|Home | About | Journals | Submit | Contact Us | Français|
The serine hydrolase α/β hydrolase domain 6 (ABHD6) has recently been implicated as a key lipase for the endocannabinoid 2-arachidonylglycerol (2-AG) in the brain. However, the biochemical and physiological function for ABHD6 outside of the central nervous system has not been established. To address this we utilized targeted antisense oligonucleotides (ASOs) to selectively knock down ABHD6 in peripheral tissues to identify in vivo substrates and to understand ABHD6's role in energy metabolism. Here we show that selective knockdown of ABHD6 in metabolic tissues protects mice from high fat diet-induced obesity, hepatic steatosis, and systemic insulin resistance. Using combined in vivo lipidomic identification and in vitro enzymology approaches we show that ABHD6 can hydrolyze several lipid substrates, positioning ABHD6 at the interface of glycerophospholipid metabolism and lipid signal transduction. Collectively, these data suggest that ABHD6 inhibitors may serve as novel therapeutics for obesity, nonalcoholic fatty liver disease, and type II diabetes.
A major challenge for drug discovery in the post genomic era is the functional characterization of unannotated genes identified by sequencing efforts. Although many unannotated gene products belong to structurally related gene or protein families, which may provide important functional clues, membership to such families does not always accurately predict the true biochemical and physiological role of proteins. Genes encoding the α/β hydrolase fold domain (ABHD) protein family are present in all reported genomes (Nardini and Dijkstra, 1999; Hotelier et al., 2004), and conserved structural motifs shared by these proteins predict common roles in lipid metabolism and signal transduction (Lefevre et al., 2001; Fiskerstrand et al., 2010; Long et al., 2011; Simon and Cravatt, 2006; Montero-Moran et al., 2009; Blankman et al., 2007; Lord et al., 2011; Brown et al., 2010). Furthermore, mutations in several members of the ABHD protein family have been implicated in inherited inborn errors of lipid metabolism (Lefevre et al., 2001; Fiskerstrand et al., 2010). Most recently, studies in cell and animal models have revealed important roles for ABHD proteins in glycerophospholipid metabolism, lipid signal transduction, and metabolic disease (Long et al., 2011; Simon and Cravatt 2006; Montero-Moran et al., 2009; Blankman et al., 2007; Lord et al., 2011; Brown et al., 2010). However, the physiological substrates and products for these lipid metabolizing enzymes and their broader role in metabolic pathways remain largely uncharacterized. Given this, functional annotation of ABHD enzymes holds clear promise for drug discovery targeting diseases of altered lipid metabolism and lipid signaling.
ABHD5, also known as CGI-58, has been studied quite extensively due to its key role in triacylglycerol (TAG) metabolism, lipid signaling, and genetic association with the human disease Chanarin-Dorfman Syndrome (CDS) (Lefevre et al., 2001; Montero-Moran et al., 2009; Lord et al., 2011; Brown et al., 2010; Lass et al., 2006; Schweiger et al., 2009). Given ABHD5's clear role in nutrient metabolism and lipid signal transduction, we aimed to test whether the closely related enzyme ABHD6 might play a similar role in lipid signaling and metabolic disease. ABHD6 has recently been described as an enzymatic regulator of endocannabinoid (ECB) signaling in the brain (Blankman et al., 2007; Marrs et al., 2010; Marrs et al., 2011). However, ABHD6 is ubiquitously expressed, and the biochemical and physiological functions of ABHD6 outside of the central nervous system have not been studied. Furthermore, unbiased identification of ABHD6 substrates in vivo has not been reported. To address this we selectively knocked down ABHD6 in peripheral tissues, allowing us to identify novel substrates in vivo and to uncover a previously underappreciated role for ABHD6 in promoting the metabolic syndrome. These studies demonstrate that ABHD6 plays a non-redundant enzymatic role in promoting the metabolic disorders induced by high-fat feeding, and suggest that ABHD6 inhibition may be effective in preventing obesity, non-alcoholic fatty liver disease, and type II diabetes.
Mouse ABHD6 is a 336 amino acid protein that shares high sequence identity with its human (94%), macaque (94%), and rat (97%) orthologues (Figure 1A). A highly conserved active site serine nucleophile is found at residue 148 (Figure 1A), which is predicted to be necessary for enzyme catalysis. ABHD6 mRNA is ubiquitously expressed (Figure 1B), with highest expression in small intestine, liver, and brown adipose tissue in mice fed standard rodent chow. Additionally, high fat diet feeding increases ABHD6 mRNA expression in the small intestine and the liver (Figure 1B). This transcriptional regulation of ABHD6 in metabolic tissue prompted us to examine whether ABHD6 may be an important mediator of high fat diet-induced metabolic disease.
