Glycerol-3-phosphate acyltransferases (GPATs) catalyze the first step in the synthesis of glycerolipids and glycerophospholipids. Microsomal GPAT, the major GPAT activity, is encoded by at least two closely related genes, GPAT3 and GPAT4. To investigate the in vivo functions of GPAT3, we generated Gpat3-deficient (Gpat3−/−) mice. Total GPAT activity in white adipose tissue of Gpat3−/− mice was reduced by 80%, suggesting that GPAT3 is the predominant GPAT in this tissue. In liver, GPAT3 deletion had no impact on total GPAT activity but resulted in a 30% reduction in N-ethylmaleimide-sensitive GPAT activity. The Gpat3−/− mice were viable and fertile and exhibited no obvious metabolic abnormalities on standard laboratory chow. However, when fed a high-fat diet, female Gpat3−/− mice showed decreased body weight gain and adiposity and increased energy expenditure. Increased energy expenditure was also observed in male Gpat3−/− mice, although it was not accompanied by a significant change in body weight. GPAT3 deficiency lowered fed, but not fasted, glucose levels and tended to improve glucose tolerance in diet-induced obese male and female mice. On a high-fat diet, Gpat3−/− mice had enlarged livers and displayed a dysregulation in cholesterol metabolism. These data establish GPAT3 as the primary GPAT in white adipose tissue and reveal an important role of the enzyme in regulating energy, glucose, and lipid homeostasis.
- lipid synthesis
acyl-coa:glycerol-3-phosphate acyltransferase (GPAT, EC 188.8.131.52) catalyzes the first and committed step in de novo triacylglycerol (TAG) and phospholipid biosynthesis, converting glycerol 3-phosphate into lysophosphatidic acid (LPA). In mammals, GPAT activities have been classified into two distinct categories: a mitochondria-associated GPAT activity, which is resistant to the sulfhydryl-modifying reagent N-ethylmaleimide (NEM), and a NEM-sensitive GPAT activity, which is detected primarily in the microsomal fraction (11, 27). We and others have previously identified two closely related members of the glycerolipid acyltransferase family as endoplasmic reticulum-associated microsomal GPAT enzymes, now named GPAT3 and GPAT4 (3, 5, 17, 22). These two proteins were also previously known as AGPAT8/9/10 or LPAAT-θ (GPAT3) (2, 23, 24, 26) or AGPAT6 (GPAT4) (2, 5, 26). Recent studies examining expression and regulation of these two genes, knockdown experiments in adipocytes, as well as analysis of GPAT4 knockout mice suggest that GPAT3 and GPAT4 may control lipid metabolism in a tissue- and stimulus-dependent fashion: GPAT4 constitutes the major microsomal GPAT activity in liver, whereas GPAT3 appears to be a critical regulator for lipid accumulation in white adipocytes (2, 5, 17, 22, 26). Whereas GPAT3 mRNA is strongly regulated by nutritional or hormonal stimulation in both adipose tissue and liver, little regulation of GPAT4 is observed, suggesting that these two enzyme may represent a constitutive (GPAT4) and an inducible (GPAT3) isoform of GPAT activity (22), respectively.
The importance of GPAT4 in mediating triglyceride biosynthesis in physiological settings was demonstrated in GPAT4-deficient (Gpat4−/−) mice that had been generated using a gene trap approach prior to identification of its enzymatic activity. Gpat4−/− mice are resistant to genetic and diet-induced obesity, exhibit decreased subcutaneous, epididymal, and inguinal fat, and have defective lipid production in the mammary gland (2, 26). Interestingly, the decreased adiposity of Gpat4−/− mice is observed in the presence of unaltered GPAT activity in white adipose tissue (WAT), suggesting that it is secondary to altered lipid metabolism in other tissues. Both liver and brown adipose tissue (BAT) show substantially decreased GPAT activity in GPAT4 knockout mice and, thus, could be the primary site of action of GPAT4.
Here, we examine the functional role of GPAT3 in vivo, using Gpat3−/− mice. Our data confirm GPAT3 as the predominant microsomal glycerol-3-phosphate acyltransferase in WAT and reveal a small, albeit significant, role for GPAT3 in energy and glucose homeostasis under conditions of high-fat feeding, which was modified by the sex of the animals. Our data also identify an unexpected effect of GPAT3 deletion on cholesterol metabolism and liver function, which could be important when one is considering this enzyme as a potential drug target.
