mTORC1 inhibition uncouples lipolysis and thermogenesis in white adipose tissue to contribute to alcoholic liver disease

Background:

Adipose tissue thermogenic activities use fatty acids from lipolysis for heat generation. Therefore, a tight coupling between lipolysis and thermogenesis is physiologically imperative in maintaining not only body temperature but also lipids homeostasis. Adipose tissue dysfunction contributes to alcoholic liver disease (ALD). Here, studies were conducted to examine how alcohol intake affects adipose tissue thermogenic activities and whether altered adipose tissue thermogenesis contributes to ALD.

Methods:

Both the Lieber-DeCarli and the NIAAA mouse models of ALD were used. Denervation surgery in epididymal fat pads was performed. CL316,243, a selective β3-adrenoceptor agonist, SR59230A, a selective β3 adrenoceptor (ADRB3) antagonist, and rapamycin, a selective mechanistic target of rapamycin complex 1 (mTORC1) inhibitor, were administrated through i.p. injection. Adipocyte-specific Prdm16 knockout mice were subjected to alcohol-containing diet chronically.

Results:

Chronic alcohol consumption, which enhances adipose tissue lipolysis, inhibits thermogenic activities of beige adipocytes in inguinal white adipose tissue (WAT), leading to an uncoupling status between lipolysis and thermogenesis in WAT at both basal and ADRB3 stimulation states. CL316,243 administration exacerbates liver pathologies of ALD. Alcohol intake inhibits mTORC1 activities in WAT. In mice, mTORC1 inhibition by rapamycin inhibits the thermogenesis of iWAT, whereas enhancing WAT lipolysis. Further investigations using adipocyte-specific Prdm16 knockout mice revealed that functional deficiency of beige adipocytes aggravates liver pathologies of ALD, suggesting that the inhibitory effect of alcohol on WAT browning/thermogenesis contributes to ALD pathogenesis.

Conclusion:

Chronic alcohol consumption induces an “uncoupling status” between lipolysis and browning/thermogenesis in WAT by inhibiting mTORC1 activation. Diminished WAT browning/thermogenesis, concomitant with enhanced lipolysis, contributes to ALD pathogenesis.

INTRODUCTION

Alcoholic liver disease (ALD) represents a spectrum of hepatic pathological alterations owning to chronic excessive alcohol intake, ranging from early steatosis to inflammatory steatohepatitis, fibrosis, and cirrhosis. Although much progress has been made over last several decades of intense investigations on its pathogenesis, the exact cellular/molecular mechanism(s) underscoring its initiation and progression remain ambiguous. Currently, ALD remains a principal health problem, ranking among the major causes of morbidity and mortality worldwide,1 with no Food and Drug Administration–approved therapies available to prevent or cure it. Thus, a better understanding of the cellular/molecular mechanism(s) underlying ALD pathogenesis is urgently required for the development of targeted therapies.

Although the mechanisms involved in ALD development are complex and evidently multifactorial, accumulated evidence supports that adipose tissue dysfunction plays an important role in disease pathogenesis. Clinically, obesity potentiates the severity of alcohol-induced liver damage and BMI is an independent risk factor for steatosis, hepatitis, and fibrosis in ALD patients.2–6 Moreover, patients with ALD display inflammation not only in the liver but also in the adipose tissue, which correlates with the severity of pathological features in the liver.7,8 Experimentally, chronic ethanol feeding is concomitant with increased macrophage infiltration into epididymal adipose tissue and augmented expression of mRNA for some proinflammatory adipocytokines, including MCP-1, TNF, and IL-6 in epididymal adipose tissue of ethanol-fed rodents.8,9 Several very recent studies, including ours, have independently documented that chronic alcohol consumption is associated with reduced white adipose tissue (WAT) mass in mice, primarily resulting from enhanced adipose tissue lipolytic activities.10–13 Two types of adipose tissues are present in humans and rodents, WAT, and brown adipose tissue (BAT). Depending on the anatomical locations, brown adipocytes can be further categorized into classical brown adipocytes, which are found in BAT, and uncoupling protein 1 (UCP1)-expressing beige adipocytes, which locate inside WAT, normally s.c. WAT, such as inguinal WAT (iWAT). WAT stores excess energy in the form of triacylglycerol (TAG) during the fed-state and releases free fatty acids (FFAs) through lipolysis to provide other tissues with substrates for energy production during the fasting state. Both brown and beige adipocytes oxidize FFAs and glucose for thermogenesis to maintain body temperature homeostasis, primarily through UCP1 expression. Beige adipocytes can be induced, in response to various stimuli, such as β3-adrenergic receptor (ADRB3) agonist and cold exposure, to attain the phenotypes of brown adipocytes through the so-called “browning” process.14,15 This property endows the induction of beige adipose tissue “browning,” an attractive therapeutic strategy for preventing and treating a variety of metabolic disorders.

In response to the release of norepinephrine (NE) from the sympathetic nervous system (SNS), the canonical lipolytic ADRB3-cAMP-PKA pathway becomes activated, which phosphorylates lipases of the lipolytic machinery associated with the lipid droplets, including adipose triglyceride lipase and hormone-sensitive lipase, to promote the release of FFAs. Chronic alcohol exposure in mice resulted in a profound enhancement of adipose tissue lipolytic activity, which is associated with the development of fatty liver and liver injury. Mice chronically exposed to an alcohol-containing diet manifest increased circulating NE concentration and activated the cAMP-PKA pathway in WAT.10–12 In accordance, both adipose triglyceride lipase and hormone-sensitive lipase in epididymal fat pads were activated in response to chronic alcohol feeding.10 Importantly, various experimental approaches that alleviated adipose tissue lipolytic response improved liver pathologies of ALD.10–13

Although alcohol-stimulated overactive adipose lipolysis has been well documented, it remains elusive as to how alcohol intake affects adipose tissue thermogenic activities, as well as whether and how this potentially implicates disease development. Given the fact that the ADRB3-cAMP-PKA pathway is “shared” by both lipolysis and thermogenesis, we initially postulated that chronic alcohol exposure would promote adipose tissue thermogenesis. Unexpectedly, our data revealed that chronic alcohol consumption differentially regulates thermogenic activities in BAT and WAT. Whereas it activated the thermogenic activities in BAT, which is in line with several previous studies,16,17 chronic alcohol feeding in mice inhibited the browning/thermogenesis process of beige adipocytes. Further mechanistic investigations uncovered that the inhibitory regulation of alcohol on mechanistic target of rapamycin complex 1 (mTORC1) activities contributes to the observed uncoupling between lipolysis and thermogenesis in WAT in response to chronic alcohol consumption.

