Disruption of hepatic small heterodimer partner induces dissociation of steatosis and inflammation in experimental nonalcoholic steatohepatitis

Nonalcoholic steatohepatitis (NASH) is a leading cause of chronic liver disease worldwide and is characterized by steatosis, inflammation, and fibrosis. The molecular mechanisms underlying NASH development remain obscure. The nuclear receptor small heterodimer partner (Shp) plays a complex role in lipid metabolism and inflammation. Here, we sought to determine SHP's role in regulating steatosis and inflammation in NASH. Shp deletion in murine hepatocytes (ShpHep−/−) resulted in massive infiltration of macrophages and CD4+ T cells in the liver. ShpHep−/− mice developed reduced steatosis, but surprisingly increased hepatic inflammation and fibrosis after being fed a high-fat, -cholesterol, and -fructose (HFCF) diet. RNA-Seq analysis revealed that pathways involved in inflammation and fibrosis are significantly activated in the liver of ShpHep−/− mice fed a chow diet. After having been fed the HFCF diet, WT mice displayed up-regulated peroxisome proliferator-activated receptor γ (Pparg) signaling in the liver; however, this response was completely abolished in the ShpHep−/− mice. In contrast, livers of ShpHep−/− mice had consistent NF-κB activation. To further characterize the role of Shp specifically in the transition of steatosis to NASH, mice were fed the HFCF diet for 4 weeks, followed by Shp deletion. Surprisingly, Shp deletion after steatosis development exacerbated hepatic inflammation and fibrosis without affecting liver steatosis. Together, our results indicate that, depending on NASH stage, hepatic Shp plays an opposing role in steatosis and inflammation. Mechanistically, Shp deletion in hepatocytes activated NF-κB and impaired Pparg activation, leading to the dissociation of steatosis, inflammation, and fibrosis in NASH development.

Nonalcoholic fatty liver disease (NAFLD) 2 is rapidly becoming a major health concern worldwide due to the epidemic of obesity and diabetes (1). Whereas simple steatosis is generally considered a relatively benign condition, ϳ20 -30% of these patients will develop nonalcoholic steatohepatitis (NASH), a more serious condition characterized by steatosis with hepatocyte injury, liver inflammation, and fibrosis, which can further progress to irreversible cirrhosis and hepatocellular carcinoma with no effective therapies (2)(3)(4). Despite the increasing prevalence of NASH and its burden to the healthcare system, the molecular mechanisms underlying NASH development still remain obscure, leading to a lack of mechanism-based targeted treatment options for NASH. Thus, understanding the molecular machinery that causes NASH will provide new insights into developing effective prevention and therapeutic strategies for this unresolved disease.
Small heterodimer partner (Nr0b2, Homo sapiens SHP; Mus musculus Shp) is an atypical nuclear receptor due to its lack of a DNA-binding domain (5). As an orphan nuclear receptor without a known ligand, SHP represses bile acid synthesis (6, 7), controls energy metabolism (8), and regulates glucose homeostasis (9, 10) via direct interactions with numerous nuclear receptors and transcription factors (11). In support of its important role in metabolic diseases, mutations in the Shp gene have been correlated with obesity and susceptibility to type 2 diabetes (12).
SHP is involved in lipid metabolism. Earlier studies demonstrate that hepatic SHP overexpression increases liver lipid levels (13). Consistently, Shp-deficient mice are protected from high-fat diet-induced hepatic steatosis (14,15). These studies indicate that antagonizing SHP may be beneficial to treat liver steatosis. In contrast, a recent study reported that Shp was not suppressed in the steatotic livers but was decreased in NASH livers, such as tetracycline-treated livers, methionine cholinedeficient diet-induced NASH livers, and glycine N-methyl-transferase-deficient (Gnmt Ϫ/Ϫ ) mouse livers (16). Similarly, our recent study also found that SHP was markedly decreased in the livers of patients with NASH and in diet-induced mouse NASH (17). Additionally, hepatic SHP overexpression after steatosis development dramatically prevented steatosis progression to NASH by attenuating liver inflammation and fibrosis (17). These results indicate that the role of SHP in NAFLD development is complicated and could be multilayered.
A diet containing high fat, cholesterol, and fructose (HFCF; 40 kcal% fat, 2% cholesterol, 20 kcal% fructose) has been utilized to induce NASH in mice (18,19). In this diet, excess fat induces the development of steatosis, whereas the addition of fructose and cholesterol increases hepatic damage, predisposing animals to inflammation and fibrogenesis (20). Therefore, this dietary model replicates many of the biochemical and histopathological hallmarks of clinical NAFLD progression from steatosis to NASH (17). In the current study, we fed mice an HFCF diet to dissect the role of SHP in steatosis and inflammation during NASH development.
Hepatic macrophage activation is a hallmark of NASH and is correlated with disease progression by producing proinflammatory and profibrogenic mediators (21). Although the mechanism underlying macrophage activation in NASH is incompletely understood, numerous studies demonstrate that lipotoxicity-induced hepatocyte damage triggers the release of damage-associated molecular pattern molecules (DAMPs) into the microenvironment, which stimulates macrophage activation (22,23). Therefore, steatosis is generally positively correlated with liver inflammation and liver injury. Strikingly, in the current study, we revealed that hepatic Shp deletion ameliorated liver steatosis but aggravated macrophage activation in HFCF-fed mice. Mechanistically, we demonstrated that Shp disruption in hepatocytes activated NF-B signaling and impaired peroxisome proliferator activated receptor ␥ (Pparg) activation, leading to the dissociation between steatosis and inflammation in NAFLD. Interestingly, Shp deletion after steatosis development accelerated hepatic inflammation and fibrosis without affecting liver steatosis. These findings indicate that, depending on NAFLD stage, SHP could play distinct roles in modulating steatosis, inflammation, and fibrosis.