MATERIALS AND METHODS
GPAT gene targeting and production of Gpat3−/− mice.
The knockout strategy was designed to replace exons 4 through 7. Deletion of exon 4 (271 bp), exon 5 (75 bp), exon 6 (90 bp), and exon 7 (94 bp), including some flanking intron, total of 2830 bp of genomic DNA, including 530 bp in total of exon 4 through 7 (143 aa, covering most of the acyltransferase domain) will produce a GPAT3 open reading frame shift and therefore was expected to produce a null mutation. Gene targeting for the Gpat3 allele was carried out in C57BL6 embryonic stem (ES) cells. To generate the Gpat3 targeting construct, a fragment of 14875 bp of genomic DNA covering exons 4 through 12 was retrieved from a BAC clone by recombineering technology. The final Gpat3 knockout construct was created by replacing exons 4 through 7 with a loxP-Neo-loxP cassette (Fig. 1A), also by BAC recombineering. The resulting Gpat3 knockout construct contains a 4576-bp 5′-homologous arm and a 6950-bp 3′-homologous arm. ES cell targeting and Southern screening were completed at Taconic (Hudson, NY). Eight resistant ES clones were identified positive in primary screening with 5′ probe and StuI digestion from 192 clones. Four of the eight ES cell clones were expanded and confirmed by both 5′ and 3′ probe Southern blot analysis (Fig. 1B). Two positive ES cell clones were used for blastocyst injection, and six chimera F0 founders were generated from a total of 15 live births from two injections. Germline transmission was confirmed for both ES cell lines. We generated Gpat3−/− and wild-type littermates from breeding heterozygous mice. All animal experimental procedures were performed in accordance with regulations and established guidelines and were approved by Pfizer's Institutional Animal Care and Use Committee.
Animal maintenance and metabolic measurements.
Age-matched wild-type and Gpat3−/− littermates were singly housed under controlled temperature (22 ± 2°C) with free access to food and water. The mice were either kept on the low-fat standard chow diet or switched to the high-fat diet containing 60% calories from fat (D12492; Research Diets, New Brunswick, NJ) at 10 wk of age. Body weight and food intake were monitored weekly. Body composition was measured by an EchoMRI-3in1 machine (EchoMRI, TX). Physical activity, oxygen consumption, and CO2
Fasting blood glucose was measured with a one-touch ultraglucometer. Insulin levels were analyzed using the Ultrasensitive rat insulin ELISA kit. All other plasma biochemical parameters, including the fed glucose, were measured in samples collected at the conclusion of the study (under fed condition) by Roche Hitachi clinical analyzer and associated kits.
Glucose tolerance test.
Mice fasted overnight were given a bolus of glucose via gavage (40% glucose, 10 ml/kg). Glucose in tail vein blood was measured immediately before administration of glucose (time 0 min) and at 15, 30, 60, and 120 min after glucose challenge. At the same time points, blood samples were also collected from tail vein, and insulin levels were determined.
GPAT and AGPAT enzyme assays.
A GPAT activity assay was performed as described (3), using [14C]glycerol 3-phosphate (G3P) and lauroyl-CoA as substrates. Briefly, liver and fat pad tissues isolated from wild-type or Gpat3−/− mice fed a high-fat diet were homogenized in cold phosphate-buffered saline, and the total protein level in each sample was measured with a BCA protein assay kit (Thermo Scientific, Rockford, IL). The enzyme assay reaction was conducted in 75 mM Tris·HCl, pH 7.5, 4 mM MgCl2, 1 mg/ml fatty acid-free bovine serum albumin with 150 μM [14C]G3P (55 mCi/mmol), and 50 μM lauryl-CoA for 15 min at room temperature in the presence of 50 (liver) or 100 (adipose tissue) μg of total lysate protein in a total volume of 100 μl. The 1-acylglycerol-3-phosphate O-acyltransferase (AGPAT) assay was performed under similar conditions with LPA and [14C]oleoyl-CoA as substrates. Lipids were extracted by using chloroform-methanol (2:1, vol/vol), dried, and separated by TLC with chloroform-methanol-water (65:25:4, vol/vol/vol) followed by exposure to a PhosphorImager screen. The radioactivity of the [14C]-labeled product was also measured by scraping the corresponding bands in the TLC plate into scintillation vials for quantitation by a Beckman LS 6500 scintillation system (Fullerton, CA). The specific GPAT or AGPAT activities were calculated by converting the radioactivity into the mass of formed radiolabeled product. Where indicated, cell lysates were incubated with or without 0.4 μM NEM for 15 min on ice before the initiation of reaction to differentiate the NEM-resistant and NEM-sensitive GPAT activity.
