Research ArticleMetabolism

TRIF-dependent Toll-like receptor signaling suppresses Scd1 transcription in hepatocytes and prevents diet-induced hepatic steatosis

See allHide authors and affiliations

Science Signaling  08 Aug 2017:
Vol. 10, Issue 491, eaal3336
DOI: 10.1126/scisignal.aal3336

TRIF against fatty liver

Viral detection by TLR pathways triggers inflammatory responses, which generally aggravate metabolic diseases. However, Chen et al. found that the TLR signaling adaptor TRIF in hepatocytes, rather than myeloid cells, limited diet-induced hepatic steatosis in mice. TRIF activation downstream of TLR3 led to the transcriptional suppression of Scd1, which encodes a key lipogenic enzyme. Viral RNA and mimetics activate TLR3, and application of RNA generated by adipose tissue prevented the increase in SCD1 abundance and the enhanced triglyceride accumulation that normally occurs in hepatocytes exposed to palmitic acid, a saturated fatty acid that is enriched in high-fat diets. The authors note that the hepatitis C virus co-opts host lipogenesis to ensure its replication and that TRIF-mediated suppression of Scd1 transcription may therefore serve to limit viral infection of hepatocytes.


Nonalcoholic fatty liver disease (NAFLD) includes a spectrum of diseases that ranges in severity from hepatic steatosis to steatohepatitis, the latter of which is a major predisposing factor for liver cirrhosis and cancer. Toll-like receptor (TLR) signaling, which is critical for innate immunity, is generally believed to aggravate disease progression by inducing inflammation. Unexpectedly, we found that deficiency in TIR domain–containing adaptor-inducing interferon-β (TRIF), a cytosolic adaptor that transduces some TLR signals, worsened hepatic steatosis induced by a high-fat diet (HFD) and that such exacerbation was independent of myeloid cells. The aggravated steatosis in Trif−/− mice was due to the increased hepatocyte transcription of the gene encoding stearoyl–coenzyme A (CoA) desaturase 1 (SCD1), the rate-limiting enzyme for lipogenesis. Activation of the TRIF pathway by polyinosinic:polycytidylic acid [poly(I:C)] suppressed the increase in SCD1 abundance induced by palmitic acid or an HFD and subsequently prevented lipid accumulation in hepatocytes. Interferon regulatory factor 3 (IRF3), a transcriptional regulator downstream of TRIF, acted as a transcriptional suppressor by directly binding to the Scd1 promoter. These results suggest an unconventional metabolic function for TLR/TRIF signaling that should be taken into consideration when seeking to pharmacologically inhibit this pathway.


Nonalcoholic fatty liver disease (NAFLD) is a spectrum of diseases that ranges from hepatic steatosis to steatohepatitis. The major characteristic of NAFLD is excessive hepatic lipid accumulation, and with the presence of inflammation, steatohepatitis predisposes patients to liver cancer and cirrhosis. In recent years, the increased prevalence of obesity and diabetes has led to the concern in the NAFLD epidemic (1). Toll-like receptor (TLR) pathway is a key part of the innate immune response that is associated with various metabolic diseases (2). It is generally believed that activation of the TLR pathway in immune cells is the source of local and systemic inflammatory molecules that aggravate the disease progression from simple steatosis to steatohepatitis (26). Moreover, TLR pathway can regulate various proteins involved in metabolism in immune cells such as macrophages and dendritic cells during activation and differentiation (79). Different TLRs are detected in metabolic cells including hepatocytes, but whether TLR pathway also has metabolic functions in these cells remains to be investigated.

The signals of most TLRs are mediated by two key cytosolic adaptors: TIR domain–containing adaptor-inducing interferon-β (TRIF) and myeloid differentiation primary response 88 (MyD88). MyD88 transduces signaling mediated by all TLRs except TLR3, and TRIF is a cytosolic adaptor for only TLR3 and TLR4. Unlike TLR3, which can directly bind to TRIF, TLR4 requires another adaptor, TRIF-related adaptor molecule (TRAM), to activate the TRIF-dependent signal. Many studies have demonstrated the pathological role of TLR4 and MyD88 in metabolic diseases; however, the effects of TLR3 and TRIF are inconclusive (5, 1013). Activation of TRIF by low-dose lipopolysaccharides (LPS) prevents chemical endoplasmic reticulum (ER) stressor–induced hepatic steatosis (12). TRIF can also restore protein translation and delay apoptosis by activating eukaryotic initiation factor 2B (14). Interferon regulatory factor 3 (IRF3), a downstream transcription factor of TRIF signaling, can decrease the abundance of retinoid X receptor–α (RXR-α), which controls the transcription of various metabolism-related genes upon challenge with vesicular stomatitis virus (VSV), LPS (a TLR4 ligand), or polyinosinic:polycytidylic acid [poly(I:C)] (a TLR3 ligand) in macrophages (15). These studies imply that TLR/TRIF pathway may play a distinct role in metabolic pathways and diseases (16).

Here, we investigated the function of TRIF pathway in the development of diet-induced hepatic steatosis. Our results showed that Trif−/− mice under high-fat diet (HFD) exhibited more severe hepatic steatosis compared with wild-type mice, and the TRIF deficiency–induced aggravation was independent of myeloid cells. Activation of TRIF inhibited lipid accumulation in hepatocytes by suppressing the expression of the gene encoding the rate-limiting enzyme of lipogenesis, namely, stearoyl–coenzyme A (CoA) desaturase 1 (SCD1), and such inhibition was facilitated by IRF3, which acted as a transcriptional suppressor of Scd1.