Hepatic Shp disruption in adult mice induces liver injury, inflammation, and fibrogenesis
To generate hepatocyte-specific Shp knockout (Shp HepϪ/Ϫ ) mice and WT controls, we injected 2-month-old Shp flox/flox mice with either AAV8-Tbg-Cre or AAV8-Tbg-null control for 1 or 12 weeks (Fig. 1A). The use of the Tbg promoter ensures that AAV8 specifically targets hepatocytes (24). Hepatic Shp deletion specifically in adult mice achieved a model in which a direct assessment of hepatic Shp function in the adult livers could be made. This is an advantage over using mice with a germ line mutation in Shp, which could interfere with proper embryonic or neonatal development and alter adult hepatic function. This acute deletion approach also made it possible to distinguish the direct effects of Shp deletion from any secondary or compensatory effects.
The knockdown of Shp mRNA in the liver after Cre injection was confirmed by RT-PCR (Fig. 1A). The body weight gain, lean mass, and fat mass were comparable between Shp HepϪ/Ϫ mice and WT controls after 1 week or 12 weeks of AAV8 injection (Fig. 1, B and C). To investigate whether ablation of hepatic Shp could affect glucose metabolism, we performed a glucose tolerance test (GTT). As shown in Fig. 1D, no difference in GTT was observed between Shp HepϪ/Ϫ mice and WT controls, which was further confirmed by the quantification of areas under the curves (AUC) of GTT (Fig. 1D, right). Additionally, the fasting serum triglycerides, cholesterol, and glucose were similar between Shp HepϪ/Ϫ mice and WT controls; however, the liver injury marker alanine aminotransferase (ALT) was significantly increased after 12 weeks of Shp deletion (Fig. 1E).
We next examined the liver. Both liver weight and liver/body weight ratios were comparable between Shp HepϪ/Ϫ mice and WT controls after AAV8 injection (Fig. 1F). No obvious histological change was observed 1 week after Shp deletion; however, after 12 weeks, a massive infiltration of inflammatory cells occurred in the livers of Shp HepϪ/Ϫ mice (Fig. 1G). Meanwhile, an increase in cell death was observed in Shp HepϪ/Ϫ liver that was detected by terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) staining (Fig. 1H), supporting the elevated serum ALT shown in Fig. 1E. Taken together, the above data indicate that hepatic Shp deletion induces liver injury.

Hepatic Shp deletion reduces diet-induced liver steatosis but exacerbates liver inflammation and fibrosis
Our previous study has described a mouse NASH model induced by HFCF diet (17). As shown in Fig. 3A, the HFCF diet feeding for 12 weeks did not alter the expression of hepatic Shp and its target gene cytochrome P450 family 7 subfamily A member 1 (Cyp7a1). We then superimposed hepatic Shp deletion by AAV8-Tbg-Cre on an HFCF feeding regimen (Fig. 3B). Both Shp HepϪ/Ϫ mice and WT mice developed glucose intolerance after HFCF diet for 12 weeks (Fig. 3C). Meanwhile, body weight gain and fasting serum triglycerides were similar in Shp HepϪ/Ϫ and WT mice after 12 weeks on the HFCF diet (Fig. 3, D and E). However, a decrease in fasting serum cholesterol level was observed in HFCF-fed Shp HepϪ/Ϫ (Fig. 3E). In contrast, serum

Dissociation of steatosis from inflammation by Shp deletion
ALT was significantly increased in HFCF-fed Shp HepϪ/Ϫ mice compared with HFCF-fed WT mice (Fig. 3E).
Whole-body Shp Ϫ/Ϫ mice and hepatocyte-specific Shp Ϫ/Ϫ mice are resistant to high-fat diet-induced liver steatosis (14,15). Consistently, a decrease in liver weight and liver/body weight ratio was observed in HFCF-fed Shp HepϪ/Ϫ mice compared with HFCF-fed WT controls (Fig. 3F). Additionally, both Oil Red O staining and liver lipid extraction quantification revealed reduced liver steatosis, triglycerides, and cholesterols in HFCF-fed Shp HepϪ/Ϫ mice (Fig. 3, G and H), supporting the overall resistance to diet-induced liver steatosis triggered by hepatic Shp deletion.
Shp is a critical suppressor for bile acid synthesis through the direct inhibition of Cyp7a1 (6, 7), a critical rate-limiting enzyme that converts cholesterol to 7␣-hydroxycholesterol for bile acid synthesis (26). Knockout of Shp increases bile acid synthesis and decreases hepatic accumulation of cholesterol (27). Consistently, HFCF-fed Shp HepϪ/Ϫ mice displayed increased bile acid pool size and fecal bile acid excretion rate (Fig. 3I) in parallel with the reduction of hepatic and serum cholesterol levels compared with HFCF-fed WT controls ( Fig. 3, E and H). The data indicate that hepatic Shp deficiency increases cholesterol conversion for bile acid synthesis and decreases hepatic and serum cholesterol levels subsequently.
In contrast to the attenuation of liver steatosis, Shp deletion surprisingly exacerbated liver inflammation and fibrosis during NASH development, evidenced by an increase in macrophage infiltration as well as collagen formation in HFCF-fed Shp HepϪ/Ϫ liver compared with HFCF-fed WT controls (Fig. 4,  A and B). At mRNA levels, an increase in expression of genes involved in inflammation (Il6, Tnf␣, Ccl2, Il1, Ifng, and Nos2) and fibrosis (Col1a1, Col1a2, Col5a1, Mmp12, and Mmp13) was observed in HFCF-fed Shp HepϪ/Ϫ liver (Fig. 4C). Taken together, these data indicate that hepatocyte-specific Shp depletion in adult mice attenuates HFCF diet-induced steatosis but exacerbates liver inflammation and fibrosis.