Total RNA was isolated from livers or adipose tissues using a commercially available kit (Qiagen RNeasy Kit; Qiagen, Valencia, CA). Total RNA (1 μg) was subjected to quantitative RT-PCR analysis with gene-specific primer pairs, which was carried out in 96-well plates using the ABI PRISM 7900HT System (Applied Biosystems, Foster City, CA). The Taqman gene expression assay ID for mGPAT3 used in this study was Mm00554802_m1, which results in a replicon of 57 bp covering boundary regions of exons 7 and 8. The sequences of additional Taqman primer-probe pairs can be found in Table 1. All other Taqman assays were purcahsed from Applied Biosystems.
Total lipids were extracted from the livers using the Folch method (10). Frozen liver aliquots (150–250 mg) were added to 4 ml of ice-cold methanol (MeOH) and homogenized for 20 s using a Polytron homogenizer (Kinematica, Switzerland). Cold chloroform (CHCl3, 4 ml) was added, and the samples were homogenized for a second time for 30 s. The samples were transferred to extraction tubes, a further 4 ml of CHCl3 was added, and the homogenate was mixed thoroughly. Samples were held at −20°C until analysis, at which point, 2.8 ml of 1 M aqueous potassium chloride was added (8:4:3 ratio CHCl3-MeOH-H2O). The samples were vortexed for 30 s and centrifuged at 4°C to separate the phases. The entire bottom phase was removed and dried down under nitrogen followed by resolubilization in CHCl3. All extraction solvents contained 50 μM butylated hydroxytoluene as antioxidant. Neutral lipids were isolated from the total lipid extracts using aminopropyl solid-phase extraction columns as previously described (14). The neutral lipid fraction was in turn dried down under nitrogen and resuspended in isooctane-isopropanol (98:2). Free and esterified cholesterol, TAG, and DAG were separated and detected using normal-phase cyanopropyl high-performance liquid chromatography (HPLC) and a charged aerosol detector. Analysis for lysophosphatidylcholine (LPC), phosphatidylcholine (PC), and sphingomyelin (SM) in total extracts was performed by standard mass spectrometic methods. Data are expressed as milligrams of lipid per gram wet weight of liver.
Unless otherwise indicated, all results are expressed as means ± SE. All assay results were evaluated using GraphPad Prism software, and statistical differences between different groups were performed by repeated-measures analyses of variance (RM-ANOVA) (for time-dependent measurements of mobility, O2 consumption, CO2 production, and respiratory exchange ratio) or t-test (for means of two independent groups).
Generation of GPAT3-deficient mice.
To generate Gpat3−/− mice, we designed a targeting strategy that replaces exons 4–7 of the Gpat3 gene with a loxP-neomycin cassette (Fig. 1A). This strategy deletes the majority of the acyltransferase domain of GPAT3 and introduces a frame-shift mutation after aa 69, resulting in a functional null allele. ES cells carrying the mutant allele were identified by Southern blot analysis of genomic DNA digested with Stul or BstXI (Fig. 1B). The mouse genotype was confirmed by PCR using genomic DNA from mouse tails (data not shown). The absence of intact GPAT3 mRNA transcripts was confirmed by Taqman analysis with total RNA samples isolated from both liver and WAT (Fig. 1, C and D). The Gpat3−/− mice were born at expected Mendelian frequencies and were viable and fertile with no obvious abnormalities. When maintained on standard laboratory chow, the levels of nonesterified fatty acids (NEFA), ketones [β-hydroxybutyrate (β-HBA)], triglycerides, glucose, and transaminases [alanine aminotransaminase and aspartate aminotransaminase (ALT and AST, respectively)] were similar in wild-type and Gpat3−/− mice. However, total plasma cholesterol levels were significantly decreased in Gpat3−/− mice (78 ± 1.2 mg/dl) compared with wild-type animals (87 ± 2.1 mg/dl). In overnight-fasted animals, there were no significant changes in any parameters examined (Table 2).
