If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Correspondence Address correspondence to: Yingfu Liu, MD, PhD, Department of Basic Medical Sciences, School of Medicine, Xiamen University, Xiamen 361102, China.
The matricellular protein periostin plays a critical role in liver inflammation, fibrosis, and even carcinoma. Here, the biological function of periostin in alcohol-related liver disease (ALD) was investigated.
Methods
We used wild-type (WT), Postn-null (Postn-/-) mice and Postn-/- mice with periostin recovery to investigate the biological function of periostin in ALD. Proximity-dependent biotin identification analysis identified the protein that interacted with periostin, and coimmunoprecipitation analysis validated the interaction between protein disulfide isomerase (PDI) and periostin. Pharmacological intervention and genetic knockdown of PDI were used to investigate the functional correlation between periostin and PDI in ALD development.
Results
Periostin was markedly upregulated in the livers of mice that were fed ethanol. Interestingly, periostin deficiency severely aggravated ALD in mice, whereas the recovery of periostin in the livers of Postn-/- mice significantly ameliorated ALD. Mechanistic studies showed that the upregulation of periostin alleviated ALD by activating autophagy through inhibition of the mechanistic target of rapamycin complex 1 (mTORC1) pathway, which was verified in murine models treated with the mTOR inhibitor rapamycin and the autophagy inhibitor MHY1485. Furthermore, a protein interaction map of periostin was generated by proximity-dependent biotin identification analysis. Interaction profile analysis identified PDI as a key protein that interacted with periostin. Intriguingly, periostin-mediated enhancement of autophagy by inhibiting the mTORC1 pathway in ALD depended on its interaction with PDI. Moreover, alcohol-induced periostin overexpression was regulated by transcription factor EB.
Conclusions
Collectively, these findings clarify a novel biological function and mechanism of periostin in ALD and the periostin-PDI-mTORC1 axis is a critical determinant of ALD.
This study demonstrates that the upregulation of periostin in liver parenchymal cells in response to alcohol exposure is regulated by transcription factor EB and prevents the development of alcohol-related liver disease by activating hepatocyte autophagy by inhibiting the mechanistic target of rapamycin complex 1 pathway through interactions with protein disulfide isomerase.
Alcohol-related liver disease (ALD) is the leading cause of chronic liver disease worldwide, and its characteristics are lipid accumulation and liver injury.
ALD includes a spectrum of clinical disorders and changes in liver tissue, ranging from hepatosteatosis to alcoholic hepatitis, cirrhosis, and ultimately hepatocellular carcinoma.
Over 90% of heavy drinkers develop hepatic steatosis during early alcohol consumption, whereas approximately 30% to 35% develop severe ALD, indicating that many factors are involved in ALD occurrence.
Currently, there is no effective therapy for ALD, and so identifying new therapeutic targets in ALD has become a clinically urgent need.
Periostin, which is a multifunctional matricellular protein, is generally expressed at low levels in most adult tissues except for sites of inflammation or injury.
Periostin deficiency reduces diethylnitrosamine-induced liver cancer in mice by decreasing hepatic stellate cell activation and cancer cell proliferation.
However, whether periostin is involved in the progression of ALD remains unclear.
Multiple factors contribute to ethanol pathogenesis. Long-term alcohol intake promotes metabolite accumulation in the liver, which, in turn, impairs hepatocyte metabolism and causes hepatocyte damage and death.
Tremendous efforts have been expended to investigate the pathogenesis and development of ALD, but the underlying molecular mechanisms of ethanol in ALD remain unclear. Recent studies have indicated that autophagy is activated in response to ethanol exposure and plays a vital role in alleviating ethanol-induced pathology in the liver. For instance, autophagy protects against ethanol-induced liver toxicity in mice, which is characterized by increased autophagosome formation and autophagic flux in the livers of mice fed ethanol and in primary hepatocytes after ethanol treatment.
These studies indicate that autophagy might be an avenue for the development of therapeutics for patients with ALD. As reported, periostin can regulate autophagy in colorectal cancer and chronic kidney failure
; however, the regulatory effect of periostin on autophagy in ALD remains unknown. Transcription factor EB (TFEB) can regulate autophagy-related gene transcription, and overexpression of TFEB in the liver reduces the severity of ethanol-induced liver injury in mice,
but whether TFEB regulates periostin overexpression under alcohol stimulation needs to be further elucidated.
Herein, the expression of periostin was found to be significantly upregulated in ethanol-fed mouse livers, and periostin deficiency severely exacerbated alcoholic liver disease in mice. In contrast, the recovery of periostin in the livers of Postn-/- mice alleviated their ALD phenotypes. Moreover, we elucidated that periostin as mainly derived from liver parenchymal cells in ALD tissues, and its expression as regulated by TFEB. Furthermore, periostin interacted with protein disulfide isomerase (PDI) to activate hepatocyte autophagy by inhibiting the mechanistic target of rapamycin complex 1 (mTORC1) signaling pathway, thereby playing a protective role in ALD. These results suggest that periostin may be an effective therapeutic target in ALD.
Results
Periostin is Significantly Upregulated in Hepatocytes in Response to Alcohol Stimulation in Vivo and in Vitro
To investigate periostin expression in ALD, we used the National Institute on Alcohol Abuse and Alcoholism model to construct a murine alcohol-related liver disease model
(Figure 1A). Compared with pair-fed mice, the mRNA and protein levels of periostin in the liver and serum periostin levels were significantly elevated in ethanol-fed mice (Figure 1B–D). Immunohistochemical and immunofluorescent staining showed that periostin levels were markedly higher in ethanol-fed mice than in pair-fed mice (Figure 1E–F). We also found that periostin expression was notably increased in the mouse hepatocyte cell line AML12 after treatment with alcohol for 6 hours 12 hours, and 24 hours (Figure 1I). Notably, the mRNA and protein levels of periostin were increased in isolated primary mouse hepatic parenchymal cells after alcohol stimulation (Figure 1G–H).
Figure 1Periostin is markedly upregulated in hepatocytes. (A) Schematic image of a murine model with chronic-plus-binge ethanol feeding. C57/BL6 male mice were fed with the Lieber-DeCarli control diet for 5 days, and then they were fed with an isocaloric Lieber-DeCarli control diet or Lieber-DeCarli ethanol diet for 10 days, followed by a single gavage of dextrin (9 g/kg) or ethanol (5 g/kg), respectively. Mice were euthanized at 9 hours after gavage. (B–C) The expression of periostin in liver tissues from the pair-fed and EtOH-fed mice (n = 5–8 mice per group) by Western blot (B) and qRT-PCR (C) analyses. (D) Enzyme-linked immunosorbent assay analysis of serum periostin levels in pair-fed and EtOH-fed mice (n = 5–8 mice per group). (E) The expression of periostin in liver tissues from pair-fed and EtOH-fed mice (n = 5–8 mice per group) by immunohistochemical staining analyses. (F) Immunofluorescence analysis of periostin, HNF4α, F4/80, α-SMA, and vimentin in mouse livers after control or EtOH feeding. (G–H) qRT-PCR analyses (G) and immunofluorescence staining (H) of periostin in primary hepatic parenchymal cells after 100 mM alcohol treatment for 6 hours (n = 3). (I) Immunofluorescence staining of periostin in AML12 cells treated with 100 mM alcohol for the indicated times. Data represent the mean ± standard deviation (3 independently repeated experiments); *P < .05; **P < .01 (Student t test). Scale bars, 50 μm.
