Diabetologia

, Volume 58, Issue 1, pp 149–157

Deletion of the gene encoding G0/G1 switch protein 2 (G0s2) alleviates high-fat-diet-induced weight gain and insulin resistance, and promotes browning of white adipose tissue in mice

  • Wissal El-Assaad
  • Karim El-Kouhen
  • Amro H. Mohammad
  • Jieyi Yang
  • Masahiro Morita
  • Isabelle Gamache
  • Orval Mamer
  • Daina Avizonis
  • Nicole Hermance
  • Sander Kersten
  • Michel L. Tremblay
  • Michelle A. Kelliher
  • Jose G. Teodoro
Article

DOI: 10.1007/s00125-014-3429-z

Cite this article as:
El-Assaad, W., El-Kouhen, K., Mohammad, A.H. et al. Diabetologia (2015) 58: 149. doi:10.1007/s00125-014-3429-z

Abstract

Aims/hypothesis

Obesity is a global epidemic resulting from increased energy intake, which alters energy homeostasis and results in an imbalance in fat storage and breakdown. G0/G1 switch gene 2 (G0s2) has been recently characterised in vitro as an inhibitor of adipose triglyceride lipase (ATGL), the rate-limiting step in fat catabolism. In the current study we aim to functionally characterise G0s2 within the physiological context of a mouse model.

Methods

We generated a mouse model in which G0s2 was deleted. The homozygous G0s2 knockout (G0s2−/−) mice were studied over a period of 22 weeks. Metabolic variables were measured including body weight and body composition, food intake, glucose and insulin tolerance tests, energy metabolism and thermogenesis.

Results

We report that G0s2 inhibits ATGL and regulates lipolysis and energy metabolism in vivo. G0s2−/− mice are lean, resistant to weight gain induced by a high-fat diet and are glucose tolerant and insulin sensitive. The white adipose tissue of G0s2−/− mice has enhanced lipase activity and adipocytes showed enhanced stimulated lipolysis. Energy metabolism in the G0s2−/− mice is shifted towards enhanced lipid metabolism and increased thermogenesis. G0s2−/− mice showed enhanced cold tolerance and increased expression of thermoregulatory and oxidation genes within white adipose tissue, suggesting enhanced ‘browning’ of the white adipose tissue.

Conclusions/interpretation

Our data show that G0s2 is a physiological regulator of adiposity and energy metabolism and is a potential target in the treatment of obesity and insulin resistance.

Keywords

Adipose tissue browning ATGL Energy expenditure G0s2 High-fat diet Insulin resistance Lipolysis Obesity Thermogenesis 

Abbreviations

ATGL

Adipose triglyceride lipase

BAT

Brown adipose tissue

DAG

Diacylglycerol

FA

Fatty acid

G0S2

G0/G1 switch protein 2

H&E

Haematoxylin and eosin

HFD

High-fat diet

HSL

Hormone-sensitive lipase

IPGTT

Intraperitoneal glucose tolerance test

IPITT

Intraperitoneal insulin tolerance test

MAG

Monoacylglycerol

MAGL

Monoacylglycerol lipase

PPARγ

Peroxisome proliferator-activated receptor γ

RQ

Respiratory quotient

TG

Triacylglycerol

UCP1

Uncoupling protein 1

\( \overset{\cdot }{V}{\mathrm{O}}_2 \)

Oxygen consumption

WAT

White adipose tissue

WT

Wild-type

Introduction

Obesity is recognised as a global epidemic with serious implications for worldwide health. The medical condition is closely associated with several metabolic diseases known as the metabolic syndrome (reviewed in [1]), characterised by dyslipidaemia, hypertension, nonalcoholic fatty liver disease, insulin resistance and others. These metabolic complications brought about by increased fat mass, particularly in the intra-abdominal area, are closely associated with insulin resistance, which subsequently leads to type 2 diabetes [2]. During states of fasting or high energy demand (e.g. exercise), fat stored in adipocytes is mobilised by a series of lipolytic enzymes, leading to the release of NEFA into the circulation. Adipose triglyceride lipase (ATGL) [3, 4, 5], the first and rate-limiting enzyme in this series, acts on triacylglycerol (TG) and hydrolyses it into diacylglycerol (DAG). Hormone-sensitive lipase (HSL) [6, 7] and monoacylglycerol lipase (MAGL) further hydrolyse DAG and monoacylglycerol (MAG) leading to NEFA release into the circulation. Increased circulating NEFA levels are strongly associated with insulin resistance and play a major role in the pathogenesis of type 2 diabetes [8]. Thus, lipid homeostasis is governed by genes and signalling molecules responsible for maintaining energy homeostasis.