RNA-Seq reveals activation of inflammatory processes and pathways by hepatic Shp deletion
To pursue an unbiased investigation of mechanisms by which hepatocyte-specific Shp deletion induces the disconnec- tion between steatosis and inflammation in NASH, we conducted RNA-Seq analysis of liver obtained from WT and Shp HepϪ/Ϫ mice after 12 weeks on the chow and HFCF diet (GEO number: GSE133566). The FPKM was calculated to quantify the expression of genes in four groups, including WT chow, Shp HepϪ/Ϫ chow, WT HFCF, and Shp HepϪ/Ϫ HFCF. The differentially expressed genes (DEGs) were determined using a -fold change cutoff of 1. 50 (Fig. 5A).
A heat map with hierarchical clustering was used to determine the expression patterns of DEGs in WT and Shp HepϪ/Ϫ mice fed a chow or HFCF diet. The idea is that genes with similar expression patterns may have similar functions or may be involved in the same biological process or cellular pathway; therefore, they are clustered into classes. In this study, we found that Shp HepϪ/Ϫ mice were distinctly separated from WT on both chow and HFCF conditions (Fig. 5B). The DEGs in four groups were clustered into two major types, one with higher expression in the WT groups but lower expression in the Shp HepϪ/Ϫ groups and the other with lower expression in the WT groups but higher expression in the Shp HepϪ/Ϫ groups. Original magnification, ϫ40. Right, liver collagen content was determined by a hydroxyproline assay. C, the relative mRNA levels of genes related to inflammation and fibrosis in liver tissues were determined by qPCR. D, left, Western blot analysis of cytosolic (Cyto) and nuclear (Nuc) proteins (Pt) in the liver. Right, the band intensities of p65 were normalized by loading controls. The expression of p65 in Shp HepϪ/Ϫ mice relative to that in WT controls is plotted. Data are represented as mean Ϯ S.E. (error bars); n ϭ 5 mice/group. *, p Ͻ 0.05 Shp HepϪ/Ϫ versus WT.

Dissociation of steatosis from inflammation by Shp deletion
We next conducted gene ontology (GO) enrichment and KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway analysis of DEGs using Enrichr Tools. The DEGs in each group were assigned to GO terms describing biological processes, cellular components, and molecular functions. The enriched GO terms and KEGG pathways with adjusted p values Ͻ 0.05 were selected and summarized in the supporting information. As shown in Fig. 6A, the GO enrichment analysis indicated that a major effect of genes involved in proinflammatory and fibrogenic processes, such as extracellular matrix organization, cytokine, type I interferon, neutrophil degranulation, and cellular response to interferon ␥, were induced by Shp deletion under a chow diet. Furthermore, KEGG pathway analysis of DEGs demonstrated that multiple pathways were differentially altered in Shp HepϪ/Ϫ chow compared with WT chow, with the cytokinecytokine receptor interaction, hematopoietic cell lineage, chemokine signaling, T cell receptor signaling, and NF-B signaling as the top five enriched KEGG pathways based on statistical significance (Fig. 6A).
The top five biological processes and pathways altered by HFCF diet in WT or Shp HepϪ/Ϫ mice are displayed in Fig. 6 (B and C, respectively). Among them are some common metabolic processes and pathways, such as steroid and cholesterol biosynthesis, regulation of triglyceride, and PPAR signaling.

Activation of NF-B and suppression of Pparg signaling in HFCF-fed Shp Hep؊/؊ mice
To further characterize HFCF diet-induced biological pathways and disease processes that are directly influenced by Shp deletion, we compared gene expression between WT HFCF and Shp HepϪ/Ϫ HFCF (Fig. 6D). As expected, biological processes and pathways involved in inflammation (type I interferon, cytokine, exocytosis, chemokine, T cell receptor, NF-B signaling) were distinctly affected by hepatic disruption of Shp. In addition, HFCF-induced PPAR signaling, the key pathway involved in lipid metabolism, was largely impacted by the absence of Shp (Fig. 6D).
To clarify the causative role of NF-B signaling in Shp deletion-induced inflammation, we isolated hepatocytes from Shp flox/flox mice and deleted Shp (Shp Ϫ/Ϫ ) using adenovirus expression Cre recombinase. The infection of adenovirus vector control was used to generate WT hepatocytes. The realtime PCR confirmed the successful knockdown of Shp in hepatocytes (Fig. 7B, left). As expected, loss of Shp in hepatocytes significantly increased the expression of Tnf␣, an established NF-B target (Fig. 7B, right). Most importantly, the induction of Tnf␣ was markedly inhibited by the treatment of a specific NF-B inhibitor, BAY 11-7082 (Fig. 7B, right). The above data indicate that NF-B activation plays a critical role in Shp deletion-induced initiation of inflammation.
SHP has been shown to interact and alter the function of many nuclear receptors in lipid metabolism, including Pparg, Ppar␣, Lrh1, Hnf4␣, Fxr, Lxr␣, and Lxr␤ (11). Strikingly, hepatic Shp disruption specifically decreased Pparg expression in both chow and HFCF conditions (Fig. 8C). In contrast, hepatic Shp deletion abolished HFCF-induced up-regulation of Ppar␣, but did not impact Ppar␣ expression in the chow-fed mice. Moreover, neither HFCF diet nor Shp deletion altered hepatic expression of Lrh1, Hnf4␣, Fxr, Lxr␣, and Lxr␤ (Fig.  8C). PPARg is expressed at a low level in normal liver but markedly increased in fatty liver (28). PPARg overexpression up-regulates various lipogenic genes and promotes liver steatosis, whereas hepatocyte Pparg deletion ameliorates liver steatosis (29,30). Consistently, here we show that hepatic Shp deletion results in Pparg down-regulation that is correlated with the decrease of liver steatosis.

