Regulation of Fasting Fuel Metabolism by Toll-Like Receptor 4

TLR4^−/− mice with a C57BL/6 background (16) were kindly provided by Prof. Vincent Deubel and Prof. Baoxue Ge (Institute Pasteur of Shanghai, Chinese Academy of Sciences). Wild-type (C57BL/6) mice were purchased from the Shanghai Laboratory Animal Co. Mice were housed under a 12-h dark/light cycle with free access to standard chow and water. For experiments, 6- to 8-week-old male wild-type and TLR4^−/− mice were housed individually and fasted at 9:00 a.m. Twenty-four hours later, mice were killed and serum, liver, epididymal fat, and gastrocenimus muscle were collected and snap-frozen in liquid nitrogen for further analysis. All experimental procedures and protocols were approved by the Institutional Animal Care and Use Committee of the Institute for Nutritional Sciences, Chinese Academy of Sciences.

For C75 (Sigma) treatment, mice received intraperitoneal injection of 20 mg/kg C75 dissolved in 200 μl RPMI-1640 right before fasting; control mice were injected with 200 μl RPMI accordingly. For the insulin tolerance test, mice were fasted for 24 h and blood glucose was assessed before insulin (novolin) injection (0.25 units/kg i.p.) and at 15, 30, 45, and 60 min after injection. For the pyruvate tolerance test, mice were fasted for 24 h and blood glucose were assessed before pyruvate injection (2g/kg i.p.) and at 20, 40, 60, and 80 min after injection.

O₂ and CO₂ consumption were determined in the comprehensive laboratory animal monitoring system (CLAMS; Columbus Instruments) according to the manufacturer's instructions. Animals were acclimated to the system for 18–24 h with free access to food and water, and the measurements were conducted for 24 h from 9:00 a.m. without food supply.

Total fat mass was measured by nuclear magnetic resonance with a Minispec Mq7.5 Analyzer (Bruker, Germany). Blood glucose was determined using a glumeter (Freestyle). Serum levels of TG and FFA were determined by an enzymatic triglyceride assay kit (Applygen, Beijing) and a LabAssay non-esterified fatty acids (NEFA) kit (Wako), respectively. Serum levels of lactate, alanine and pyruvate were determined by enzymatic kits (Biovision). Serum insulin levels were determined by an ELISA kit (Millipore). Tissue TG content was measured as previously described (17). Briefly, frozen liver and gastrocenimus muscle were weighted, homogenized in isopropanol, incubated at 4°C for 1 h, and centrifuged. The supernatants were collected to measure TG concentrations by the enzymatic kit.

Lipolysis studies were performed in explants from freshly isolated edipidymal fat as previously described (18). Briefly, fat explants from mice fasted for 24 h were incubated in Krebs-Ringer buffer (12 mmol/l HEPES, 121 mmol/l NaCl, 4.9 mmol/l KCl, 1.2 mmol/l MgSO₄, and 0.33 mmol/l CaCl₂) with 3.5% fatty acid–free BSA and 0.1% glucose. Glycerol (Applygen, Beijing) and NEFA (Wako) contents were measured after one-hour incubation.

Fatty acid oxidation rate was determined in muscle homogenates as previously described (19) with modifications. Briefly, muscle homogenates were incubated with reaction mixture, in which [³H]oleate was used as substrate. After incubation, reaction medium was added with 2.5 ml methanol:chloroform (1:2) and 1 ml 2 mol/l KCl/2 mol/l HCl to separate the aqueous phase, which contains ³H₂O and was taken for scintillation counting.

Actual PDC activity in gastrocenimus muscle was measured by an assay coupling with arylamine N-acetyltransferase as previously described (20). One unit of PDC activity corresponds to the acetylation of 1 μmol p-(paminophenylazo)benzene sulfonic acid/min at 30°C.

