63 on: (b) hematoxylin-eosin (upper panel) and oil red O staining (lower panel) of mice liver sections (n=3 per group); (c) total liver TG content, serum AST, ALT and TG levels (n=8 per group); (d) liver protein levels of FAS, pJNK/JNK, pIRE/IRE, XBP1s, pPERK, peIF2α/eIF2α, cleaved caspase 3, ApoB100 and ApoB48 (n=5 per group); and (e) mRNA expression of CPT1, ACADM, ACADL and FATP2. (f) p63 protein levels in the liver of mice fed a HFD injected with lentiviruses encoding shRNA p63; (g) effect of the hepatic down-regulation of p63 in mice fed a HFD in the hematoxylin-eosin (upper panel) and oil red O staining (lower panel) of mice liver sections (n=4 per group); (h) total liver TG content, serum AST, ALT and TG levels (n=11 GFP and 13 sh-RNA p63 mice); (i) liver protein levels of FAS, pJNK/JNK, pIRE/IRE, XBP1s, pPERK, peIF2α/eIF2α, cleaved caspase 3, ApoB100 and ApoB48; and (j) mRNA expression of CPT1, ACADM, ACADL and FATP2. Protein GAPDH or transferrin levels were used to normalize protein levels, and control values (AAV8 GFP and lentivirus Turbo GFP) were normalized to 100%. Dividing lines indicate splicings in the same gel (n=7 per group). Uncropped blots of this Figure accompanied by the location of molecular weight markers are shown in Supplementary Fig. 14. Data are presented as mean±standard error mean (s.e.m.). Statistical significance, *P<0.05 and **P<0.01, was tested using Student t-test.

Figure 6. Hepatic over-expression of TAp63α causes steatosis via ER stress.

(a) p63 protein levels in the liver of mice after a tail vein injection of an AAV8 over-expressing either GFP or TAp63α isoform. (b) Representative photomicrographs of hematoxylin-eosin (upper panel) and oil red O staining (lower panel) of mice liver sections (n=4 per group); (c) total liver TG content, serum AST, ALT and TG levels (n=9 AAV8-GFP and 10 AAV8-TAp63α); (d) protein levels of FAS, pJNK/JNK, pIRE/IRE, XBP1s, pPERK, peIF2α/eIF2α, cleaved caspase 3, cleaved caspase 7, ApoB100 and ApoB48; and (e) mRNA expression of CPT1, ACADM, ACADL and FATP2 in the liver of mice after hepatic over-expression of TAp63 and IP TUDCA administration (n=7 per group). (f) GRP78 protein levels in the liver of mice after a tail vein injection of an Ad over-expressing either GFP or GRP78 (n=6 per group). (g) Representative photomicrographs of hematoxylin-eosin (upper panel) and oil red O staining (lower panel) of mice liver sections (n=4 per group); (h) total liver TG content, serum AST, TG and ketone bodies levels (n=9 AAV8-GFP and 10 AAV8-TAp63α); (i) protein levels of FAS, pIRE/IRE, XBP1s, ApoB100 and ApoB48 (n=6 per group); and (j) mRNA expression of CPT1, ACADM, ACADL and FATP2 in the liver of mice after hepatic over-expression of TAp63α and Ad GRP78 administration (n=8 per group). Protein GAPDH or transferrin levels were used to normalize protein levels and control values (AAV8 GFP) were normalized to 100%. Dividing lines indicate splicings in the same gel (n=7 per group). Uncropped blots of this Figure accompanied by the location of molecular weight markers are shown in Supplementary Fig. 15 and Supplementary Fig. 16. Data are presented as mean±standard error mean (s.e.m.). Statistical significance, *P<0.05, **P<0.01 and ***P<0.001. For multiple comparison (c–e,h–j) a one way ANOVA followed by Bonferroni or Kruskal-Wallis test was performed. Student t-test was used in TAp63alpha and GRP78 liver protein levels (a,f).

Figure 7. TAp63α-induced ER stress precedes changes in FAS expression and IKKβ links p63 to ER stress.