Dissociation of steatosis from inflammation by Shp deletion
We next examined Pparg and its target gene Cd36 in WT and Shp Ϫ/Ϫ hepatocytes. As shown in Fig. 8D, the mRNA levels of Pparg and Cd36 were significantly decreased in Shp Ϫ/Ϫ hepatocytes compared with WT hepatocytes. We then treated WT and Shp Ϫ/Ϫ hepatocytes with oleic acid (0.5 mM) for 24 h to assess steatosis. Oleic acid dramatically increased steatosis in WT hepatocytes. Strikingly, the lipid accumulation was largely inhibited in Shp Ϫ/Ϫ hepatocytes (Fig. 8E). Interestingly, overexpressing Pparg by adenovirus greatly increased lipid contents in oleic acid-treated WT hepatocytes, and this response was maintained in Shp Ϫ/Ϫ hepatocytes overexpressed with Pparg (Fig. 8E). Taken together, the above data clearly indicate that Pparg plays a critical role in Shp deficiency-induced resistance to steatosis. The overexpression of Pparg in WT and Shp Ϫ/Ϫ hepatocytes was confirmed by real-time PCR shown in Fig. 8F. Collectively, the above data suggest that Shp disruption in hepatocytes activated NF-B signaling and impaired Pparg activation, leading to the progression of liver inflammation and fibrosis with an attenuation of steatosis.

Ablation of hepatic Shp after the development of steatosis enhances liver inflammation and fibrosis in HFCF-fed mice without affecting liver steatosis
In our previous study, we have described that mice fed an HFCF diet for 4 weeks develop liver steatosis without appreciable inflammation or fibrosis (17). We also noticed that 4 weeks of HFCF feeding does not alter hepatic Shp expression (17). To investigate the role of Shp specifically in the transition of steatosis to NASH, Shp flox/flox mice were fed an HFCF diet for 4 weeks followed by Shp deletion in hepatocytes by injection of AAV8-Tbg-Cre and remained on the HFCF diet for an additional 8 weeks (Fig. 9A). The knockdown of hepatic Shp was confirmed by real-time PCR (Fig. 9A). No differences in body weight, serum fasting triglycerides, cholesterol, and glucose

Dissociation of steatosis from inflammation by Shp deletion
were noticed between Shp HepϪ/Ϫ and WT controls (Fig. 9, B and C). However, serum ALT was consistently increased in Shp HepϪ/Ϫ mice (Fig. 9C). Interestingly, hepatic Shp disruption after steatosis development did not alter liver weight and liver/body weight ratio compared with WT controls (Fig. 9D).
Both hematoxylin and eosin (H&E) staining and lipid extraction revealed the similar extent of liver steatosis developed in

Dissociation of steatosis from inflammation by Shp deletion
Shp HepϪ/Ϫ mice and WT controls in this steatosis-to-NASH transition model (Fig. 9E). However, the livers of Shp HepϪ/Ϫ mice displayed a marked increase in F4/80 ϩ macrophages and collagen formation (Fig. 9, F and G). Consistently, the expression of genes involved in inflammation (Ccl2, Tnf␣, and Nos2) and fibrosis (Col1a1 and Col1a2) were significantly increased in the livers of HFCF-fed Shp HepϪ/Ϫ mice (Fig. 9H). Next, we examined hepatic Pparg expression. Surprisingly, although there were similar levels of liver steatosis in HFCF-fed Shp HepϪ/Ϫ mice and WT controls, the expression of hepatic Pparg was consistently lower in Shp HepϪ/Ϫ mice (Fig. 9H), supporting a strong regulation of Shp on Pparg's expression that is independent of liver steatosis. Meanwhile, our data also indicate that after steatosis development, disruption of hepatic Shp decreases Pparg expression but could not alter liver steatosis; instead, other Pparg-independent mechanisms may contribute significantly to liver steatosis. Taken together, the above data support Shp as a key player that prevents NASH progression by inhibiting both liver inflammation and fibrosis.

