|Home | About | Journals | Submit | Contact Us | Français|
Reducing triacylglycerol (TAG) in the liver continues to pose a challenge in states of nonalcoholic hepatic steatosis. Monoacylglycerol O-acyltransferase (MOGAT) enzymes convert monoacylglycerol (MAG) to diacylglycerol, a precursor for TAG synthesis, and are involved in a major pathway of TAG synthesis in selected tissues, such as small intestine. MOGAT1 possesses MGAT activity in in vitro assays, but its physiological function in TAG metabolism is unknown. Recent studies suggest a role for MOGAT1 in hepatic steatosis in lipodystrophic [1-acylglycerol-3-phosphate O-acyltransferase (Agpat)2−/−] and obese (ob/ob) mice. To test this, we deleted Mogat1 in the Agpat2−/− and ob/ob genetic background to generate Mogat1−/−;Agpat2−/− and Mogat1−/−;ob/ob double knockout (DKO) mice. Here we report that, despite the absence of Mogat1 in either DKO mouse model, we did not find any decrease in liver TAG by 16 weeks of age. Additionally, there were no measureable changes in plasma glucose (diabetes) and insulin resistance. Our data indicate a minimal role, if any, of MOGAT1 in liver TAG synthesis, and that TAG synthesis in steatosis associated with lipodystrophy and obesity is independent of MOGAT1. Our findings suggest that MOGAT1 likely has an alternative function in vivo.
Obesity (1) and lipodystrophy (2–4) are two extreme ends of the disorders of adipose tissue. Expansion of adipose tissue (in obesity) or loss of adipose tissue (in lipodystrophy) present with similar clinical abnormalities, such as insulin resistance (IR), diabetes, hepatic steatosis, and hyperlipidemia (5, 6). Although it is still debated whether IR is the cause for hepatic steatosis or excessive liver fat causes IR (7), it has been shown that reducing liver fat is beneficial overall, improving both diabetes and IR (8, 9).
Almost all cell types are able to synthesize triacylglycerol (TAG), although some cells synthesize more than others. Adipose tissue is normally the major site of TAG synthesis and storage, whereas tissues like the liver and skeletal muscle are not. Any increase in TAG levels in the liver or skeletal muscle will result in metabolic disruption in these tissues (10). TAG can be synthesized by several routes. The two most commonly known pathways are the glycerol-3-phosphate (G-3-P; also known as Kennedy) pathway, in which G-3-P is sequentially acylated and dephosphorylated to produce TAG, and the monoacylglycerol (MAG) pathway, in which MAG is acylated to diacylglycerol (DAG), which is then further acylated to TAG, as in the G-3-P pathway (11, 12). The MAG pathway is thought to be prominent in re-esterification of hydrolyzed TAG. A third pathway that has not yet been elucidated in mammalian tissue is the conversion of phospholipids to DAG, which is accomplished by an enzyme, phospholipid DAG acyltransferase (13). These TAG pathways are reviewed in (11, 12, 14).
There are currently three mammalian isoforms of monoacylglycerol O-acyltransferase (Mogat isoforms 1–3)6 cloned from rodents and humans encoded from different genes (15–17). While Mogat1 and Mogat2 are present in both humans and rodents, expression of Mogat3 has only been observed in rats and in humans, but not in mice, where it is shown to be a pseudogene (18). Based on conservation of Mogat1 and Mogat2 in rats and humans, it seems unlikely that mice lack an active Mogat3. Thus, it will require further in silico data mining to identify the Mogat3 gene in mice. The Mogat1 mRNA in normal adult mouse livers is either undetectable or expressed at a very low level (19). However, hepatic MGAT enzymatic activity in rodents has been measured in the suckling rat, which declined after postnatal day 8 (20). MGAT enzymatic activity was also measurable in the livers of streptozotocin-induced adult diabetic rats and was increased by ~2-fold (21). Although we can measure MGAT enzymatic activity in a cultured cell model system and in tissues like mouse and human intestine, a more robust MGAT enzyme assay needs to be developed for measuring its activity in other mammalian tissues, like mouse kidney and liver. Additionally, the measured MGAT enzymatic activity represents the total enzymatic activities of all the MOGAT isoforms known and/or yet to be identified. Almost all previous studies have determined the role of MOGATs in the context of enterocytes showing their action in fat absorption and chylomicron secretion (17, 22). An additional role of MOGATs has been implicated in the recycling of TAG and re-esterifying fatty acids to MAG (23, 24).