(a) Representative photomicrographs of hematoxylin-eosin (upper panel) and oil red O staining (lower panel) of mice liver sections (n=3 per group); (b) total liver TG content and protein levels of TAp63α, XBP1s and FAS in the liver of mice 2 weeks after hepatic over-expression of TAp63α (n=6 per group). (c) Representative photomicrographs of hematoxylin-eosin (upper panel) and oil red O staining (lower panel) of mice liver sections (n=3 per group); (d) total liver TG content (n=8 per group); (e) hepatic FAS activity (n=8 per group); and (f) protein levels of XBP1s in the liver of mice injected with AAV over-expressing TAp63α treated with the FAS inhibitor C75 (n=7 per group). (g) pIKKα, pIKKβ, IKKβ protein and mRNA levels in the liver of mice after 2 or 4 weeks of the tail vein injection of an AAV8 over-expressing either GFP or TAp63α isoform (n=7 per group). (h) pIKKα, pIKKβ, IKKβ protein and mRNA levels in the liver of mice after a tail vein injection of a lentiviral particle that encodes either GFP or a shRNA p63 (n=7 per group). Protein GAPDH levels were used to normalize protein levels and control values (AAV8 GFP) were normalized to 100%. Dividing lines indicate splicings in the same gel (n=7 per group). Uncropped blots of this Figure accompanied by the location of molecular weight markers are shown in Supplementary Fig. 17. Data are presented as mean±standard error mean (s.e.m.). Statistical differences are denoted by *P<0.05, **P<0.01 and ***P<0.001. For multiple comparison (d) a one way ANOVA followed by Bonferroni or Kruskal-Wallis test was performed. Student t-test was used in the other panels.

Figure 8. Genetic manipulation of TAp63α regulates lipid content in THLE2 hepatocytes.

(a) Representative dual channel fluorescent photomicrograph of THLE2 cells showing staining of lipids for oleic acid (BODIPY 493/503, green) and nuclei (DAPI, blue) of THLE2 cells after the over-expression of TAp63α or ΔNp63α. Magnifications 63X. Right panel shows total lipid content (green area) (n=4 per group). (b) Protein levels of TAp63α, ΔNp63α, XBP1s and FAS in THLE2 cells after the over-expression of TAp63α or ΔNp63α (n=4 per group). (c) ER fragments in HepG2 cells transfected with empty plasmid control and plasmid encoding TAp63α. (d) Representative dual channel fluorescent photomicrograph of THLE2 cells showing staining of lipids for oleic acid (BODIPY 493/503, green) and nuclei (DAPI, blue) of THLE2 cells cultured in oleic acid medium, treated with empty siRNA control (left image) or siRNA TAp63α isoform (right image). Magnifications 63X. Right panel shows total lipid content (green area). (e) Protein levels of TAp63α, pIKKβ, XBP1s and FAS in THLE2 cells exposed to oleic acid, after treatment with siRNA against either control or TAp63α (n=3 per group). (f) Representative dual channel fluorescent photomicrograph of THLE2 cells showing staining for oleic acid (BODIPY 493/503, green) and nuclei (DAPI, blue) of THLE2 cells transfected with empty plasmid control (left image), plasmid encoding TAp63α (middle image) and co-transfected with plasmid encoding TAp63α and siRNA IKKβ 24 h after transfection (right image). Magnifications 63X. Right panel shows total TG (green area). (g) Protein levels of TAp63α, pIKKβ, XBP1s and FAS in THLE2 cells after TAp63 over-expression followed by IKKβ silencing. GAPDH was used to normalize protein levels (n=3 per group). Uncropped blots of this Figure accompanied by the location of molecular weight markers are shown in Supplementary Fig. 18. Data are presented as mean±standard error mean (s.e.m.). Statistical differences are denoted by *P<0.05, **P<0.01 and ***P<0.001. For multiple comparison (a,b,f,g) a one way ANOVA followed by Bonferroni or Kruskal-Wallis test was performed. Student t-test was used in the other panels.

Figure 9. TAp63α stimulates lipogenesis in THLE2 cells.

(a) De novo triglyceride (TG), diacylglicerol (DG), fatty acid, phospholipid, esterified cholesterol and free cholesterol lipogenesis (n=4 per group); (b) palmitate oxidation (n=4 per group); and (c) lipid turnover in THLE2 cells transfected with empty plasmid control and plasmid encoding TAp63α (n=4 per group). Data are presented as mean±standard error mean (s.e.m.). Statistical differences are denoted by *P<0.05, **P<0.01 and ***P<0.001, was tested using Student t-test.

Figure 10. TAp63α is up-regulated in the liver of obese patients with NAFLD/NASH.

Correlations between (a) BMI and TAp63α; (b) BMI and NAS score; (c) TAp63α and NAS score; (d) Liver mRNA levels of TAp63, IKKβ and XBP1s in subjects without NAFLD (n=18) and with NAFLD (n=48). HRPT was used to normalize mRNA levels; and (e) Representative photomicrographs of an immunohistochemistry against TAp63 in human liver slices from lean (n=11), NAFLD (n=23) and NASH (n=19) patients. (f) Schematic representation of the pathway proposed to modulate lipid metabolism in liver. Data are presented as mean±standard error mean (s.e.m.). Statistical differences are denoted by *P<0.05, **P<0.01 and ***P<0.001, was tested using Student t-test.