Despite the importance of lipid and energy homeostasis to human health, these processes are not yet fully understood. Recently, G0S2 emerged as a potential player in lipid homeostasis regulating the rate-limiting step of lipid catabolism [9]. G0/G1 switch protein 2 (G0S2) was shown to directly interact with ATGL and inhibit its catalytic activity [10]. G0S2 is ubiquitously expressed and has been shown to localise to different cellular compartments including endoplasmic reticulum, mitochondria and lipid droplets [9, 10, 11, 12]. Moreover, G0S2 was reported to be a peroxisome proliferator-activated receptor γ (PPARγ) target gene containing a PPAR response element [11] and a carbohydrate response element in its promoter [13]. Thus, G0S2 is emerging as a key regulator of intracellular lipid metabolism but its function has yet to be placed within the physiological context of the whole organism.

In this study, we focused on the metabolic aspects of G0s2 and its regulation of ATGL in a mouse model in which the G0s2 gene is deleted. We investigated the role of G0s2 in lipid metabolism and energy homeostasis in vivo and examined the fate of the released NEFA.

Methods

Generation of G0s2 knockout mice

The homozygous G0s2 knockout (G0s2−/−) mice were generated by homologous recombination in embryonic stem cells. A schematic of the targeting strategy is shown in Fig. 1a and additional information is described in electronic supplementary material (ESM Methods). All animal protocols were approved by the Canadian Council on Animal care and the McGill University Animal Care committee. 129SvEv (wild-type [WT]) and G0s2−/− mice were housed in a controlled environment at 21°C with a 12 h light/dark cycle and free access to food and water.
Fig. 1

Effects of G0s2 deletion on weight gain. (a) Schematic representation of the G0s2 knockout (KO) gene with locations of primers used for genotyping indicated. (b) PCR genotyping of G0s2 WT, heterozygote and KO genes as described in the Methods. (c) G0s2 expression levels in liver, fat and muscle tissues of WT (white bars) and KO (black bars) mice. (d) G0S2 protein levels in liver lysates. (e) Comparison of body length (nose–anus length) of WT and KO mice. (f) Comparison of body weights of male and female mice at 6 weeks on chow diet (n = 11 for WT and 18 for KO male mice; n = 6 for WT and n = 14 for female KO mice). (g) Body mass composition of 22-week-old male WT and KO mice on chow diet (n = 8 for WT and 7 for KO mice). (h, i) Percentage of weight gain of male WT (white circles) and KO (black squares) mice housed on either chow (h) or HFD (i) for 22 weeks (n = 8 for WT mice on both diets; n = 13 on chow diet and n = 12 for KO mice on HFD). Data are presented as means ± SEM. **p < 0.01, ***p < 0.001

Weight studies

At 6 weeks of age, mice were placed on either a HFD (24.6% carbohydrates, 21% protein and 54.3% fat on the basis of energy; Harlan Teklad, Toronto, ON, Canada) or a chow diet (60% carbohydrates, 24% protein and 16% fat on the basis of energy; Harlan, Teklad) for 22 weeks. Body weight was measured and recorded on a weekly basis and expressed as percentage of weight gain. At the end of the experiment, mice were killed and tissues were rapidly collected, weighed and either frozen in liquid nitrogen and stored at −80°C or processed for histological analysis.

Adipocyte measurements

Histological analyses were performed using haematoxylin and eosin (H&E) stained cross-sectional images using ImageJ software (NIH, Bethesda, MD, USA). More than 1,500 cells on 10× images and 9,000 nuclei on 20× images were counted manually using the cell counter plugin available on ImageJ from each genotype for white adipose tissue (WAT) and brown adipose tissue (BAT), respectively.