To determine which cell types are the major source of periostin in the alcohol-induced liver, specific markers were used to label different cell types. Interestingly, immunofluorescence staining of ethanol-fed mouse liver tissues showed that periostin mainly colocalized with HNF4α (hepatocyte marker) rather than F4/80 (Kupffer cell marker), α-SMA (activated hepatic stellate cell marker), and vimentin (fibroblast cell marker) (Figure 1F). Taken together, these results suggest that periostin is markedly upregulated in ALD and mainly expressed in liver parenchymal cells.
Periostin Deficiency Severely Exacerbates ALD in Mice
To determine the effect of periostin on ALD, we subjected wild-type (WT) and Postn-/- mice to ethanol feeding. During the process of ethanol feeding, Postn-/- and WT mice showed similar body weight changes and food intake (Figure 2A–B). However, ethanol-fed Postn-/- mice exhibited higher ratios of liver-to-body weight than ethanol-fed WT mice (Figure 2C). Moreover, as shown in Figures 2D and 2E, the livers of ethanol-fed Postn-/- mice had more lipid accumulation than those of ethanol-fed WT mice. Hepatic and serum levels of triglyceride (TG) and total cholesterol (T-CHO) were significantly increased in ethanol-fed Postn-/- mice compared with ethanol-fed WT mice (Figure 3A, B, E, and F). Moreover, hepatic and serum levels of alanine transaminase (ALT) and aspartate transaminase (AST) in ethanol-fed Postn-/- mice were also markedly higher than those in ethanol-fed WT mice or pair-fed mice (Figure 3C, D, G, and H).
Figure 2Periostin deficiency aggravates ALD in mice. (A–C) The effect of EtOH feeding on body weight changes (A), food intake (B), and the ratios of liver-to-body weight (C) (n = 5–6 mice per group). (D–E) H&E staining (D) and Oil red staining (E) of pair-fed and EtOH-fed in WT and Postn-/- mouse livers. Data represent the mean ± standard deviation (3 independently repeated experiments); *P < .05; ***P < .001; n.s., not significant (Student t test). Scale bars, 50 μm.
Figure 3Periostin deficiency exacerbates ALD in mice. (A–H) Serum (A–D) and hepatic (E–H) levels of TG, T-CHO, AST, and ALT in WT and Postn-/- mice under control diet or EtOH feeding (n = 5–6 mice per group). (I–J) The relative mRNA levels of inflammation-related genes (I) and fibrosis-related genes (J) in WT and Postn-/- mouse livers (n = 5–6 mice per group). (K–L) Sirius red staining (K) and Masson staining (L) of WT or Postn-/- mouse livers. Data represent the mean ± standard deviation (3 independently repeated experiments); *P < .05; **P < .01; ***P < .001; n.s., not significant (Student t test). Scale bars, 50 μm.
We further found that inflammatory-related genes such as Tnf, Il1b, Il6, and Ccl2 and fibrosis-related genes such as Col6a1, Col6a3, Col8a1, and Col1a1 were all higher in Postn-/- mice than in WT mice under ethanol feeding (Figure 3I–J). Additionally, Sirius red staining and Masson staining showed that Postn-/- mice exhibited more fiber deposition in their livers than WT mice in response to ethanol stimulation (Figure 3K–L). However, there was no significant difference in any of the above factors between control diet-fed WT and Postn-/- mice. Collectively, these data indicate that deletion of periostin severely exacerbates liver injury induced by ethanol in the development of murine ALD.
Supplementation With Periostin in Postn-/- Mouse Livers Attenuates Murine ALD
Because periostin deletion in mice exacerbated ALD, we explored whether periostin recovery in Postn-/- mouse livers in vivo could attenuate ALD pathology by injecting adenoviruses expressing murine periostin (Ad-Postn) through the tail vein (Figure 4A). As shown in Figure 4B, periostin re-expression was detectable in Postn-/- mouse livers 10 days after Ad-Postn injection but not in the Ad-GFP-injected group. Body weight changes and food intake were comparable among the 4 groups, indicating that ethanol feeding and the adenoviruses did not affect the growth of the mice (Figure 4C–D). However, the Ad-Postn-injected Postn-/- mice (Ad-Postn/Postn-/- mice) showed lower ratios of liver-to-body weight than Ad-GFP-injected Postn-/- mice (Ad-GFP/Postn-/- mice) in response to ethanol feeding (Figure 4E). Ad-Postn/Postn-/- mice exhibited much less hepatic lipid droplet accumulation than Ad-GFP/Postn-/- mice after ethanol treatment (Figure 4F–G). Moreover, hepatic and serum TG and T-CHO levels were significantly decreased when periostin was re-expressed in ethanol-fed Postn-/- mouse livers (Figure 5A, B, E, and F).
Figure 4Supplementation with periostin in Postn-/-mouse livers attenuates murine ALD. (A) Schematic image of Postn-/- mice injected with adenovirus containing Postn (Ad-Postn) or GFP (Ad-GFP) through tail vein injection on the third day of control food feeding. Male mice were fed with an isocaloric Lieber-DeCarli control diet (Pair-fed) or Lieber-DeCarli ethanol diet (EtOH-fed) for 10 days, followed by a single gavage of dextrin (9 g/kg) or ethanol (5 g/kg), respectively. Mice were euthanized at 9 hours after gavage. (B) Confirmation of periostin in Postn-/- mouse livers after injection of Ad-Postn or Ad-GFP by Western blot. (C–E) The effect of EtOH feeding on body weight changes (C), food intake (D), and the ratios of liver-to-body weight (E) in Ad-Postn/Postn-/- and Ad-GFP/Postn-/- mice (n = 6–8 mice per group). (F–G) H&E staining (F) and Oil red staining (G) of Ad-Postn/Postn-/- and Ad-GFP/Postn-/- mouse livers. Data represent the mean ± standard deviation (3 independently repeated experiments); *P < .05; **P < .01; ****P < .0001; n.s., not significant (Student t test). Scale bars, 50 μm.
Figure 5Supplementation with periostin in Postn-/-mouse livers alleviates murine ALD. (A–H) Serum (A–D) and hepatic (E–H) levels of TG, T-CHO, AST, and ALT in Ad-Postn/Postn-/- and Ad-GFP/Postn-/- mice under control diet or EtOH feeding (n = 6–8 mice per group). (I–J) The relative mRNA levels of inflammation-related genes (I) and fibrosis-related genes (J) in Ad-Postn/Postn-/- and Ad-GFP/Postn-/- mouse livers (n = 6–8 mice per group). (K–L) Sirius red staining (K) and Masson staining (L) of Ad-Postn/Postn-/- and Ad-GFP/Postn-/- mouse livers. Data represent the mean ± standard deviation (3 independently repeated experiments); *P < .05; **P < .01; ***P < .001; n.s., not significant (Student t test). Scale bars, 50 μm.