MATERIALS AND METHODS

Reagents and chemicals

CL316,243 disodium salt (CL), a selective β3-adrenoceptor agonist, and SR59230A hydrochloride (SR), a selective β3 adrenoceptor antagonist, were purchased from TOCRIS. Rapamycin was obtained from AdipoGen Life Sciences (Cat#: AG-CN2-0025, San Diego, CA). Tween 80 (Cat#: S6702) and PEG300 (Cat#: S6704) were both from SELLECKCHEM (Radnor, PA). DMSO was from Sigma Aldrich (Cat#: D8418, Saint Louis, MO). The antibody for β-actin was from ABclonal (Cat#: AC026, Woburn, MA) and antibody for PGC1α was from NOVUS Biologicals (Cat#: NBP1-04676, Centennial, CO). The following antibodies were from Cell Signaling Biotechnology (Danvers, MA): p-S6 (Ser 235/236) (Cat#: 4858S), S6 (Cat#: 2217S), p-S6K1 (Thr389) (Cat#: 9205S), p-Akt (Thr 308) (Cat#: 4056S), Akt (Cat#: 9272S), p-AMPKα (Thr 172) (Cat#: 2535T), AMPKα (Cat#: 5832T), UCP1 (Cat#: 14670S), and p-eIF2α (Ser 51) (Cat#: 3398T).

Animal studies

All mice were housed in a temperature of 18 °C–23 °C with 40%–60% humidity and 12-hour light/dark cycle environment unless otherwise stated. All mice studies were approved by the Institutional Animal Care and Use Committee at the University of Illinois at Chicago (Chicago, IL) and consistent with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

All animal studies were approved by the Institutional Animal Care and Use Committee at the University of Illinois at Chicago (Chicago, IL) and consistent with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Ucp1^Cre+/mTmG+ mice were generated through breeding Ucp1^Cre+ mice and ROSA26mTmG/mTmG mice and were kindly provided by Dr. Yuwei Jiang (University of Illinois at Chicago). The adipocyte-specific Prdm16 knockout mice (Adipo-Prdm16 KO) were generated by crossing Prdm16lox/lox mice (B6.129-Prdm16tm1.1Brsp/J mice; Stock No: 024992) with adiponectin-Cre mice (B6.129-Prdm16tm1.1Brsp/J mice; Stock No: 024992) (both obtained from Jackson Laboratory, Bar Harbor, ME). Prdm16lox/lox mice were used as control. Both Adipo-Prdm16 KO mice and their corresponding littermates control mice (Prdm16lox/lox) were exposed to either isocaloric pair-fed (PF) or Lieber-De Carli alcohol-containing liquid diet for 5 weeks. Both traditional Lieber-DeCarli model of ALD and NIAAA model18 were used (see Supplementary Information for details, http://links.lww.com/HC9/A154). Both sham surgery and surgical denervation (DN) of epididymal fat pads were performed in male C57BL/6J mice (10-week old), as described19 before alcohol feeding (see Supplementary Information for details, http://links.lww.com/HC9/A154). For rapamycin and CL combined administration, male C57BL/6N mice (10-week old) were i.p. injected with rapamycin at a dose of 5 mg/kg BW per day for 9 days. Rapamycin was dissolved in vehicle solution containing 2% DMSO, 5% Tween 80, and 30% PEG300 dH₂O. Control group was given isovolumic vehicle solution. Two days later, mice were given a combined administration of rapamycin and CL (1 mg/kg BW/d through i.p. injection) for following 7 days.

Cold tolerance test

Mice were placed in a cold environment (6 °C) for last 3 days of PF or alcohol-fed (AF) feeding. Rectal temperature was monitored daily.

Western blot

Total proteins were isolated using radioimmunoprecipitation lysis buffer with phenylmethylsulfonyl fluoride, protease inhibitor cocktail and sodium orthovanadate (all from Santa Cruz Biotechnology, Dallas, TX) and the target proteins were detected by western blotting using primary antibodies and IRDye secondary antibodies (Licor, Lincoln, NE). The signals were detected by Odyssey DLx Imaging System (Licor, Lincoln, NE) (see Supplementary Information for details, http://links.lww.com/HC9/A154).

Quantitative real-time PCR

Total RNA isolation, generation of cDNA, and real-time quantitative PCR were conducted, as previously reported. Primer sequences used are listed in Table 1. The comparative threshold cycle (or 2 Ct) method was used to calculate relative mRNA expression levels (see Supplementary Information for details, http://links.lww.com/HC9/A154).

TABLE 1.

Primer sequences

Gene	Sequence (5′-3′)
M-Rn18s-F	GTAACCCGTTGAACCCCATT
M-Rn18s-R	CCATCCAATCGGTAGTAGCG
M-Srebf1-F	CGACTACATCCGCTTCTTGCAG
M-Srebf1-R	CCTCCATAGACACATCTGTGCC
M-Srebf2-F	AGAAAGAGCGGTGGAGTCCTTG
M-Srebf2-R	GAACTGCTGGAGAATGGTGAGG
M-Acaca-F	GTTCTGTTGGACAACGCCTTCAC
M-Acaca-R	GGAGTCACAGAAGCAGCCCATT
M-Acacb-F	AGAAGCGAGCACTGCAAGGTTG
M-Acacb-R	GGAAGATGGACTCCACCTGGTT
M-Fasn-F	CACAGTGCTCAAAGGACATGCC
M-Fasn-R	CACCAGGTGTAGTGCCTTCCTC
M-Ppargc1a-F	GAATCAAGCCACTACAGACACCG
M-Ppargc1a-R	CATCCCTCTTGAGCCTTTCGTG
M-Pparg-F	GTACTGTCGGTTTCAGAAGTGCC
M-Pparg-R	ATCTCCGCCAACAGCTTCTCCT
M-Acly-F	AGGAAGTGCCACCTCCAACAGT
M-Acly-R	CGCTCATCACAGATGCTGGTCA
M-Acsl1-F	ATCAGGCTGCTTATGGACGACC
M-Acsl1-R	CCAACAGCCATCGCTTCAAGGA
M-Acsl4-F	CCTTTGGCTCATGTGCTGGAAC
M-Acsl4-R	GCCATAAGTGTGGGTTTCAGTAC
M-Cd36-F	GGACATTGAGATTCTTTTCCTCTG
M-Cd36-R	GCAAAGGCATTGGCTGGAAGAAC
M-Scd1-F	GCAAGCTCTACACCTGCCTCTT
M-Scd1-R	CGTGCCTTGTAAGTTCTGTGGC
M-Hmgcr-F	GCTCGTCTACAGAAACTCCACG
M-Hmgcr-R	GCTTCAGCAGTGCTTTCTCCGT