Adipocyte differentiation and lipolysis measurement

Pre-adipocytes isolated from 3- to 4-week-old mice were differentiated as described previously [14]. Glycerol (BioVision, Milpitas, CA, USA) and NEFA (Wako Chemicals, Richmond, VA, USA) were measured according to the manufacturer’s instructions using appropriate standards. Data were normalised to protein content.

Real-time quantitative PCR analysis

Real-time quantitative PCR analysis and the primers used are listed in ESM Methods and in ESM Table 1.

Statistical analyses

Data are presented as means ± SEM, as calculated by GraphPad Prism (San Diego, CA, USA) software using unpaired two-tailed Student’s t test. p < 0.05 was considered significant.

Details of the tolerance tests, in vivo metabolic and thermogenesis measurements, body mass composition, TG content, TG lipase activity, GC-MS and western blot analysis are available in the ESM Methods.

Results

G0s2−/− mice are resistant to diet-induced obesity

To understand the role of the G0s2 gene in lipolysis and whole body metabolism, we generated mice with targeted G0s2 deletions using homologous recombination (Fig. 1a, b). G0s2 mRNA and protein levels are shown in Fig. 1c, d. All experiments were performed in the 129SvEv genetic background. The G0s2−/− mice were viable and fertile; however, they were leaner than littermate controls (ESM Fig. 1a), with no difference in body length (Fig. 1e). Figure 1f shows weights of mice at 6 weeks of age on a chow diet. Both male and female G0s2−/− mice were leaner than littermate controls; however, only in male G0s2−/− mice was the reduction in weight statistically significant (21.36 ± 0.8 g for WT vs 18.47 ± 0.3 g for G0s2−/− mice, p = 0.001). Analysis of body composition using MRI showed an increase in whole body lean mass and a decrease in fat mass by around 15% and 50%, respectively, in 18-week-old G0s2−/− mice compared with WT mice (Fig. 1g). To further elucidate the effects of G0s2 on body weight, WT and G0s2−/− mice were placed either on a chow or HFD and weights were recorded weekly for 22 weeks. There was no difference in weight gain between WT and G0s2−/− mice on the chow diet (Fig. 1h); however, on the HFD G0s2−/− mice were clearly leaner and resistant to obesity induced by the HFD (Fig. 1i and ESM Fig. 1b, c). The resistance to weight gain became apparent at around 12 weeks of HFD feeding and persisted until the end of the experiment. Although 129SvEv background mice are known to be somewhat resistant to diet-induced obesity [15, 16], G0s2−/− male mice showed increased resistance to HFD-induced obesity compared with WT littermates.

Blood chemistry analyses of WT and G0s2−/− mice are shown in Table 1. There were no significant changes in any of the measured variables in either fed or fasted states, including glucose, insulin or TG. Sera TG levels were decreased in knockout mice under HFD conditions but the difference did not achieve statistical significance.
Table 1

Plasma levels of glucose, insulin, NEFA, TG, cholesterol and HDL in 22-week-old fed or overnight-fasted male WT and G0s2−/− mice on chow and HFD

Variable

Chow

HFD

WT

G0s2−/−

WT

G0s2−/−

Glucose (mmol/l)

4.8 ± 0.18 (n = 8)

4.26 ± 0.18 (n = 10)

5.57 ± 0.32 (n = 8)

5.8 ± 0.2 (n = 11)

Glucose (mmol/l) (fed)

8.3 ± 0.13 (n = 8)

8.1 ± 0.12 (n = 10)

8.3 ± 0.13 (n = 8)

8.3 ± 0.18 (n = 10)

Insulin (pmol/l) (fed)

178.5 ± 28 (n = 6)

171.5 ± 10.5 (n = 6)

490 ± 98 (n = 8)

437.5 ± 105 (n = 9)

NEFA (mmol/l)

1.41 ± 0.28 (n = 8)

1.22 ± 0.16 (n = 13)

1.35 ± 0.14 (n = 8)

0.94 ± 0.13 (n = 12)

TG (mmol/l)

ND

ND

9.23 ± 2.2 (n = 7)

8.83 ± 1.1 (n = 11)

Cholesterol (mmol/l)

3.6 ± 0.08 (n = 6)

3.49 ± 0.15 (n = 11)

3.9 ± 0.21 (n = 8)