We also found that Ad-Postn/Postn-/- mice exhibited much lower AST and ALT levels than Ad-GFP/Postn-/- mice in response to alcohol treatment (Figure 5C, D, G, and H). Notably, the expression of inflammatory-related genes and fibrosis-related genes was markedly decreased in Ad-Postn/Postn-/- mice compared with Ad-GFP/Postn-/- mice under ethanol feeding, but no differences were observed in pair-fed mice (Figure 5I–J). Moreover, there were fewer fibers in the Ad-Postn group than in the Ad-GFP group after ethanol feeding, and there were no significant differences in control diet-fed mice (Figure 5K–L). These data suggest that periostin recovery in Postn-/- mouse livers effectively rescues severe ALD caused by periostin deletion in mice.
Periostin Enhances Autophagy by Inhibiting the mTORC1 Pathway in Response to Alcohol Exposure
Next, we explored the mechanism by which periostin participates in the pathogenesis of ALD. Current evidence indicates that autophagy plays an important role in reducing acute ethanol-induced hepatotoxicity and steatosis in mice.
Although endoplasmic reticulum (ER) stress, fatty oxidation, and fat synthesis might be involved in our murine ALD model (Figure 6A–C), we focused on autophagy. Interestingly, the protein levels of autophagy-related genes, including Atg7, Beclin-1, and LC3-II, were decreased in Postn-/- mice compared with WT mice in response to ethanol feeding (Figure 6D), whereas periostin re-expression in the Postn-/- mouse livers restored these autophagy-related protein levels (Figure 6E). Moreover, p62 was significantly upregulated in Postn-/- mice fed ethanol, whereas periostin re-expression resulted in a decrease despite no statistical significance (Figure 6D–E). The mTORC1 signaling pathway is not only a key factor in ALD
DEP domain-containing mTOR-interacting protein suppresses lipogenesis and ameliorates hepatic steatosis and acute-on-chronic liver injury in alcoholic liver disease.
Then, we examined whether periostin affects the mTORC1 signaling pathway. Compared with ethanol-fed WT mice, the phosphorylated levels of the mTORC1 substrate S40 ribosomal protein S6 Ser235/Ser236 sites and 4E-BP1 Thr37/Thr46 sites were remarkedly increased in Postn-/- mice in response to ethanol exposure (Figure 6F), whereas the phosphorylated levels of S6 and 4E-BP1 were markedly decreased when periostin was re-expressed in Postn-/- mouse livers (Figure 6G). Moreover, the changes of mTORC1 and autophagy in primary hepatocytes under ethanol treatment were similar to those of livers in the murine ALD model (Figure 7A). Collectively, these data reveal that periostin deficiency activates the mTORC1 signaling pathway and impairs autophagy in murine ALD.
Figure 6Periostin activates autophagy by inhibiting mTORC1 pathway in response to alcohol exposure. (A–C) The mRNA levels of genes related to ER stress (A), fatty oxidation (B), and fat synthesis (C) in WT and Postn-/- mouse livers after pair diet or EtOH feeding (n = 4–6 mice per group). (D, F) Western blot analysis of autophagy and mTORC1 pathway-related proteins in WT and Postn-/- mouse livers after control or ethanol feeding. (E, G) Western blot analysis of the indicated proteins in Ad-Postn/Postn-/- and Ad-GFP/Postn-/- mouse livers after control or EtOH feeding. Data represent the mean ± standard deviation (3 independently repeated experiments); *P < .05; **P < .01; ***P < .001; n.s., not significant (Student t test).
Figure 7Periostin enhances autophagy by inhibiting mTORC1 pathway under alcohol exposure. (A–D) Western blots analysis of the indicated proteins in primary hepatic cells isolated from the WT and Postn-/- mice (A), rapamycin-treated primary Postn-/- mouse hepatocytes (B), MHY1485-administrated primary WT mouse hepatocytes (C), and periostin-overexpressed AML12 cells (D) under EtOH exposure or not. (3 independently repeated experiments)
To further determine whether periostin regulates autophagy through the mTOR pathway, we used the mTOR inhibitor rapamycin or activator MHY1485 combined with ethanol and treated isolated primary hepatocytes from Postn-/- or WT mice. As expected, rapamycin treatment significantly inhibited the mTORC1 signaling pathway and activated autophagy in primary Postn-/- murine hepatocytes, although alcohol treatment alone had no effect (Figure 7B). However, in primary WT hepatic cells, alcohol treatment markedly inhibited the mTORC1 signaling pathway and effectively activated autophagy, while the mTOR activator MHY1485 significantly blocked the activation of autophagy (Figure 7C). Moreover, overexpressing periostin in AML12 cells inhibited the mTORC1 signaling pathway and enhanced autophagy after alcohol stimulation (Figure 7D). Overall, periostin promotes autophagy by inhibiting the mTORC1 signaling pathway in hepatocytes in response to alcohol exposure.
Periostin Alleviates Murine ALD by activating autophagy and inhibiting the mTORC1 pathway
Furthermore, we investigated the function of periostin in regulating autophagy in ALD in vivo. We first treated WT mice with the mTOR inhibitor rapamycin and the autophagy inhibitor MYH1485 to assess the effect of autophagy on ALD (Figure 8). To further determine the role of periostin, we treated Postn-/- mice with rapamycin and Ad-Postn/Postn-/- mice with MHY1485 (Figures 9 and 10). Unexpectedly, rapamycin administration significantly enhanced autophagy and alleviated ALD in WT mice (Figure 8). In Postn-/- mice, rapamycin treatment markedly inhibited the liver mTORC1 pathway, resulting in a significant decrease in 4E-BP1 and S6 phosphorylation (Figure 9B), and prominently increased the expression of the autophagy-related proteins Atg7, Beclin-1, and LC3-II (Figure 9C). During rapamycin treatment, the body weight and food intake of mice did not change (Figure 9D–E). However, rapamycin administration significantly attenuated ALD phenotypes of Postn-/- mice, which were characterized by a decreased ratio of liver-to-body weight and the number of lipid vacuoles in the livers (Figure 9F, K). Hepatic and serum contents of TG, T-CHO, AST, and ALT, hepatic inflammation, and fiber deposition were decreased in the rapamycin-treated Postn-/- group (Figure 9G–L). These results suggest that autophagy activation by rapamycin mitigates the aggravation of hepatic steatosis and liver injury caused by periostin deficiency.
Figure 8Periostin alleviates ALD by activating autophagy and inhibiting the mTORC1 pathway in WT mice. (A) Schematic images of WT mice administrated with rapamycin and MHY1485. (B–C) Western blot analysis of the mTORC1 and autophagy-related proteins in WT mouse livers treated with rapamycin and MHY1485. (D–E) Body weight changes and food intake in WT mice with rapamycin and MHY1485 administration (n = 10 mice per group). (F) The ratios of liver-to-body weight in the livers of WT mice treated with MHY1485. G, H&E, Oil red, Sirius red, and Masson staining of the livers from WT mice injected with rapamycin and MHY1485. (H–I) The relative mRNA levels of inflammation (H) and fibrosis-related genes (I) in the livers of WT mice treated with rapamycin and MHY1485. (n = 10 mice per group). (J–K) Serum (J) and hepatic (K) levels of TG, T-CHO, AST, and ALT in WT mice administrated with rapamycin and MHY1485. Data represent the mean ± standard deviation (3 independently repeated experiments); *P < .05; **P < .01; ***P < .001; n.s., not significant (Student t test). Scale bars, 50 μm.
