Thyrotropin increases hepatic triglyceride content through upregulation of SREBP-1c activity

Background & Aims: Hallmarks of non-alcoholic fatty liver disease (NAFLD) are increased triglyceride accumulation within hepatocytes. The prevalence of NAFLD increases steadily with increasing thyrotropin (TSH) levels. However, the underlying mech- anisms are largely unknown. Here, we focused on exploring the effect and mechanism of TSH on the hepatic triglyceride content. Methods: As the function of TSH is mediated through the TSH receptor (TSHR), Tshr—/— mice (supplemented with thyroxine) were used. Liver steatosis and triglyceride content were analysed in Tshr—/— and Tshr+/+ mice fed a high-fat or normal chow diet, as well as in Srebp-1c—/— and Tshr—/—Srebp-1c—/— mice. The expres- sion levels of proteins and genes involved in liver triglyceride metabolism was measured.

Results: Compared with control littermates, the high-fat diet induced a relatively low degree of liver steatosis in Tshr—/— mice. Even under chow diet, hepatic triglyceride content was decreased in Tshr—/— mice. TSH caused concentration- and time-dependent effects on intracellular triglyceride contents in hepatocytes in vitro. The activity of SREBP-1c, a key regulator involved in  triglyceride metabolism and in the pathogenesis of NAFLD, was significantly lower in Tshr—/— mice. In Tshr—/—Srebp-1c—/— mice, the liver triglyceride content showed no significant difference compared with Tshr+/+Srebp-1c—/— mice. When mice were injected with forskolin (cAMP activator), H89 (inhibitor of PKA) or AICAR (AMPK activator), or HeG2 cells received MK886 (PPARa inhibi- tor), triglyceride contents presented in a manner dependent on SREBP-1c activity. The mechanism, underlying TSH-induced liver triglyceride accumulation, involved that TSH, through its receptor TSHR, triggered hepatic SREBP-1c activity via the cAMP/PKA/ PPARa pathway associated with decreased AMPK, which further increased the expression of genes associated with lipogenesis. Conclusions: TSH increased the hepatic triglyceride content, indicating an essential role for TSH in the pathogenesis of NAFLD.

Introduction

Non-alcoholic fatty liver disease (NAFLD) is the non-alcohol- induced accumulation of extra fat in the liver, resulting in dis- eases ranging from benign steatosis to advanced cirrhosis and cancer. Importantly, people with NAFLD have an increased risk of developing cardiovascular problems, such as heart attack and stroke [1]. The prevalence of NAFLD ranges from 9 to 36.9% of the population in different regions of the world [2]. Approximately 20% of the United States population suffers from NAFLD, and the prevalence of this condition is increasing [3]. The exact cause of NAFLD remains unknown. However, endo- crine disorders such as obesity, insulin resistance, and hypothy- roidism likely play a strong role in the pathology of this disease [4,5].
Subclinical hypothyroidism (SCH), characterized by elevated thyrotropin (TSH) and normal free thyroxine (T4) levels, has recently been demonstrated to be a risk factor for NAFLD [6,7]. The prevalence of NAFLD gradually increased with rising TSH lev- els [6]. Moreover, a prospective case-control study showed that over, clinical studies indicated a positive association between TSH and serum TG level [8,9]. As TSH is the only thyroid function component affected in SCH, we endeavoured to investigate whether TSH induces the accumulation of extra fat in liver cells in the context of NAFLD development and aimed to determine the underlying molecular mechanism. Notably, no related studies have been reported to date.