Total RNA was extracted from tissues using TRIzol reagent (Invitrogen, Carlsbad, CA) and treated with RNase-Free DNase. First-strand cDNA was synthesized with M-MLV reverse transcriptase and oligo (dT) primers. Real-time quantitative PCR was performed on an ABI Prism 7900 sequence detection system (Applied Biosystems), using SYBR Green PCR Master Mix (Applied Biosystems). The results were normalized against 36B4 gene expression. The primer sequences are available upon request.

Tissue was homogenized in radioimmunoprecipitation assay lysis buffer and centrifuged to remove the debris. Proteins were separated by SDS-PAGE, transferred to polyvinylidene fluoride membrane, and blocked in 5% nonfat milk at room temperature for 1 h. The membrane was incubated overnight at 4°C with the following primary antibodies: mouse anti-fatty acid synthase (FAS) (BD Transduction Laboratories) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (KANGCHEN, China), rabbit anti–ATP citrate lyase (ACL), IκBα, phosphorylated levels of insulin receptor, phosphorylated levels of Akt, and phosphorylated levels of glycogen synthase kinase (GSK)3β (Cell Signaling). After washing, the blots were incubated with horseradish peroxidase–conjugated anti-mouse or rabbit immunoglobulin G secondary antibody for 1 h at room temperature and then developed with SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific).

Data are presented as means ± SEM. Statistical significance was determined using two-tailed unpaired Student's t test. A value of P < 0.05 was considered significant.

We first examined the involvement of TLR4 in glucose metabolism in response to fasting. After fasted for 24 h, TLR4^−/− mice exhibited significantly lower blood glucose levels than wild-type controls (wild type: 82 ± 2.42 mg/dl, TLR4^−/−: 65.11 ± 2.48 mg/dl; P < 0.001), whereas no difference was observed in the fed state (Fig. 1 A), suggesting an essential role for TLR4 in maintaining fasting blood glucose levels.

To explore the mechanisms underlying the severe fasting hypoglycemia, we first focused on hepatic glucose production. We examined the expression levels of key enzymes governing FAO and gluconeogenesis in liver. No differences in the mRNA levels of medium-chain acyl-CoA dehydrogenase (MCAD) or PPARa were observed between wild-type and TLR4^−/− mice (Fig. 1,B). Surprisingly, TLR4 deficiency increased the mRNA levels of carnitine palmitoyltransferase 1 (CPT1) in the liver (Fig. 1,B) and the β-hydroxybytyrate levels in the serum (Fig. 1,C). mRNA levels of PEPCK in TLR4^−/− mice were comparable with those in wild-type controls, while G6Pase only showed a tendency to decrease in TLR4^−/− mice (P = 0.071) (Fig. 1,D). To further compare the abilities of gluconeogenesis between mice, we performed a pyruvate tolerance test. As expected, wild-type and TLR4^−/− mice showed similar glycemic responses to pyruvate after fasting (Fig. 1,E), indicating comparable capacities of gluconeogenesis. Also, no differences in the serum levels of gluconeogenic substrates, including lactate, alanine, and pyruvate, were observed between wild-type and TLR4^−/− mice in the fasted state (Fig. 1 F). These data suggest that the fasting hypoglycemia in TLR4^−/− mice does not result from impaired hepatic gluconeogenesis.

During fasting, glucose could also be spared by blocking the irreversible net loss of glucose in skeletal muscle through PDC inhibition (3,4). PDC consists of three enzymes: pyruvate dehydrogenase (PDHA1), dihydrolipoamide acetyltransferase (DLAT), and dihydrolipoamide dehydrogenase (DLD). We found that fasting significantly inhibited PDC activity in skeletal muscle of wild-type mice. Notably, the inhibition was largely reversed by TLR4 deletion (Fig. 2,A). In addition, fasting inhibited the mRNA expression of all three enzymes of PDC in wild-type mice, whereas it only inhibited the expression of DLAT in TLR4^−/− mice (Fig. 2,B). These results suggest that unsuppressed PDC activity in skeletal muscle contributes at least partly to the severe fasting hypoglycemia in TLR4^−/− mice. We also found that TLR4 deficiency reversed the mRNA levels of hexokinase 2 (HK2) and muscle phosphofructokinase (PFKM), two rate-limiting enzymes of glycolysis, in skeletal muscle during fasting (Fig. 2 C), suggesting that glycolysis may also be reversed by TLR4 deficiency.