Statistics analysis

Prism GraphPad (Version 8.3, San Diego, CA) was used to analyze the results. All data were analyzed by Student t test or 1-way ANOVA test. Differences between groups were considered statistically significant if p<0.05.

RESULTS

Unrestrained adipose tissue lipolysis contributes to ALD development

Using the well-established Lieber-DeCarli mouse model of ALD20, we examined the role of adipose tissue lipolysis in ALD disease development. After 5 weeks of feeding with either isocaloric control (PF) or alcohol-containing (AF) diet, mice in AF group exhibited a significant increase of hepatic fat (TAG, triacylglycerol) accumulation, assessed through both biochemical measurement and histologic examination (Supplemental Figure 1A, B, http://links.lww.com/HC9/A148). Chronic alcohol consumption elevated plasma levels of alanine aminotransferase (Supplemental Figure 1C, http://links.lww.com/HC9/A148), an indicator of liver injury, and enhanced adipose tissue lipolytic activities, evidenced by increased plasma FFAs and glycerol levels (Supplemental Figure 1D, E, http://links.lww.com/HC9/A148). As adipose tissue lipolysis is primarily triggered by locally released NE from nerve fibers of the SNS,21 the most direct way to inhibit adipose tissue lipolysis is to remove the neural supply to the tissue. To determine the pathological role of adipose tissue lipolysis in ALD development, we performed surgical DN procedure in epididymal fat pads. Mice were, after either sham or DN surgery, subjected to either PF or AF feeding for 5 weeks. As expected, the surgical DN in epididymal fat pads (DN) prevented the reduction of epididymal fat to body weight ratio and adipocyte size induced by chronic alcohol feeding (Figure 1A–C). DN blunted alcohol-induced augment of plasma glycerol and FFAs levels (Figure 1D, E) and attenuated alcohol-induced increase of liver to body weight ratio (Figure 1F). Importantly, DN ameliorated plasma alanine aminotransferase elevation and hepatic TAG accumulation induced by chronic alcohol feeding (Figure 1G–I). In rodents, the β3-adrenergic receptor (ADRB3) is the predominant cell membrane receptor in adipocytes to drive NE-stimulated lipolysis.22,23 To further solidify our notion, another well-established ALD mouse model, the NIAAA chronic plus binge ethanol feeding model (10-d feeding with alcohol-containing liquid diet plus 1 gavage, hereafter referred to as “the NIAAA model”) was used. Alcohol-fed animals were administrated with and without SR59230A (hereafter referred to as SR), a specific ADRB3 antagonist, through daily i.p. injection (1 mg/kg BW). As shown in Figure 1J, K, SR administration rescued plasma glycerol elevation and blunted hepatic fat accumulation in the setting of chronic alcohol exposure, whereas liver cholesterol was not significantly affected (Supplemental Figure 1F, http://links.lww.com/HC9/A148). Collectively, these results support that unrestrained adipose tissue lipolysis contributes to ALD development.

FIGURE 1.

Adipose tissue lipolysis contributes to alcoholic liver disease (ALD) development. Denervation (DN) or sham surgeries in epididymal fat pads were conducted in male C57BL6/J mice (10-week old) before the Lieber-DeCarli liquid diet exposure for 5 weeks. (A) Surgical denervation of eWAT; green arrow: eWAT; blue arrow: vas deferens; red arrow: testis; circle: nerves. (B) eWAT to body weight ratio. (C) Histological examination (H&E staining) of eWAT. (D) Plasma glycerol concentrations. (E) Plasma FFAs levels. (F) Liver to body weight ratio. (G) Plasma ALT levels. (H) Histological examination (H&E staining). (I) Liver TAG contents. Male C57BL6/N mice (10-week old) were subjected to the NIAAA model of ALD with/without administration of SR59230A (1 mg/kg BW/d) or isovolumic vehicle through i.p. injection. (J) Plasma glycerol levels. (K) Liver TAG contents. All values are denoted as means±SD (n=4–6). Differences between 2 groups were determined using Student t test (**p<0.01, ***p<0.001 vs. corresponding control). Differences between groups were determined using 1-way ANOVA analysis (*p<0.05, **p<0.01, ***p<0.001 vs. corresponding control). Abbreviations: AF, alcohol-fed; ALT, alanine aminotransferase; FFA, free fatty acid; PF, pair-fed; TAG, triacylglycerol; SR, SR59230A hydrochloride.