3.9 ± 0.16 (n = 12)

HDL-cholesterol (mmol/l)

3.58 ± 0.12 (n = 6)

3.4 ± 0.19 (n = 11)

4.02 ± 0.16 (n = 8)

3.9 ± 0.13 (n = 12)

Data are presented as means ± SEM (n)

ND, not determined

Although no overall differences in NEFA levels were observed, we determined if specific NEFA species were affected. Studies of MAGL have suggested that alterations in lipase activity can lead to differences in the profile of specific NEFA species [17]. We therefore determined if this was also the case in G0s2−/− mice. Fatty acid (FA) profiling using GC-MS on lipids from visceral fat (ESM Fig. 2) was performed but did not show statistical differences between genotypes.

G0s2−/− mice have reduced fat deposition in the liver

Analysis of organ weights and morphology showed no significant changes between genotypes under conditions of the chow diet (Fig. 2a); however, under the HFD significant increases were observed in weights of liver and WAT of WT animals that were not apparent in knockout mice (Fig. 2a). Histological analysis of liver sections showed increased fat deposition in the form of lipid droplets in WT mice, which were completely absent in G0s2−/− mice (Fig. 2b). TG levels in livers of WT mice were on average 60% higher than those of the G0s2−/− mice when fed chow diet. Under the HFD, differences in TG levels were even greater, with WT mice having ∼80% more TG relative to knockout mice (p < 0.0001; Fig. 2c). Surprisingly, no differences were observed in NEFA levels in liver (Fig. 2d), suggesting that NEFA may be rapidly used or oxidised in G0s2−/− mice. Consistent with the observation showing less fat deposition in livers of G0s2−/− mice, we generally observed reduced expression of genes encoding lipogenic factors such as Scd1 and Srebp1c (also known as Srebf1) and elevated levels of lipid oxidative genes (Fig. 2e).
Fig. 2

TG accumulation in livers of G0s2−/− mice. (a) Analysis of body organs of WT (white bars) and G0s2−/− (KO) (orange) mice on chow diet and WT (blue) and KO (green) mice on HFD. (b) Representative images of H&E staining of liver sections of WT and KO mice on HFD (n = 4 per genotype). (c, d) Liver TG and NEFA levels of WT (white bars) and KO (black bars) mice on chow and HFD. (e) Expression analysis of lipogenic and oxidative genes in livers of HFD-fed mice normalised to cyclophilin. Values represent means ± SEM. n = 8 for WT and 11 for G0s2−/− mice on chow diet; n = 5 for WT and 8 for G0s2−/− mice on HFD. *p < 0.05, **p < 0.01, ****p < 0.0001

G0s2−/− mice have smaller adipocytes and elevated lipolysis

Figure 2a indicated that reduced body mass of the G0s2−/− mice is due primarily to differences in the weight of visceral fat. On HFD, but not chow, the weight of visceral fat was significantly elevated in WT animals relative to G0s2−/− mice. Further analyses of histological sections of visceral fat (Fig. 3a) showed a decrease in the average area of individual adipocytes in G0s2−/− mice (Fig. 3b). This again suggests that in G0S−/− mice, adipocytes are smaller due to enhanced lipolysis and a decrease in the accumulation of TG. Similar trends were observed with BAT, although the values did not reach statistical significance (Fig. 3a, d). No changes were observed in the size of fat pads between WT and G0s2−/− mice (Fig. 3a, c).
Fig. 3

Decreased adipocytes size and increased lipase activity in G0s2−/− (KO) mice. (a) H&E staining of WAT, fat pads and BAT of mice on an HFD for 22 weeks (scale bar, 100 μm). (b) Average WAT area, (c) fat pad area and (d) BAT area of WT (white bars) and KO (black bars) mice. (e) TG hydrolase activity measured in visceral fat extracts of male WT (white bars) and KO (black bars) mice. (f) Protein levels of ATGL and HSL in visceral lysates of WT and KO mice. Values represent means ± SEM. n = 4 per genotype for adipocyte area; n = 6 per genotype for TG hydrolase activity and n = 4 per genotype for protein levels. **p < 0.01, ***p < 0.001