We next examined the involvement of TLR4 in lipid metabolism. Compared with wild-type mice, TLR4^−/− mice exhibited higher levels of TG and FFA in circulation after fasting and similar serum levels of TG and FFA in the fed state (Fig. 3,A and 3B). In liver, TG levels were increased in response to fasting in both wild-type and TLR4^−/− mice, with a tendency to be higher in TLR4^−/− mice (Fig. 3,C). In skeletal muscle, a 2.5-fold increase in TG content was observed in TLR4^−/− mice in the fasted state only (Fig. 3 D). These findings indicate that TLR4 deficiency elevates lipid levels in skeletal muscle and circulation in response to starvation.

The lipid abnormalities in TLR4^−/− mice might result from increased TG mobilization in WAT, decreased FAO, and/or increased de novo lipogenesis in key metabolic tissues. We found that fasting significantly reduced total fat mass in both wild-type and TLR4^−/− mice but with less loss in TLR4^−/− mice (Fig. 4,A). Consistently, TLR4^−/− mice exhibited less epididymal fat loss during fasting (Fig. 4,B). We also performed lipolysis assay in epididymal fat explants from fasted mice. Wild-type and TLR4^−/− mice exhibited similar glycerol release rates (Fig. 4,C), indicating comparable lipolysis. Interestingly, TLR4^−/− mice showed lower NEFA release than wild-type mice, resulting in a significant reduction in NEFA-to-glycerol ratio (Fig. 4 C), indicating increased fatty acid reesterification. These data suggest that the higher lipid levels in TLR4^−/− mice do not result from increased fat mobilization.

As to muscle FAO, TLR4 deficiency increased the mRNA levels of CPT1 and MCAD after fasting and had no effect on the expression of PPARa (Fig. 4,D). The mRNA levels of acetyl-CoA carboxylase 2 (ACC2), which inhibits CPT1 through its product mitochondrial malonyl-CoA (21), were similar between two sets of mice (Fig. 4,D), indicating that ACC2-related regulation may not be involved in the augmentation of CPT1 in TLR4^−/− mice. More importantly, TLR4^−/− mice exhibited increased muscle FAO rates compared with wild-type mice after fasting (Fig. 4 E). These observations suggest that FAO does not contribute to the higher lipid levels in TLR4^−/− mice.

We then examined the expression of enzymes regulating de novo fatty acid synthesis, including ACC1, ACL, and FAS. Fasting decreased the expression of these enzymes in liver, WAT, and muscle (Fig. 5,A–C). Surprisingly, TLR4 deficiency resulted in a complete reversion of the mRNA levels of ACC1 and ACL and a partial reversion of FAS in muscle but not liver or WAT (Fig. 5,A-5C). TLR4 deficiency also fully reversed the protein levels of ACL and FAS in muscle but not liver (Fig. 5,D-5E). In addition, the mRNA expression of elongation of very long–chain fatty acids, family member 6 (Elovl6) and stearoyl-CoA desaturase 1 (SCD1), two other important enzymes regulating fatty acid synthesis, showed a pattern simiar to that of FAS in skeletal muscle (Fig. 5,C). Consistently, the mRNA levels of glycerol-3-phosphate acyltransferase 1 (GPAT1) and diacylglycerol acyltransferase 1 (DGAT1), two enzymes controlling TG synthesis, were also higher in skeletal muscle of TLR4^−/− mice (Fig. 5,C). The mRNA levels of SREBP1c and PPARγ, two transcription factors controlling lipogenesis (22,23), were all decreased in the skeletal muscle of wild-type and TLR4^−/− mice after fasting, but TLR4^−/− mice had a higher levels of PPARγ (Fig. 5 F), suggesting that PPARγ might be involved in TLR4-dependent inhibition of lipogenic genes in skeletal muscle. Collectively, these data suggest that TLR4 is required for the inhibition of lipogenic genes in skeletal muscle during fasting and provide a possible explanation for the relatively higher TG levels in skeletal muscle but not liver or WAT of TLR4^−/− mice.