Alcohol differentially regulates thermogenic activities in WAT and BAT

Other than regulating lipolysis, the SNS, through locally releasing NE, which binds to and activates cell membrane ADRB3, also induces thermogenic program activation in both beige (thermogenic adipocytes in WAT) and brown adipocytes, primarily through upregulating UCP1 expression.24,25 The general paradigm is that adipose tissue lipolysis and thermogenesis are physiologically “coupled” through sharing the SNS/NE/ADRB3 pathway.26 This notion prompted us to postulate that chronic alcohol consumption could lead to enhanced adipose tissue thermogenic activities in mice. To test our hypothesis, we first examined expression of critical thermogenic genes in both BAT and inguinal WAT (beige fat, iWAT). As shown in Figure 2A and C, UCP1 expression in BAT was significantly increased at both mRNA and protein levels in AF mice relative to these in PF mice. In contrast, chronic alcohol exposure significantly decreased the gene expression of 3 signature thermogenic genes, Ucp1, Ppargc1a, and Prdm16, in iWAT (Figure 2B). Correspondingly, markedly reduced UCP1 protein abundance in iWAT was observed in AF mice when compared with that in PF mice (Figure 2C). The inhibitory effect of chronic alcohol exposure on iWAT UPC1 expression was further substantiated using Ucp1^Cre+/mTmG+ mice (Supplemental Figure 2A, http://links.lww.com/HC9/A149) subjective to either PF or AF for 5 weeks (Figure 2D, E, Supplemental Figure 2B–E, http://links.lww.com/HC9/A149). In accordance, when placed in a cold temperature (6°C) for 3 days, mice from AF group exhibited a moderate, but significant, decrease in core body temperature (Figure 2F).

FIGURE 2.

Alcohol differentially regulates thermogenic activities in brown adipose tissue (BAT) and inguinal white adipose tissue (iWAT). (A–C) Male C57BL/6J mice (10-week old) were exposed to either isocaloric control (PF for pair-fed) or ethanol-containing Lieber-DeCarli liquid diet (AF for alcohol-fed) for 5 weeks. (A) Expression of thermogenic genes in BAT. (B) Expression of thermogenic genes in iWAT. (C) UCP1 protein abundance in BAT and iWAT. Ucp1^Cre+/mTmG mice were exposed to either isocaloric control (PF for pair-fed) or ethanol-containing Lieber-DeCarli liquid diet (AF for alcohol-fed) for 5 weeks. (D) Frozen sections of iWAT were directly examined for green and red fluorescence. (E) iWAT (upper part and lower part) and eWAT were directly examined for whole mount, RFP fluorescence, and GFP fluorescence. (F) C57BL/6J mice (10-week old) were exposed to either isocaloric control (PF for pair-fed) or ethanol-containing Lieber-DeCarli liquid diet (AF for alcohol-fed) for 5 weeks, followed by cold temperature exposure (6 °C) for 3 days before sacrifice. Rectal temperatures were measured daily. (G, H) C57BL/6J mice (10-week old) were exposed to either isocaloric control (PF for pair-fed) or ethanol-containing Lieber-DeCarli liquid diet (AF for alcohol-fed) for 5 weeks with/without CL316,243 administration at a dose of 1 mg/kg BW/d through i.p. injection during the last 7 days of experimental feeding. (G) Ucp1 mRNA expression in BAT. (H) Ucp1 mRNA expression in iWAT. (I, J) C57BL/6J mice (10-week old) were exposed to either isocaloric control (PF for pair-fed) or ethanol-containing Lieber-DeCarli liquid diet (AF for alcohol-fed) for 5 weeks, followed by cold temperature exposure (6 °C) for 3 days before sacrifice. (I) Ucp1 mRNA expression in iWAT. (J) Ucp1 mRNA expression in BAT. All values are denoted as means±SD (n=4–6). Differences between 2 groups were determined using Student t test (*p<0.05, ***p< 0.001 versus corresponding control, ns: not significant). Differences between groups were determined using 1-way ANOVA analysis. Bars with different characters differ significantly (p<0.05). Abbreviations: AF, alcohol-fed; eWAT, epididymal white adipose tissue; PF, pair-fed; SR, SR59230A hydrochloride.

ADRB3 activation and cold exposure are 2 classic maneuvers to trigger thermogenic process in beige and brown adipocytes.25 To further validate our results observed at basal conditions, CL316,243 (hereafter referred to as CL), a specific ADRB3 agonist, were injected (i.p.) in a dose of 1 mg/kg BW/d for the last 7 days during the experimental diet feedings and UCP1 expression in both inguinal and brown fat were examined. As shown in Figure 2G, similar to that observed at basal conditions, AF mice showed a higher Ucp1 expression in BAT than in PF mice and CL administration enhanced Ucp1 gene expression in BAT from both PF and AF mice. In contrast, in iWAT, although CL administration resulted in a robust increase of Ucp1 expression in both PF and AF mice, alcohol-feeding blunted CL-triggered Ucp1 increase when compared with PF mice (Figure 2H). When exposed to cold temperature (6°C) for 3 days, mice from AF group demonstrated blunted Ucp1 upregulation when compared with PF mice (Figure 2I). Unlike CL administration, cold exposure showed a tendency to compromise Ucp1 upregulation in BAT, similar to that in iWAT (Figure 2J). As a negative control, we also measured Ucp1 expression in SR-treated PF and AF mice. SR administration was concomitant with markedly decreased Ucp1 expression in both iWAT and BAT (Supplemental Figure 2F, G, http://links.lww.com/HC9/A149). Collectively, these results indicate that chronic alcohol consumption in mice leads to an “uncoupling” status between lipolysis and thermogenesis in WAT.

Thermogenesis induction exacerbates hepatic pathologies of ALD in mice

The finding that chronic alcohol consumption uncouples lipolytic and thermogenic activities in WAT prompted us to examine whether thermogenic induction can impact ALD progression. The adipose tissue thermogenic process was elicited through either CL administration or cold exposure. As shown in Figure 3, CL administration significantly increased plasma FFAs in PF group (Figure 3A), concomitant with lowered epididymal fat pad weight to body weight ratio (Supplemental Figure 3A, http://links.lww.com/HC9/A150). CL administration caused a subtle, but significant, elevation of liver TAG content (Figure 3B, C), without affecting plasma liver enzyme activities in PF mice (Figure 3D, Supplemental Figure 3B, http://links.lww.com/HC9/A150). In AF group, CL administration further increased plasma FFAs (Figure 3A). Remarkably, CL aggravated liver fat accumulation and liver injury, as determined both biochemically and histologically (Figure 3B–D), which were associated with the significantly increased liver to body weight ratio (Supplemental Figure 3C, http://links.lww.com/HC9/A150). Similarly, a 3-day cold exposure enhanced adipose tissue lipolytic activity in AF mice (Figure 3E, F) and exacerbated alcohol-induced fatty liver development (Figure 3G, H) and liver injury (Figure 3I, Supplemental Figure 3D, http://links.lww.com/HC9/A150). Cold exposure increased the liver to body weight ratio in alcohol-fed mice (Supplemental Figure 3E, http://links.lww.com/HC9/A150). In contrast, the epididymal fat pad to body weight ratio was significantly decreased (Supplemental Figure 3F, http://links.lww.com/HC9/A150).