Figure 9Inhibition of mTORC1 pathway alleviates ALD in Postn-/-mice. (A) Schematic images of Postn-/- mice treated with rapamycin. (B–C) Western blot analysis of the mTORC1 and autophagy-related proteins in the livers of Postn-/- mice treated with rapamycin. (D–E) Body weight changes and food intake in Postn-/- mice with rapamycin treatment (n = 6–8 mice per group). (F) The ratios of liver-to-body weight in the livers of Postn-/- mice with rapamycin treatment. (G–H) Serum (G) and hepatic (H) levels of TG, T-CHO, AST, and ALT in Postn-/- mice administrated with rapamycin. (I–J) The relative mRNA levels of inflammation (I) and fibrosis-related genes (J) in the livers of Postn-/- mice treated with rapamycin. (n = 6–8 mice per group). (K) H&E staining and Oil red staining in the livers of Postn-/- mice with rapamycin treatment. (L) Sirius red and Masson staining of the livers of Postn-/- mice injected with rapamycin. Data represent the mean ± standard deviation (3 independently repeated experiments); *P < .05; **P < .01; ***P < .001; n.s., not significant (Student t test). Scale bars, 50 μm.
Figure 10Autophagy inhibition abrogates the positive effect of periostin in Ad-Postn/Postn-/-mice. (A) Schematic images of Ad-Postn/Postn-/- mice administrated with MHY1485. (B–C) Western blot analysis of the mTORC1 and autophagy-related proteins in the livers of Ad-Postn/Postn-/- mice treated with MHY1485. (D–E) Body weight changes and food intake in Ad-Postn/Postn-/- mice with MHY1485 administration (n = 6–8 mice per group). (F) The ratios of liver-to-body weight in the livers of Ad-Postn/Postn-/- mice treated with MHY1485. (G–H) Serum (G) and hepatic (H) levels of TG, T-CHO, AST, and ALT in Ad-Postn/Postn-/- mice administrated with MHY1485. (I–J) The relative mRNA levels of inflammation (I) and fibrosis-related genes (J) in the livers of Ad-Postn/Postn-/- mice treated with MHY1485. (n = 6–8 mice per group). (K) H&E staining and Oil red staining in the livers of Ad-Postn/Postn-/- mice treated with MHY1485. (L) Sirius red and Masson staining of the livers from Ad-Postn/Postn-/- mice injected with MHY1485. Data represent the mean ± standard deviation (3 independently repeated experiments); *P < .05; **P < .01; n.s., not significant (Student t test). Scale bars, 50 μm.
Conversely, treatment with the autophagy inhibitor MHY1485 significantly impaired autophagy and aggravated ALD in WT mice (Figure 8). To elucidate the role of periostin in regulating autophagy, we further treated the Ad-Postn/Postn-/- mouse model with MHY1485. After MHY1485 treatment, the phosphorylated levels of 4E-BP1 and S6 were markedly increased, whereas the expression of the autophagy-related proteins Atg7, Beclin-1, and LC3-II was significantly reduced in Ad-Postn/Postn-/- mice (Figure 10B–C). The body weight and food intake of Ad-Postn/Postn-/- mice did not change during MHY1485 treatment (Figure 10D–E). In contrast to rapamycin treatment, MHY1485 administration increased the ratio of liver-to-body weight and the number of lipid droplets in Ad-Postn/Postn-/- mouse livers (Figure 10F, K). Moreover, hepatic and serum levels of TG, T-CHO, AST, and ALT, hepatic inflammation, and fiber accumulation were increased in ethanol-fed Ad-Postn/Postn-/- mice treated with MHY1485 (Figure 10G–L), suggesting that autophagy inhibition by MHY1485 abrogated the positive effect of periostin in Ad-Postn/Postn-/- mice. Briefly, the protective effect of periostin on murine ALD involves enhancing autophagy by inhibiting the mTORC1 signaling pathway.
Proximity-dependent Biotin Identification Technique Identifies PDI as the Interacting Protein of Periostin
To further reveal the molecular mechanism by which periostin regulates ALD, we used a proximity-dependent biotin identification (BioID) technique to explore the interactome of periostin. After construction and confirmation of stable expression of recombinant periostin-BirA∗ in AML12 cells (Figure 11A–B), we successfully determined promiscuous biotinylation activity by incubating the transfected AML12 cells with biotin (Figure 11B–C). After propitious capturing of the biotinylated proteins by streptavidin affinity purification and liquid chromatography with tandem mass spectrometry (LC‒MS/MS) analysis, a total of 330 proteins and 352 proteins were identified in the BirA∗ and periostin-BirA∗ groups, respectively (Figure 11D). Given that only the proteins identified in the periostin-BirA∗ group were considered to interact with periostin, 153 candidate proteins were obtained (Table 1). Gene Ontology (GO) annotation further classified these identified candidate proteins into cellular component, molecular function, and biological process categories. As shown in Figure 11E, among cellular components, these proteins were mainly distributed in the cytoplasm and membrane, suggesting that periostin takes part in connecting the inside and outside of the cell; among molecular functions, these candidates were mainly involved with protein binding and RNA binding; and among biological processes, a majority of these proteins were associated with translation and transport, whereas several were implicated in protein folding and cell‒cell adhesion. Moreover, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis indicated that these candidate proteins significantly participated in the metabolic pathway and protein processing (Figure 11F).
Figure 11BioID and bioinformatics analysis of candidate interacting proteins of periostin. (A) Schematic representation of the constructed periostin-BirA*-HA and BirA*-HA plasmids. (B) Analysis of promiscuous biotinylation by periostin-BirA* and BirA*. The black triangle indicates the target proteins. (C) Evaluation of the biotinylated proteins by Western blot analysis. AML12 cells were separately transfected with the periostin-BirA*-HA or pBirA*-HA plasmids. Cells were biotinylated and lysed in vivo for pull-down, and the biotinylated proteins were eluted from beads and analyzed by Western blot. (D) The number of proteins identified in samples transfected with periostin-BirA*-HA or BirA*-HA plasmids was analyzed by Venn diagram. (E) Classification of the candidate proteins by GO in terms of cellular component, molecular function, and biological process. (F) KEGG enrichment analysis of the candidate proteins.
Based on the results of MS/MS identification, the top 10 identified candidate proteins that interacted with periostin are listed in Figure 12A. Referring to the parameters of the identified protein, physiological function, and literature review, several candidate proteins, including protein disulfide isomerase (PDI, encoded by P4hb), which is an enzyme in the eukaryote ER, were chosen for further validation.