Discussion
Steatosis alone is considered to be little to no risk for progressive liver disease, whereas NASH can progress into irreversible cirrhosis and hepatocellular carcinoma without effective treatments. Given the clinical significance of NASH, identifying key factors and pathways that promote the progression of steatosis to NASH is critically important for developing effective prevention and therapeutic strategies. Using a clinically relevant dietary mouse NASH model, we investigated the role of SHP in NASH development. We demonstrate that disruption of

Dissociation of steatosis from inflammation by Shp deletion
hepatic Shp activates NF-B signaling and impairs Pparg activation, leading to the dissociation of steatosis, inflammation, and fibrosis during NAFLD development. Interestingly, ablation of hepatic Shp after the development of steatosis exacerbates liver inflammation and fibrosis without affecting steatosis (Fig. 9I). Our finding that hepatic Shp plays a critical role in the steatosis-to-NASH transition provides some new mechanistic insights into our understanding of NASH pathogenesis and progression and may present a new target for NASH treatment.
An important finding from our current study is that Shp disruption in hepatocytes induces liver inflammation, which is supported by several pieces of evidence. First, immunohistochemistry staining confirmed an increase in CD4 ϩ T cells, B cells, macrophages, and neutrophils in the liver after 12 weeks of Shp deletion. Second, deletion of Shp led to a robust induction of genes involved in liver inflammation, such as Ly6d, Il6, Il1, Tnf␣, Ccl2, Ifn␥, and Nos2. Most importantly, RNA-Seq revealed that absence of Shp significantly altered biological pro-cesses and pathways involved in inflammation. Finally, loss of Shp induced NF-B signaling activation in hepatocytes that can be blocked by a specific NF-B inhibitor.
SHP plays an important anti-inflammatory role in monocytes by negatively regulating inflammation induced by Tolllike receptor 4 (TLR4) or NLRP3 inflammasome (31,32). Here, we found that in hepatocytes, SHP also plays an anti-inflammatory role. Our finding is important, as the hepatocyte is gradually recognized as a key cell type involved in innate immunity by secreting innate immunity proteins, such as bactericidal proteins, opsonins, iron-sequestering proteins, coagulation factor fibrinogen, and cytokines (33). The dysregulation of these innate immunity proteins contributes significantly to the pathogenesis and progression of chronic liver diseases including NASH. For instance, palmitic acid induces production of proinflammatory cytokine interleukin-8 from hepatocytes, contributing to hepatic inflammation and liver injury subsequently (34). Lipotoxicity induces the release of chemokine Depending on NAFLD stage, hepatic Shp plays an opposing role in regulating steatosis, inflammation, and fibrosis. Shp disruption in hepatocytes activates NF-B signaling and impairs Pparg activation, leading to the dissociation of steatosis, inflammation, and fibrosis in NAFLD development. Shp deletion after steatosis development exacerbates hepatic inflammation and fibrosis without affecting liver steatosis.