To understand the role of G0s2 in lipolysis in vivo, we measured TG hydrolase activity in visceral fat from WT and G0s2−/− mice. TG hydrolase activity was significantly higher in G0s2−/− mice than WT littermate controls (about a twofold increase), consistent with G0S2 being a lipase inhibitor (Fig. 3e) [9, 18]. Moreover, activity was even further enhanced in G0s2−/− mice fed HFD (more than 2.5-fold) compared with WT controls (Fig. 3e). In addition to the increased lipase activity in G0s2−/− mice, ATGL protein levels were also increased under both chow and HFD conditions in G0s2−/− mice (Fig. 3f), suggesting that G0S2 may inhibit ATGL activity by decreasing steady state levels of the enzyme. No changes were observed in levels of HSL under both diets (Fig. 3f). No changes were observed at the expression level of these key lipolytic genes (ESM Fig. 3).

We further analysed pre-adipocytes isolated from 3-week-old WT and G0s2−/− mice differentiated in vitro. Differentiated adipocytes were stimulated with the β-adrenergic agonist isoprenaline (10 μmol/l) for 2 h. Glycerol and NEFA levels were measured in media of the differentiated cells. Figure 4 shows increased basal and stimulated lipolysis as measured by glycerol release (Fig. 4a) and NEFA levels (Fig. 4b) in G0s2−/− adipocytes. In vivo lipolysis was also carried out in fasted 10–12 week WT and G0s2−/− mice treated with isoprenaline. As expected, treating mice with isoprenaline increased serum-free glycerol levels in all mice indicating hormone stimulated lipolysis; however, there were no significant differences between WT and G0s2−/− mice glycerol levels (data not shown).
Fig. 4

Deletion of G0s2 enhances lipolysis. In vitro lipolysis of differentiated adipocytes from 3-week-old mice under basal and isoprenaline-stimulated conditions. (a) Glycerol levels, (b) NEFA levels of WT (white bars) and G0s2−/− (KO) (black bars) mice (n = 9–10 mice per condition). (c) Gene expression analysis of adipogenic differentiation markers normalised to cyclophilin. *p < 0.05, **p < 0.01, ****p < 0.0001

To determine whether the decrease in adipocyte size is related to G0s2 regulation of adipocyte differentiation, gene expression analysis of adipocyte transcription factors and differentiation genes was performed [19, 20, 21]. Figure 4c shows no significant changes between genotypes, indicating that deletion of G0s2 does not interfere with adipocyte differentiation. Thus, the lean phenotype and decrease in adipocyte size observed in absence of G0s2 is likely to be due to elevated ATGL activity and enhanced lipolysis.

G0s2−/− mice displayed improved glucose and insulin tolerance on high-fat feeding

To examine whether insulin resistance developed in mice on the HFD, intraperitoneal glucose (IPGTT) and insulin (IPITT) tolerance tests were performed on overnight-fasted male WT and G0s2−/− mice (Fig. 5). On chow diet, glucose and insulin response curves were not different (Fig. 5a, c). However, in high-fat-fed mice, plasma glucose levels were higher in WT mice, suggesting increased glucose tolerance in the knockout mice (Fig. 5b). Moreover, G0s2−/− mice exhibited lower levels of hyperglycaemia in response to insulin load (Fig. 5d), suggesting that these mice are more insulin tolerant. On either diet, when IPITT or IPGTT were performed on female mice, no differences between genotypes were found (ESM Fig. 4a–d). Insulin levels during IPGTT in male mice are shown in ESM Fig. 4e.
Fig. 5

Improved glucose and insulin sensitivity in G0s2−/− (KO) male mice. Blood glucose levels during (a, b) IPGTT and (c, d) IPITT in fasted WT (black squares) and KO (black triangles) mice on (a, c) chow diet and (b, d) HFD. Values represent means ± SEM. n = 8 for WT and n = 13 for KO mice. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Enhanced energy expenditure in G0s2−/− mice