TRIF deficiency caused aggravation of HFD-induced hepatic steatosis in mice, which was independent of myeloid cells

Six-week-old male wild-type and Trif−/− mice in C57BL/6J background fed a normal chow diet (NCD) or a 60% fat kcal HFD for 10 weeks did not show differences in body weight or fat mass (fig. S1, A and B). Hematoxylin and eosin (H&E) and Oil Red O staining showed more lipid vacuoles in the livers of Trif−/− than wild-type mice on an HFD (Fig. 1, A and B). Quantification of extracted hepatic lipids revealed that both triglyceride and total cholesterol were significantly higher in the Trif−/− mice compared with the wild-type control (Fig. 1C). Furthermore, Trif−/− mice had significantly greater serum aspartate transaminase (AST) and alanine transaminase (ALT) activities and fasting blood glucose (Fig. 1D and fig. S1C). On an NCD, Trif−/− mice displayed AST and ALT activities, hepatic triglyceride and cholesterol content, and fasting blood glucose similar to those in wild-type mice (Fig. 1, A to D, and fig S1C). However, these NCD-fed Trif−/− mice had lower glucose tolerance, an impairment that was further aggravated by HFD feeding (fig. S1D). The effects of TLR pathway on metabolic dysfunction are mediated by immune cells such as macrophages (5, 17). We used bone marrow transplantation to find out whether the observed phenotypes were mediated by myeloid cells (5). Irradiated wild-type and Trif−/− mice received bone marrow isolated from either wild-type or Trif−/− donors before being fed an HFD. Successful transplantation was confirmed by DNA sequencing to detect the presence or absence of guanine at codon 708 of Trif in the bone marrow of recipients at the end of HFD feeding. Similar to previous data, mice undergoing bone marrow transplantation had lower body weights and fat mass without significant differences between the two genotypes (fig. S1, E to F) (18). The Trif−/− mice that received either wild-type or Trif−/− bone marrow had increased hepatic triglyceride and cholesterol content, degree of steatosis, serum AST and ALT activities, and fasting blood glucose compared with the corresponding wild-type recipients (Fig. 1, E to H, and fig. S1G).

Fig. 1 The aggravation of diet-induced hepatic steatosis in Trif−/− mice was mediated by nonmyeloid cells.

(A to D) Six-week-old wild-type (WT) and Trif−/− mice were fed an NCD or HFD (60% fat kcal) for 10 weeks. (E to H) Irradiated eight-week-old WT and Trif−/− mice received bone marrow (BM) transplantation and were fed an HFD for 10 weeks as in (A) to (D). Livers were subjected to H&E (A and E) and Oil Red O staining (B and F). Scale bars, 100 μm (A and E) and 20 μm (B and F). Hepatic total triglyceride (TG) and total cholesterol (TC) content (C and G) and serum AST and ALT activities were measured (D and H). For (A) to (D), n = 8 to 10 mice per genotype and diet; for (E) to (H), n = 4 to 6 mice per genotype and treatment. *P < 0.05 and **P < 0.01. Representative images are shown.

Inflammation aggravates metabolic dysfunction. Nevertheless, in spite of the worsened hepatic steatosis, hepatic inflammation was decreased in a nonstatistically significant manner in Trif−/− mice compared with the wild-type littermates (fig. S2A). However, serum insulin, triglyceride and cholesterol, and low-density lipoprotein and high-density lipoprotein did not differ between HFD-fed wild-type and Trif−/− mice (fig. S2, B to D). Together, these results suggest that deficiency in TRIF in nonmyeloid cells worsened the degree of HFD-induced hepatic steatosis.

Increased hepatic lipid content in Trif−/− mice was due to augmented Scd1 expression in hepatocytes

Activation of the TRIF pathway inhibits the ER stress–induced increase in the expression of Chop [which encodes CCAAT/enhancer binding protein (C/EBP) homologous protein] (12), and CHOP can dimerize and inhibit C/EBPs, resulting in the disruption of lipid regulation (19). We speculated that the increased hepatic steatosis in Trif−/− mice might be due to an increase in Chop expression. However, HFD-fed Trif−/− and wild-type mice did not show increased CHOP abundance at the mRNA or protein levels (Fig. 2A and fig. S3, A to B), and hepatic apoptosis was also absent (fig. S3C). Because hepatic triglyceride content was increased in both whole-body knockout and bone marrow transplantation models, we next examined the expression of several genes involved in lipogenesis in liver (Fig. 2A). Trif−/− mice fed an NCD did not show any significant changes in lipogenic gene expression (fig. S4). However, upon HFD feeding, Scd1 expression and SCD1 protein abundance were significantly increased in Trif−/− mice compared with wild-type mice (Fig. 2, A and B). Furthermore, palmitic acid, a long-chain saturated fatty acid, stimulated Scd1 expression and SCD1 protein in hepatocytes (Fig. 2, C and D), and the TLR3 agonist poly(I:C) alleviated such induction in wild-type but not Trif−/− hepatocytes (Fig. 2, C and D). TLR3/TRIF signaling results in the activation of IRF3 by phosphorylation (20, 21), and poly(I:C) induced IRF3 phosphorylation at Ser396 in hepatocytes (fig. S5A). The inhibition of the palmitic acid–induced increase in SCD1 abundance by poly(I:C), which was TRIF-dependent, also correlated with lower triglyceride and cholesterol content of hepatocytes as shown by Oil Red O staining (Fig. 2E) and quantification of hepatic lipids (Fig. 2F). The ability of poly(I:C) to inhibit the palmitic acid–induced increase in Scd1 expression in hepatocytes was blocked by knockdown of TLR3 with small interfering RNA (siRNA) (fig. S5, B to D) but not by treating the cells with LPS to activate TLR4, which can stimulate TRIF (fig. S5D). This result might be due to the different extents of TRIF activation elicited by TLR3 and TLR4 in hepatocytes. Compared with other cell types in the liver, hepatocytes produce less cytokines and downstream signaling molecules upon LPS stimulation, implying that hepatocyte TLR4 is less sensitive (22). These results show that activation of TLR3/TRIF pathway by poly(I:C) could inhibit SCD1 in hepatocytes at both mRNA and protein levels and prevent lipid accumulation.