According to the results of external co-immunoprecipitation (Co-IP) in HEK293T cells and semi-internal Co-IP in AML12 cells, we confirmed that PDI could coprecipitate with periostin (Figure 12B–D). The structure of the periostin protein is composed of an N-terminal EMI domain, four tandem Fasciclin (FAS1) domains, and C-terminal domains (Figure 12E). The EMI domain is mainly involved in protein‒protein interactions or protein multimerization.
To further identify the functional motifs by which periostin interacts with PDI, we constructed full-length (FL) periostin and 2 truncated vectors, one containing a signal peptide, EMI domain, and four FAS1 domains (D1) and the other composed of a signal peptide and EMI domain (D2). We ultimately found that the EMI domain of periostin showed a high affinity for PDI (Figure 12F). Moreover, a series of vectors expressing PDI FL, truncated PDI lacking Sig_peptide (D1), thioredoxin_like (D2), PDI_b_family (D3), and PDI_b'_family (D4) were also constructed (Figure 12G). Interestingly, PDI FL, D1, and D2 but not D3 or D4 exhibited a high affinity for periostin (Figure 12H). These results suggest that the EMI domain of periostin interacts with the PDI_b_family domain.
Figure 12The interaction between periostin and PDI. (A) The top 10 identified candidate interacting proteins of periostin. (B–C) Co-IP analysis of HER293T cells co-transfected with HA-tagged periostin and Flag-tagged PDI. (D) Co-IP analysis of AML12 cells transfected with HA-tagged periostin plasmid. (E, G) The model of the periostin (E) and PDI (G) truncated plasmids. (F, H) Co-IP analysis of HER293T cells co-transfected with Flag-tagged PDI and the indicated HA-tagged periostin truncations (F) and co-transfected with HA-tagged periostin and the indicated Flag-tagged PDI truncations (H) (3 independently repeated experiments).
Periostin-mediated Activation of Autophagy Depends on its Interaction With PDI in Murine ALD
Immunofluorescence staining showed that periostin markedly colocalized with PDI in ALD mouse livers, alcohol-stimulated AML12 cells, and primary mouse hepatocytes (Figure 13A–C), even though the expression of PDI in mouse livers was not affected by ethanol exposure (Figure 13D, F). The colocalization and Co-IP results validated the interaction between periostin and PDI. To further ascertain whether periostin-mediated regulation of autophagy depends on PDI, we used shRNAs to knock down PDI in AML12 cells (Figure 13E, G). As shown in Figure 13H, PDI knockdown relieved the phosphorylation of 4E-BP1 and S6 caused by periostin and downregulated the expression of Atg7, Beclin-1, and LC3-II in AML12 cells after alcohol stimulation. Notably, lipid droplet accumulation was significantly increased in PDI-knockdown AML12 cells in response to alcohol treatment (Figure 13I–J). In addition, pharmacological intervention of AML12 cells and primary WT mouse hepatocytes with the PDI inhibitor CCF642 also resulted in consistent findings (Figure 13K–L).
Figure 13Periostin-mediated activation of autophagy depends on its interaction with PDI in murine ALD. (A) Immunofluorescence staining of periostin and PDI in the livers of mice with ALD. (B–C) Immunofluorescence staining of periostin and PDI in AML12 cells (B) and hepatic parenchymal cells isolated from WT mice (C) treated with 100 mM alcohol for 6 hours. (D–E) The relative mRNA levels of P4hb in WT and Postn-/- mouse livers after pair-fed or EtOH-fed feeding (n = 4–6 mice per group) (D) and in AML12 cells transfected with control or shP4hbs (n = 4) (E). (F–G) Western blot analyses of PDI in WT and Postn-/- mouse livers (F) and in AML12 cells with or without shP4hbs (G). (H) Western blot analysis of the indicated proteins in AML12 cells transfected with shCT, shP4hb1, and shP4hb2 with or without alcohol treatment. (I–J) The effects of alcohol on lipid accumulation (I) and lipid positive cells (J) in control or PDI knockdown AML12 cells. (K–L) Western blot analysis of the indicated proteins in primary WT mouse hepatocytes (K) and in AML12 cells (L) treated with alcohol and PDI inhibitor CCF642. Data represent the mean ± standard deviation (3 independently repeated experiments); *P < .05; **P < .01; ***P < .001; ****P < .0001; n.s., not significant (Student t test). Scale bars, 50 μm.
To further determine the functional correlation between periostin and PDI in ALD in vivo, we observed the development of ALD in Ad-Postn/Postn-/- mice with or without CCF642 treatment (Figure 14A). After CCF642 administration, the phosphorylated levels of 4E-BP1 and S6 were prominently increased, whereas Atg7, Beclin-1, and LC3-II were markedly reduced (Figure 14B). Although the body weight and food intake of mice were not significantly altered (Figure 14C–D), CCF642 administration increased the liver-to-body weight ratio and lipid droplet accumulation in the livers of ethanol-fed Ad-Postn/Postn-/- mice (Figure 14E, J). Moreover, hepatic and serum contents of TG, T-CHO, AST, and ALT, hepatic inflammation and fiber accumulation were all increased in ethanol-fed Ad-Postn/Postn-/- mice in response to CCF642 intervention (Figure 14F–K). These results suggest that periostin fails to protect against ALD after PDI is blocked. In conclusion, our data elucidate that periostin activation of autophagy depends on its interaction with PDI in murine ALD.
Figure 14PDI inhibitor CCF642 treatment exacerbates ALD in Ad-Postn/Postn-/-mice. (A) Schematic image of the Ad-Postn/Postn-/- mice administrated with PBS or PDI inhibitor CCF642 (n = 6–8 mice per group). (B) Western blot analysis of the indicated proteins in the livers of Ad-Postn/Postn-/- mice treated with PBS or CCF642 after EtOH feeding. (C–E) The effect of ethanol and CCF642 on body weight changes (C), food intake (D), and the ratios of liver-to-body weight (E) in Ad-Postn/Postn-/- mice (n = 6-8 mice per group). (F, H) Serum (F) and hepatic (H) levels of TG, T-CHO, AST, and ALT in Ad-Postn/Postn-/- mice treated with CCF642 or not (n = 6–8 mice per group). (G, I) The relative mRNA levels of inflammation (G) and fibrosis-related genes (I) in the livers of Ad-Postn/Postn-/- mice treated with CCF642 or not (n = 6–8 mice per group). (J–K) H&E staining, Oil red staining (J) and Sirius red staining, Masson staining (K) of the livers from Ad-Postn/Postn-/- mice treated with or without CCF642. Data represent the mean ± standard deviation (3 independently repeated experiments); *P < .05; **P < .01; ***P < .001; n.s., not significant (Student t test). Scale bars, 50 μm.
Thus, we examined how periostin is upregulated in ALD. Several studies have shown that TFEB plays a crucial role in the development of ALD. Overexpression of TFEB in the liver alleviates alcohol-induced liver injury in mice.