G0s2−/− mice are leaner and accumulate significantly less fat, suggesting enhanced energy expenditure. To further study this phenotype, mice fed either chow or HFD were placed in metabolism cages (Columbus Instruments, Columbus, OH, USA) for 4 days with free access to food and water. Oxygen consumption, activity and body temperature were recorded. Oxygen consumption measurements (\( \overset{\cdot }{V}{\mathrm{O}}_2 \)) showed an increase in G0s2−/− mice both in light and dark cycles (about 20%); however, the measurement was only statistically significant in dark cycles (Fig. 6a). No significant differences were observed in activity or food intake throughout the course of the metabolism cage analysis (Fig. 6b, c), confirming that the increased energy expenditure is not linked to physical activity or food intake, but is rather due to intrinsic changes in metabolism in G0s2−/− mice.
Fig. 6

Increased energy expenditure in G0s2−/− (KO) mice. Mice were housed individually for 4 days in metabolism cages and fed either chow or HFD. (a) \( \overset{\cdot }{V}{\mathrm{O}}_2 \) and (b) ambulatory activity of mice on HFD measured as counts per 12 h cycle. (c) Food intake on chow diet. WT (white bars) and KO (black bars). Respiratory quotient in WT (white squares) and KO (black hexagons) mice fed (d) chow or (e) HFD. Values represent means ± SEM. n = 7 per genotype. *p < 0.05; in (d) p < 0.001 WT vs KO at all time points

Intriguingly, on chow diet, the respiratory quotient (RQ), which represents the ratio of (\( \overset{\cdot }{V}{\mathrm{CO}}_2 \)/\( \overset{\cdot }{V}{\mathrm{O}}_2 \)), was shifted towards 0.7 in G0s2−/− mice, indicating that fat is a preferred substrate in these animals (Fig. 6d). The lower RQ is consistent with the G0s2−/− animals having enhanced lipolysis and fat metabolism. The RQ values of WT and G0s2−/− mice were nearly identical once animals were placed on a HFD since this forces the use of lipids in both groups (Fig. 6e).

Enhanced thermogenesis in G0s2−/− mice

Brown adipocytes, unlike white adipocytes, produce heat as a means to counteract hypothermia, obesity and diabetes. Brown fat has a high mitochondrial content and mitochondrial uncoupling protein 1 (UCP1) can be used to uncouple respiration and dissipate chemical energy as heat. Analysis of G0s2−/− mice housed in metabolism cages, showed an increase in heat production (Fig. 7a, b) on 14 weeks of HFD conditions (Fig. 7b). mRNA and protein levels of ATGL, HSL and UCP1 were measured in BAT tissue and showed a slight increase in UCP1 levels under HFD conditions (ESM Fig. 5). These data prompted us to investigate the adaptation of G0s2−/− mice to cold exposure. Mice (12 weeks old) on chow diet were housed at 4°C and their rectal temperature was recorded during the first 5 h and then at 30 h (Fig. 7c). G0s2−/− mice had an approximately 0.7°C higher body temperature than WT mice. Moreover, abdominal temperature was measured at 30 h and was on average 0.8°C higher in G0s2−/− animals (Fig. 7d). These data suggest that G0s2−/− mice are overall more thermogenic and the heat generated is able to counteract obesity.
Fig. 7

Increased thermogenicity and oxidation in G0s2−/− (KO) mice. Body temperature measured indirectly as heat production in metabolism cages of WT (white bars) and KO (black bars) mice on (a) chow and (b) HFD. (c, d) Rectal and abdominal temperatures of WT (white) and KO (black) mice on chow diet housed for 30 h at 4°C. (eh) Expression levels of thermogenic and oxidation genes in WAT of WT (white bars) and KO (black bars) mice on HFD housed at (e, g) room temperature or (f, h) 4°C for 8 h. n = 7–8 per genotype. *p < 0.05, **p < 0.01, ***p < 0.001

We next examined whether BAT molecular markers were increased in WAT of G0s2−/− mice at normal and cold temperatures. HFD-fed G0s2−/− mice expressed increased levels of brown fat genes at normal temperature (Fig. 7e), and these increases were enhanced and significant under cold temperature (4°C) (Fig. 7f). We observed an increase in the expression levels of key thermogenic genes, such as Ucp1 and Dio2, in G0s2−/− mice (Fig. 7e, f). This overall increase in selective brown fat genes in G0s2−/− WAT suggests an overall browning of these tissues. In addition, there was a significant increase in genes that encode proteins for both mitochondrial and peroxisomal FA oxidation (Fig. 7g, h) such as Cox8b, Acox, Pparα, Cpt1β, Pgc1α, Pgc1β and Ech1 in HFD-fed mice (Fig. 7h). These results suggest that the increased thermogenesis and browning of WAT are likely to be responsible for the enhanced energy dissipation in the knockout mice.