The present study stems from our previous observation made while analyzing the mouse model of congenital generalized lipodystrophy type 1 in which the 1-acylglycerol-3-phosphate O-acyltransferase 2 (Agpat2) gene was deleted. AGPAT2 converts lysophosphatidic acid to phosphatidic acid in the second step of the G-3-P pathway. We observed a robust increase in the expression of Mogat1 in the severely steatotic livers of Agpat2−/− mice, both at the mRNA (25- to 50-fold increase) and protein (6-fold) levels (19). We followed this observation and found that the expression of Mogat1 was also upregulated in ob/ob mice (unpublished observation). The ob/ob is a leptin-deficient mouse model and is a widely studied model of obesity, diabetes, hepatic steatosis, and IR (25). These mouse models closely recapitulate many features of common forms of human nonalcoholic hepatic steatosis (26). It is interesting to note that both murine models lack leptin. In Agpat2−/− mice, lack of adipose tissue results in leptin deficiency, and in ob/ob mice, a mutation in the leptin (Lep) gene results in nonfunctional leptin.
The findings of increased Mogat1 expression occurring in the severely steatotic livers of Agpat2−/− and ob/ob mice suggest that Mogat1 activity contributes to hepatic steatosis. The goal of this study was to test this hypothesis by determining whether deletion of functional Mogat1 gene expression in Agpat2−/− mice or in ob/ob mice by genetic ablation would reduce liver TAG, improve insulin sensitivity, or reduce hyperinsulinemia and hyperglycemia. Surprisingly, our results indicate that Mogat1 does not play a major role in hepatic steatosis in these models and suggest that the enzyme has an alternative activity in vivo.
All animal studies were approved by the University of Texas Southwestern Medical Center, Dallas or the Gladstone Institute for Cardiovascular Disease, San Francisco, by their respective Institutional Animal Care and Use Committee.
A targeting vector was engineered to replace exon 2 of murine Mogat1 with a 1.9 kb neo cassette, allowing for positive selection with neomycin-resistance and negative selection for thymidine kinase. The vector was constructed by amplifying129/SvJae mouse genomic DNA and cloning this into pNTKloxP (a gift from Joachim Herz, University of Texas Southwestern, Dallas, TX). The PCR primers used to amplify the 1.2 kb fragment upstream of exon 2 and introduce XhoI sites were: forward, 5′-AAA CTC GAG CAA GTC ACC ACA CTC CAC TCT CTG-3′ reverse, 5′-ACA CTC GAG CTG TAA GAA AGG GAA AAG CCA GGT AG-3′). Primers to amplify the 6.6 kb fragment downstream of exon 2 and introduce NotI sites were: forward, 5′-CTC GCG GCC GCC TCC TTT CCA GAC AGA ACG TGC CTT AGG TC-3′ reverse, 5′-ATC GCG GCC GCG ACT GAA TGA CCC TGT CAC AGG-3′. This targeting vector was used to generate Mogat1-targeted 129/SvJae RF8 embryonic stem cells and, subsequently, mice carrying this deletion. Disruption of Mogat1 was screened by PCR and confirmed by Southern blotting of genomic DNA digested with HindIII using a 32P-labeled oligonucleotide probe (5′-CAG GGT TCC TAT TGT TGT GAA C-3′) upstream of the sequence included in the targeting vector and by PCR using three primers (which are located upstream of the targeting vector, in the targeted region of the WT allele, and in the targeting vector, respectively, as shown in Fig. 1; P1, 5′-CTG GAG CAA GCA GGG CCA GAA TGA G-3′ P2, 5′-GGA CCT AAG GCA CGT TCT GTC TG-3′ P3: 5′-CGT TGA CTC TAG AGG ATC CGA C-3′). For subsequent genotyping of Mogat1−/− mice, we designed additional primers to amplify the WT allele individually: Mogat1_WT_F, 5′-CCA GTG GCT GTG AAG TAC AG; Mogat1_WT_R, 5′-CAT CTT CTG CCT CCT TGC TC. The product size for the WT PCR amplicon is 191 bp and for the targeted amplicon is 322 bp. Both amplification products were confirmed by DNA sequencing. Exon 2-3 deletion was confirmed by amplifying Mogat1 from liver cDNA from Agpat2−/− and Mogat1−/−;Agpat2−/− mice using primers in exon 1 (forward: 5′-GCA GTG GGT CCT GTC CTT C-3′) and exon 4 (reverse: 5′-ACA TTG CCA CCT GGA TCC T-3′). This Mogat1−/− mouse model was developed in the laboratory of Dr. Robert V. Farese, Jr. (Gladstone Laboratory, San Francisco, CA)
Agpat2−/− mice were generated as described previously (19). Mice were genotyped using the following allele discriminating primer sets: A15, CGG CTA GGT AAG CAG TTT GA; A8, AAA GCT GTG CCA GGG TGG GT; and S175, GAT TGG GAA GAC AAT AGC AGG CAT GC. Genomic DNA amplified with A15+A8 will produce the WT allele of 733 bp and A8+S175 will produce the knockout allele of 614 bp.