Fig. 2 Activation of TRIF by poly(I:C) inhibited palmitic acid–induced hepatic SCD1 expression and lipid accumulation.

(A and B) Six-week-old WT and Trif−/− mice were fed an NCD or HFD for 10 weeks. (A) Hepatic Chop, Acc1, Fasn, Scd1, and Dgat2 expression was examined by real-time polymerase chain reaction (PCR) and normalized to Actb. (B) Western immunoblotting was performed to detect hepatic SCD1 and β-actin proteins. (C to F) Primary hepatocytes from WT and Trif−/− mice were treated with or without palmitic acid (PA; 250 μM) (ctrl, control) and poly(I:C) (2.5 μg/ml) for (C) 6 hours or (D to F) 8 hours. Scd1 mRNA (C) and SCD1 protein abundance (D) were measured. Oil Red O staining (E) and quantification of hepatic TG and TC content (F) were performed. Scale bar, 40 μm. Representative images are shown. (G) The concentrations of serum RNA from WT mice fed an NCD or HFD for 10 weeks were measured. (H and I) WT and Trif−/− hepatocytes were treated with PA and RNA (2.4 μg/ml) isolated from adipose tissue and complexed with Lipofectamine as in (D). (H) SCD1 protein abundance and (I) intracellular TG content were measured. For (A), (B), and (G), n = 4 to 7 mice per genotype or diet; for (C) to (F) and (H) and (I), n = 3 independent experiments. *P < 0.05 and **P < 0.01.

Endogenous mRNA released by apoptotic or necrotic cells and gut-derived viromes can activate TLR3 under noninfectious conditions (23, 24). We observed an increase in circulating RNAs in HFD-fed mice (Fig. 2G), and these circulating RNAs consisted of ribosomal RNAs (fig. S6A). As a proof of concept, we applied freshly isolated RNAs complexed with Lipofectamine and palmitic acid to primary mouse hepatocytes and found that the cells took up these RNAs (fig. S6B). Similar to poly(I:C), these isolated RNAs partially inhibited the palmitic acid–induced increase in SCD1 at the mRNA and protein levels and triglyceride accumulation in wild-type but not Trif−/− hepatocytes (Fig. 2, H to I).

Reconstitution of TRIF by adenovirus gene delivery inhibited the increase in SCD1 and hepatic steatosis in Trif−/− mice

To verify the specific role of hepatic TRIF in lipogenesis, we reconstituted TRIF expression in Trif−/− mice using adenovirus gene delivery (Ad-TRIF). When wild-type and Trif−/− hepatocytes were infected with the control virus [Ad–green fluorescent protein (GFP)], we observed that poly(I:C) failed to suppress the palmitic acid–induced increase in SCD1 and lipid accumulation in Trif−/− cells. Infection with Ad-TRIF restored poly(I:C)–mediated suppression in Trif−/− hepatocytes (Fig. 3, A to C). Moreover, the reconstitution of TRIF in vivo reversed the increase in SCD1 at the mRNA and protein levels and reduced hepatic triglyceride accumulation in Trif−/− mice to amounts comparable to those of wild-type mice infected with Ad-GFP (Fig. 3, D to G).

Fig. 3 Ad-TRIF reversed the increased SCD1 and hepatic steatosis in Trif−/− mice.

(A to C) WT and Trif−/− hepatocytes infected with adenovirus encoding GFP (Ad-GFP) or TRIF (Ad-TRIF) were treated with or without PA (250 μM) and poly(I:C) (2.5 μg/ml) for 8 hours. GFP was detected in cells transfected with Ad-GFP and Ad-TRIF (A). Scale bar, 40 μm. SCD1 protein (B) and intracellular TG content (C) were determined. (D to G) Six-week-old WT and Trif−/− mice were fed an HFD for 4 weeks, injected once with 109 viral particles (vp) per mouse of Ad-GFP or Ad-TRIF, and continued on the HFD for two more weeks. Livers were analyzed for GFP fluorescence (D), Scd1 mRNA expression (E), SCD1 protein abundance (F), and TG content (G). Scale bar, 20 μm. For (A) to (C), n = 3 independent experiments; for (D) to (G), n = 4 to 6 mice per genotype and/or infection condition. *P < 0.05 and **P < 0.01. Representative images are shown.

Silencing SCD1 decreased intracellular triglyceride content in Trif−/− hepatocytes

Scd1−/− mice are resistant to HFD-induced hepatic steatosis (25). To determine whether attenuation of SCD1 prevented TRIF-mediated aggravation of hepatic steatosis, we treated wild-type and Trif−/− hepatocytes that had been infected with lentivirus encoding SCD1 siRNA (Lv-SCD1-siRNA) or scrambled RNA (Lv-Scr-RNA) with palmitic acid (Fig. 4, A to C, and fig. S7, A and B). The palmitic acid–induced increase in Scd1 expression was successfully inhibited by Lv-SCD1-siRNA in both wild-type and Trif−/− hepatocytes (Fig. 4A). Similar to the above finding, poly(I:C) inhibited the palmitic acid–induced increase in Scd1 expression and triglyceride accumulation in infected wild-type but not Trif−/− cells (Fig. 4, A to C). Conversely, infection with Lv-SCD1-siRNA abolished the palmitic acid–induced increase in intracellular triglyceride content in hepatocytes from both wild-type and Trif−/− mice (Fig. 4, B and C). Scd1 expression and hepatic triglyceride accumulation were decreased in HFD-fed Trif−/− mice infected with Lv-SCD1-siRNA compared to those infected with Lv-Scr-RNA (Fig. 4, D to F). Together, these findings suggest that silencing SCD1 alone was sufficient to reverse hepatic lipid accumulation in Trif−/− mice.

Fig. 4 Silencing SCD1 abolished lipid accumulation in Trif−/− hepatocytes.