Discussion

The current study demonstrates the role of G0s2 as an in vivo regulator of lipase activity. Mice lacking G0s2 expression had elevated TG hydrolase activity, were leaner and more resistant to HFD weight increase than WT littermates. In addition, lipolysis was enhanced in differentiated adipocytes isolated from G0s2−/− mice and markedly enhanced in response to stimulation by isoprenaline. G0s2−/− mice had reduced TG levels in the liver, and both WAT and BAT cells were smaller than those in WT controls. In addition, G0s2−/− mice had improved glucose tolerance and insulin sensitivity and were more thermogenic due to enhanced fat catabolism. Taken together, the data provide strong evidence that G0S2 is a key regulator of lipid metabolism in vivo.

Our results are consistent with recent findings in a transgenic mouse model in which Atgl (also known as Pnpla2) was overexpressed in adipose tissue causing reduced TG content and enhanced lipolysis [22]. By contrast, Atgl−/− mice accumulate large amounts of fat in multiple tissues and die at around 12 weeks of age [23]. Correlative studies have indicated that the G0S2 regulatory effects on lipid metabolism may also extend to humans. For example, fasting in humans was shown to reduce G0S2 levels and elevate ATGL [24].

Despite the enhanced lipolysis and decreased TG levels in tissues and sera, serum NEFA levels in the G0s2−/− mice were surprisingly unchanged. Moreover, we noticed that on both diets, NEFA levels in the liver and sera of G0s2−/− mice tended to be even lower than those of WT mice, suggesting enhanced fat use/uptake. Under normal conditions, a balance between hydrolysis and use/uptake regulates NEFA levels by peripheral tissues. Many factors can regulate levels of NEFA in serum. Thus, the lack of increased serum NEFA levels suggests alternative mechanisms by which NEFA are used. We have observed enhanced glucose tolerance and insulin sensitivity in G0s2−/− mice, which could suggest increased FA uptake by peripheral tissues. Oxidation and/or re-esterification of NEFA are potential mechanisms by which NEFA are used. NEFA can also be recycled back to DAG or MAG through the TG/NEFA cycle bypassing the TG step, perhaps favouring shorter cycles to ensure the availability of different signalling intermediates (discussed in [25]).

G0s2−/− mice had higher body temperatures, increased oxygen consumption and lower RQ ratio, indicating more fat usage. Moreover, knockout mice were more tolerant to cold temperatures and maintained higher body temperatures than WT mice. We speculate that elevated fat oxidation, and possibly uncoupling, are the mechanisms by which G0s2−/− mice are able to use fat more rapidly and thus generate heat. Increased expression of brown fat-like genes in WAT of G0s2−/− mice further argues for enhanced uncoupling and energy dissipation in these animals as a means to counteract obesity. Our findings are in agreement with other mouse models that display enhanced FA oxidation and decreased serum NEFA levels [26, 27]. Overexpression of Ucp1 in WAT of mice for example, resulted in enhanced uncoupling, leanness and increased FA oxidation [26] with improved insulin sensitivity [28]. Prolonged cold exposure is known to induce WAT into a BAT-like state, a process termed browning [29, 30]. Interestingly, a recent study demonstrated that in response to cold exposure, expression levels of ATGL became elevated, whereas levels of G0S2 were decreased in fat deposits of SvEv mice [31]. These changes in gene expression would enhance the lipolytic potential of adipocytes. Our findings suggest that lack of G0S2 activity is able to accelerate the process of browning. Inducing the browning of WAT has been proposed as a potential therapy for obesity and type 2 diabetes and our findings suggest that G0S2 may be an effective therapeutic target in this regard [32].