Dissociation of steatosis from inflammation by Shp deletion
(CXC motif) ligand 10 (CXCL10)-bearing vesicles from hepatocytes, resulting in macrophage recruitment into the liver (35). Moreover, in a recent study, hepatocytes were found to secret CXCL1/interleukin-8, which causes neutrophil infiltration and alcoholic liver injury (36).
One well-known powerful pro-inflammatory chemokine is CCL2, or monocyte chemoattractant protein 1 (MCP1), which is responsible for attracting monocytes/macrophages and T cells during liver injury. Increase in CCL2 secretion from hepatocytes has been shown in both alcohol fatty liver disease (37) and obesity-associated NAFLD (38), which exacerbates liver injury and inflammation. Similarly, our recent study demonstrated that disruption of Shp in hepatocytes induces NF-B signaling activation, resulting in CCL2 production and secretion from hepatocytes, leading to macrophage recruitment and pro-inflammatory M1 polarization (17). Consistently, in the current study, Shp deletion in hepatocytes also causes liver inflammation in vivo. Therefore, we speculate that the increase in CCL2 secretion from Shp-deficient hepatocytes may contribute to the liver inflammation developed in Shp HepϪ/Ϫ mice. Further investigation is warranted to explore whether blocking CCL2 production from hepatocytes or repressing CCL2 function could inhibit or ameliorate liver inflammation developed in Shp HepϪ/Ϫ mice.
The potential role of SHP in NASH patients has been recognized recently. One study (39) showed that SHP is down-regulated in NASH patients. Similarly, our recent study has determined a decrease of SHP expression during the progression of fatty liver to NASH in patients (17). Additionally, increasing SHP expression by the FXR agonist obeticholic acid has been used in the treatment of NAFLD (40), supporting the notion that approaches inducing SHP would be beneficial for NASH treatment.
However, the role of hepatic Shp in the development of mouse NAFLD is somehow controversial. Studies showed that both whole-body Shp knockout mice and hepatic Shp knockout mice displayed resistance to high-fat diet-induced hepatic steatosis (14,15,41). Given that all of these fatty liver models are associated with little or no inflammation and fibrosis, the role of hepatic SHP in mouse NASH still remains obscure. Here, we fed mice a high-fat, -cholesterol, and -fructose diet that induces many biochemical and histopathological hallmarks of NASH, to study the role of SHP in modulating liver steatosis, inflammation, and fibrosis during NASH development. The most striking observation was that hepatic Shp disruption induces the dissociation of steatosis, inflammation, and fibrosis during NASH development.
Although the molecular mechanism underlying NASH pathogenesis is incompletely understood, numerous studies have demonstrated that NASH development is attributable to multiple hits, including excessive accumulation of toxic lipids within hepatocytes and other insults, such as macrophage-mediated liver inflammation and hepatic stellate cell (HSC)-mediated liver fibrosis. Because toxic lipid metabolites can promote the overproduction of reactive oxygen species that damage hepatocytes, leading to the release of DAMPs into the microenvironment and resulting in macrophage activation, steatosis is generally believed to be positively associated with liver inflam-mation and liver injury. However, recent studies indicate that the storage of excessive free fatty acids (FFAs) as triglycerides is unlikely to be the cause of hepatocyte injury in NASH (42). Instead, the triglyceride accumulation within hepatocytes acts as a protective mechanism against fatty acid-induced lipotoxicity, which is in response to lipid overload due to the increases in dietary intake of FFA, de novo lipogenesis, adipose lipolysis, and impaired FFA oxidation (43). In support of this notion, inhibiting triglyceride synthesis improves hepatic steatosis but exacerbates liver damage and fibrosis in NASH (44). Genetic or pharmacological inhibition of stearoyl-CoA desaturase-1 (SCD1), the enzyme that converts saturated fatty acids to monounsaturated fatty acids for triglyceride synthesis, sensitizes hepatocytes to saturated fatty acid-induced apoptosis. Consequently, Scd1 Ϫ/Ϫ mice displayed decreased steatosis but markedly increased liver injury and fibrosis during NASH development (42). PPARg, a key modulator of lipid synthesis, is markedly up-regulated in fatty liver (28). Disruption of Pparg in hepatocytes reduces liver steatosis but dramatically increases hepatic inflammation after high-fat diet plus binge ethanol (36). Fascinatingly, in the current study, disruption of Shp in hepatocytes attenuates steatosis but exacerbates liver inflammation and fibrosis in HFCF-fed mice. Mechanistically, the dissociation between steatosis and inflammation in Shp HepϪ/Ϫ is accompanied by inactivation of Pparg and activation of NF-B, indicating that hepatic SHP plays an opposing role in liver steatosis and inflammation through modulating both Pparg and NF-B signaling.
To further elucidate the role of SHP specifically in the disease transition from steatosis to NASH, we developed a steatosis-to-NASH transition model where mice were fed an HFCF diet for 4 weeks to induce liver steatosis, followed by disruption of Shp in hepatocytes. Interestingly, although Pparg expression was consistently decreased in this model, hepatic Shp deletion after steatosis development exacerbated liver inflammation and fibrosis without affecting liver steatosis. The above observation is in line with our previous finding that hepatic SHP overexpression in HFCF-fed mice does not affect hepatic lipid contents but greatly attenuates liver inflammation and fibrosis during NASH progression (17). Collectively, our current study reconfirms the critical anti-inflammatory and antifibrotic role of SHP in the progression of fatty liver to NASH.
Shp deletion in hepatocytes causes down-regulation of Pparg, suggesting a tight regulatory link between Shp and Pparg in hepatocytes. The mechanism by which Shp regulates the expression of Pparg remains incompletely understood. One earlier study demonstrated that SHP increases Pparg gene expression through a transcriptional cascade where SHP inhibits hairy and enhancer of split 6 (Hes6), a transcriptional repressor that suppresses Hnf4␣-induced Pparg gene expression (41). Another study showed that in rat immortalized HSC cell line HSC-T6, SHP is recruited to the Pparg gene promoter and activates Pparg transcription (45). Additionally, increasing SHP expression by 6-ethyl cheno-deoxycholic acid (INT-747) stimulates the expression and activity of Pparg in adipocytes (46). These data indicate that SHP is a strong transcriptional regulator of Pparg, and multiple mechanisms are implicated in this regulation.

Dissociation of steatosis from inflammation by Shp deletion
Loss of hepatic Shp affects the fibrogenic process in the liver, which is another important observation from our current study. The critical antifibrotic role of Shp has been documented in the literature. An earlier study (47) demonstrated that disruption of Shp exacerbates bile duct ligation-induced cholestatic liver fibrosis, whereas Shp overexpression or increasing Shp expression by pharmacological compounds attenuates hepatic fibrosis induced either by hepatitis C virus infection (48) or by carbon tetrachloride and ␣-naphthyl-isothiocyanate (49). In addition, increasing Shp mRNA levels in HSCs by FXR ligands abrogates thrombin-and TGF-␤1-induced up-regulation of ␣1 collagen mRNA (50). Furthermore, Shp overexpression in HSCs inhibits the expression of tissue metalloproteinase inhibitor 1 (Timp1) and promotes a quiescent phenotype of HSCs (51). All of these findings indicate that SHP acts as a critical antifibrotic factor in various liver diseases. Our results, obtained from hepatic Shp-deficient animals, support this notion.
In summary, our study provides compelling evidence that hepatic Shp disruption in adult mice induces the dissociation of steatosis, inflammation, and fibrosis during NASH development. The complex effects of SHP on hepatic steatosis, inflammation, and fibrosis compromise the potential use of SHP as a therapeutic target for NAFLD treatment. Approaches targeting SHP should be used cautiously, because inhibition of hepatic SHP, although beneficial for the amelioration of liver steatosis, could exacerbate liver inflammation and fibrosis.