GPAT3-deficient mice had reduced NEM-sensitive GPAT activity in both WAT and liver.
There are at least four mammalian GPAT isoforms (11, 27). It has been well documented that mitochondrial GPAT1 contributes ∼10% of total GPAT activity in most tissues and 30–50% in liver. Analysis of Gpat4−/− mice suggests that GPAT4, one of the two microsomal GPAT enzymes identified so far, contributes ∼50% of total GPAT activity in the liver but does not contribute significantly to WAT GPAT activity (17). To determine the contribution of GPAT3 to GPAT activity in vivo, we measured total GPAT activity in WAT depots and liver from both wild-type and Gpat3−/− mice. As shown in Fig. 2A, left, total GPAT activity in Sub. WAT was reduced by ∼80% in Gpat3−/− mice, as determined by the conversion of [14C]G3P to [14C]phosphatidic acid. In Vis. WAT, Gpat3−/− mice had an ∼50% decrease in total GPAT activity compared with wild-type mice (Fig. 2A, middle). In contrast, total liver GPAT activity was not significantly changed in Gpat3−/− mice (Fig. 2A, right). The NEM-resistant GPAT activity in Gpat3−/− Sub. WAT remained unchanged (Fig. 2B), indicating that the decrease in total GPAT activity can be attributed entirely to decreased NEM-sensitive microsomal GPAT activity (Fig. 2C). In the liver, deficiency of GPAT3 caused an ∼30% increase in NEM-resistant liver GPAT activity (Fig. 2B), and as a consequence, liver NEM-sensitive microsomal GPAT activity, calculated by subtracting NEM-resistant GPAT activity from total GPAT activity, was decreased by ∼32% (Fig. 2C). We also measured AGPAT activity in Sub. WAT and liver in wild-type and Gpat3−/− wild-type animals by monitoring incorporation of [14C]-labeled oleoyl moiety from [14C]oleoyl-CoA into PA with LPA as an acceptor substrate. As shown in Fig. 2D, whereas liver AGPAT activity in Gpat3−/− mice remained unchanged, the absence of GPAT3 resulted in an ∼25% increase in AGPAT activity in subcutaneous fat depots. Collectively, these data show that GPAT3 is responsible for the majority of GPAT activity in Sub. WAT and a substantial amount of GPAT activity in Vis. WAT. These data also reveal that GPAT3 deficiency in liver caused a small decrease in NEM-sensitive GPAT activity and a compensatory increase in NEM-resistant GPAT activity, allowing total GPAT activity to remain unchanged.
GPAT3-deficient female, but not male, mice were protected from diet-induced obesity.
To evaluate the impact of GPAT3 deletion on energy and glucose homeostasis, we measured body weight, food intake (FI), body composition, fat pad mass, and plasma metabolites from Gpat3−/− mice and wild-type littermates. When fed a high-fat diet containing 60 kcal% from fat starting at 10 wk of age, the Gpat3−/− female mice began to gain less weight than wild-type littermates (Fig. 3A). The divergence in body weight between the GPAT3-null and wild-type littermates became greater as mice aged and reached statistical significance in the 18th wk of high-fat feeding (Fig. 3A). By 40 wk of age (30 wk on high-fat diet), the female GPAT3-deficient mice weighed 50.5 ± 1.8 g, which is 11% lower than the weight of wild-type littermates (56.9 ± 2.1 g, P < 0.01; Fig. 3, A and B). The difference in body weight between the two groups of mice was largely due to reduced fat mass in Gpat3−/− mice (Fig. 3, B and C). As measured by MRI at the 28th wk on high-fat diet, the whole body fat content in Gpat3−/− female mice was ∼28% lower than in wild-type littermates (25.6 ± 1.2 vs. 32.3 ± 1.3 g, P < 0.01), whereas the lean mass between the two groups was unchanged (Fig. 3C). Further analysis of the fat pad weight in different locations in the body at the conclusion of the study revealed that the reduction in fat mass was observed in both subcutaneous and visceral fat depots (Fig. 3D). The GPAT3 deletion on male mice had little effect on weight gain on high-fat diet, although a nonsignificant decrease in fat mass was noted in male Gpat3−/− mice after 30 wk on high-fat diet (Fig. 4).