To examine the role of ABHD6 in lipid metabolism in peripheral tissues without altering expression in the brain, we utilized antisense oligonucleotide (ASO) targeting (Crooke et al., 1997). Initially, we tested four separate ASOs targeting murine ABHD6, and found that all reduced hepatic ABHD6 mRNA by >80% after 4 weeks of administration (Figure S1A). Following 12 weeks of ASO knockdown with two of our ABHD6 ASOs (ASOα and ASOβ), we observed tissue selective knockdown of ABHD6 protein with the following rank order: liver (>90%) > white adipose tissue (>90%) > kidney (~50%) (Figure 2A and Figure S1B). In contrast, ABHD6 protein expression in the brain, spleen, and brown adipose tissue was relatively unaffected by ASO treatment (Figure S1B). Given that ABHD6 has been described as a key enzymatic regulator of endocannabinoid signaling in the brain (Blankman et al., 2007; Marrs et al., 2010; Marrs et al., 2011), we carefully examined whether ASO treatment reduced ABHD6 in brain regions relevant to metabolic disease. Importantly, ABHD6 was not altered in the hippocampus (Figure S1D), whereas livers from the same mice showed marked reductions in ABHD6 expression in the liver (Figure S1C). Furthermore, the mRNA expression of ABHD6 in the hypothalamus was likewise not altered by ASO treatment (Figure S1E).
ABHD6 ASO treatment did not alter body weight (Figure 2B), adiposity (Figure 2E), or food intake (Figure 2H) in mice fed a standard rodent chow diet. However, whenchallenged with a high fat diet, ABHD6 ASO-treated mice were protected from diet-induced body weight gain (Figure 2B and 2C), which was in large part due to a reduction in adipose tissue (Figure 2D). Magnetic resonance imaging (MRI) showed that ABHD6 ASO treatment significantly reduced body fat mass (Figure 2E) without altering lean body mass (Figure 2F). Furthermore, ABHD6 ASO treatment did not result in general growth retardation, given that snout to anus lengths were unaffected (Figure 2G). The protection from high fat diet-induced obesity in ABHD6 ASO-treated mice could not be explained by reductions in food intake (Figure 2H) or by reductions in intestinal fat absorption (Figure 2I). In fact the absorption of several long chain fatty acids was actually slightly increased in ABHD6 ASO-treated mice (Table SI). Gene expression analysis in epididymal white adipose tissue revealed that ABHD6 knockdown caused increased expression of lipolyticgenes such as HSL and MAGL (Figure 2J), with very minor changes in lipogenic gene expression such as SREBP1c and SCD1 in white adipose tissue (Figure 2J). The ability of ABHD6 ASO treatment to protect against high fat diet-induced obesity could be explained in part by increases in physical activity during the dark cycle (Figure 2K) and increases in energy expenditure during both the diurnal and nocturnal phases (Figure 2L). There were no detectable alterations in the respiratory exchange ratio (RER) in control and ABHD6 ASO-treated mice fed a high fat diet (Figure 2M).
ABHD6 knockdown significantly reduced high fat diet-induced accumulation of total hepatic triacylglycerol (TAG) (Figure 3A) without altering total hepatic levels of diacylglycerols (DAG) or monoacylglycerols (MAG) in either dietary setting (Figure 3B). This reduction in neutral lipids was observed in almost all molecular species of TAG, with the exception of 56:6 TAG (Figure S3A). ABHD6 ASO treatment significantly blunted high fat diet-driven increases in some hepatic DAG species (32:2, 32:1, 34:3, 34:2, 34:1, and 38:6), but not in others (36:2) (Figure S3B). In parallel, ABHD6 knockdown protected mice from high fat diet-induced hyperglycemia (Figure 3C), hyperinsulinemia (Figure 3D), and improved both glucose and insulin tolerance (Figure 3H and 3I). Interestingly, reducing ABHD6 did not alter plasma TAG levels on either diet (Figure 3E), yet knockdown increased plasma non-esterified fatty acid (NEFA) levels specifically in high fat diet fed mice (Figure 3G). Additionally, ABHD6 ASO treatment did not alter VLDL triacylglycerol secretion rates (Figure 3J). However, ABHD6 knockdown protected against high fat-dietinduced hypercholesterolemia (Figure 3F), which was reflected as a significant decrease in LDL and a modest increase in HDL levels (data not shown). Importantly, hepatic short chain TAG species were significantly decreased by ABHD6 ASO treatment (50:4, 52:5,52:4, and 54:7) in chow-fed mice, where systemic insulin sensitivity and plasma triacylglycerol levels were similar to control mice (Figure S7).