(A to C) Isolated WT and Trif−/− hepatocytes infected with lentivirus encoding either SCD1 siRNA (Lv-SCD1-siRNA) or scrambled RNA (Lv-Scr-RNA) were treated with or without PA (250 μM) and poly(I:C) (2.5 μg/ml) for 8 hours. Scd1 mRNA expression (A) and TG content (B) were measured, and Oil Red O staining (C) was performed. Scale bar, 40 μm. (D to F) Six-week-old Trif−/− mice were fed an HFD for 10 days, injected with either Lv-SCD1-siRNA or Lv-Scr-RNA, and continued on the HFD for another 3 weeks. Hepatic Scd1 mRNA (D), SCD1 protein (E), and TG content (F) were determined. A group of age-matched Trif−/− mice fed an NCD served as controls. n = 4 independent experiments for (A) to (C); n = 6 mice per genotype, diet, and/or injection condition for (D) to (F). *P < 0.05 and **P < 0.01. Representative images are shown.

IRF3 acted as a transcriptional suppressor of Scd1

Because TRIF inhibited Scd1 expression, we next tested whether this effect occurred at the promoter region of the Scd1 gene. First, we delivered a bioluminescent reporter plasmid containing the promoter region of Scd1 (26) into both wild-type and Trif−/− mice fed an HFD (27). Trif−/− mice displayed a threefold more intense chemiluminescent signal in the liver area compared with wild-type mice (Fig. 5A). Activation of TLR pathway can induce transcriptional activity of several IRFs (28). We speculated that IRF3, which is selectively responsive to TRIF, was involved in poly(I:C)–induced transcriptional suppression of Scd1 (28). In HeLa cells, SCD1 abundance was reduced by overexpression of a phosphomimetic IRF3(5D) [in which the five Ser/Thr sites in the region between amino acid residues 395 and 407 (ISNSHPLSLTSDQ) at the C terminus of IRF3 were substituted with Asp (21)] but not wild-type IRF3 (Fig. 5B). Bioinformatics analysis identified two putative IRF3 binding sites (GAAANN) at the −248 and −593 base pair of the 5′ flanking region of Scd1 gene. Two different lengths of the upstream segments of Scd1 gene (−816 + 229 and −469 + 229) covering the trans- and/or cis-regulatory elements were inserted into pcDNA3.1/V5-His/lacZ reporter vector. Palmitic acid or poly(I:C) treatment did not alter reporter activity in cells transfected with the plasmid containing the sequence between −469 and +229 of the Scd1 gene promoter (Fig. 5C). However, palmitic acid treatment stimulated reporter activity in cells transfected with the plasmid containing the region between −816 and +229, an effect that was blocked by poly(I:C) (Fig. 5C). In this region, the IRF3 binding motif is located between −599 and −593 and is highly conserved across different species (table S1). Mutation of this sequence did not affect the palmitic acid–mediated induction but abolished the poly(I:C)–mediated repression of Scd1 promoter activity (Fig. 5C). Chromatin immunoprecipitation analysis further demonstrated that in poly(I:C)–treated cells, IRF3 directly bound to the wild-type version of the Scd1 promoter but not to the mutant version of the promoter (Fig. 5D). In addition, poly(I:C) did not suppress palmitic acid–induced Scd1 promoter activity in HepG2 human liver cells expressing the mutant promoter (Fig. 6A). Overexpression of IRF(5D) inhibited the palmitic acid–induced increase in Scd1 promoter activity, SCD1 protein abundance, and intracellular triglyceride content in HepG2 cells (Fig. 6, B to D). Moreover, the suppressive effect of poly(I:C) was abolished after silencing IRF3 (Fig. 6, E and F). These data suggest that IRF3 is a transcriptional suppressor of Scd1 in both rodents and humans.

Fig. 5 Suppressive action on the Scd1 promoter by TRIF was mediated by IRF3.

(A) Six-week-old WT and Trif−/− mice fed an HFD for 10 weeks were hydrodynamically injected with the luciferase-expressing plasmid pGL3/–1537+155mSCD1. Luciferase activity was analyzed 1 hour after administration of d-luciferin. (B) HeLa cells expressing IRF3(WT) or IRF3(5D) were immunoblotted for IRF3 and SCD1 with β-tubulin as loading control. (C) β-Galactosidase activity was measured in HeLa cells expressing WT pcDNA3.1/–469+229mSCD1 and WT or mutated pcDNA3.1/–816+229mSCD1 that were treated with or without PA (250 μM) and poly(I:C) (2.5 μg/ml) for 8 hours. (D) HeLa cells transfected with WT or mutated pcDNA3.1/–816+229mSCD1 were treated with PA (250 μM) and poly(I:C) (2.5 μg/ml) for 2 hours. Chromatin immunoprecipitation with IRF3 antibody and PCR amplification using specific primers against the Scd1 promoter region were performed. IgG, immunoglobulin G. For (A), n = 4 mice per genotype; for (B) to (D), n = 3 to 4 independent experiments. *P < 0.05 and **P < 0.01. Representative images are shown.

Fig. 6 IRF3-mediated SCD1 suppression was validated in human hepatic cells.

(A) β-Galactosidase activity was measured in HepG2 cells transfected with WT or mutated pcDNA3.1/–816+229mSCD1 followed by the treatment of PA and poly(I:C). (B) HepG2 cells expressing WT pcDNA3.1/–816+229mSCD1 and vector or IRF3(5D) plasmids were treated with or without PA (250 μM) and poly(I:C) (2.5 μg/ml) for 8 hours, and reporter activity was measured. (C and D) HepG2 cells were treated as in (B), and SCD1 protein with β-tubulin as loading control (C) and TG content (D) were measured. (E and F) HepG2 cells transfected with scrambled RNA or IRF3 siRNA were treated with or without PA and poly(I:C) for 8 hours. IRF3 and SCD1 protein abundance (E) and TG content (F) were measured. n = 3 to 4 independent experiments for (A) to (F). *P < 0.05 and **P < 0.01. Representative images are shown.