Two strategies were used to obtain the Mogat1−/−;Agpat2−/− mice. In the first strategy, Agpat2+/− mice were mated with Mogat1−/− mice to produce 50% Mogat1+/−;Agpat2+/+ and 50% Mogat1+/−;Agpat2+/− (double heterozygous for Agpat2 and Mogat1 alleles) mice. In the second cross of this strategy, the above double heterozygous mice (Mogat1+/−;Agpat2+/−) were crossed to generate all the genotypes required for this study: Mogat1+/+;Agpat2+/+ (wt/wt), Mogat1−/−;Agpat2+/+ (Mogat1−/−), Mogat1+/+;Agpat2−/− (Agpat2−/−), and Mogat1−/−;Agpat2−/− [double knockout (DKO)]. In another strategy, the Mogat1−/−;Agpat2+/− were crossed with Mogat1−/−;Agpat2+/− to generate Mogat1−/−;Agpat2+/+ (Mogat1−/−), Mogat1−/−;Agpat2+/−, and Mogat1−/−;Agpat2−/− (DKO). The DKOs obtained from this strategy were also used in the experiments.
Here again, two strategies were used to obtain the Mogat1−/−;ob/ob mice. To distinguish the various ob/ob alleles (WT, heterozygous, and null genotype), we have labeled them as wt/wt (WT), wt/ob (heterozygous), and null ob/ob. In the first strategy, wt/ob mice were mated with Mogat1−/− mice to produce 50% Mogat1+/−;wt/wt and 50% Mogat1+/−;wt/ob (double heterozygous for ob and Mogat1 genotypes) mice. In the second cross, the above double heterozygous mice (Mogat1+/−;wt/ob) were crossed to generate all the genotypes required for this study: Mogat1+/+;wt/wt (WT), Mogat1−/−;wt/wt (Mogat1−/−), Mogat1+/+;ob/ob (ob/ob), and Mogat1−/−;ob/ob (DKO). In another strategy, the Mogat1−/−;wt/ob were crossed with Mogat1−/−;wt/ob to generate Mogat1−/−;wt/wt (Mogat1−/−), Mogat1−/−;wt/ob, and Mogat1−/−;ob/ob (DKO). The DKOs obtained from this strategy were also used in the experiments.
Livers from 16-week-old mice were fixed in 4% paraformaldehyde overnight, embedded in paraffin, sectioned, and stained with hematoxylin and eosin (H and E). All images were acquired with a Leica DM 2000 compound microscope with an Optronics MicroFire camera. The acquisition software was Picture Frame 2.0. All tissues were processed at the pathology core laboratory at the University of Texas Southwestern Medical Center in Dallas, Texas.
Total RNA was prepared from mouse livers (100–200 mg) using RNA STAT-60 (Tel-Test Inc., Friendswood, TX). RNA (20 μg) was DNaseI treated using the DNase-free kit from Ambion (Grand Island, NY). RNA was subjected to RT-PCR using the reverse transcription kit from ABI (Carlsbad, CA). All quantitative real-time PCR reactions were carried out in 20 μl volume in 96-well plates using the ABI PRISM 7700 sequence detection system (Applied Biosystems), as reported previously (10). Quantitative real-time PCR was performed in duplicate and the transcript levels were normalized to cyclophilin. Primers are listed in Table 1.
For indirect calorimetry, individual mice were placed in metabolic cages and acclimated for 72 h before the start of the experiments. A comprehensive animal metabolic monitoring system (Columbus Instruments) was used to collect data for Mogat1−/−;Agpat2−/− mice, or TSE LabMaster (TSE Systems GmbH, Germany) was used to collect data for Mogat1−/−;ob/ob mice. Oxygen consumption (VO2), carbon dioxide production (VCO2), and total fluid and food intake were collected continuously over a 72 h period. All values are compared with the lean body mass.