When comparing global hepatic gene expression between high fat diet-fed control vs. ABHD6 ASO-treated mice, we found a surprisingly small number of genes that were differentially expressed by greater than 2-fold (167 upregulated genes and 138 downregulated genes; p <0.005) (Figure 4A). The most highly enriched gene ontology (-log[p-value] = 10) regulated by ABHD6 knockdown was fatty acid metabolism (Figure 4A). ABHD6 ASO treatment resulted in a 50% reduction in hepatic ABHD5 expression, which is unlikely to be from direct ASO-mediated silencing due to lack of sequence homology between the two mRNAs. Knockdown of ABHD6 also increased ATGL expression, while reducing HSL expression in the liver (Figure 4B). More consistently, we observed coordinate downregulation of genes involved in de novo fatty acid synthesis and lipogenesis (SREBP1c, FAS, ACC-1, and SCD-1) in ABHD6 ASO-treated mouse liver (Figure 4B and 4C). In agreement, the in vivo rate of de novo fatty acid synthesis wasreduced by 62% in ABHD6 ASO-treated mice (Figure 4D). Consistent with this, primary hepatocytes isolated from ABHD6 ASO-treated livers showed decreased rates of 3H-oleate esterification into triacylglycerol (Figure 4E) as well as decreased de novo lipogenesis rates as measured by the kinetic conversion of 14C-acetate into 14C-triacylglycerol (Figure 4F). It is important to note that all four ASOs targeting the knockdown of ABHD6 at different sites consistently decreased hepatic lipogenic gene expression (Figure S1B, S1C, and S1D), without altering adipose tissue (Figure 2J) or hypothalamic (Figure S1E) lipogenic gene expression. We also examined thephosphorylation state of both AMPKα and AMPKβ in mouse liver, but noted no obvious differences in activation state (Figure S2E).
Given that ABHD6 was previously described as a monoacylglycerol lipase in the brain (Blankman et al., 2007; Marrs et al., 2010; Marrs et al., 2011), we carefully examined whether ABHD6 knockdown resulted in accumulation of hepatic MAG species, or whether ABHD6 knockdown resulted in hyperactivation of the ECB system in mouse liver. ABHD6 knockdown did not alter the total hepatic levels of MAG in mice fed either chow or a high fat diet (Figure 3B), although two species of MAG containing oleate (18:1) or linoleate (18:2) did modestly increase (Figure 5A). In contrast, hepatic levels of endocannabinoid lipids (2-AG and anandamide) were not changed by ABHD6 ASO treatment (Figure 5B and 5C). There was also no apparent difference in total hepatic MAG lipase activity with ABHD6 knockdown (data not shown). Given that previous studies have demonstrated that the CB1-dependent signaling is largely desensitized when 2-AG builds up as a result of monoacylglycerol lipase (MAGL) inhibition or genetic deficiency (Schlosburg et al., 2010; Taschler, et al., 2012), we carefully examined acute CB1 signaling in ABHD6 ASO-treated mice. To examine the effect of ABHD6 knockdown on CB1 receptor desensitization, we administered the cannabinoid receptor (CB1) agonist CP-55,940 or vehicle directly into the portal vein of ASO-treated mice and followed downstream signaling. ABHD6 knockdown did not alter CP-55,940-induced ERK-MAPK activation compared to control ASO-treated mice (Figure 5D). However, basal activation of ERKMAPK was lower in ABHD6 ASO-treated mice (Figure 5D), indicating that ABHD6 may be involved in regulating other inputs into ERK-MAPK activation.