Therefore, we further explored whether TFEB is involved in regulating ethanol-induced periostin overexpression. Interestingly, the mRNA levels of Tefb, as well as its downstream molecules Atp6v1h, Ppargc1a, Lamp1, and Lamp2, were elevated in the livers of mice treated with alcohol (Figure 15A and D–G). Alcohol stimulation also resulted in the overexpression of TFEB and Lamp1 proteins (Figure 15B) and increased nuclear localization of TFEB in the liver (Figure 15C). Then, in vitro assays demonstrated that the mRNA and protein levels of periostin, TFEB, and its downstream factors were increased in AML12 cells treated with alcohol (Figure 15H–N). Moreover, the upregulated Postn mRNA levels were positively associated with the mRNA levels of Tefb in AML12 cells treated with ethanol (Figure 15O). We further found that TFEB knockdown not only thoroughly blocked ethanol-induced periostin overexpression (Figure 16A–C) but also markedly increased lipid droplet accumulation in AML12 cells in response to alcohol treatment (Figure 16D–E). Taken together, these results indicate that alcohol-stimulated periostin upregulation can be mediated by TFEB in hepatocytes.
Figure 15Alcohol-induced periostin overexpression is regulated by TFEB. (A–C) The expression of TFEB in WT and Postn-/- mouse livers (n = 4-6 mice per group) by qRT-PCR (A), Western blot (B), and immunofluorescence staining (C). (D–G) The relative mRNA levels of Tefb downstream molecules Atp6v1h, Ppargc1a, Lamp1, and Lamp2 in the livers of WT and Postn-/- mice after pair-fed or EtOH-fed feeding (n = 4-6 mice per group). (H–K) The relative mRNA levels of Tfeb, Postn, Atp6v1h and Ppargc1a in AML12 cells treated with the indicated concentrations for 6 hours (n = 3 per group). (L–M) The relative mRNA levels of Tfeb and Postn in AML12 cells treated with 100 mM alcohol at the indicated times (n = 3–4 per group). (N) Western blot analysis of periostin and TFEB expression in AML12 cells incubated with 100 mM alcohol at the indicated times. (O) Correlation of Postn mRNA with Tefb mRNA (n = 27). Data represent the mean ± standard deviation (3 independently repeated experiments); *P < .05; **P < .01; ***P < .001; ****P < .0001; n.s., not significant according to Student t test (A and D–G) or ordinary one-way analysis of variance multiple comparison test (H–M). Scale bars, 50 μm.
Figure 16Periostin overexpression is regulated by TFEB in response to alcohol exposure. (A–C) qRT-PCR (A and B) (n = 3 biological replicates) and Western blot (C) analyses of periostin and TFEB expression in AML12 cells transfected with shCT, shTfeb1, or shTfeb2. (D) Immunofluorescence staining of lipid and periostin in AML12 cells with or without shTfeb. (E) Statistical analysis of panel D. (F–G) Predicted binding sequences (F) and binding sites (G) of TFEB in the human periostin promoter. (H) The analysis of periostin promoter activity after co-transfection with TFEB vector and variable lengths of the human periostin promoter by Luciferase assays in the HER293T cell line (n = 3 per group). (I) The analysis of the activity of periostin FL, deletion 2 (D2) promoter construct, and its mutant promoter construct (n = 3 per group). Data represent the mean ± standard deviation (3 independently repeated experiments); *P < .05; **P < .01; ***P < .001; ****P < .0001; n.s., not significant (Student t test). Scale bars, 50 μm.
To further elucidate whether TFEB regulates periostin overexpression at the transcriptional level, we predicted the potential binding sites of the Postn promoter that can be regulated by TFEB through JASPAR analysis (a database of transcription factor binding profiles). According to the predicted results (Figure 16F), luciferase reporter constructs harboring −1643 bp to +757 bp (FL) of the mouse Postn promoter and a series of truncations (D1–D5) were successfully generated (Figure 16G). The dual-luciferase reporter assay demonstrated that TFEB induction markedly increased the relative luciferase activity of the periostin promoter (FL, D1, D2, and D3 but not the truncations D4 and D5 in HEK293T cells (Figure 16H). These results suggested that +20 bp to +757 bp of the Postn promoter might be the main binding region for TFEB. We then mutated the predicted site CTCACATGAT into CTTTACCCTA by a point mutation technique and found that the relative luciferase activity of the Postn promoter induced by TFEB was significantly suppressed (Figure 16I). Taken together, our results reveal TFEB as a novel transcription factor regulating periostin expression.
Discussion
Accumulating evidence indicates that periostin plays a vital role in various liver diseases, such as hepatic inflammation, hepatosteatosis, fibrosis, nonalcoholic steatohepatitis, and hepatocellular carcinoma.
Periostin deficiency reduces diethylnitrosamine-induced liver cancer in mice by decreasing hepatic stellate cell activation and cancer cell proliferation.
However, the function of periostin and its underlying molecular mechanisms in ALD remain unknown. Herein, we chose the chronic-plus-binge model, which induces hepatosteatosis, inflammation, and liver injury and fully mimics acute-on-chronic alcohol-related liver disease in patients,
and demonstrated that periostin was upregulated in ethanol-induced mouse livers. In our experimental models, serum periostin levels were significantly elevated in ethanol-fed mice, as shown by enzyme-linked immunosorbent assay. As a secreted matricellular protein, periostin is usually elevated in serum in response to various pathological conditions, including acute or chronic hepatitis, liver regeneration, and even hepatocellular carcinoma (HCC).
Periostin deficiency reduces diethylnitrosamine-induced liver cancer in mice by decreasing hepatic stellate cell activation and cancer cell proliferation.
Although serum periostin levels were elevated in many disease models, the potential application of serum periostin in the diagnosis, staging, and monitoring of diseases still needs to be assessed in a large clinical cohort in the future. To further determine the effect of periostin on ALD, Postn-/- mice were subjected to ethanol feeding. Unexpectedly, we found that periostin deficiency exacerbated murine ALD, whereas periostin recovery in Postn-/- mouse livers reversed the ALD phenotype, suggesting that periostin plays a protective role during the development of murine ALD. This result is contrary to previous findings by us and others about the function of periostin in nonalcoholic fatty liver disease,
which indicates that periostin performs different functions in ALD through distinct molecular mechanisms.
Autophagy, which is a self-degradative process, is vital for maintaining energy supply in response to nutrient stress and development. Autophagy is also involved in liver diseases and protects against ethanol-induced pathology in the liver.
In this study, our data showed that periostin activated autophagy by inhibiting the mTORC1 signaling pathway in ALD. Overexpression of periostin effectively inhibited the mTORC1 signaling pathway and activated autophagy, thus reducing ethanol-induced liver steatosis and injury in mice. Interestingly, blocking mTORC1 activity or activating autophagy with rapamycin mitigated the severe liver injury caused by periostin deletion. In contrast, mTORC1 activation and autophagy inhibition with MHY1485 eliminated the effect of periostin recovery on Postn-/- mice, ultimately aggravating murine ALD. Therefore, periostin protects against ALD by activating autophagy by inhibiting the mTORC1 signaling pathway.
To further reveal the molecular mechanism of periostin in the course of murine ALD, we used the BioID technique to screen the proteins that interacted with periostin. Although BioID is a perfect high-throughput technology for screening interacting proteins, it still has some potential pitfalls. For instance, the BioID method relies on biotin ligase BirA, which only biotinylates proteins within an enzyme-labeled radius of ∼10 nm.