FIGURE 3.

Thermogenesis induction aggravates hepatic pathologies of ALD in mice. C57BL/6J mice (10-week old) were exposed to either isocaloric control (PF for pair-fed) or ethanol-containing Lieber-DeCarli liquid diet (AF for alcohol-fed) for 5 weeks with/without CL316,243 administration at a dose of 1 mg/kg BW/d through i.p. injection during the last 7 days of experimental feeding. (A) Plasma FFAs levels. (B) Liver TAG contents. (C) H&E staining of liver samples. (D) Plasma ALT levels. (E–I) C57BL/6J mice (10-week old) were exposed to either isocaloric control (PF for pair-fed) or ethanol-containing Lieber-DeCarli liquid diet (AF for alcohol-fed) for 5 weeks, followed by cold temperature exposure (6 °C) for 3 days before sacrifice. (E) Plasma glycerol concentrations. (F) Plasma FFAs levels. (G) Liver TAG contents. (H) H&E staining of liver samples. (I) Plasma ALT levels. All values are denoted as means±SD (n=4–5). Differences between 2 groups were determined using Student t test (*p<0.05 vs. corresponding control). Differences between groups were determined using 1-way ANOVA analysis. Bars with different characters differ significantly (p<0.05). Abbreviations: AF, alcohol-fed; ALT, alanine aminotransferase; CL, CL316,243 disodium salt; FFA, free fatty acid; PF, pair-fed; TAG, triacylglycerol.

Alcohol differentially regulates mTORC1 activities in BAT and WAT

To gain insight into the molecular events that underpin the uncoupling between lipolysis and thermogenesis in the setting of chronic alcohol exposure, we examined the effects of chronic alcohol feeding on mTORC1 activation in different adipose depots based on emerging evidence that adipose tissue mTORC1 activity is essential for lipolysis inhibition and thermogenesis induction.27–29 The activation status of mTORC1 was determined by Western blot detection of phospho-p70 S6 kinase (S6K1) at Th389, which is activated directly by mTORC1 activation, and phopho-S6 ribosomal protein, a direct target of S6K1. As shown in Figure 4, chronic alcohol feeding has no effect on BAT mTORC1 activation (Figure 4A, B), however, mTORC1 activities in both epididymal WAT (eWAT) (Figure 4C, D) and iWAT (Figure 4E, F) were markedly inhibited in AF mice relative to PF animals. Insulin activates mTORC1 through activating Akt. Thus, we subsequently examined the effect of chronic alcohol consumption on Akt activation/phosphorylation status in iWAT. As shown in Figure 4G, H, chronic alcohol feeding indeed showed a propensity of activating Akt in iWAT, presumably owning to the alleviation of feedback inhibition of mTORC1 on Akt activation. Furthermore, AMPK is a well-established physiological inhibitor of mTORC1, however, our Western blot result showed that chronic alcohol feeding had no effect on AMPK activation in iWAT (Figure 4G, I).

FIGURE 4.

Alcohol differently regulates mTORC1 activities in BAT and WAT. C57BL/6J mice (10-week old) were exposed to either isocaloric control (PF for pair-fed) or ethanol-containing Lieber-DeCarli liquid diet (AF for alcohol-fed) for 5 weeks. (A) Protein abundance of p-S6, total S6, and p-S6K1 in BAT. (B) Quantitative analysis of the ratios of p-S6 and p-S6K1 to β-actin in brown adipose tissue (BAT). (C) Protein abundance of p-S6 and p-S6K1 in epididymal fat pad (eWAT). (D) Quantitative analysis of the ratios of p-S6 and p-S6K1 to β-actin in eWAT. (E) Protein abundance of p-S6 and p-S6K1 in inguinal WAT (iWAT). (F) Quantitative analysis of the ratios of p-S6 and p-S6K1 to β-actin in iWAT. (G) Protein abundance of p-Akt, total Akt, p-AMPK, and total AMPK in iWAT. (H) Quantitative analysis of the ratio of p-Akt to β-actin in iWAT. (I) Quantitative analysis of the ratio of p-AMPK to β-actin in iWAT. All values are denoted as means±SD. Differences between 2 groups were determined using Student t test (*p<0.05, **p<0.01 vs. corresponding control). Abbreviations: AF, alcohol-fed; PF, pair-fed;

mTORC1 inhibition uncouples CL-induced lipolysis and thermogenesis in WAT

To directly examine how mTORC1 inhibition affects the WAT lipolytic and thermogenic activities, we treated mice with either vehicle or rapamycin injection, with/without CL, for 7 days. Rapamycin administration at 5 mg/kg BW efficiently inhibited mTORC1 in all 3 adipose depots, namely BAT, iWAT, and eWAT (Figure 5C–E). In comparison to vehicle control, rapamycin enhanced CL-induced adipose tissue lipolysis, evidenced by a significant elevation of plasma glycerol levels (Figure 5A), which was concomitant with incremental liver TAG accumulation (Figure 5B). Intriguingly, rapamycin failed to alter expression of 2 critical thermogenic molecules, UCP1 and PGC1α, in BAT (Figure 5C), whereas both proteins were substantially downregulated in WAT, both iWAT and eWAT (Figure 5D, E). Rapamycin had no effect on plasma TAG concentrations (Supplemental Figure 4A, http://links.lww.com/HC9/A151); however, it dramatically increased plasma total cholesterol contents (Supplemental Figure 4B, http://links.lww.com/HC9/A151), which was accompanied with a marked reduction of hepatic total cholesterol levels (Supplemental Figure 4C, http://links.lww.com/HC9/A151). To determine whether hepatic de novo lipogenesis activation potentially contributes to the observed rapamycin-elicited liver TAG accumulation (Figure 5B), the expression of some critical lipogenic genes in the liver were subsequently examined and results are shown in Figure 5F. When compared with control mice, CL injection activated hepatic PPARγ, as demonstrated by increased gene expression of Pparg and Cd36, 2 signature targets of PPARγ activation. Overall, rapamycin inhibited hepatic de novo lipogenic activities in response to CL administration, as shown by reduced expression of de novo lipogenic genes Acaca, Fasn, and Srebf1, while TAG synthetic activities were enhanced, evidenced by upregulated gene expression of Acsl4 and Scd1. These data altogether demonstrate that mTORC1 inhibition uncouples lipolysis and thermogenesis in WAT.