An oral glucose tolerance test (OGTT) was performed in mice fasted for 6 or 16 h and then orally gavaged with D-glucose (2 g/kg body weight). Blood glucose was determined using a ReliOn blood glucometer (Walmart) immediately before the glucose gavage and at various time points after the administration of the glucose. Approximately 40 μl blood was collected from the tail vein in Microvette tubes (Sarstedt, Nümbrecht, Germany) containing clot activator. Tubes were spun at 10,000 g for 5 min and serum was collected for insulin measurement. Serum insulin was measured by ELISA method (Crystal Chem, Downers Grove, IL)
A meal test was performed in mice fasted for 6 or 16 h and then orally gavaged with 0.8 ml of Ensure (Abbott Nutrition, Abbott Laboratories, Columbus, OH). Tail vein blood was obtained at the start of the oral gavage and then at 60 min. Approximately 40 μl blood was collected from the tail vein, as described above for OGTT. Blood glucose was measured by ReliOn blood glucometer and insulin levels were measured as described above for OGTT.
Total RNA was prepared from mouse kidneys of WT, Mogat1+/−, and Mogat1−/− mice by standard techniques (15). Blots were probed with32P-CTP labeled full length Mogat1 cDNA (by random priming; Amersham Pharmacia), and were exposed to X-ray film.
Total MGAT activity in the liver of 16-week-old male WT, Mogat1−/−, Agpat2−/−, and Mogat1−/−;Agpat2−/− mice was measured using tissue homogenates, as previously described (16, 27). Reactions were started by adding homogenates to the assay mixture and were stopped after 5 min by adding chloroform:methanol (2:1, v:v). The lipids were extracted, dried, and separated by TLC on silica gel G-60 plates with the solvent system hexane:diethyl ether:acetic acid (80:20:1, v:v:v). Lipid bands were visualized with iodine vapor, and products were identified by comparison with the migration of lipid standards. The incorporation of radioactive substrates into lipid products was visualized by an imaging scanner (Typhoon FLA 7000; GE Healthcare Life Sciences, Piscataway, NJ) followed by scraping the region of interest and counting in a scintillation counter (Packard Tri-Carb 2200 CA liquid scintillation counter analyzer).
Assay conditions for Mogat1 enzymatic activity expressed in Spodoptera frugiperda (Sf9) insect cells and the construction of recombinant baculovirus for WT Mogat1 have been described before (15). The Mogat1 mutant cDNA was amplified from the kidneys of Mogat1−/− mice.
We have described methods for biochemical measurements in detail elsewhere (28). Briefly, to determine TAG concentration, approximately 100–150 mg of frozen liver was homogenized in Folch solution (chloroform:methanol 2:1, v/v). The organic phase was separated and collected by adding buffered saline and dried. The extract was reconstituted in the same extraction buffer and TAG and cholesterol levels were measured using reagents from Infinity liquid stable reagents (Fisher Diagnostics, Middletown, VA). Plasma cholesterol, TAG, and glucose were measured using Dry-slide technology (Vitros 250 analyzer from Ortho Clinical Diagnostic). All slides were purchased from Cardinal Health, Inc., and their catalog numbers are: SP1707801 for glucose, SP1669829 for cholesterol, and SP1336544 for TAG. All measurements were carried out at the Mouse Metabolic Phenotyping Core at University of Texas Southwestern Medical Center. Because plasma insulin concentrations were extremely high, they were measured with the kit obtained from Crystal Chem (Downers Grove, IL) in its high-range mode, which allows measurements up to 64 ng/ml.
Statistical significance was calculated either by Dunnett’s test in Excel (for two groups) or by one-way ANOVA in Graphpad Prism (for multiple groups). To compare the four genotype groups in the meal tolerance test (MTT), mixed effects model repeated measures analysis was employed to test the insulin and glucose responses to the meal test at 0 and 60 min. Separate analyses were performed for males and females. Statistical analysis was performed with SAS 9.4 (SAS Institute, Cary, NC). P < 0.05 was considered significant.