In addition, the binding of biotin ligase BirA to the N-terminus or C-terminus of the target protein can capture the different interacting proteins of different domains. In this study, we identified 153 potential proteins that interacted with periostin through the BioID technique and further confirmed the protein PDI as a key interactor of periostin. PDI, which is known as collagen prolyl 4-hydroxylase subunit b (P4hb), is one of the most abundant enzymes in the ER and protects against ER stress-induced cell death.
Here, we elucidated that periostin-mediated inhibition of mTORC1 and autophagy activation depended on its interaction with PDI. Moreover, the key interaction domains between periostin and PDI were identified. Using shRNAs to knock down PDI or inhibit PDI with CCF642 reduced the inhibitory effect of periostin on mTORC1 and autophagy, thus worsening murine ALD phenotypes. Therefore, PDI, which is a novel protein that interacts with periostin, is required for periostin-mediated autophagy activation and mTORC1 inhibition in ALD.
Recent studies have demonstrated that various stimuli, such as interleukin (IL)-4, IL-13, and TGF-β, can induce periostin overexpression.
Our previous studies identified that CCL2 in acute B-cell lymphoblastic leukemia and IL-6 in colorectal cancer induce periostin overexpression by binding to their corresponding receptors CCR2 and IL-6R, respectively.
In this study, we attempted to identify the factors that triggered periostin upregulation in response to alcohol exposure. The lysosomal biogenesis regulator TFEB translocates from the lysosome into the nucleus, inducing the transcription of related genes.
Interestingly, TFEB expression and nuclear entry were markedly increased after ethanol stimulation. Our results demonstrated that TFEB is a novel regulator that induces the upregulation of periostin by binding to its promoter in ALD. Moreover, previous reports have shown that TFEB expression and entry into the nucleus can be regulated by the mTOR pathway.26 In this study, alcohol stimulation enhanced TFEB expression, while periostin expression and mTOR activity did not affect TFEB expression (Figure 15A–B), which indicates that the increase in TFEB expression is regulated by other mechanisms.
In conclusion, we revealed that ethanol exposure promoted the upregulation of periostin by activating TFEB in hepatocytes. Upregulated periostin alleviated the development of ALD by activating autophagy by inhibiting mTORC1 in a PDI-dependent manner. In summary, this study provides novel insights into the biological function and mechanism of periostin in ALD, and periostin is a promising therapeutic target for ALD.
Materials and Methods
Animal Experiments
Heterozygous B6;129-Postntm1/Jmol/J (Postn+/-) mice were obtained from the Jackson Laboratory (Bar Harbor, ME) and then crossed with C57BL/6J WT mice for more than 10 generations to obtain C57BL/6J Postn+/- mice. Postn+/- mice were further crossed to generate WT and Postn-/-mice as described in our previous study.
Periostin deficiency reduces diethylnitrosamine-induced liver cancer in mice by decreasing hepatic stellate cell activation and cancer cell proliferation.
Genotyping of the WT, Postn-/-, and Postn+/- mice were validated according to the manufacturer’s instructions. As described in the previous study, 10-week-old male mice received chronic-plus-binge ethanol feeding.
The details are shown in Figure 1A; the male mice were fed with a Lieber-DeCarli control diet (pair-fed) (Trophic Animal Feed High-Tech, Haian, Jiangsu, China) for 5 days, then changed into a Lieber-DeCarli ethanol diet (EtOH-fed) (5% ethanol, Trophic Animal Feed High-Tech, Haian, Jiangsu, China) for 10 days and given a single gavage of 5 g/kg ethanol on the 16th day. The control group was fed with a control diet for 15 days and then given a single gavage of 9 g/kg dextrin on the 16th day. Rapamycin was intraperitoneally injected at a dose of 1 mg/kg/day from the first day of ethanol feeding.
DEP domain-containing mTOR-interacting protein suppresses lipogenesis and ameliorates hepatic steatosis and acute-on-chronic liver injury in alcoholic liver disease.
The mTOR activator MHY1485 was dissolved in dimethyl sulfoxide and intraperitoneally administrated at a dose of 2 mg/kg/day following the first day of ethanol feeding. The PDI-inhibited murine model was obtained by intraperitoneal injection with CCF642 at a dose of 10 mg/kg/2-days on the second day of ethanol feeding. All mice were housed under specific pathogen-free conditions at the Xiamen University Laboratory Animal Center, and all animal experiments were carried out in strict accordance with the animal protocols as defined by the Institutional Animal Care and Use Committee of Xiamen University.
Cell Lines and Primary Cell Culture
The HER293T cell line was purchased from the American Type Culture Collection and maintained in Dulbecco’s Modified Eagle Medium (DMEM) with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin (Gibco, Grand Island, NY). AML12 cells were presented by Professor Yunqing Yang (Xiamen University School of Life Science) and cultured in DMEM-F12 containing 10% FBS and 1% penicillin-streptomycin (Gibco). Primary hepatocytes were isolated by differential centrifugation on Percoll (Sigma-Aldrich) solutions from male mice (8–10 weeks) according to the previous method.
In detail, the portal vein was cannulated and perfused with perfusion buffer (0.5 mM EDTA, 25 mM HEPES in Hank’s Balanced Salt Solution [HBSS] buffer without calcium, magnesium, and phenol red), followed by the digestion buffer (25 mM HEPES, 0.5 mg/mL collagenase type IV in HBSS buffer with calcium, magnesium, and phenol red) at 37 °C for 20 minutes. The cell suspension was filtered and then centrifuged at 50 g for 2 minutes at room temperature to isolate hepatocytes. Further, 40% Percoll solutions were used to purify isolated hepatocytes, and the suspension was centrifuged at 200 g to obtain primary hepatocytes. All cells were cultured in William’s medium E (Gibco) with 10% FBS (Gibco), 1% penicillin-streptomycin, and L-Glutamine (2 μM; Gibco).
Recombinant Adenoviruses
Adenoviruses expressing murine Postn (Ad-Postn) or GFP (Ad-GFP) were constructed using the vector pAdEasy-1 with a full-length Postn or GFP complementary DNA (cDNA) coding sequence. The recombinant adenoviruses were presented by Professor Xiaoying Li (Zhongshan Hospital of Fudan University, Shanghai, China). As previously described, adenoviruses were injected into mice via tail vein on the third day of control food feeding to rescue periostin or GFP (1 × 109 plaque-forming units, 150 μL) in Postn-/- mouse livers.
Ocular blood was collected, clotted, and centrifuged at 3000 rpm for 10 minutes, then serum was collected. The serum AST, ALT, TG, and T-CHO were tested by commercial reagent kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). The AST, ALT, TG, and T-CHO of livers were measured from 100 mg of liver tissues dissolving in phosphate buffered saline (PBS) (1:9 ratio).
BioID and LC-MS/MS Analysis
The BioID analysis was performed according to the previous protocol.