Poly(I:C) prevented the development of HFD-induced hepatic steatosis in mice

Because poly(I:C) inhibited palmitic acid–induced Scd1 expression and lipid accumulation in hepatocytes, we next tested whether administration of poly(I:C) prevented HFD-induced hepatic steatosis in mice. Without affecting body weight and fat mass (fig. S8, A and B), administration of poly(I:C) significantly decreased Scd1 expression, SCD1 protein abundance, hepatic triglyceride and cholesterol accumulation, and fasting blood glucose in wild-type mice, protective effects of poly(I:C) that were markedly diminished in Trif−/− mice (Fig. 7, A to C, and fig. S8C). This dosage of poly(I:C) did not aggravate hepatic inflammation (as assessed by analysis of the expression of inflammatory cytokine–encoding mRNAs) or serum ALT activities in mice of both genotypes but lowered AST activity in Trif−/− mice (Fig. 7D and fig. S8, D to F). These results suggest a possible pharmacological option for treating hepatic steatosis by selective activation of TRIF in hepatocytes.

Fig. 7 Poly(I:C) ameliorated HFD-induced lipid accumulation.

(A to D) Six-week-old WT and Trif−/− mice were injected with phosphate-buffered saline (PBS) or poly(I:C) (5 μg/g) intraperitoneally three times a week while being fed an HFD for 4 weeks. Hepatic Scd1 expression (A), SCD1 protein abundance (B), TG and TC content (C), and serum AST and ALT activities (D) were measured. n = 8 to 10 mice per genotype and treatment condition for (A) to (D). *P < 0.05 and **P < 0.01. (E) Working model. IRF3 in response to TLR/TRIF signaling serves as a transcriptional suppressor of Scd1 in hepatocytes, resulting in the suppression of lipogenesis.


Here, we show that the increased Scd1 expression in hepatocytes in TRIF-deficient mice was attributed to the worsened hepatic steatosis under HFD. Activation of TRIF pathway stimulated phosphorylation and transcriptional activity of IRF3, which acted as a suppressor on the promoter region of Scd1 gene. Treatment with poly(I:C) resulting in phosphorylation of IRF3 suppressed both HFD- and palmitic acid–induced SCD1 expression and lipid accumulation in liver and hepatocytes (Fig. 7E).

NAFLD is a commonly underdiagnosed medical problem because, in most situations, it is asymptomatic. Currently, there is no definitive treatment for NAFLD. Preventing the progression to nonalcoholic steatohepatitis (NASH) or fibrosis using anti-inflammatory and antioxidant therapies is the goal of treatment for susceptive patients with obesity or diabetes. TLRs play a key role in the progression of simple hepatic steatosis to NASH by driving the hepatic inflammation, and their antagonists have been considered as potential treatments (2, 5, 29, 30). Hepatocyte-specific deficiency of TLR4 improves glucose tolerance and insulin sensitivity and ameliorates hepatic steatosis and adipose tissue inflammation in HFD-induced obesity (10). Here, we showed that suppression of SCD1 in hepatocytes was mediated by TLR3/TRIF but not TLR4/TRIF (fig. S5D). The hepatocyte-specific ablation of TLR4 would yield a double benefit by suppressing inflammation through the TLR4/MyD88 pathway and inhibiting lipogenesis through the intact TLR3/TRIF pathway. The TRIF pathway can be activated through TLR3 by endogenous RNA and gut-derived virome during HFD-induced obesity (23, 24). High-fat intake can weaken the gut barrier, resulting in penetration of the intestinal bacteria and viruses and their products into the bloodstream, providing a panel of TLR ligands (24, 31, 32). Adipose tissue expansion during obesity is associated with necrosis and apoptosis of adipocytes, and cellular RNA released by necrotic tissue can activate TLR3 (23, 33, 34). We found that circulating RNAs were increased in HFD-fed mice (Fig. 2G), which may be derived from the turnover of adipocytes. Circulating RNAs are encapsulated in microvesicles or exosomes (35). We showed that Lipofectamine-complexed RNA suppressed palmitic acid–induced SCD1 expression and triglyceride accumulation in a TRIF-dependent manner (Fig. 2, H and I). The activation of TLR3/TRIF during HFD-induced obesity may intrinsically provide a protective mechanism against hepatic lipid accumulation. The absence of such a pathway aggravated HFD-induced hepatic steatosis (Fig. 1, A to C).

The prevalence of NAFLD in diabetic and obese individuals is high, and their risk of developing into NASH and fibrosis is greater than that of the general public (36). Because there is no confirmative diagnostic tool to identify the initiation of NASH, intervention at the early stage of benign hepatic steatosis would be beneficial in these patients. Several Scd1-deficient mouse models do not develop diet-induced hepatic steatosis (25, 37), and inhibition of SCD1 using antisense oligonucleotide can also prevent HFD-induced hepatic insulin resistance and obesity (38, 39). Pharmacological approaches to inhibiting SCD1 for treating metabolic diseases have been explored but have been unsuccessful due to the potential nonspecific actions on other peripheral tissues such as vascular wall and pancreas (40). Here, we found that Scd1 expression was inhibited by intraperitoneal injection of poly(I:C), which is a mimetic of double-stranded RNA (dsRNA). A cell- and pathway- specific synthetic ligand mimicking our observed poly(I:C) effect would be an ideal SCD1 inhibitor. Hepatocyte-specific delivery of siRNA has been explored, and, for example, siRNA conjugated to triantennary N-acetylgalactosamine is effectively taken up by the liver due to binding to a liver-specific asialoglycoprotein receptor (41). Specific delivery of a dsRNA mimetic to hepatocytes appears to be feasible. Moreover, a stereochemically altered mimetic of LPS, CRX-527, can activate the TRIF pathway without stimulating the production of MyD88-mediated proinflammatory molecules (42). The architecture of the TIR domain signalosome reveals that MyD88- and TRIF-dependent pathways cannot be simultaneously induced upon TLR4 activation because of the shared binding site on a single TLR4 dimer (43). Because deficiency of TLR4 yields a protective anti-inflammatory effect in metabolic diseases, a TLR4 antagonist with intact TRIF-dependent effects could ameliorate NAFLD and prevent progression to NASH. The detailed understanding of how the known ligands stereochemically interact with different TLRs and trigger the downstream pathway would help to precisely design small-molecule drugs.