To further test the contribution of lipid synthesis to the serum lipids in TLR4^−/− mice, we treated mice with C75, a potent FAS inhibitor (24), to suppress de novo fatty acid synthesis. After fasting, vehicle-treated TLR4^−/− mice showed elevated serum levels of TG and FFA (Fig. 5,G and 5H)—similar to untreated mice. Notably, C75 treatment fully abolished these elevations (Fig. 5 G and 5H). Together with our observation that FAS expression was only reversed in muscle, but not liver and WAT, these data indicate that TLR4 may be involved in fasting serum lipids regulation by governing muscle fatty acid synthesis.

Because de novo fatty acid synthesis uses acetyl-CoA oxidized from pyruvate as substrate, we asked whether TLR4 controls PDC activity and blood glucose levels through fatty acid synthesis regulation. As expected, we found that C75 treatment abolished the increase in muscle PDC activity and the severe fasting hypoglycemia resulting from TLR4 deficiency (Fig. 6,A and 6B). These findings indicate that excess acetyl-CoA from increased muscle pyruvate oxidative decarboxylation in TLR4^−/− mice may enter lipogenesis instead of TCA cycle. Thus, to better understand the effect of TLR4 deficiency on systemic fuel utilization, energy expenditure and respiratory quotient (RQ) were determined. During fasting, total energy expenditure was comparable in TLR4^−/− mice and wild-type mice (Fig. 6,C). The RQ of wild-type and TLR4^−/− mice exhibited a similar pattern, with a rapid decrease after fasting followed by a steady level, indicating the shift of fuel utilization from carbohydrates to fat. However, the RQ of TLR4^−/− mice decreased to a lower level than that of wild-type mice (Fig. 6 D), indicating that more fat was used as energy. The lower RQ level in TLR4^−/− mice is consistent with their increased FAO rates and supports that the products of pyruvate oxidative decarboxylation may be used for lipogenesis. Collectively, these data suggest an important role for TLR4 in maintaining the homeostasis between glucose and lipid fuel during fasting.

TLR4 has been reported to modulate insulin action by enhancing proinflammatory cytokine expression (13). However, TLR4 deficiency had no effect on circulating insulin levels and insulin sensitivity during fasting (Fig. 7,A and B). In addition, phosphorylated levels of insulin receptor, Akt, and GSK3β, key molecules of insulin signaling, were comparable in skeletal muscle between wild-type and TLR4^−/− mice upon insulin stimulation in the fasted state (Fig. 7,C). We also observed no differences in the mRNA levels of tumor necrosis factor α (TNFα) or interleukin-6 (IL-6) in skeletal muscle between wild-type and TLR4^−/− mice (Fig. 7,D). Nuclear factor-κB is the key molecule mediating TLR4 activation–induced proinflammatory cytokine expression. Consistently, TLR4^−/− mice had protein levels of IκBα comparable with those of wild-type controls in skeletal muscle after fasting (Fig. 7 E). These results suggest that TLR4 regulates fasting metabolic adaptations in skeletal muscle independent of insulin and local inflammation.

The goal of this study was to investigate the physiological role of TLR4 in metabolism. We found that TLR4 played an important role in glucose and lipid metabolism in response to starvation through regulation of critical metabolic pathways in skeletal muscle, including pyruvate oxidative decarboxylation and de novo lipogenesis.