Our study is the first to show that G0S2 regulates WAT browning by acting as a negative regulator of lipolysis. A recent report by Zhang et al has also described the phenotype of G0s2−/− mice [33]. Both our study and that of Zhang et al [33] show that G0s2−/− mice are leaner and resistant to HFD feeding in 129SvEv and C57BL/6J genetic background, respectively. Both studies show a decrease in TG accumulation in the liver and enhanced lipase activity in visceral fat and liver. However, in our study, effects of G0S2 on lipolysis, resistance to HFD obesity, and improvement in glucose and insulin sensitivity were more significant in male than in female G0s2−/− mice, whereas in the study by Zhang et al female mice displayed greater resistance to HFD-induced obesity [33]. These variations may be due to the different mouse strains used in the two studies or to metabolic adaptation/compensation in female 129SvEv mice in our study. Sex differences have been previously reported to have different effects in metabolic studies [34]. Our data are also in agreement with a recently published report of G0s2 overexpression in adipocytes [35]. In this model, overexpression of G0s2 in adipocytes rendered the mice unable to maintain their body temperature at 4°C, probably due to increased fat deposition in these mice. A new report also showed G0S2 regulation of thermogenesis and an increase in brown-like and oxidation genes in BAT of G0s2−/− mice [36].

In conclusion, our study highlights the role of G0S2 in whole body lipid metabolism and suggests that G0S2 may have a role in metabolic disorders such as obesity, insulin resistance and type 2 diabetes.

Acknowledgements

We thank S. Hardy (Goodman Cancer Research Centre, McGill University, Montreal, QC, Canada) for reading the manuscript and A. L. Loy (Goodman Cancer Research Centre) for assistance with mouse experiments.

Funding

This work was supported by grants from the CIHR (MOP-115195) and Cancer Research Society (CRS) to JGT and NIH grant AI075118 to MAK.

Duality of interest

The authors declare that there is no duality of interest associated with this manuscript.

Contribution statement

WE-A, MAK, SK, MLT and JGT designed the study. WE-A, KE-K, JY, IG, and NH contributed to the acquisition and interpretation of data. AHM, MM, OM and DA analysed the data. WE-A and JGT wrote the manuscript with all other authors contributing to the review and revision of intellectual content. All authors approved the final manuscript. JGT is the guarantor of this work.

Supplementary material

125_2014_3429_MOESM1_ESM.pdf (110 kb)
ESM Methods(PDF 109 kb)
125_2014_3429_MOESM2_ESM.pdf (77 kb)
ESM Table 1(PDF 77 kb)
125_2014_3429_MOESM3_ESM.pdf (27.6 mb)
ESM Fig. 1(PDF 28267 kb)
125_2014_3429_MOESM4_ESM.pdf (249 kb)
ESM Fig. 2(PDF 249 kb)
125_2014_3429_MOESM5_ESM.pdf (98 kb)
ESM Fig. 3(PDF 97 kb)
125_2014_3429_MOESM6_ESM.pdf (167 kb)
ESM Fig. 4(PDF 167 kb)
125_2014_3429_MOESM7_ESM.pdf (468 kb)
ESM Fig. 5(PDF 467 kb)

Copyright information

© Springer-Verlag Berlin Heidelberg 2014

Authors and Affiliations

  • Wissal El-Assaad
    • 1
  • Karim El-Kouhen
    • 1
  • Amro H. Mohammad
    • 2
  • Jieyi Yang
    • 2
  • Masahiro Morita
    • 1
  • Isabelle Gamache
    • 1
  • Orval Mamer
    • 3
  • Daina Avizonis
    • 3
  • Nicole Hermance
    • 4
  • Sander Kersten
    • 5
  • Michel L. Tremblay
    • 1
    • 2
  • Michelle A. Kelliher
    • 4
  • Jose G. Teodoro
    • 1
    • 2
    • 6
  1. 1.Goodman Cancer Research CentreMcGill UniversityMontréalCanada
  2. 2.Department of BiochemistryMcGill UniversityMontréalCanada
  3. 3.Metabolomics Core Facility, Goodman Cancer Research CentreMcGill UniversityMontréalCanada
  4. 4.Department of Cancer BiologyUniversity of MassachusettsWorcesterUSA
  5. 5.Nutrition, Metabolism, and Genomics Group, Division of Human NutritionWageningen UniversityWageningenthe Netherlands
  6. 6.Department of Microbiology and ImmunologyMcGill UniversityMontréalCanada

Personalised recommendations