The murine Mogat1 exon 2 deletion strategy is shown in Fig. 1A. Deletion of exon 2 was chosen because it is an early exon that contains highly conserved sequences and deletion would result in a frame shift and an aberrant protein (p.Val32Alafs*66) (WT mouse Mogat1 GenBank accession AAI06136.1). Southern blots revealed the expected genomic deletion (Fig. 1B). Northern blots revealed a smaller mRNA transcript in the kidney of the knockout mice (Fig. 1C). Upon sequencing of the transcript cloned from the kidney of Mogat1−/− mice, we found that the transcripts had an exon skipping event such that exon 3 was also deleted, with exon 1 spliced to exon 4. The deletion of exons 2-3 yields an mRNA with in-frame splicing such that any expressed truncated protein would be ~32% shorter (128 of 335 amino acids) and would delete a presumed catalytic site: a highly conserved “HPHG” sequence. Thus, any protein produced from this locus would be predicted to be enzymatically inactive. Indeed, expression of the truncated (exon 2-3 deleted) Mogat1 protein in Sf9 insect cells showed no significant MGAT activity above background (Fig. 1D). We further confirmed the exon 2-3 deletion in the livers of Mogat1−/−;Agpat2−/− mice (Fig. 1E). Of note, we also could not detect MGAT enzymatic activity in tissues such as kidney (data not shown), where the expression of Mogat1 is high, nor in the fatty livers of Agpat2−/− lipodystrophic mice (data not shown), which also have markedly increased Mogat1 expression levels. Taken together, our results indicate that the Mogat1 gene is disrupted in this knockout model, but call into question whether Mogat1 has appreciable MGAT activity in murine tissues.
The growth curves for Agpat2−/− and Mogat1−/−;Agpat2−/− mice were followed at 8 through 15 weeks. The body weight of Mogat1−/−;Agpat2−/− mice of both sexes remained unchanged, as compared with Agpat2−/− mice (Fig. 2A, B). Likewise, when we analyzed the 16-week-old Mogat1−/−;Agpat2−/− mice of both sexes, their body and liver weights were unchanged compared with Agpat2−/− mice (Fig. 2C, D). This continuity is reflected in the liver TAG level as well. There was no difference in the liver TAG levels (6.72% decrease, P = 0.961) between Agpat2−/− and Mogat1−/−;Agpat2−/− in male mice (Fig. 2E) or in female mice (~30% increase, P = 0.503) (Fig. 2F). This slight increase in liver TAG is reflected in liver histology (Fig. 2G). Upon estimating the liver lipid droplet sizes and number in the liver images, we observed no significant differences in the lipid droplet sizes among Agpat2−/− and Mogat1−/−;Agpat2−/− mice (data not shown).
There was also no decrease in the plasma TAG level between Mogat1−/−;Agpat2−/− and the other three genotypes (WT, Mogat1−/−, and Agpat2−/−) (Fig. 3A) in either male or female mice. However, we did note a statistically significant decrease in plasma cholesterol levels in Mogat1−/−;Agpat2−/− compared with Agpat2−/− mice, but only in female mice (Fig. 3B). As we noticed a small increase in the liver TAG in the female mice only, we also saw an increase in the plasma glucose level in female Mogat1−/−;Agpat2−/− mice compared with WT and Mogat1−/− (Fig. 3C). At this age (16 weeks), mice of Agpat2−/− and Mogat1−/−;Agpat2−/− show normal skeletal muscle and kidney function, as determined by the creatine kinase and creatinine levels (data not shown).
In indirect calorimetry experiments, no statistically significant change in VO2 consumption (Fig. 3D) and VCO2 output (Fig. 3E) between Agpat2−/− and Mogat1−/−;Agpat2−/− mice was seen. There was no statistically significant change in water or food consumption when Mogat1 was deleted in the Agpat2−/− background (Mogat1−/−;Agpat2−/−) (Fig. 3F, G). The respiratory exchange ratio (RER), a measure of the consumption of fuel source, also remained unchanged between Agpat2−/− and Mogat1−/−;Agpat2−/− genotypes (Fig. 3H, I). The physical activity measured in two dimensions (x- and z-axis) was also unremarkable between the Agpat2−/− and the Mogat1−/−;Agpat2−/− mice (data not shown).