Interestingly, a number of phospholipids and lysophospholipids accumulated to varying degrees in ABHD6 ASO-treated livers (Figure 6), implicating them as potential physiologically relevant substrates. ABHD6 knockdown significantly increased total hepatic levels of phosphatidylcholine (PC), lysophosphatidylcholine (LPC), phosphatidylethanolamine (PE), lysophosphatidylethanolamine (LPE), phosphatidylglycerol (PG), lysophosphatidylglycerol (LPG), phosphatidylinositol (PI), lysophosphatidylinositol (LPI), and phosphatidylserine (PS) (Figure 6B, 6C, 6D, 6E, 6F, 6H, 6I, 6J, 6K, and 6L), without altering hepatic levels of phosphatidic acid or lysophosphatidic acid species (Figure 6A, 6G, and Figure S4C). Of interest, the most prominent accumulation was observed for nearly all species of LPG (Figure 6J, Figure S6A) and PG (Figure 6D, Figure S5C) in ABHD6 knockdown mice. Inhibition of ABHD6 also promoted the accumulation of several ether-linked glycerophospholipids (plasmalogens) including all detected molecular species of plasmanylcholines (Figure S4A) and plasmenylethanolamines (Figure S4B). In parallel, nearly all species of LPE, LPG, LPI, and LPS were increased with ABHD6 knockdown regardless of diet (Figure S6). We subsequently tested whether ABHD6 can hydrolyze phospholipid and neutral lipid substrates in vitro. To accomplish this we expressed a GST-tagged ABHD6 in S. cerevisiae (Figure 6M), and the purified protein was incubated in the presence of a panel of lipid substrates (Figure 6M-Q). As previously demonstrated with ABHD6-expressing cell homogenates (Blankman et al., 2007; Marrs et al., 2010; Marrs et al., 2011; Navia-Paldanius et al., 2012) ABHD6 hydrolyzes MAG substrates including 1,(3)-rac--oleoylglycerol and 2-oleoylglycerol, and its ability to hydrolyze MAG substrates is lost when the active site serine is mutated to alanine (S148A) (Figure 6N). However, in saturation kinetic experiments recombinant MAGL exhibited a 11-fold higher specific activity (Vmax = 5359.3 μmol/h*mg) than ABHD6 (Vmax = 478.6 μmol/h*mg) (Figure 6O), and under the applied conditions the apparent Km was lower for MAGL (Km = 1.2) compared to ABHD6 (Km = 1.9), indicating that ABHD6 has a lower affinity for MAG.
Next, we investigated whether affinity-purified ABHD6 hydrolyzes the other glycerophospholipid substrates identified by in vivo lipidomics (Figure 6A-6L). Recombinant ABHD6 showed considerable lipase activity towards several lysophospholipids including LPG, LPA, and LPE but not LPC (Figure 6P). In contrast, ABHD6 exhibited no lipase activity against major phospholipid classes including PG, PE, PA, PS, and PC (Figure 6P). The highest activity was observed using LPG as a substrate (Figure 6P and 6Q), and this activity was even higher in a detergent-containing (5 mM CHAPS) buffer system. Under these conditions, we determined a Vmax of 93.2 μmol/h*mg and a Km of 0.75 (Figure 6Q). Furthermore, purified ABHD6 exhibited low activity against retinyl palmitate (RP) and rac-dioleoylglycerol, whereas no activity was observed in the presence of trioleoylglycerol or cholesteryl oleate (data not shown). Notably, ABHD6's activityagainst 1,3-diacylglycerol was 5-fold higher in comparison to 1,2(2,3)-diacylglycerol (data not shown). The ability of ABHD6 to hydrolyze neutral lipid and lysophospholipid substrates was completely lost when the active site serine was mutated to alanine (Figure 6 and data not shown). It is important to note that MAGL did not hydrolyze any of the tested phospholipid substrates (data not shown) annotating MAGL, but not ABHD6, as a specific MAG hydrolase. Taken together, these observations suggest that ABHD6 can act both as a monoacylglycerol lipase and lysophospholipase exhibiting a preference for LPG among the tested lysophospholipids.