Accumulation of triglycerides within hepatocytes is the hall- mark of NAFLD and the liver is a central organ in the regulation of triglyceride metabolism [2,10]. Moreover, sterol regulatory element which directly activates the expression of more than 30 genes, ded- icated to fatty acid uptake and triglyceride synthesis [11]. Increased SREBP-1c levels were found in patients with histologically diag- nosed NAFLD [12]. SREBP-1c was regulated at multiple levels, such as the proteolytic cleavage of SREBP-1c precursors and post-transla- tional modification of mature SREBP-1c [13]. Studies have reported that SREBP-1c expression decreased in peroxisome proliferator-activated receptor a (PPARa)-null mice compared with wild type, suggesting the PPARa-dependent induction of hepatic fatty acid synthesis and SREBP-1c activation [14,15]. Moreover, it is known that PPARa agonists enhance the activity of the Srebp-1c promoter through direct binding with the DR1 motif [16], and AMP-activated protein kinase (AMPK) induced SREBP-1c phosphorylation at Ser372 associated with the accumulation of nuclear SREBP-1c [17]. We hypothesized that TSH might play a novel role in the regula- tion of the triglyceride content in liver, which, at least partially, involves the development of NAFLD. Therefore, in the present study, we characterized the effect and mechanism of TSH on the hepatic triglyceride content in TSH receptor knockout (Tshr—/—), Srebp-1c knockout (Srebp-1c—/—) and Tshr—/—Srebp-1c—/— double knockout mice. Compared with wild type (Tshr+/+) mice, Tshr—/— mice exhib- ited a relatively low degree of liver steatosis. Moreover, TSH increased the hepatic triglyceride content via the thyrotropin recep- tor (TSHR), in which SREBP-1c activation induced by the cAMP/PKA/ PPARa signalling pathway, associated with declining AMPK activity,played an indispensable role. These findings highlight a novel phy- sio-pathological role for TSH in the regulation of triglyceride metab- olism in the liver and suggest that TSH or hepatic TSHR might have key therapeutic importance in preventing fatty liver disease.

Materials and methods

See Supplementary Materials and methods section.

Results

The function of TSH is mediated through the highly specific TSHR [18]. In a previous study, we demonstrated the presence of func- tional TSHRs in hepatocytes [19]. To examine the effect of TSH on liver steatosis, we generated a Tshr knockout mouse model.

Liver steatosis is attenuated in Tshr—/— mice and TSH promotes triglyceride accumulation

First, we fed mice with a high-fat diet. After 20 weeks, Tshr—/— mice were generally lean compared with Tshr+/+ mice (Supplementary Fig. 1A). Tshr+/+ mice exhibited a uniformly pale yellow fatty liver, while Tshr—/— mice had a relatively normal liver (Fig. 1A and Sup- plementary Fig. 1B). Importantly, the hepatic triglyceride content (Fig. 1B) and serum total triglyceride levels (Supplementary Fig. 1C) were decreased in Tshr—/— mice. Correspondingly, Tshr—/— mice showed reduced fat accumulation in the hepatic intracellular vacuoles (Fig. 1C and D). Notably, both THRb protein expression and mRNA levels of genes that are under direct control of thyroid hormones (deiodinase iodothyronine type 1 [Dio1] and the thyroid hormone-inducible hepatic protein [Spot14] proved not to be detectably changed between Tshr+/+ and Tshr—/— mice (Supplemen- tary Fig. 1D and E). These results indicated that the absence of TSHR effectively improved high-fat diet-induced obesity and liver steatosis independent of thyroid hormones.

We next explored the role of TSH in liver triglyceride metab-
olism in Tshr—/— and Tshr+/+ mice under chow diet. Similarly, less
lipid droplet accumulation (Fig. 1E) and a reduction in triglycer- ide content of approximately 50% were observed in the liver of Tshr—/— mice (Fig. 1F). Therefore, the absence of TSHR, even with a chow diet, still reduced liver steatosis.
HepG2 is a human hepatoblastoma cell line with a wide vari- ety of liver-specific metabolic responses to different types of stimuli, such as drugs. TSH caused concentration- and time- dependent effects on the intracellular triglyceride contents in HepG2 cells (Supplementary Fig. 1F and G). Accordingly, electron microscopy analysis revealed only a few small lipid droplets in the control cells. TSH treatment, however, induced the produc- tion of a large number of medium-to-large-sized lipid droplets (Supplementary Fig. 1H). These results confirmed that TSH directly increased the intracellular triglyceride content.

Simultaneously, we proved the indispensable role of SREBP-1c in vitro using both the dominant negative form of SREBP-1c (DN-SREBP-1c) and SREBP-1 siRNA. When a plasmid encoding DN-SREBP-1c was used, the TSH-mediated increase in hepatic tri- glyceride content was completely prevented (Fig. 2F). Similarly, SREBP-1 siRNA resulted in a 60% reduction in SREBP-1 protein expression (Supplementary Fig. 2E) and a disappearance of the TSH-mediated increase in intracellular triglycerides (Supplemen- tary Fig. 2F). Hence, SREBP-1c was necessary for TSH-induced liver triglyceride accumulation.