Expression of mRNA for various metabolic pathwaysremained largely unchanged in the livers of Mogat1−/−;Agpat2−/− compared with those of Agpat2−/− mice (Table 2). Because we observed a modest increase in liver TAG in the livers of Mogat1−/−;Agpat2−/− female mice, which might be due to an increase in de novo lipogenesis in the liver, we measured the expression of key enzymes involved in fatty acid synthesis (Acc1, Fas, and Scd1), TAG synthesis (Dgat1 and Dgat2), glucose metabolism (G6P and Pepck), and fatty acid oxidation. The mRNA expression of all enzymes examined remained unchanged between Agpat2−/− and Mogat1−/−;Agpat2−/− genotypes (Table 2). Because we did not observe a marked decrease in liver TAG content as expected (more than 50% reduction in hepatic TAG levels in Mogat1−/−;Agpat2−/− mice), we then measured the expression level of hepatic lysophosphatidylglycerol acyltransferase 1 (Lpgat1), shown to possess MGAT activity (29). Again, we did not see any difference between the Agpat2−/− and Mogat1−/−;Agpat2−/− mice (Table 2).
We next examined whether the lack of Mogat1 in Agpat2−/− mice results in changes in insulin sensitivity. The OGTT was carried out in 6 h-fasted mice of all genotypes. It was intriguing to note that, upon 6 h of fasting, the plasma glucose of Agpat2−/− and Mogat1−/−;Agpat2−/− mice was similar and was no different than WT and Mogat1−/− mice [Fig. 4A (male); Fig. 4D (female)]. However, the plasma insulin levels in the Agpat2−/− and Mogat1−/−;Agpat2−/− mice remained elevated [Fig. 4B (male); Fig. 4E (female)]. In an OGTT test, the response to the single bolus of glucose had similar glucose disposal among all the genotypes. In fact, the male Agpat2−/− and Mogat1−/−;Agpat2−/− mice were more sensitive than WT and Mogat1−/− mice [Fig. 4C (male); Fig. 4F (female)]. Similar observations were made when an OGTT was performed in 16 h-fasted mice (data not shown).
While the OGTT is conducted to determine the glucose tolerance by providing a bolus of glucose, most meals are a mix of carbohydrate, fat, and protein. To determine whether Mogat1 had any specific role in modulating glucose and insulin in a mixed meal assay, we fasted groups of mice for 6 h and provided a known quantity of liquid meal by oral gavage and measured their glucose and insulin before (0 min) and after 1 h (60 min). The plasma levels of glucose and insulin are displayed in Fig. 5A–D. In this assay, we noted a small, but statistically significant, increase in plasma glucose levels upon feeding, but only in WT and Agpat2−/− male mice and in Mogat1−/−;Agpat2−/− mice of both sexes (Fig. 5A, B). The increase in glucose levels was consistent with a similar increase in plasma insulin levels (Fig. 5C, D). However, there was no difference between glucose or insulin levels between Agpat2−/− and Mogat1−/−;Agpat2−/− mice at either 0 min or 60 min. This observation in Mogat1−/−;Agpat2−/− mice is different than that of Mogat1−/−;ob/ob mice (described below), who are more sensitive in a mixed meal test. It is unclear how deleting Mogat1 in ob/ob and Agpat2−/− mice shows different responses to a mixed meal test. It could be due to a total loss of adipose tissue in Agpat2−/− mice compared with ob/ob mice, which have excess adipose tissue. Further studies are required to examine the role of Mogat1 in modulating the response to a mixed meal.
We examined the liver TAG levels in mice where the expression of Mogat1 was ablated by homologous gene deletion in an ob/ob background. We initially followed the body weight of all four genotypes: WT, Mogat1−/−, ob/ob, and Mogat1−/−;ob/ob. As expected, a significant increase in the ob/ob body weight compared with WT and Mogat1−/− was observed (Fig. 6A, B). However, removing the functional Mogat1 expression in the ob/ob genetic background did not affect the body weight of Mogat1−/−;ob/ob mice for either sex as compared with ob/ob (Fig. 6A, B), nor were there any statistically significant differences in their body or liver weights at 16 weeks (Fig. 6C, D). Interestingly, we observed a modest decrease (~20%, P = 0.049) in the liver TAG in female Mogat1−/−;ob/ob mice (Fig. 6F), but no change in the male Mogat1−/−;ob/ob mice (~10% decrease, P = 0.161) (Fig. 6E). This modest change in the liver TAG could also be noted in the livers of 16-week-old female Mogat1−/−;ob/ob mice stained with H and E (Fig. 6G). We also measured the lipid droplet size and noted no significant changes between ob/ob mice and Mogat1−/−;ob/ob mice (data not shown)
We then measured the plasma TAG levels to determine whether a modest decrease in liver TAG levels in Mogat1−/−;ob/ob mice leads to decreased plasma levels. However, we observed no difference in the TAG levels between the ob/ob and Mogat1−/−;ob/ob mice (Fig. 7A), nor was there any difference in their plasma cholesterol (Fig. 7B) or glucose levels of either sex (Fig. 7C).