GPAT3 deletion had no detectable effect on food intake but increased energy expenditure.
To understand the mechanisms accounting for the protection of Gpat3−/− mice from diet-induced obesity, food intake was monitored on a weekly basis in wild-type and knockout mice fed a high-fat diet. Gpat3−/− mice ate the same amount of food as age- and sex-matched wild-type mice (Fig. 5, A and B). We also measured locomotor activity, oxygen consumption, and carbon dioxide production (as an indirect measurement for energy expenditure) in wild-type and GPAT3-null mice in the 28th wk after initiation of high-fat diet by placing mice in a metabolic monitoring system for 18–24 h. Gpat3−/− mice appeared to be more active than wild-type littermates, with the difference being most pronounced in the initial course of dark phase and in female mice (Fig. 5C; P < 0.01 vs. wild-type, RM-ANOVA). Consistently, both male and female Gpat3−/− mice also showed increased oxygen consumption (Fig. 6A) and carbon dioxide production (Fig. 6B), which were also most pronounced during the early phase of the dark period (P < 0.01 vs. wild type, RM-ANOVA). When measurements were analyzed depending on the time of day and significance assessed by Student's t-test, significant increases in energy expenditure were apparent for both male and female Gpat3−/− mice, primarily during the active food-seeking period in the early dark phase (Fig. 6, insets). Respiratory exchange ratio was largely unchanged between wild-type and Gpat3−/− mice (Fig. 6C), suggesting that the altered energy expenditure in these mice was not accompanied by alterations in fatty acid vs. carbohydrate utilization. The increased energy expenditure observed especially during the first part of the dark phase in female mice appeared to be well correlated with the increased locomotion, suggesting a contribution of increased locomotion to the lean phenotype of Gpat3−/− mice under high-fat feeding conditions.
GPAT3 deletion improved glucose homeostasis.
Deficiency of enzymes in the triglyceride biosynthetic pathway can have either beneficial or detrimental effects on glucose homeostasis, despite lowered body weight, depending on the exact enzyme and, in some cases, the dietary intervention (4, 6, 9). To understand the role of GPAT3 in maintaining glucose homeostasis and insulin sensitivity, we first analyzed changes in fasted and fed glucose and insulin level in the high-fat-fed Gpat3−/− mice relative to the sex-matched wild-type siblings. In line with animals maintained on standard laboratory chow (Table 2), fasted glucose levels were not significantly different between Gpat3−/− and the wild-type mice when fed a high-fat diet (Fig. 7A). However, the fed blood glucose levels in the Gpat3−/− mice were significantly lower than in the wild-type mice (Fig. 7B); these changes were statistically significant in both male and female mice. Serum insulin levels under both conditions were similar between Gpat3−/− and wild-type controls (Fig. 7, C and D). In response to an oral glucose tolerance test (OGTT, performed after 26 wk on the high-fat diet), Gpat3−/− female mice showed a similar glucose excursion curve (Fig. 7, E and F) compared with wild-type controls. In a similar OGTT performed in male mice, GPAT3 deletion appeared to improve the glucose disposal at 30 min after glucose challenge (Fig. 7F). The mean value of area under the glucose excursion curve (AUC) from male Gpat3−/− mice was 15% less than from wild-type controls, although the change did not reach statistical difference (P = 0.066; Fig. 7G). The AUC values between female Gpat3−/− and wild-type mice were similar (Fig. 7G). During the OGTT, the insulin levels in Gpat3−/− and wild-type mice were also similar (data not shown).
GPAT3-deficient mice under high-fat diet have enlarged liver, decreased fat pad size, and altered serum lipid profile.