FIGURE 5.

mTORC1 inhibition uncouples CL-induced lipolysis and thermogenesis in WAT. Male C57BL6/N (10-week old) mice were treated (i.p. injection) with rapamycin (5 mg/kg BW/d) or isovolumic vehicle for 9 days. The combined administration of rapamycin and CL started on day 3 after rapamycin treatment began, and CL (1 mg/kg BW/d, i.p. injection) or isovolumic vehicle were given for the remaining 7 days. (A) Plasma glycerol levels. (B) Liver TAG contents. (C) Protein abundance of p-S6, total S6, PGC1α, UCP1 in brown adipose tissue (BAT). (D) Protein abundance of p-S6, total S6, PGC1α, UCP1 in epididymal fat pad (eWAT). (E) Protein abundance of p-eIF2α, p-S6, total S6, PGC1α, UCP1 in inguinal fat (iWAT). (F) Expression of key lipogenic genes in liver samples. All values are denoted as means±SD (n=4–6). Differences between 2 groups were determined using Student t test (*p<0.05, **p<0.01 vs. corresponding control). Abbreviations: CL, CL316,243 disodium salt; TAG, triacylglycerol;

Rapamycin exacerbates liver pathologies of ALD

Next, we set out to determine whether mTORC1 inhibition in WAT, which uncouples lipolysis and thermogenesis, contributes to ALD progression. Mice chronically exposed to alcohol-containing liquid diet for 5 weeks were treated with either vehicle or rapamycin for last 7 days before sacrifice. The effects of rapamycin administration on fatty liver and liver injury development were assessed. In iWAT, rapamycin inhibited mTORC1 activity, evidenced by obviously decreased p-S6 protein abundance and AMPK activation (Figure 6A). Both UCP1 and PGC1α protein abundance were reduced by rapamycin treatment (Figure 6A), indicative of inhibitory thermogenic activities in iWAT. In BAT, rapamycin caused a subtle reduction of p-S6 protein abundance without affecting AMPK activation (Figure 6B). BAT UCP1 protein abundance was reduced in response to rapamycin treatment, however, PGC1α protein expression was not affected. Moreover, the body weight was not affected by rapamycin treatment (Figure 6C). Rapamycin treatment exhibited a tendency to enhance adipose tissue lipolysis, as demonstrated by elevated plasma glycerol levels (Figure 6D). Notably, rapamycin worsened ALD pathologies, demonstrated by higher plasma liver enzymes, alanine aminotransferase, and AST (Figure 6E, Supplemental 6A, http://links.lww.com/HC9/A153), as well as liver TAG accumulation (Figure 6F, G, Supplemental Figure 5B, http://links.lww.com/HC9/A152).

FIGURE 6.

Rapamycin exacerbates liver pathologies of ALD. C57BL/6J mice (10-week old) were exposed to ethanol-containing Lieber-DeCarli liquid diet (AF for alcohol-fed) for 5 weeks, with wither rapamycin (5 mg/kg BW/d) or isovolumic vehicle administration (i.p. injection) for the last 7 days of experimental feedings. (A, B) Western blot detection of protein abundance of p-S6, total S6, p-AMPK, total AMPK, PGC1α, and UCP1 in iWAT and BAT, respectively. (C) Body weight changes. (D) Plasma glycerol levels. (E) Plasma ALT levels. (F) Liver TAG contents. (G) Liver morphological alterations. All values are denoted as means±SD (n=5–6). Differences between 2 groups were determined using Student t test (*p<0.05 vs. corresponding control). Abbreviations: AF, alcohol-fed; ALT, alanine aminotransferase; PF, pair-fed.

Adipocyte-specific Prdm16 deletion aggravates ALD

Prdm16 is a transcriptional regulatory protein. Despite expressed in both beige and brown adipocytes, Prdm16 has been shown to play a key role in the development and function of beige adipocytes. The adipocyte-specific deletion of Prdm16 gene (Adipo-Prdm16 KO) ablates the thermogenic program of beige fat cells without affecting BAT function.30,31 The differential regulation of chronic alcohol exposure on the thermogenic activities in beige and brown adipocytes encouraged us to use Adipo-Prdm16 KO mice to determine whether compromised browning of WAT in the setting of chronic alcohol feeding, as observed by us, contributes to ALD pathogenesis. Age- and gender-matched control and Adipo-Prdm16 KO mice were subjected to ethanol-containing Lieber-DeCarli liquid diet for 5 weeks. As shown in Figure 7A–C, adipocyte-specific Prdm16 deletion drastically decreased Prdm16 gene expression in both iWAT and BAT without affecting its hepatic expression. Adipo-Prdm16 KO is concomitant with a dramatic reduction of thermogenic genes, Ucp1 and Ppargc1a, in iWAT, without affecting their expression in BAT (Figure 7A, B). Interestingly, UCP1 protein abundance was downregulated in both iWAT and BAT (Figure D, E). After a 5-week alcohol-containing diet challenge, Adipo-Prdm16 KO tended to worsen liver pathologies of ALD, as reflected by the trend toward incremental plasma alanine aminotransferase and hepatic TAG in KO mice relative to age-matched and sex-matched control animals (Figure 7F, G). Moreover, higher plasma glycerol levels were observed in Adipo-Prdm16 KO mice on alcohol-diet exposure when compared with control animals (Figure 7H), indicative of enhanced adipose tissue lipolysis in KO mice. No differences in hepatic cholesterol, plasma TAG, as well as plasma cholesterol were observed between 2 groups (Supplemental Figure 6A–C, http://links.lww.com/HC9/A153). There was no alteration in the expression of key lipogenic genes in the liver between control and KO mice (Figure 7I), ruling out the potential implication of hepatic de novo lipogenesis pathway activation in the observed liver TAG increase in the KO mice.