TLR4^−/− mice exhibited exacerbated fasting hypoglycemia along with unimpaired hepatic gluconeogenesis. Strikingly, we observed that starvation inhibited the PDC activity in skeletal muscle of wild-type mice and that this inhibition was markedly reversed in TLR4-deficient mice. Inhibition of PDC activity is essential for glucose preservation because acetyl-CoA could not be reconverted to pyruvate. The decline in PDC activity permits cycling of carbon between glycolysis and gluconeogenesis and thus maintains blood glucose even if glycolysis is active (3). Besides, Jeoung et al. (25) have reported that partially reversed fasting PDC activity by pyruvate hydrogenase kinase 4 deletion, a critical PDC kinase, results in exacerbated fasting hypoglycemia in mice. Thus, although other mechanism(s) may also be involved, our results clearly suggest an essential role for TLR4 in sparing glucose during starvation, at least in part, through PDC inhibition in skeletal muscle. Interestingly, the reversed PDC activity is accompanied with the reversed mRNA expression of two of its components, PDHA1 and DLD. Thus, although the PDC activity is generally controlled by reversible phosphorylation (26), regulation at the transcriptional level may also be involved. Intriguingly, the protein levels of PDHA1 in mitochondrial extracts do not change during fasting (25); we therefore measured the mRNA levels of superoxide dismutase 2 (SOD2), a reference mitochondrial marker. We found that fasting did not affect the expression of SOD2 in wild-type mice (data not shown), indicating that the suppressive effect is specific to components of PDC.

During fasting, PDC inhibition is closely paralleled by a reduced rate of glycolysis. Intriguingly, the mRNA expression of glycolytic enzymes, HK2 and PFKM, was also reversed in TLR4^−/− mice, which is indicative of reversed glycolysis. These results indicate that TLR4 may be involved in inhibiting muscle glucose disposal through coordinate suppression of glycolysis and pyruvate oxidative decarboxylation during fasting. In addition, we found that, in murine primary muscle cells, activation of TLR4 by lipopolysaccharide had no effect on either PDC activity or expression of HK2 and PFKM (data not shown), indicating that the regulation of these enzymes by TLR4 during fasting may not be direct.

TLR4^−/− mice also exhibited higher lipid levels in muscle and circulation in the fasted state. These mice showed less fat loss but similar adipose lipolysis capacity compared with wild-type mice, ruling out the contribution of fat mobilization to serum lipid elevation. In fat lipolysis assay, we unexpectedly observed a decreased ratio of NEFA to glycerol in fasted TLR4^−/− mice, indicating that the higher serum lipid may contribute to less fat loss, possibly through increased fatty acid reesterification. TLR4^−/− mice also showed increased expression of FAO genes and FAO rates in muscle, as well as reduced RQ levels during fasting, arguing against decreased FAO as a contributor to lipid abnormalities.

In terms of lipogenesis, FAS inhibitor C75 treatment abolished the elevations of serum lipids in TLR4 mice. Although the effect of C75 is not tissue specific, TLR4 deficiency reversed FAS expression in skeletal muscle only, not liver or WAT, during fasting, indicating that C75 effect on muscle FAS may account for its abolishment on serum lipid increases. It is therefore likely that reversed muscle lipogenesis in TLR4^−/− mice may contribute to blood lipid increases during fasting, although other tissue (like the skeleton) may also be involved. We also found that TLR4 deletion increased the mRNA levels of CD36 but had no effect on LPL expression in the skeletal muscle during fasting (data not shown), arguing against the possibility that TLR4 deficiency increases circulating lipid levels by inhibiting muscle FFA or TG uptake. Thus, how dysregulated muscle lipogenesis under a fasting condition contributes to the lipid disturbance in circulation requires further investigation.