Accumulating evidences suggest that PPARa activation promotes lipogenesis and SREBP-1c expression [14,21,22]. Subsequently, we determined whether PPARa was involved in the regulation of SREBP-1c by TSH.

PPARa mediates the regulation of SREBP-1c by TSH and PKA mediates phosphorylation of PPARa

First, we observed PPARa levels in vivo. Regardless if mice were fed with high-fat (Supplementary Fig. 3A) or chow (Fig. 3A) diet, liver PPARa protein expression was remarkably decreased in Tshr—/— mice, consistent with a reduced expression of SREBP-1c protein. Cpt1a mRNA levels, directly regulated through PPARa, also decreased in Tshr—/— mice (Supplementary Fig. 3B).

In either Tshr+/+ mouse primary hepatocytes or HepG2 cells, TSH stimulated the expression of PPARa at the mRNA and protein level (Fig. 3B, Supplementary Fig. 3C and D). Strong PPARa fluo- rescence was observed in both the nucleus and cytoplasm in pri- mary hepatocytes (Supplementary Fig. 3E). However, the effects of TSH were not observed in hepatocytes obtained from Tshr—/— mice due to the absence of TSHR (Fig. 3B).

To determine the role of PPARa in TSH-induced hepatic lipogenesis, we applied MK886, a specific PPARa inhibitor. Notably, pretreatment of HepG2 cells with MK886 attenuated the effect of TSH on the activation of SREBP-1 by approximately 70% (Fig. 3C). The stimulation of downstream signals induced by TSH were also partly reversed (Supplementary Fig. 4A). The luciferase activity assay showed a significant increase in SREBP-1c promoter activity in HepG2 cells upon TSH treatment. When cells were preincubated with MK886, this activity was sig- nificantly attenuated, but not completely abolished (Fig. 3D). These results implicated PPARa mediated TSH-induced triglycer- ide synthesis in hepatocytes and further supported the conclu- sion that PPARa promotes hepatic lipogenesis.

The activation of the TSHR through TSH via G-protein-coupled receptor pathways results in the activation of the cAMP/PKA pathway [18]. To determine whether TSH-mediated PPARa acti- vation occurred via the cAMP pathway, we explored the effects of forskolin (cAMP activator) on the expression of PPARa in vivo. After Tshr—/— or Tshr+/+ mice were treated with forskolin, respectively, the mRNA levels of hepatic SREBP-1c, PPARa and their downstream genes were induced (Supplementary Fig. 4B). Accordingly, the protein expression of PPARa and SREBP-1 was enhanced in a time-dependent manner (Fig. 3E). However, when mice were treated with H89, a selective and potent inhibitor of PKA, hepatic SREBP-1 protein expression exhibited the opposite trend (Fig. 3F).

PPARa is exclusively phosphorylated on serine residues in vivo with no observed activity on threonine or tyrosine residues [23]. As shown in Supplementary Fig. 4C, using co-immunoprecipita- tion, forskolin increased serine phosphorylation in a dose-depen- dent manner in HepG2 cells. Therefore, the phosphorylation of serine residues was involved in the upregulation of PPARa activa- tion. Together with our previous results that TSH elevates intra- cellular cAMP levels via TSHR [24], we hereby demonstrated that TSH induced SREBP-1c expression via the TSHR/cAMP/PKA/ PPARa pathway.

AMPK is also involved in TSH-mediated hepatic triglyceride accumulation

The fact that preincubation with MK886 did not completely abol- ish the TSH-induced activation of SREBP-1 (Fig. 3E) implicated that pathways, other than PKA/PPARa, might be involved in TSH-mediated triglyceride accumulation. As AMPK is a major reg- ulator of lipid metabolism and SREBP-1 is negatively regulated by AMPK via inhibition of SREBP-1c translocation into nuclei [17], we decided to assess the potential role of AMPK in TSH-mediated triglyceride accumulation. The injection of AICAR (AMPK activa- tor) into mice resulted in an approximate 40% decrease in hepatic triglyceride content (Fig. 4A), as well as in the reduced expression of SCD1, the target gene of SREBP-1c. The expression of apolipo- protein B (ApoB), which is responsible for fatty acid export, was also reduced (Supplementary Fig. 5A).