To determine whether the modest decrease seen in the livers of Mogat1−/−;ob/ob mice was due to decreased de novo lipogenesis in the liver, we measured the mRNA expression of key enzymes involved in fatty acid synthesis (Acc1, Fas, and Scd1), TAG synthesis (Dgat1 and Dgat2), glucose metabolism (G6P and Pepck), and fatty acid oxidation (PPARα). The mRNA expression of all the enzymes examined remained unchanged (Table 3) because we did not observe a marked decrease in liver TAG content (more than 50% reduction in hepatic TAG levels in Mogat1−/−;ob/ob mice), we measured the expression levels of hepatic Lpgat1, shown to possess MGAT activity (29). Although we did notice an increase in its expression by about 1.5-fold in the livers of ob/ob and Mogat1−/−;ob/ob mice compared with WT mice, we did not see any difference between the ob/ob and Mogat1−/−;ob/ob mice.
Indirect calorimetric studies did not reveal any changes in the O2 consumption, CO2 exhaled, or water or food consumption, expressed as cumulative values normalized to the lean body mass (Fig. 7D–G). There was also no change in the RER, which is indicative of any switching of fuel type, for the duration of the study, which lasted 72 h (Fig. 7H, I). The physical activity measured in all three dimensions (x-, y-, and z-axis) were also unremarkable between the ob/ob and the Mogat1−/−;ob/ob mice (data not shown)
Next we determined whether the Mogat1−/−;ob/ob mice become more insulin sensitive upon removing the expression of Mogat1. Groups of mice were fasted for 6 h and a single bolus of glucose was given by oral gavage and their blood glucose level was followed for 180 min. Again, no difference in plasma glucose disposal was observed between the ob/ob and Mogat1−/−;ob/ob genotypes (Fig. 8A, B).
We followed the OGTT in these mice with a MTT and measured plasma glucose and insulin levels, as displayed in Fig. 8C–F. We noticed a decrease in blood glucose levels at 60 min compared with those at 0 min for Mogat1−/−;ob/ob mice of both sexes and ob/ob females, but not for ob/ob males or WT and Mogat1−/− mice of either sex (Fig. 8C, D). This decrease in plasma glucose levels at 60 min upon mixed meal ingestion is consistent with the increase in plasma insulin levels at the same time points in ob/ob and Mogat1−/−;ob/ob mice. However, there was an increase in the plasma insulin levels in WT and Mogat1−/− mice despite no significant decrease in the plasma glucose level in these mice (Fig. 8E, F). This MTT indicated that the Mogat1−/−;ob/ob mice are more sensitive to a mixed meal than to a single bolus of glucose (OGTT) and points to a possible role of stomach Mogat1 in glucose homeostasis. Mogat1 is highly expressed in the stomach in addition to kidney and adipose tissue.
Understanding the molecular basis for the accumulation of liver TAG is a key to our efforts in developing strategies to ameliorate hepatic steatosis. Both Agpat2−/− and ob/ob mice have markedly increased expression of Mogat1 in their livers. Additionally, the deletion of Mogat1 alone did not result in any detectable metabolic abnormalities (this study), which made it a good candidate as a therapeutic target to inhibit and decrease hepatic fat. Therefore, in the current study, we tested the hypothesis of whether reducing the expression of Mogat1 will result in amelioration of hepatic steatosis and improved insulin sensitivity. We crossed Mogat1−/− mice to Agpat2−/− and ob/ob mice to generate DKOs (Mogat1−/−;Agpat2−/− and Mogat1−/−;ob/ob mice) and expected to reduce the liver TAG. In contrast to our expectations, neither DKO mouse model had reduced hepatic TAG levels. This was also reflected in the lack of decrease in their plasma glucose or insulin levels. Accordingly, both of the DKOs remained insulin resistant and diabetic.