At the conclusion of the high-fat feeding study, a variety of metabolically active tissues were isolated from wild-type and Gpat3−/− mice, including liver, WAT (including gonadal and subcutaneous fat pads), BAT, heart, and kidney. Consistent with the in-life MRI data, both gonadal and subcutaneous fat pad weights were significantly decreased in GPAT3-deficient female mice compared with wild-type controls (Table 3). There was also a numerical decrease in WAT mass in male GPAT3-deleted mice (Table 3). Unexpectedly, we found that the liver weight was significantly greater in both female and male Gpat3−/− mice (Table 3). BAT mass was slightly decreased in male, but not in female, Gpat3−/− mice (Table 3). There was no difference in the weight of kidney and heart between the two groups (Table 3). We also determined various plasma metabolic parameters in both male and female mice (Table 4). Notably, total cholesterol levels were significantly increased in both female and male GPAT3-deficient mice, and the low-density lipoprotein cholesterol levels were numerically elevated in GPAT3-deficient mice. Triglyceride levels tended to be decreased in GPAT3-deleted mice. Free fatty acids, HDL-cholesterol, and total protein, did not differ between the two genotypes. As indicators for liver function, ALT and AST levels were significantly elevated in male GPAT3−/− mice, possibly indicating a hepatic pathogenesis in this group of mice.
GPAT3 deletion causes an alteration in liver cholesterol metabolism.
To investigate the effects of GPAT3 absence on lipid metabolism in vivo, we further analyzed the content of different lipids in the liver. To this end, we used both HPLC- and LC-MS-based lipid analysis to determine several common lipid species, including TAG, DAG, cholesterol and cholesteryl ester (CE), LPC, PC, and SM. Unexpectedly, we found that free cholesterol and CE were significantly increased in the liver of Gpat3−/− male mice (Fig. 8). The contents of TAG, DAG, LPC, PC, and SM in the livers of Gpat3−/− mice remained unchanged (Fig. 8). The increase in hepatic free cholesterol in the livers of male Gpat3−/− mice was accompanied by increased ALT/AST serum levels; interestingly, increases in hepatic cholesterol and CE, but not hepatic triglycerides, have been previously linked to increased ALT/AST levels (13). Using a mass spectrometry-based approach, we also examined the composition of specific lipid species in liver and epididymal and subcutaneous fat depots but were unable to discern any significant alterations in the composition of TAG, DAG, SM, and PC (data not shown).
Analysis of gene expression profile in Gpat3−/− mice.
To investigate the molecular changes associated with GPAT3-deficient mice, we performed quantitative real-time PCR analysis on a number of genes, including the GPAT3 homologs GPAT1 and GPAT4, several genes involved in adipogenesis and adipocyte function such as peroxisome proliferator-activated receptor-γ coactivator-1α and -1β (PGC-1α and PGC-1β), as well as the lipogenic genes stearoyl-CoA desaturase-1 (SCD1) and sterol regulatory element-binding protein-1 (SREBP-1c), using several metabolically active tissues, including WAT, BAT, and liver. We did not find any significant changes in the expression of any of these genes in liver (Table 5) or BAT (Fig. 9). Interestingly, in the epididymal WAT depot, PGC-1α was significantly upregulated, whereas SCD1 expression was numerically downregulated; a significant downregulation of SCD1 was also observed in subcutaneous adipose tissue (Fig. 9). This regulation would be consistent with the increased energy expenditure observed in Gpat3−/− mice, although additional studies will need to be conducted to further examine this relationship. Expression of the lipogenic transcription factor SREBP-1c was significantly reduced in both epididymal and subcutaneous adipose depots. GPAT4 expression was modestly but significantly increased in epididymal but not subcutaneous WAT (Fig. 9); however, given that the absolute mRNA levels of GPAT4 in adipose tissue are significantly lower compared with GPAT3, the functional significance of this regulation is unclear. In the liver, profiling of ∼40 genes involved in lipid metabolism revealed a modest but significant increase in SREBP-1c and its downstream target genes acetyl-CoA carboxylase-1 (ACACA) and GAPDH. The expression of ATP-citrate lysase (ACLY) also tended toward an increase. In addition, numerical increases in the ATP-binding cassette subfamily G members 5 and 8 (ABCG5 and ABCG8, respectively) were also apparent (Table 5).