Antibodies
The following antibodies were used for Western blotting and

Animal studies
Shp flox/flox mice were generously provided by Drs. Johan Auwerx and Kristina Schoonjans at the Ecole Polytechnique de Lausanne and backcrossed into C57BL/6J background for 10 generations (17). Mice were maintained in a 12-h light/dark cycle (light on from 6 a.m. to 6 p.m.), temperature-controlled (23°C), and virus-free facility with free access to food and water. Experiments on mice were performed on males at the age of 8 -10 weeks unless stated otherwise (n ϭ 5-6/group). To generate hepatocyte-specific Shp knockout (Shp HepϪ/Ϫ ) and WT controls, Shp flox/flox mice were administered with either adeno-associated virus serotype 8 (AAV8) expressing Cre recombinase driven by the thyroxine-binding globulin (Tbg) promoter (AAV8-Tbg-Cre) or control AAV8 (AAV8-Tbg-null) at a dose of 2 ϫ 10 11 genome copies/mouse through tail vein injection. Both AAV8-Tbg-Cre (AV-8-PV1090) and AAV8-Tbg-null (AV-8-PV0148) were obtained from the University of Pennsylvania Vector Core. In dietary NAFLD models, mice were placed on a diet enriched in high fat, cholesterol, and fructose (HFCF; Research Diet, D09100301, 40 kcal% fat, 2% cho-lesterol, 20 kcal% fructose) for 12 weeks. In a steatosis-to-NASH transition model, Shp flox/flox mice were fed an HFCF diet for 4 weeks to develop liver steatosis, followed by tail vein injection of AAV8-Tbg-Cre or AAV8-Tbg-null control. Mice remained on the HFCF diet for an additional 8 weeks. Mice fed normal chow served as controls. Blood and liver samples were collected after mice had been fasted for 16 h. All experiments were performed in accordance with relevant guidelines and regulations approved by the Institutional Animal Care and Use Committee at the University of Kansas Medical Center.

Glucose tolerance test
Mice were fasted for 16 h, followed by intraperitoneal injection of glucose (2 g/kg body weight). Blood was collected by tail vein puncture. Glucose levels were determined before and at 0.5, 1, 1.5, and 2 h after glucose administration.

Liver histological examinations of steatosis, cell death, inflammation, and fibrosis
Fresh liver tissues were fixed with 10% formalin (Fisher, SF100). Paraffin sections at 4 m were stained with H&E. TUNEL staining for detection of cell death in the liver was performed using an in situ cell death detection kit-alkaline phosphatase (Sigma, 11684809910) according to the manufacturer's suggestions. For Oil Red O staining of liver lipids, fresh liver tissue was immediately embedded in Tissue-Tek O.C.T. compound (VWR, 25608-930), and frozen sections were cut at 8 m, fixed by 10% formalin, and stained with Oil Red O (Sigma, O0625). Images were acquired with a BX60 microscope (Olympus, Lake Success, NY). For immunohistochemical staining, 6-m frozen sections were fixed by 10% formalin and treated with 0.3% H 2 O 2 in methanol for 15 min to block endogenous peroxidase activity. Slides were then treated with 10% normal serum for 30 min, followed by incubation with primary antibody overnight at 4°C. The ImmPRESS peroxidase polymer detection kit (Vector Laboratories, MP-7444) and ImmPACT 3,3Ј-diaminobenzidine peroxidase substrate (Vector Laboratories, SK-4105) were used for the final detection. Sections were then counterstained with hematoxylin, dehydrated, cleared, and mounted. Images were acquired with a BX60 microscope. For Picrosirius Red staining of liver fibrosis, 4-m paraffin sections were rehydrated and incubated in 0.1% Sirius Red F3B (Sigma, Direct Red 80, 365548) containing saturated picric acid (Sigma, p6744) for 1 h. After washing three times in 0.5% glacial acetic acid, sections were briefly dehydrated, cleared, and mounted. Images were acquired with a BX60 microscope, and collagen density was quantified using ImageJ software.

Hepatic triglycerides and cholesterol measurements
According to our previous publication (17), 100 mg of liver tissues were homogenized in 300 l of chloroform/methanol

Dissociation of steatosis from inflammation by Shp deletion
(1:2, v/v) for 2 min, followed by a second homogenization for 30 s with an addition of 300 l of chloroform. The homogenates were mixed with 100 l of H 2 O and homogenized again for 30 s. The lipid layer (ϳ600 l) was separated via centrifugation at 800 ϫ g for 10 min at room temperature. The lower phase enriched in lipid was transferred and dried using nitrogen gas. The lipid extract was suspended in 300 l of 5% Triton X-100 in PBS (pH 7.4). The measurement was performed using respective kits for triglycerides (Pointe Scientific, T7532) and cholesterol (Pointe Scientific, C7510). The hepatic triglycerides or cholesterol content was defined as g of triglycerides or cholesterol per mg of liver tissue.

Analysis of bile acid (BA) pool size and fecal BA extraction rate
To determine BA pool size, fresh mouse tissues including gallbladder, liver, and entire small intestine were minced and extracted in 75% ethanol at 50°C for 2 h. The extract was then centrifuged, diluted with 75% ethanol, and further diluted with 25% PBS before the BA measurement using a BA colorimetric assay BQ kit (Thermo Fisher Scientific, BQ092A-EALD). The pool size was expressed as mol of bile acid/g of body weight. To determine fecal bile acid excretion, the feces from individually housed mouse over a 72-h period were collected, weighed, dried, and extracted in 75% ethanol. The extract was then diluted with 25% PBS and subjected to bile acid measurement. The daily fecal output (g/day/g of body weight) and fecal bile acid content (mol/g) were used to calculate the rate of bile acid excretion (mol/day/g of body weight).