FIGURE 7.

Adipocyte-specific Prdm16 ablation aggravates ALD progression. Adipo-Prdm16 KO mice (10-week old), and their age-matched and sex-matched littermates were subjected to the Lieber-DeCarli ALD model. Gene expression of Prdm16, Ucp1, and Ppargc1a in iWAT (A) and BAT (B) were examined. (C) Liver Prdm16 gene expression. Protein abundance of PGC1α and UCP1 in iWAT (D) and BAT (E). (F) Plasma ALT levels. (G) Liver TAG contents. (H) Plasma glycerol levels. (I) Hepatic expression of key lipogenic genes. All values are denoted as means±SD (n=4–5). Differences between 2 groups were determined using Student t test (*p<0.05, ****p<0.0001 vs. corresponding control). Abbreviations: AF, alcohol-fed; BAT, brown adipose tissue; eWAT, epididymal white adipose tissue; iWAT, inguinal white adipose tissue; TAG, triacylglycerol; WT, wild type.

DISCUSSION

Thermogenic brown adipocytes in BAT and beige adipocytes in s.c. WAT, such as iWAT, are capable of dissipating chemical energy as heat through mitochondrial UCP1 expression. Given the fact that fatty acids, primarily originated from lipolysis, are the major substrates for thermogenesis, adipose tissue lipolytic and thermogenic processes must be tightly “coupled” to maintain both lipids and body temperature homeostasis.32,33 In the present study, we provide initial evidence that chronic alcohol consumption causes an “uncoupled status” between lipolysis and thermogenesis in WAT. In contrast to markedly enhanced lipolytic activities, significantly suppressed thermogenic activities were observed in WAT on chronic alcohol exposure. Our further mechanistic investigations pointed to diminished mTORC1 activities as a mechanism underlying the opposite effects of alcohol intake on thermogenesis and lipolysis in WAT in that chronic alcohol exposure inhibited mTORC1 activation in WAT and rapamycin, a selective inhibitor of mTORC1, mirrored the opposite effects observed in alcohol-fed mice. Remarkably, rapamycin administration exacerbates liver pathologies of ALD. Using Adipo-Prdm16 KO mice, our results demonstrate that the functional deficiency of beige adipocytes accelerated disease progression of ALD, suggesting that suppressed browning/thermogenesis of beige adipocytes plays an important role in ALD pathogenies.

It is well documented that unrestrained adipose tissue lipolytic activity contributes to ALD development, although the exact mechanisms by which chronic alcohol consumption triggers overactive adipose tissue lipolysis remain to be clarified.10 It is the general paradigm that adipose tissue lipolysis is elicited by NE, locally released by the SNS, through specifically binding with β3-adrenergic receptor (ADRB3) in adipocytes.14,21–24 Recently, 2 research groups independently reported that alcohol intake results in the SNS activation in mice, although different mechanisms were proposed. Central administration of a low dose of alcohol directly activated the SNS.16 On the other hand, Zhao et al.11 demonstrated that FGF21 overproduction by the liver contributes to the SNS activation in the setting of chronic alcohol exposure. Here, our studies provided additional lines of evidence confirming that the canonical SNS/NE/ADRB3-initated lipolytic pathway activation plays a central role in activating adipose tissue lipolytic activities on chronic alcohol exposure, including: (1) fatty liver and liver injury development in the setting of chronic alcohol exposure are associated with increased circulatory FFAs and glycerol levels, indicative of enhanced adipose tissue lipolysis; (2) epididymal fat pad DN abrogated alcohol-induced augments of plasma FFAs and glycerol contents, suggesting that the SNS activation plays a key role in mediating alcohol-induced adipose tissue lipolytic activation; (3) ADRB3 antagonism with SR59230A, a selective ADRB3 antagonist, prevented alcohol-induced adipose tissue lipolysis activation. Remarkably, both WAT DN and ADRB3 antagonism alleviated liver pathologies of ALD, confirming that adipose tissue FFAs mobilization through the SNS/NE/ADRB3 pathway activation contributes the pathogenesis of ALD.

The SNS/NE/ADRB3 system is not only for lipolysis induction but also a canonical mechanism for the thermogenic program activation in both brown adipocytes in BAT and beige adipocytes in WAT, such as iWAT. After binding to and activating ADRB3 in adipocytes, locally released NE in adipose tissue by the SNS increases the intracellular concentration of cAMP through activating adenylate cyclase, leading to subsequent activation protein kinase A (PKA).22,23 In addition to phosphorylating several lipases/proteins in lipid droplet, including adipose triglyceride lipase, hormone-sensitive lipase, and perilipin, to release FFAs from stored triglycerides (lipolysis), the SNS/NE/ADRB3/cAMP/PKA pathway also leads to increased expression of UCP1 and other thermogenic genes, leading to browning/thermogenesis.10,22,34,35 On the basis of this paradigm, we initially expected to see an induction of thermogenic activities in both BAT and iWAT of mice chronically exposed to an alcohol-containing diet. As expected, we indeed found that chronic alcohol consumption upregulated UCP1 expression in BAT at both basal and adrenergic stimulation conditions. These observations were consistent with previous studies that alcohol intake activated thermogenic programs in BAT in both moderate alcohol drinking mouse model17 and the Lieber-DeCarli ALD mouse model.16 Unexpectedly, we observed that chronic alcohol feeding was associated with inhibited thermogenic activities in iWAT. At both basal and adrenergic stimulation conditions, the expression of UCP1 and several other thermogenic genes were significantly reduced on chronic alcohol exposure. In accordance, mice chronically exposed to alcohol diet manifested a lower body temperature compared with isocaloric control animals when exposed to a cold environment, indicative of an insufficiency of whole-body thermogenic capabilities. It is interesting and noteworthy that alcohol-fed mice, when exposed to cold temperature, showed downregulated UCP1 expression in not only iWAT but also BAT, which contrasted with that on β3-adrenergic stimulation. We do not currently have a clear explanation for the discrepancy. A previous study by Jiang and colleagues showed that cold temperatures and ADRB3 agonists activate distinct cellular populations that express different β-adrenergic receptors to induce beige adipogenesis. Instead of ADRB3, cold-induced beige adipocyte formation requires ADRB1.25 It is unclear whether this scenario also holds true for BAT and further investigations are warranted.