Several key metabolic transcription factors can independently regulate Scd1 expression including sterol regulatory element–binding protein transcription factor 1c (SREBP-1c), peroxisome proliferator–activated receptor α (PPARα), and liver X receptor (26, 44). We found that IRF3 directly bound to the Scd1 promoter and suppressed its transcription. The transcription-suppressing activity of IRF3 can occur by different mechanisms (15, 45). For example, poly(I:C) can suppress the transcription of Rxra through IRF3, which stimulates a transcriptional suppressor, Hes1 (Hes family BHLH transcription factor 1), leading to the recruitment of transcriptional repression machinery to the Rxra promoter (15). Activated IRF3 also interacts with Smad3 and disrupts functional Smad3 transcription complexes by competing with co-regulators (45). PPARα and SREBP-1c binding sites are in close proximity of the IRF3 binding motif on the Scd1 promoter. The binding of IRF3 could physically obstruct the attachment of other transcription activators on the promoter region, resulting in inhibition of transcription. Global knockout of IRF3 promotes HFD-induced hepatic insulin resistance and steatosis, suggesting that the protective effect of IRF3 is mediated by suppressing inflammation because IRF3 can prevent nuclear factor κB (NFκB) activation by interacting with inhibitor of NFκB kinase β (46). These studies suggest that activation of IRF3 in liver results in both metabolic and anti-inflammatory effects. Moreover, our finding of the attenuation of SCD1 by TLR3/TRIF/IRF3 pathway may also be considered as an immune response. Hepatitis C virus (HCV), a single-stranded RNA virus able to generate dsRNA, can hijack host lipogenesis for its own viral assembly (47). Host SCD1 activity can facilitate HCV viral replication, and supplementation of the products of SCD1 such as oleate and palmitoleate restores HCV replication in Scd1-knockdown hepatocytes (48). TRIF/IRF3-induced suppression of SCD1 may limit viral infection by counteracting the virus-hijacked metabolic pathway. Of the HCV genotypes, infection with HCV genotype 3 (GT3) is highly associated with hepatic steatosis that is more prone to develop into fibrosis (49). The degree of steatosis is proportional to the virus load (50). A dsRNA mimetic targeted to the liver would be an ideal adjuvant treatment for this particular HCV genotype. By activating the TRIF pathway, it would inhibit hepatic lipogenesis while maintaining interferon production, yielding a dual benefit against HCV GT3 infection. Activation of TRIF signaling can counterbalance the ER stress–induced protein translation suppression and delay apoptosis, and such mechanism is essential to maintain proper secretory functions of both immune and metabolic cells in the early stages of metabolic stress (12, 14).

Activation of TLR pathways by microbial and endogenous stimuli in metabolic disorders (2, 3, 51) is detrimental because of the release of inflammatory molecules from immune cells (2, 17, 22). TLRs are present in various nonimmune cell types and have unique functions (16, 52). Here, we showed that activation of the TRIF-dependent TLR pathway suppressed Scd1 expression and prevented hepatic lipid accumulation and that the absence of this pathway exacerbated hepatic steatosis under metabolic stress. It is important to take this metabolic function of TLRs into consideration when using antagonists to ameliorate metabolic inflammation in chronic diseases.



C57BL/6J wild-type mice and Trif−/− male mice (stock no. 005037) were purchased from the Jackson Laboratory. All animal experimental procedures were approved by the Committee on the Use of Live Animals in Teaching and Research of the University of Hong Kong. For the HFD model, mice were co-housed and fed a 60% kcal fat diet purchased from Research Diets Inc. for 10 weeks. In one set of experiment, poly(I:C) (5 μg/g) (P9582, Sigma-Aldrich) was injected intraperitoneally into mice three times a week starting on the day when the HFD was introduced until euthanasia after 4 weeks of HFD. For experiments using lentivirus, mice were fed the HFD for 10 days before lentiviral infection (3.6 × 105 transduction units of Lv-SCD1-siRNA or Lv-Scr-RNA per mouse) and continued on the HFD for another 3 weeks because the effectiveness of lentivirus lasts for about 3 weeks. For experiments using adenovirus, mice were fed the HFD for 4 weeks, and the wild-type and half of the Trif−/− mice received Ad-GFP (5 × 109 vp per mouse), whereas the other half of Trif−/− mice received Ad-TRIF (5 × 109 vp per mouse). The HFD was continued for two more weeks.

Bone marrow transplantation

Mice received total body irradiation of 9 gray in a Lucite ionization chamber. Bone marrow cells were isolated from the femurs of wild-type and Trif−/− mice and resuspended in serum-free RPMI medium. A total of 5 × 106 cells were injected into each of the irradiated recipient mice through the lateral tail vein on the day of irradiation. The mice were kept on antibiotic-supplemented water for 2 weeks after irradiation (5). To verify that irradiation was successful, two mice were not injected with bone marrow and died 8 days after initial irradiation. HFD feeding was initiated after the recovery. Successful transplantation was further confirmed by genotyping the bone marrow collected after euthanasia. The irradiated mice were co-housed based on the genotypes of the recipients due to the large number of mice and cage size limitations, and the cages of mice with different background genotypes were swapped periodically to achieve co-housed conditions of four groups of mice (WT/WT-BM, Trif−/−/WT-BM, WT/Trif−/−-BM, and Trif−/−/Trif−/−-BM).