Determination of hepatic collagen content by hydroxyproline assay
Liver tissues (10 mg) were homogenized in 100 l of H 2 O. The homogenates were mixed with 100 l of 12 M HCl and incubated at 120°C for 3 h for acid hydrolysis. The homogenates were then centrifuged at 10,000 ϫ g for 10 min. Aliquots of the hydrolyzed samples (10 l) were incubated with 100 l of chloramine T solution (1.27% chloramine T and 10% isopropyl alcohol in acetate-citrate buffer, pH 6.0) at room temperature for 25 min, followed by a second incubation with 100 l of Ehrlich's solution (Sigma, 03891) at 60°C for 35 min. A plate reader measured sample absorbance at 550 nm. The hepatic collagen content was defined as g of collagen per mg of liver tissue.

Western blotting
Nuclear and cytoplasmic protein extraction was carried out using a commercial kit (Fisher, PI78833). Protein lysates (60 g) were resolved by SDS-PAGE and transferred to nitrocellulose membranes. Membranes were then blocked and incubated with primary antibodies, followed by horseradish peroxidaseconjugated corresponding secondary antibody incubation. Antibody binding was visualized using either SuperSignal West Pico Plus Chemiluminescent Substrate (Fisher, PI34580) or SuperSignal West Femto Chemiluminescent Substrate (Fisher, PI34094). Images were captured by LI-COR, and equal loading of protein was verified by loading controls such as ␣-tubulin and histone H3. Quantitative analysis of band intensity was performed by Image Studio Lite software, and relative expression levels were normalized to the loading controls.

RNA isolation and RNA-Seq
Total RNA for Illumina sequencing was extracted from mouse liver tissues using a Direct-zol RNA kit (Zymo Research, R2071) according to the manufacturer's protocol. During RNA isolation, DNA was removed by treating the samples with RNase-free DNase to avoid DNA contamination. The purity, concentration, and integrity of the RNA were examined using a NanoDrop 1000 spectrophotometer (Thermo Fisher) and an Agilent Bioanalyzer 2100 system (Agilent Technologies). Three biological replicates in each group for a total of 12 samples were submitted for RNA-Seq. The RNA integrity number values of all samples used for RNA-Seq were Ͼ6.0. The high-throughput genomics core at the Huntsman Cancer Institute, University of Utah, performed library preparation and sequencing. cDNA libraries were prepared using the Illumina TruSeq stranded RNA kit with Ribo-Zero Gold. Fifty-cycle single-read sequencing was performed with an Illumina (San Diego, CA) HiSeq 2500, and reads were aligned to a mouse reference sequence genome mm10 using the Novalign short-read alignment software. Sample reads were visualized, and DEGs were identified based on the log-transformed false discovery rate of Ͼ1.3 and ՆϮ1.5-fold change in expression relative to controls using the USeq application as described previously (52).

GO and KEGG enrichment analysis of differentially expressed genes
DEGs were used for GO and KEGG (Kyoto Encyclopedia of Genes and Genomes) enrichment analyses on the Enrichr web server to identify biological functions and significantly enriched pathways (53,54). Enrichr is freely available at http://amp. pharm.mssm.edu/Enrichr. 3 In this study, both GO terms and KEGG pathways with adjusted p value Ͻ0.05 were considered significantly enriched. An Enrichr combined score was used to select top altered GO terms and KEGG pathways.

Real-time quantitative PCR
The real-time quantitative PCR (qPCR) was carried out using the SYBR Green PCR master mix (Applied Biosystems) as described previously (55). The specific primers are shown in the supporting information. The amount of PCR products was measured by threshold cycle (Ct) values, and the relative ratio of specific genes to the housekeeping gene hypoxanthine guanine phosphoribosyl transferase 1 (Hprt1) was calculated and presented as -fold change in the tested group relative to the control group.

Mouse primary hepatocyte culture and adenovirus infection
The University of Kansas Medical Center Cell Isolation Core conducted hepatocyte isolation from Shp flox/flox mice using the method described previously (56) with a slight modification. In brief, mouse liver was perfused with 25 ml of solution I (9.5 g/liter Hanks' balanced salt solution, 0.5 mmol/liter EGTA, pH 7.2), followed by 50 ml of solution II (9.5 g/liter Hanks' balanced salt solution, 0.14 g/liter collagenase IV, and 40 mg/liter trypsin inhibitor, pH 7.5). After digestion, single-cell suspension was filtered through a 100-m Falcon cell strainer (Fisher, 08-771- 19), and the cells were centrifuged at 50 ϫ g for 5 min at 4°C to pellet hepatocytes. Hepatocytes were then seeded in collagen type 1-coated dishes. After a 2-h incubation, cell culture medium was replaced by fresh William E medium (Sigma, W4128) with various adenoviruses at a multiplicity of infection of 20. The Pparg adenovirus (catalog no. 1354), Cre adenovirus (catalog no. 1045), and vector control adenovirus (catalog no. 1240) were purchased from Vector Biolabs. At the second day, hepatocytes were treated either with 5 M NF-B inhibitor BAY 11-7082 (Sigma, B5556) for 6 h or with 0.5 mM oleic acid (Sigma, O1008) conjugated with BSA (Fisher, BP9704-100) for 24 h. Cells were then collected for RNA isolation or Oil Red O staining.

Statistical analysis
Quantitative data are presented as the mean Ϯ S.E. The statistically significant difference in data obtained between two groups was determined by Student's t test. Multiple groups were compared by one-way ANOVA, followed by Duncan's test. Statistical significance was accepted within 95% confidence limits.