In brief, the cells transfected with periostin-BirA∗-HA or pBirA∗-HA vectors were cultured in a medium containing 50 μM biotin (Sigma) for 18 hours, then the incubated cells were harvested, washed, and lysed. Next, the supernatant (100 μL) was applied to analyze periostin-BirA∗ or BirA∗ expression by Western blot. The biotinylated proteins were detected by anti-HA (3724S; CST) and HRP-streptavidin (3999S; CST) antibodies. Meanwhile, the remaining supernatant was further incubated with 80 μl NeutrAvidin agarose beads (Thermo Scientific) overnight at 4 °C. The beads were collected, washed, resuspended, and boiled. The eluted bound proteins were further analyzed by Western blot and LC-MS/MS assay. Before the MS analysis, the proteins were digested and desalted.
Then, the LC-MS/MS analysis was performed by the Core Facility of Biomedical Sciences of Xiamen University.
Alcohol, Activator, and Inhibitor Treatment
Primary mouse hepatocytes and AML12 cells were cultured in FBS-free medium for 12 hours. Rapamycin (100 nM) (HY-10219; MedChemExpress (MCE), Monmouth Junction, NJ), MHY1485 (10 μM) (B5853; Achieve Perfection, APExBIO, Houston, TX), and CCF642 (5 μM) (s8281; Selleck Chemicals, Houston, TX) were incubated for 6 hours with alcohol treatment (100 mM).
Histological Analysis
For hematoxylin and eosin (H&E) staining, Sirius red staining, and Masson staining, liver tissue sections were prepared. The methods of staining were performed as previously described.
In brief, for H&E staining, the sections were dewaxed, rehydrated, and treated with hematoxylin for 20 seconds and eosin for 15 seconds. For Sirius red staining, the deparaffinized sections were dyed using picrosirius red solution. Masson staining was performed according to the manufacturer's instructions (G1006-100ML, Wuhan Servicebio Technology, Wuhan, Hubei, China). Images were obtained using a Leica DM4B microscope.
Quantitative Polymerase Chain Reaction
Total RNA in livers or cells was extracted with TRIzol reagent (Thermo Fisher Scientific) and then reversed into cDNA using a ReverTra Ace qPCR RT kit (FSQ-201, TOYOBO, Japan). Quantitative real-time polymerase chain reaction (qRT-PCR) was performed by TransStart Top Green qPCR SuperMix (Transgene, China). The mRNA relative levels of the indicated genes were normalized to Gapdh expression. The primers used in this work are listed in Table 2.
Table 2Primer Sequences for RT-PCR Used in This Study
For natural lipid staining, AML12 cells were seeded on cover glasses and were fixed with 4% paraformaldehyde after alcohol treatment, then incubated with 1 ng/mL Bodipy 493/503 (D3922; Invitrogen, Carlsbad, CA) at room temperature for 20 minutes. The cells were stained with DAPI before mounting and imaging on Zeiss LSM 780 confocal microscope.
Plasmids and Transfection
We purchased pGL3-Luc and Renilla from Promega (Madison, WI). Professor Qingsong Lin (National University of Singapore) gifted the plasmid MCS-BIOID2-HA. Professor Yunqing Yang (Xiamen University) presented pLV-tagII. pCMV5 and pCDH-CMV-EF1-GFP are stored in our laboratory. Fragments encoding periostin, TFEB, and PDI were cloned from mouse liver cDNA via PCR, then inserted into pCDH-CMV-EF1-GFP or MCS-BIOID2-HA, pCMV5, and pLV-tagII, respectively. The primer pairs are listed in Table 3. For stable knockdown of TFEB and PDI in AML12 cells, a lentivirus-mediated packaging system was used. The shRNAs sequences targeting TFEB and PDI were listed below: shTfeb-1, 5'-CCAAGAAGGATCTGGACTTAA-3'; shTfeb-2, 5'-GCAGGCTGTCATGCATTATAT-3'; shP4hb-1, 5'- GCATTTCATCTGTGAGGCATT-3'; shP4hb-2, 5'- CCCAAGAGTGTATCTGACTAT-3'.
HER293T cells were transfected with pCMV5-TFEB, pGL3-Postn-Luc reporter constructs (Promega), and Renilla luciferase vector (Promega). Luciferase activity was determined according to the manufacturer’s instructions (Promega). Renilla activity was used as the control transfection efficiency. Three biological replicates were performed for each experiment.
Co-IP Assay
The interacting relationship between periostin and PDI was validated by semi-internal and external IP. For semi-internal IP, AML12 cells were transfected with periostin-BirA∗-HA vectors. After 36 hours of transfection, cells were washed, collected, and then lysed in 600 μL lysis buffer (150 mM NaCl, 50 mM Tris, pH 7.4, 1 % Triton X-100, 1 mM PMSF [Sangon Biotech, Shanghai, China] and protease inhibitor cocktail [Thermo Scientific]) for 30 minutes at 4 °C. The samples were centrifuged, and the collected supernatant was pre-incubated with protein A/G agarose beads (Thermo Scientific) for about 1 hour at 4 °C. Aliquots of the supernatant (36 μL) were left as input samples, and the remaining supernatant was further equally divided into 2 parts incubated with 5 μg PDI or IgG antibody for 6 hours, and then incubated with beads overnight at 4 °C. At the end, beads were washed and boiled with 40 μl loading buffer for 10 minutes. The bound proteins were further performed to Western blot analysis. For external IP, the periostin-BirA∗-HA and pLV-PDI-flag plasmid were transfected into HER293T cells.
Bioinformatics Analysis
GO annotation analysis and KEGG pathway enrichment analysis were performed by DAVID 6.8 (https://david.ncifcrf.gov/).
Statistical Analysis
Quantitative data (mean ± standard deviation) were subjected to the Student t test or ordinary one-way analysis of variance multiple comparison test analysis, using GraphPad Prism 8 software (GraphPad Software, San Diego, CA).
Periostin deficiency reduces diethylnitrosamine-induced liver cancer in mice by decreasing hepatic stellate cell activation and cancer cell proliferation.
DEP domain-containing mTOR-interacting protein suppresses lipogenesis and ameliorates hepatic steatosis and acute-on-chronic liver injury in alcoholic liver disease.
Conflicts of interest The authors disclose no conflicts.
Funding This research was supported by the National Natural Science Foundation of China (82170600, 82070607, 81871679, 82172932, and 81972748) and the Natural Science Foundation of Fujian Province of China (2022J01015, 2019J02002, and 2021J05010).
00025-5&id=fx1.jpg){kind=link}
00025-5&id=gr1.jpg){kind=link}
00025-5&id=gr2.jpg){kind=link}
00025-5&id=gr3.jpg){kind=link}
00025-5&id=gr4.jpg){kind=link}
00025-5&id=gr5.jpg){kind=link}
00025-5&id=gr6.jpg){kind=link}
00025-5&id=gr7.jpg){kind=link}
00025-5&id=gr8.jpg){kind=link}
00025-5&id=gr9.jpg){kind=link}
00025-5&id=gr10.jpg){kind=link}
00025-5&id=gr11.jpg){kind=link}
00025-5&id=gr12.jpg){kind=link}
00025-5&id=gr13.jpg){kind=link}
00025-5&id=gr14.jpg){kind=link}
00025-5&id=gr15.jpg){kind=link}
00025-5&id=gr16.jpg){kind=link}