Isolation of primary hepatocytes and cell culture

Primary hepatocytes were isolated from six- to eight-week-old wild-type or Trif−/− mice as previously mentioned with modification (53). Briefly, liver was perfused with PBS through the portal vein with an outlet at the inferior vena cava, followed by digestion with type I collagenase (~40 collagen digestion units/ml, Sigma-Aldrich, C2674). Digested liver was minced and filtered, and cell suspension was collected and further separated using Percoll (Sigma-Aldrich). Hepatocytes were collected after centrifugation at 100g for 5 min and cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal bovine serum (FBS) on collagen-coated plate (BD BioCoat Collagen). Cells were incubated overnight at 37°C under 5% CO2 and 95% relative humidity. HeLa and HepG2 cells (American Type Culture Collection) were routinely maintained in DMEM with 10% FBS. Mouse TLR3 siRNA and human IRF3 siRNA were purchased from Santa Cruz Biotechnology Inc. and Qiagen Inc., respectively.

Isolation of serum RNA and RNA for treatment of hepatocytes

Serum RNA and adipose tissue RNA for in vitro experiment were isolated using a commercially available kit (Zymo Research). RNA concentrations were measured by NanoDrop 2000 (Thermo Fisher).

Histochemical staining

Livers were collected and fixed in 10% buffered formalin followed by embedding in paraffin. Paraffin-embedded sections (5 μm thick) were prepared and subjected to H&E staining. Another portion of the liver was embedded in frozen Tissue-Tek OCT compound, and 7-μm frozen sections were subjected to Oil Red O staining or GFP detection.

Lipid extraction and measurement

Lipids were extracted with a chloroform/methanol/water ratio of 4:3:2, and dried lipid pellets were resuspended in 100% ethanol. Hepatic and cellular triglyceride and total cholesterol were assessed by using commercially available kits (Stanbio Diagnostics). Triglyceride and cholesterol amounts were normalized to liver wet weight or cellular protein.

Fasting blood glucose and liver function test

Mice were fasted for 5 hours for blood glucose measurement using a glucometer (Accu-Chek Performa, Roche Diagnostics). Serum AST and ALT activities were measured using commercially available kits (Stanbio Diagnostics).

Real-time PCR

Total RNA from hepatocytes or liver tissues were isolated using TRIzol reagent (Invitrogen). RNA was reverse-transcribed into complementary DNA (cDNA) using oligo(dT) and ImProm-II Reverse Transcription System (Promega). Real-time PCR was conducted using the SYBR Green PCR reagent (Roche) and StepOnePlus System (Applied Biosystems). The mRNA expression of target genes was normalized to Actb. The sequences of the primers are listed in table S2.

Western immunoblotting

Immunoblots were conducted as described previously (12). Briefly, cultured cells were lysed with Laemmli sample buffer (Bio-Rad), and livers were homogenized in a lysis buffer containing 20 mM tris (pH 7.4), 150 mM NaCl, 1% Triton X-100, and phosphatase and protease inhibitor cocktails. Protein samples were separated by electrophoresis on a 10% SDS polyacrylamide gel and transferred to polyvinylidene difluoride membranes. The membranes were probed with the indicated primary antibodies against SCD1, IRF3, or phospho-IRF3(Ser396) (Cell Signaling), and the protein bands were detected with horseradish peroxidase–conjugated secondary antibodies (Jackson ImmunoResearch) and WesternBright ECL reagent (Advansta). The membranes were reprobed with antibody against β-actin (Sigma-Aldrich) or β-tubulin (Cell Signaling) to detect differences in protein loading. Densitometry analysis of the gels was carried out using ImageJ software from the National Institutes of Health.

Adenovirus construction

Adenovirus encoding TRIF (Ad-TRIF) was constructed by Welgen Inc. Briefly, mouse Trif cDNA (GE Dharmacon) was subcloned into pEntCMV-Ef1aGFP vector predigested with Pme 1. The ligation mixture was transformed into Escherichia coli, and the positive clones were screened with Eco R5 and sequenced. The pEntCMV-mTRIF-Ef1aGFP was treated with LR Clonase II enzyme (Invitrogen) and ligated to a pAd-REP plasmid that contains the remaining adenovirus genome. The recombination products were transformed into E. coli cells. After overnight incubation, positive clones were selected, and cosmid DNA was purified. The purified cosmid DNA was digested with Pac 1 and then transfected into human embryonic kidney (HEK) 293 cells. The adenovirus plaques were seen 7 days after transfection. The low titer virus was further amplified and purified. The purified high titer virus (1012 vp/ml) was diluted into 109 vp/ml to infect HEK293 cells followed by sequencing to validate the absence of contamination.

Lentivirus construction

Vector carrying scrambled or SCD1 siRNA was purchased from Applied Biological Materials Inc. Lentiviruses were produced by transfecting HEK293TN cells with the purchased vectors expressing GFP and three helper plasmids (pVSV-G, pPACKH1-GAG, and pPACKH1-REV). The medium were harvested 48 and 72 hours after initial transfection and filtered through a 0.22-μm pore size filter. The lentivirus was precipitated by adding 5% polyethylene glycol–8000 and 0.15 M NaCl to the medium followed by incubation overnight at 4°C along with mixing every 20 min. The virus was pelleted by centrifugation at 7000g for 15 min (54). Titer was determined by performing flow cytometry on transduced HeLa cells.