Liver and WAT are widely considered major lipogenic tissues. However, muscle tissue may also actively participate in de novo lipogenesis. Several studies around the 1970s showed that liver and dissectible adipose tissue usually contribute to no more than 40% of total fatty acid synthesis rate in rodents. Instead, the rest of carcass, mainly muscle and skeleton, is the major site of fatty acid synthesis (27,,,,–32). Although there was one study that suggested that the intermuscular fat in carcass might contribute to lipogenesis (33), several following studies using dissected muscle without fat showed that muscle tissue may be an important lipogenic tissue because of its large mass (27,34,35). Interestingly, in the fasted state, liver's contribution to lipogenesis becomes almost negligible (31,36), whereas muscle may be one of the predominant tissues for lipogenesis (36). Recently, several myotube cell-based studies have also supported an active capacity of muscle in de novo lipogenesis. The expression of SREBP1c, a key lipogenic transcription factor, is clearly detectable in muscle at a level close to the liver (37). Upon insulin or glucose stimulation, muscle SREBP1c regulates the transcription of lipogenic enzymes, such as FAS and ACL (37,38). Glucose also stimulates muscle de novo lipogenesis, which in turn results in lipid accumulation in myotubes (38,39). Taken together, these studies support an active role for muscle in lipogenesis, although its role(s) under physiological or pathophysiological conditions requires further investigation.

Glucose and lipid are tightly related in metabolism. In the glucose–fatty acid cycle, FAO suppresses PDC activity by increasing the ratio of acetyl-CoA to CoA (40). Intriguingly, TLR4 may control glucose metabolism through regulation of de novo fatty acid synthesis, because FAS inhibition abolished the fasting hypoglycemia and muscle PDC activity increase in TLR4^−/− mice. Since de novo fatty acid synthesis uses acetyl-CoA as substrate, it is conceivable that de novo fatty acid synthesis regulates muscle PDC activity by altering the ratio of acetyl-CoA to CoA. These findings indicate that not only lipid oxidation but also lipid synthesis may affect muscle glucose metabolism.

Currently, it is unclear how TLR4 is activated and regulates these metabolic pathways during fasting. Unlike its reported role in insulin resistance (13), TLR4-dependent regulation of fasting metabolism is neither insulin nor inflammation dependent. During starvation, the increases in fatty acids have been reported to modulate muscle PDC activity (41,42). However, although direct FAO is generally thought to mediate the acute effect of FFA on PDC activity (41), it is not responsible for this long-term (starvation) regulation of muscle PDC (43), indicating the existence of alternative mechanism(s). Because saturated fatty acids have been reported as the agonists for TLR4 (8,–10), it is possible that TLR4 activation by FFA regulates muscle PDC activity during starvation. Hormone signaling, such as triiodothyronine, is critical for PDC activity modulation during the fed-to-starved transition (41,44), raising the possibility that TLR4 regulates fasting metabolism through cross-talk with other hormone signaling. These interesting hypotheses require further investigation.

In summary, our findings identify TLR4 as a physiological regulator of fuel metabolism independent of insulin and identify a novel physiological role for TLR4 in metabolism. Muscle lipogenesis may play an important role in maintaining fasting fuel homeostasis.

This work was supported by the National Basic Research Program of China (973 Program) (2010CB529701, 2011CB504002), the One Hundred Talents Program of the Chinese Academy of Sciences, and by the Chief Scientist Program of Shanghai Institutes for Biological Sciences (SIBS2008006), the Chinese Academy of Sciences.

No potential conflicts of interest relevant to this article were reported.

S.P. researched data, contributed to discussion, wrote the manuscript, and reviewed and edited the manuscript; H.T. researched data, contributed to discussion, and reviewed and edited the manuscript; S.Z. researched data and contributed to discussion; Y.Q.Z. contributed to discussion and reviewed and edited the manuscript; Y.L. researched data, contributed to discussion, and reviewed and edited the manuscript.

We thank Prof. Vincent Deubel and Prof. Baoxue Ge (Institute Pasteur of Shanghai, Chinese Academy of Sciences) for providing TLR4^−/− mice.