Notably, the phosphorylation of AMPKa at Thr172, of acetyl-CoA carboxylase 1 (ACC1), a downstream target protein of AMPK, at Ser79 [25] and of SREBP-1c at Ser372 was significantly increased in Tshr—/— mice relative to Tshr+/+ mice (Fig. 4B). How- ever, the expression levels of the above phosphorylated proteins in the liver of Tshr+/+ and Tshr—/— mice were both stimulated upon AICAR treatment (Fig. 4C). These results suggest that AMPK was involved in TSH-induced SREBP-1c phosphorylation via TSHR.
To evaluate the association between AMPK and PPARa, we treated Tshr+/+ mice with AICAR. Notably, AICAR treatment reduced the expression of mature SREBP-1c, whereas the expression of PPARa was not affected (Supplementary Fig. 5B).Taken together, these results demonstrate that TSH-induced SREBP-1c expression was associated with AMPK.

Discussion

Here, we discovered a novel extra-thyroidal role of TSH in regu- lating triglyceride metabolism. TSH increased hepatic triglyceride content via TSHR, in which SREBP-1c activation induced by cAMP/PKA/PPARa signalling pathway associated with AMPK played an indispensable role (Supplementary Fig. 6). The ablation of TSHR alleviated liver steatosis in both high-fat and chow-fed mice. Our findings indicate a potential role for TSH in the pathogenesis of liver steatosis and might be provide an additional strategy for treatment of NAFLD.

Thyroid hormones regulate free fatty acid and triglyceride synthesis [26,27], and there is a feedback loop between TSH and thyroid hormones in the body [28]. Therefore, the role of thy- roid hormones must be ruled out to observe the direct effect of TSH on triglyceride metabolism. We used Tshr—/— mice as our experimental model according to the study by Marians et al. [29]. Thyroid hormone (T4) was supplemented after weaning to eliminate the influence of thyroid hormones. Total serum T4 lev- els were maintained at the same level in 6- to 8-week-old Tshr+/+ and Tshr—/— mice [30], indicating that the supplemented T4 diet was appropriate. Moreover, both THRb protein expression and mRNA levels of genes that are under direct control of thyroid hor- mones (TH) (Doi1 and Spot14) were not detectably changed between Tshr+/+ and Tshr—/— mice, which indicated that there was no significant difference in the biological effects of TH between Tshr—/— mice supplemented with TH and Tshr+/+ mice. In addition, considering the complexity of the in vivo conditions, we used mouse primary hepatocytes and HepG2 cells to observe the direct effects of TSH. We obtained similar data from both the in vivo and in vitro experiments. Despite this, the administration of rats with T3 also induced a decrease in hepatic steatosis and simultaneously reduced TSHb mRNA [26]. However, Gnoni and coworkers showed that SREBP-1c protein was upregulated through non-genomic (post-transcriptional) action of T3 in HepG2 cells [31]. These findings are in line with our results as the in vivo anti-steatotic effect of T3 might be the result of a neg- ative feedback on TSH through T3, causing TSH to stimulate SREBP-1c less efficiently.

Apart from T3, TSH can stimulate the production of metabolites that could be defined as bona fide thyroid hormones, which are produced within the thyroid and are not derived from the deiodination of T4, including 3-iodothyronamine (T1AM) and 3,5-diiodo-L-thyronine (T2). Interestingly, 3,5-T2 has also been demonstrated to reduce the activity of SREBP-1c through proteo- lytic cleavage [32] or deacetylation [33] and thus to prevent hepatic steatosis. However, these metabolites may not be excreted efficiently from the thyroid follicle in absence of TSH signalling. Further studies are required to more fully define the impact of the ablation of Tshr on hepatic T1AM and 3,5-T2 levels.

The G-protein coupled receptor-mediated stimulation of the cAMP/PKA pathway has been considered the most important intracellular signalling mechanism, mediating the action of TSH [18]. In this study, an enhancement of activated PPARa was consistently observed after treating mice with forskolin and stim- ulating cells with TSH. Treating mice with H89, an inhibitor of PKA, showed the opposite effect. These changes suggest an asso- ciation between TSH and PPARa. Lazennec et al. showed that PPARa phosphorylation by the PKA pathway enhanced the acti- vation of PPARa [34]. There are several consensus phosphoryla- tion sites for PPARa, including the PKA site [23]. We examined the elevation of serine phosphorylation in PPARa proteins after PKA activation, but the precise site of the phosphorylated amino acid residues has not been further explored, indicating a limita- tion of the present study.