Plasmid construction and transfection

The plasmid pGL3/–1537+155mSCD1 was provided by J. Ntambi (University of Wisconsin). The promoter regions (−816 + 229) and (−469 + 229) were constructed by PCR using 5′-CGGGGTACCTGATTCCTTAGTCCCTTTTCTTGGA-3′ and 5′-CGCGGATCCTGGTGTAGGCGAGTGGCGGAACTGC-3′ and 5′-CGGGGTACCCCTCACTTCTTTCGATGCGATTTCC-3′ and 5′-CGCGGATCCTGGTGTAGGCGAGTGGCGGAACTGC-3′, respectively. The PCR product was subcloned into pcDNA3.1/lacZ (Invitrogen) at Kpn I and Bam HI restriction sites. The IRF3 binding site at −593 was mutated using GENSTART site-directed mutagenesis system (Invitrogen) and the primers 5′-CTGAAGGGATACACTATTACCCCTCCGGGTCAGAGCCCTGGG-3′ and 5′-CCCAGGGCTCTGACCCGGAGGGGTAATAGTGTATCCCTTCAG-3′. β-Galactosidase activity was determined by a colorimetrical method using o-nitrophenyl-β-d-galactopyranoside (Sigma-Aldrich) as substrate. The plasmids encoding wild-type and mutated IRF3 including pEGFPC1-IRF3(WT) and pEGFPC1-IRF3(5D) were provided by J. Hiscott (University of Florida). All the plasmids were verified by sequencing service provided by Beijing Genomics Institute.

Bioluminescence imaging

The luciferase reporter plasmid (pGL3/–1537+155mSCD1) was injected into mice by hydrodynamic injection. Briefly, DNA (50 μg/ml) in sterile PBS was injected into mice within 10 s through the tail vein, and the total volume injected was maintained within 8 to 10% of body weight. Mice were allowed to recover for 24 hours before receiving an intraperitoneal injection of d-luciferin (1 μg/g) (Promega) under anesthesia. Images were captured by the IVIS Spectrum In Vivo Imaging System (PerkinElmer) and analyzed with the Living Image Software 4.4 (PerkinElmer).

Chromatin immunoprecipitation

Chromatin immunoprecipitation was performed as previously described with minor modification (53). Cells were treated with 3.7% formaldehyde to cross-link DNA and protein, followed by lysis with a buffer containing 150 mM NaCl, 50 mM tris-HCl (pH 7.5), 5 mM EDTA, 0.5% (v/v) NP-40, and 1% (v/v) Triton X-100. Chromatin was prepared and sheared by sonication with Bioruptor (Diagenode, Belgium) and subjected to immunoprecipitation with an antibody against IRF3 (Cell Signaling) or rabbit IgG control and protein A magnetic beads (Biotool). An aliquot of chromatin from each sample was reserved for analysis of DNA input before immunoprecipitation. The specific protein-DNA complexes were eluted from the magnetic beads, and cross-links were reversed. DNA was precipitated from the samples, including from the previously reserved aliquots, and subjected to PCR amplification and detection of the segment of Scd1 promoter region using the primers 5′-CCAATGAGTGAGTGCAGTTGTA-3′ (forward) and 5′-GCATCGAAAGAAGTGAGGAAGA-3′ (reverse).

Statistical analysis

Statistical analyses were performed using SPSS version 23.00. Data were presented as means ± SEM. One-way analysis of variance (ANOVA) was applied for comparisons between multiple experimental groups, followed by post hoc analysis using Tukey post hoc test for data with equal variance or Games-Howell for data with unequal variance. Data with small sample size were analyzed using the Kruskal-Wallis test, a nonparametric one-way ANOVA. An unpaired Student’s t test was applied for two-group comparison with normal distribution. P values less than 0.05 were considered to indicate statistically significant differences.


Fig. S1. Trif−/− mice have similar body weight and fat mass as wild-type mice but increased fasting blood glucose after HFD feeding.

Fig. S2. Hepatic inflammation, serum insulin concentrations, and serum lipids were similar between HFD-fed wild-type and Trif−/− mice.

Fig. S3. HFD did not increase CHOP abundance or apoptosis in wild-type and Trif−/− mice.

Fig. S4. Lipogenic gene expression did not increase in wild-type and Trif−/− mice fed an NCD.

Fig. S5. Poly(I:C)–stimulated phosphorylation of IRF3 and suppressed palmitic acid–induced Scd1 expression in hepatocytes in a TLR3-dependent manner.

Fig. S6. RNA was detected in the serum of HFD-fed mice and taken up by hepatocytes.

Fig. S7. Lv-SCD1-siRNA decreased SCD1 abundance.

Fig. S8. Activation of TRIF by poly(I:C) decreased fasting blood glucose concentrations in HFD-fed wild-type mice without significantly affecting body weight, fat mass, or hepatic inflammation.

Table S1. IRF3 binding motifs in the Scd1 promoter in different species.

Table S2. Primer sequences for specific genes.


Acknowledgments: We would like to thank I. Tabas (Columbia University) for his valuable suggestions on this manuscript. The pGL/–1537+155mSCD1 and the IRF3 plasmids [pEGFPC1-IRF3(WT) and pEGFPC1-IRF3(5D)] were provided by J. Ntambi (University of Wisconsin) and J. Hiscott (University of Florida), respectively. Funding: This study was supported by a matching fund offered by the Research Centre of Heart, Brain, Hormone and Healthy Aging of the University of Hong Kong (200007252), the National Key Basic Research Development Program 973 (2015CB553603), the Hong Kong Research Grants Council/Collaborative Research Fund (C7055-14G), and a matching grant for the State Key Laboratory of Pharmaceutical Biotechnology from the University of Hong Kong. Author contributions: C.W.W. designed the experiments. J.C., J.L., and J.H.C.Y. performed the experiments. J.K.W.L assisted with the RNA labeling experiment. J.C., J.L., J.H.C.Y., C.-M.W., B.D., and C.W.W. analyzed the data. C.W.W., B.D., and A.X. wrote the paper. Competing interests: The authors declare that they have no competing interests.

Stay Connected to Science Signaling

Navigate This Article