In the present study, we report the first phenotypic characterization of mice deficient for GPAT3, one of two microsomal GPAT enzymes identified so far at the molecular level. Our data show that total and NEM-sensitive GPAT activity was significantly reduced in Gpat3−/− mice compared with wild-type controls in both subcutaneous and visceral fat depots, whereas NEM-resistant GPAT activity in adipose tissue remained unchanged. In contrast, in liver, although the total GPAT activities were comparable between wild-type and Gpat3−/− littermates, the NEM-sensitive (presumably microsomal) activity was decreased by 30% due to an unexpected increase in NEM-resistant GPAT activity in this tissue. In addition, our results show that AGPAT activity in GPAT3 knockout mice was either unchanged (liver) or upregulated (adipose tissue) when compared with wild-type controls. These results, together with our previous findings of high expression level of GPAT3 in adipose tissue and the impairment of adipogenesis by GPAT3-specific RNAi (22), provided strong evidence showing that GPAT3 encodes the major microsomal GPAT enzyme in adipose tissue. GPAT4 may account for the majority of microsomal GPAT activity in liver, as supported by a previous report (17).
Consistent with a role for GPAT3 as a lipid-metabolizing enzyme in adipose tissues, female Gpat3−/− mice fed a high-fat diet had decreased fat mass and gained less weight than wild-type controls. Further analysis of energy intake and expenditure showed no significant change in energy intake but revealed a role for GPAT3 in modulating energy expenditure in both male and female mice, which was particularly apparent during the early phase of the dark period, when the majority of feeding occurs. Given the complexity and challenges in accurate and standardized measurements of energy intake and expenditure in mice (25), it is still premature to conclusively identify increased energy expenditure as the primary cause for the decreased weight gain in female Gpat3−/− mice; however, our data to date are certainly consistent with this suggestion. Compared with other lipogenic enzymes involved in triglyceride biosynthesis (12, 18, 26) the effects of GPAT3 deletion on fat mass and body weight are relatively subtle, and it is interesting that the respiratory exchange ratio was similar in GPAT3−/− and wild-type mice, suggesting no large changes in fatty acid utilization. GPAT4, the closest homolog of GPAT3 within the family of glycerophospholipid acyltransferases, shows a much more pronounced decrease in adiposity, a selective subdermal lipodystrophy, and impaired lactation. In contrast, Gpat3−/− mice do not display lipodystrophy, and lactation remained intact (data not shown). These results suggest that the two microsomal GPAT enzymes play a complementary role in mediating triglyceride biosynthesis in a tissue-specific manner: GPAT3 may mediate triglyceride formation in WAT in a regulated manner, while GPAT4 functions as a constitutive enzyme responsible for the basal lipid storage in tissues such as liver and mammary glands. Overall, deletion of GPAT3 and GPAT4 reveals overlapping but unique phenotypes. Given the high homology and identical biochemical functions of the two proteins, it would be very interesting to generate double-knockout mice to fully assess the importance of these two microsomal GPATs in lipid metabolism and their potential as targets for disease intervention. One would expect to see a more profound impact on lipid metabolism and other metabolic outcomes with the double deletion of GPAT3 and GPAT4.