Our finding that PPARa activation resulted in an increased liver triglyceride content was consistent with our recent results and several published studies. Mice treated with fenofibrate (PPARa agonist) exhibit increased acetyl-CoA incorporation into hepatic fatty acids [14]. PPARa deficiency abolished normal diur- nal variations in lipogenic FASN and ACC1 expression levels [15]. Treatment with a PPARa agonist promotes 3H2O incorporation into hepatic lipids and reduces the concentration of serum tri- glycerides in wild type mice but not in PPARa-null mice [21]. Therefore, PPARa activation may increase hepatic triglyceride content and play an important role in TSH-induced triglyceride accumulation.

In addition, activation of AMPK by AICAR or metformin could inhibit lipid synthesis by repressing the transcriptional regulation of lipid-related genes including ACC1, FASN, SCD1 or GPAT in the liver [35]. Thus, the observation of decreased AMPK activity also gave a supporting explanation on the TSH-induced triglyceride increase in the liver and serum.

SREBP-1a and SREBP-1c isoforms affect liver triglyceride syn- thesis [36]. To determine whether one or both isoforms were involved in TSH regulation, we designed several experiments. The results showed that SREBP-1c, not SREBP-1a, was affected at the mRNA level in Tshr—/— mice and hepatocytes after TSH chal- lenge. Importantly, in the double-knockouts, Tshr and Srebp-1c could no longer mediate a decrease in liver TG content and mutating the SREBP-1c sequence diminished the TSH-induced tri- glyceride synthesis, which further confirmed the action of TSH on SREBP-1c in the liver. In addition, Gnoni et al. showed that thy- roid hormones induced SREBP-1 protein expression through non-genomic actions [31], supporting that TSH induced SREBP- 1c transcription efficiency. Mature SREBP-1c proteins target and directly bind to downstream genes encoding enzymes involved in lipogenesis, such as ACC1, FASN, SCD1, and GPAT [37]. Based on the activation of SREBP-1c through TSH, we examined the increase in the mRNA levels of these genes and quantified the
hepatic triglyceride contents, consistent with the idea that PPARa
activation promotes lipogenesis [14,21,22]. The results suggested that SREBP-1c is required for the increase in the hepatic triglyc- eride content through TSH.
The inhibition of PKA activation through H89 and the inhibi- tion of PPARa through MK886 or siRNA did not completely abolish the TSH-induced SREBP-1c activity. We propose that this effect reflects the following: (1) as stated above, SREBP-1c is reg- ulated at multiple levels [13]; (2) the transcription of SREBP-1 might be mediated through a feed forward mechanism in which mature SREBPs binds to the SRE in the promoter of the SREBP gene to induce the transcription of this gene [36]; and (3) the activity of SREBP-1c may be regulated via several pathways. Here, we discovered AMPK decreased phosphorylated SREBP-1c and increased mature SREBP-1c, which was also involved in the TSH-induced triglyceride accumulation. Other pathways, such as PKA/mTOR, may also be involved [38]. We did not perform detailed elucidation experiments to address these issues, sug- gesting another limitation of the present study.

Although the effect of TSH was weak under physiological con- ditions or low doses of TSH, which was consistent with results that the TSHR plays a marginal role in the modest increase in cAMP after TSH stimulation [39] it may be pathologically relevant and clinically significant. The results elucidated the molecular mecha- nism by which TSH increased liver triglyceride content, indicating an essential role for TSH in the pathogenesis of disorders associ- ated with lipid metabolism, particularly NAFLD. Therefore, TSH or hepatic TSHR might be a potential target in preventing fatty liver disease.

In conclusion, the present study provided evidence that TSH, through TSHR, increased hepatic SREBP-1c activity via the cAMP/PKA/PPARa pathway associated with decreased AMPK activity, leading to hepatic triglyceride accumulation. These results reveal a novel regulation of triglyceride metabolism through TSH in the liver, with pathological implications for the pathogenesis of abnormal triglyceride metabolism in SCH patients.