Mechanistic target of rapamycin (mTOR) is an evolutionarily conserved protein serine/threonine kinase and a master regulator of cell growth and anabolic metabolism. It forms 2 structurally and functional different complexes, mTORC1 and mTORC2. mTORC1 is sensitive to rapamycin and regulates anabolic metabolism through directly phosphorylating ribosomal p70S6 kinase (p70S6K) and eIF4E-binding protein.36 Evidence is emerging that mTORC1 plays a key role in regulating adipose tissue functions and several lines of evidence suggest that mTORC1 is required for β3-adrenergic stimulation of adipose browning, which involves a direct phosphorylation of RAPTOR by PKA.27,37 Meanwhile, mTORC1 activity is required to restrain adipose tissue lipolysis activation, mechanistically accounting for clinically observed dyslipidemia caused by rapamycin treatment.28,38 In accordance, mTORC1 inhibition, either genetically or pharmacologically, has been associated with enhanced adipose tissue lipolytic activity and inhibited browning/thermogenesis.27,29 Our observation that alcohol intake enhanced lipolysis while inhibiting thermogenic activities in WAT suggests that chronic alcohol exposure disrupts the physiological coupling between lipolysis and browning/thermogenesis processes in WAT at a certain point. In search of the molecular target(s) responsible for this uncoupling, we uncovered that alcohol-induced inhibition of iWAT browning/thermogenesis was concomitant with suppressed mTROC1 activities. In contrast, in BAT, wherein thermogenic activities were enhanced by alcohol exposure, mTORC1 activities were unchanged in response to chronic alcohol exposure. Remarkably, mTORC1 inhibition in WAT by rapamycin administration was associated with inhibited UCP1 expression, enhanced lipolysis, and exacerbated liver pathologies of ALD without affecting hepatic de novo lipogenic activities. These observations collectively suggest that mTORC1 inhibition in WAT plays a mechanistic role in uncoupling lipolysis and browning/thermogenesis in the setting of chronic alcohol exposure, thereby contributing to ALD development, although we cannot rule out the other effects of mTORC1 activation in hepatic lipid homeostasis. For instance, it has been reported that mTORC1 activation in the liver stimulates VLDL-TAG secretion through promoting phosphatidylcholine synthesis, and hepatic mTORC1 inhibition is associated with the development of hepatosteatosis.39 The finding that rapamycin administration exacerbates liver pathologies of ALD is in stark contrast to the report from Lin et al.40, which demonstrated that rapamycin treatment was protective against ALD through enhancing ethanol-induced macroautophagy in hepatocytes. The discrepant results between the 2 studies are not clear but may be owning to several different experimental settings, including differences in diets (Dyets Inc. vs. Bio-Serv), in rapamycin doses used (5 mg/kg BW vs. 2 mg/kg BW), and in administration protocols (once a day for 7 d before sacrifice vs. every other day for 3 times in the week before sacrifice).

The differential regulation of chronic alcohol intake on thermogenic activities in BAT and WAT has raised the question of whether the inhibitory effect of alcohol on WAT browning/thermogenesis contributes to the pathogenesis of ALD. It has been recently documented that adipocyte-specific Prdm16 knockout (Adipo-Prdm16 KO) results in a functional deficiency of beige fat cells while leaving the classical BAT functionally intact.30,31 To directly assess the pathological role of beige fat cell function in ALD development, we generated Adipo-Prdm16 KO mice and subjected them to the Lieber-DeCarli ethanol-containing diet for 5 weeks. Consistent with previous reports,30,31 Adipo-Prdm16 KO mice manifested UCP1 deficiency in iWAT, whereas UCP1 expression in BAT was only minimally affected. In response to chronic alcohol challenge, Adipo-Prdm16 KO mice showed a tendency of worsened liver pathologies of ALD. No significant changes were observed in the hepatic de novo lipogenic genes. Notably, in comparison to WT control animals, Adipo-Prdm16 KO mice demonstrated increased adipose tissue lipolytic activities on chronic alcohol feeding. This observation is in accordant with the study by Cohen et al.31 showing that adipocyte-specific deletion of Prdm16 is also associated with a s.c. to visceral fat switch. Collectively, our results demonstrate that inhibited WAT browning/thermogenesis contributes to the pathogenesis of ALD.

In conclusion, we have provided original experimental evidence that, through inhibiting mTORC1 activation, chronic alcohol consumption uncouples lipolysis from thermogenesis in WAT. Diminished WAT browning/thermogenesis, concomitant with enhanced lipolysis in the setting of chronic alcohol consumption, contribute to the pathogenesis of ALD (Figure 8). s.c. WAT, wherein the UCP1-positive beige adipocytes locate, is by far the largest adipose depot in humans, thereby playing a pivotal role in the regulation of whole-body energy homeostasis. Considering that human BAT depots share more molecular properties with rodent beige fat than with classical BAT,41–43 our results suggest that targeting adipose tissue mTORC1 to improve beige adipocyte functions could be a potential therapeutic choice for the treatment of ALD.

FIGURE 8.

Free fatty acids (FFAs) overflow due to uncoupled lipolysis and thermogenesis in white adipose tissue (WAT) contributes to alcoholic liver disease (ALD) development. Chronic alcohol consumption inhibits mTORC1 activities in WAT, leading to an uncoupling status between lipolysis and thermogenesis. Ensuing FFAs overflow to the liver promotes hepatic pathologies of ALD.

Supplementary Material

hc9-7-e0059-s001.docx ^{(375.1KB, docx)}

hc9-7-e0059-s002.docx ^{(243.6KB, docx)}

hc9-7-e0059-s003.tiff ^{(2.2MB, tiff)}

hc9-7-e0059-s004.tiff ^{(1.8MB, tiff)}

hc9-7-e0059-s005.tiff ^{(2MB, tiff)}

hc9-7-e0059-s006.tiff ^{(1.8MB, tiff)}

hc9-7-e0059-s007.docx ^{(24.2KB, docx)}