Mice of both sexes deficient for GPAT3 under diet-induced obese conditions had lower fed glucose levels compared with wild-type controls, indicating improved glucose disposal under conditions of high-fat feeding. Consistent with this, improved glucose excursions were observed in an OGTT in GPAT3-deficient male mice. However, it should be noted that fasting glucose levels, as well as insulin levels in both fed and fasting conditions were comparable between wild-type and Gpat3−/− cohorts, suggesting that GPAT3 deficiency has only limited effects on overall glucose homeostasis. It is possible that the relatively long fasting period employed during the OGTT in this study (overnight) may have obscured the effects of GPAT3 deletion, and employing shorter fasts (6 h) may allow a more sensitive detection (1). Given the high level of gastrointestinal expression of Gpat3 (3, 22), it will be important to determine whether GPAT3 deletion affects intestinal lipid or carbohydrate absorption or the meal-induced increases in gut peptides, such as Glp1 and GIP, which could possibly contribute to the altered postprandial glucose levels in GPAT3-deficient mice. Our results revealed an unexpected dysregulation of cholesterol metabolism in GPAT3-deficient mice fed a high-fat diet. In animals maintained on standard laboratory chow, circulating cholesterol levels were lower in Gpat3−/− mice than in wild-type animals. However, this observation was reversed when the mice were placed on a high-fat diet. In this setting, total circulating cholesterol levels were elevated. Moreover, male Gpat3−/− mice fed a high-fat diet had significantly increased levels of free cholesterol and CE in liver. The upregulation of hepatic cholesterol levels was accompanied by enlarged livers and impaired liver function, as inferred by elevated plasma AST and ALT levels. Additionally, several other lipid species, including triglycerides and DAG, also tended to be increased in the liver of Gpat3−/− mice. Our initial efforts in analyzing the mRNA expression profile of a selective subset of lipogenic genes provided no plausible explanation for such changes (Table 5). Interestingly, the expressions of SREBP-1c and some of its downstream target genes (ACACA and GAPDH) were significantly elevated. In parallel, expressions of the cholesterol efflux transporters ABCG5 and ABCG8 were also elevated. It is possible that the coordinate increase in SREBP-1c signaling and ABCG5/ABCG8 is mediated through liver X receptors LXRα and LXRβ as an adaptive response to increased hepatic sterol burden (19–21). We speculate that, under the chronic condition of excess nutrition (e.g., high-fat feeding) the absence of GPAT3 may lead to a deficiency in storing calories in adipose tissues and that an aberrant metabolism of fatty acids within nonadipose tissues such as liver may subsequently occur. The molecular basis for these observations is unclear but warrants further investigation. It will be interesting to determine the requirement for GPAT3 in neutral vs. phosphoglycerolipid synthesis and whether GPAT3 is associated with lipid droplets in a manner similar to what has recently been described for GPAT4 (28).
Our studies also revealed a sexual dimorphism in the phenotypes of the Gpat3−/− mice. Notably, the female but not the male Gpat3−/− mice were protected from diet-induced obesity. The changes in hepatic cholesterol metabolism appeared to be more prominent in male Gpat3−/− mice than in the females. Interestingly, the sexual dimorphism in phenotypes related to lipid, glucose, and energy homeostasis also have been observed in mice deficient for other lipid-metabolizing enzymes such as GPAT1, AGPAT2, and ALCAT1 (9, 12, 15). Although not extensively and mechanistically studied, the sexual dimorphism in human lipid metabolism has long been recognized (16). The sex difference in the phenotypes of Gpat3−/− mice may suggest the role of GPAT3 in sex hormones (e.g, testosterone and estrogen) regulated energy or lipid metabolism. Additional mechanistic and/or comparative studies are warranted in order to reveal a possible interaction between sex hormones and their receptors and GPAT3 in energy and lipid homeostasis.
In the past several years, cloning of novel lipid-metabolizing enzymes in triglyceride biosynthesis has enabled the generation and characterization of knockout mice for these enzymes, which have greatly advanced our understanding of the role of these enzymes in normal physiology and their potential as targets for human disease intervention. Our data presented in the current work underscore the importance of GPAT3 in the maintenance of glucose and lipid homeostasis and highlight the need for further studies to determine its possible contribution to human pathophysiology.
All authors are former or current employees of Pfizer Inc.
Author contributions: J.C., B.G., and R.E.G. conception and design of research; J.C., S.M.P., Q.L., H.P., A.Q., and R.W.C. performed experiments; J.C., S.M.P., B.G., Y.Z., M.P., and R.E.G. analyzed data; J.C., B.G., M.P., J.F.T., and R.E.G. interpreted results of experiments; J.C., S.M.P., B.G., Q.L., and Y.Z. prepared figures; J.C., S.M.P., B.G., and Q.L. drafted manuscript; J.C., B.G., J.F.T., and R.E.G. edited and revised manuscript; J.C., B.G., and R.E.G. approved final version of manuscript.
We thank Mary Bauchmann, Dave Koubasiak, Tiffany Gareski, and Xiangping Li for their excellent technical assistance and Dr. Simon Theodore for helpful discussion in generating GPAT3 knockout mice.
- Copyright © 2014 the American Physiological Society