As discussed above, there are several pathways to synthesize TAG. In Mogat1−/−;Agpat2−/− mice, two pathways are compromised (G-3-P pathway and MAG pathway), while in Mogat1−/−;ob/ob mice, only the MAG pathway is compromised. Thus, the question is this: with two key TAG synthesis enzymes missing in the livers of Mogat1−/−;Agpat2−/− mice, how do these mice assemble liver TAG? We previously noted an increase of several of the other Agpat isoforms in the livers of Agpat2−/− mice (19). However, we could only detect ~10% of the total AGPAT enzymatic activity in the liver homogenates of Agpat2−/− mice (19). In light of the substantial reduction in AGPAT enzymatic activity, the increase in alternate isoform expression seems insufficient to account for all the increased liver TAG synthesis. Mogat2 is undetectable in the liver of any genotype, so it cannot play a redundant role in the place of Mogat1 in synthesizing liver TAG. In addition, while we could measure MGAT enzymatic activity when overexpressed in insect cells, our repeated efforts to measure MGAT enzymatic activity in mouse livers and kidneys remained unsuccessful. This observation suggests that either the MOGAT1 protein is bound to a factor(s), either another protein or a small molecule, which inhibits its activity or MOGAT1 essentially could not participate in TAG synthesis. The latter argument is consistent with the idea that Mogats have no role in TAG synthesis in the livers of these mice. Although the phospholipid DAG acyltransferase pathway (presented in the introduction) is not fully established in mammalian livers, it is possible that such a route to synthesize TAG is present in the livers of Mogat1−/−;Agpat2−/− mice. Lastly, the liver can directly accumulate TAG from the uptake of circulating chylomicron remnants. In support of the last possibility, our previous experiments revealed that feeding the Agpat2−/− mice a fat-free diet resulted in marked improvement in hepatic steatosis (19). The results presented would now suggest that the livers of Mogat1−/−;Agpat2−/− mice and Mogat1−/−;ob/ob mice are highly flexible tissues and adapt when it comes to TAG synthesis and, thus, will require further investigation seeking additional pathway(s) for hepatic TAG synthesis.
Since our demonstration of the upregulation of Mogat1 mRNA expression in Agpat2−/− steatotic liver (10), a few studies using siRNA (30) or antisense oligonucleotide in an acute setting showed suppression of the expression of Mogat1 in mouse livers (31, 32). This resulted in a modest decrease in the liver TAG and an increase in hepatic insulin signaling and whole body glucose homeostasis. Using our genetic approach, we are unable to show a robust decrease in liver TAG or improved glucose tolerance in the Mogat1−/−;Agpat2−/− mice. In fact, in our Mogat1−/−;Agpat2−/− mice, there was either no change (male mice) or only a slight increase (female mice) in liver TAG levels. It remains unclear how siRNA and antisense oligonucleotide approaches in an acute experimental setting bring about reduced liver TAG and improved IR and type 2 diabetes mellitus. A recent study, although not related to the metabolic studies discussed here, reported that the use of antisense technology and genetic deletion resulted in widely different phenotypes (33–35). We interpret our study with an additional caution. Both the mouse models used in this study are on mixed genetic backgrounds. However, because neither mouse model resulted in a decrease in hepatic TAG, this would also indicate that the role of Mogat1 remains minimal. Our mouse models in a mixed genetic background would be much closer to the outbred human population, where most of the animal studies are translated.
In summary, although Mogat1 expression is highly upregulated in fatty livers of lipodystrophic (Agpat2−/−) and ob/ob mice, Mogat1 does not contribute significantly to hepatic TAG synthesis in these mice. In agreement with this, tissues where the Mogat1 mRNA is expressed at relatively high levels (stomach, kidney, and fatty livers) do not have prominent MGAT activity. When overexpressed, Mogat1 can catalyze the MGAT reaction (15), but the current data suggest that this enzyme likely has an alternative function in vivo and plays a different functional role in fatty liver disease.
The authors thank Paul Swinton of the transgenic mouse core at the Gladstone Institutes for injection of the Mogat1-targeted embryonic stem cells to mice and Ting-Ni Huang of University of Wisconsin-Madison for performing the MGAT assay. The authors also thank members of the mouse metabolic core at University of Texas Southwestern Medical Center for indirect calorimetric assays and John Shelton, along with the members of the histopathology core, for the staining of liver samples, and Beverley Adams-Huet for assistance with statistical analyses.
This work was supported by National Institutes of Health Grants DK54387 and DK077233, the Southwestern Medical Foundation, and the J. David Gladstone Institutes. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. All authors have no potential conflicts of interest relevant to this study to report.