ASO-mediated inhibition is a tissue-restricted therapeutic approach, targeting knockdown in metabolic tissues (liver, white adipose tissue, and kidney) without altering ABHD6 expression or activity in many other tissues (brain, spleen, brown adipose tissue). However, ABHD6 is ubiquitously expressed (Figure 1B), begging the question whether systemic inhibition would likewise protect against metabolic disease. Therefore, we determined whether a small molecule inhibitor of ABHD6, which would be predicted to target all tissues including the brain, also provides protection against the metabolic disorders driven by high fat diet feeding. Treatment with the small molecule inhibitor of ABHD6 (WWL-70) protected mice from high fat diet-induced body weight gain (Figure 7A), which was largely due to a reduction in adipose tissue mass (Figure 7B). WWL-70 treatment also protected mice against high fat diet-induced glucose intolerance. Although there was a trend towards a decrease, WWL-70 did not significantly reduce hepatic triacylglycerol levels in high fat diet fed mice (Figure 7D), which contrasts with what is observed with ASO-mediated inhibition (Figure 3). These results show that inhibition of ABHD6 using a systemic inhibitor improves some aspects of the metabolic syndrome, but does not protect against hepatic steatosis to the same degree that ASO-mediated inhibition does (Figure 3).
Although other members of the ABHD protein family have been clearly linked to lipid signaling and metabolic disease (Lefevre et al., 2001; Fiskerstrand et al., 2010; Long et al., 2011; Simon and Cravatt 2006; Montero-Moran et al., 2009; Blankman et al., 2007; Lord et al., 2011; Brown et al., 2010), this is the first study to document a role for ABHD6 in promoting the metabolic disorders driven by high fat diet feeding. The major findings of this work demonstrate the following: 1) peripheral knockdown of ABHD6 protects mice from high fat diet-induced obesity, 2) ABHD6 knockdown protects mice from high fat diet-induced hepatic steatosis and associated insulin resistance, 3) ABHD6 is a critical regulator of hepatic de novo lipogenesis, and 4) ABHD6 hydrolyzes several lipid substrates in vitro, including lysophospholipids with preference for LPG. Accordingly, we show that inhibition of ABHD6 results in the accumulation of LPG and PG in vivo. Taken together, our results reveal that ABHD6 plays a role in the development of the metabolic syndrome.
ABHD6 has been shown to regulate the ECB system in the brain due to its ability to hydrolyze specific pools of 2-AG (Blankman et al., 2007; Marrs et al., 2010; Marrs et al., 2011). We also confirmed that ABHD6 exhibits MAG lipase activity in vitro (Figure 6N and 6O). Our data suggest however, that ABHD6 is a minor contributor to MAG lipolysis and ECB signaling in mouse liver. This finding is supported by the fact that total MAG levels do not change (Figure 3B), 2-AG and anandamide levels are not elevated (Figure 5B and 5C), and acute CB1-driven activation of ERK-MAPK is unaltered with ABHD6 knockdown (Figure 5D), arguing against CB1 receptor desensitization. It is important to contrast these findings with the very striking accumulation of hepatic 2-AG, marked reduction in hepatic MAG lipase activity, and CB1 receptor desensitization seen in both MAGL−/− mice and mice treated chronically with a specific MAGL inhibitor (Taschler et al., 2011; Schlosburg et al., 2010). It is also important to note that if ABHD6 was a major regulator of hepatic 2-AG levels, one would anticipate that ABHD6 knockdown would cause hyperactivation of hepatic CB1 signaling, which has been reported to promote hepatic steatosis through increasing de novo lipogenesis (Osei-Hyiaman et al., 2005, Osei-Hyiaman et al., 2008). However, we observed that ABHD6 knockdown protected mice from high fat diet induced hepatic steatosis and suppressed de novo lipogenesis (Figure 4). Collectively, these results support the notion that MAGL is the predominant MAG lipase in mouse liver. We did, however, see minor elevations in 18:1 and 18:2 MAG (Figure 5A), making it possible that ABHD6's ability to hydrolyze these lipids may affect other signaling processes regulating hepatic de novo lipogenesis.
ABHD6 contributes to 2-AG hydrolysis and ECB signaling in the brain (Blankman et al., 2007; Marrs et al., 2010; Marrs et al., 2011). However, MAGL accounts for the vast majority (>80%) of 2-AG hydrolysis in the brain, whereas ABHD6 only accounts for less than 5% of the total 2-AG hydrolase activity (Blankman et al., 2007). Our studies suggest that ABHD6's additional ability to hydrolyze a variety of lipid substrates may be important in linking lipid mediator signaling to the metabolic adaptations resulting from high fat diet feeding. In this study we utilized a targeted lipidomics approach to identify new substrates of ABHD6 in mouse liver, and verified LPG as a bona fide substrate in vitro and in vivo. LPG is considered a bioactive lipid, although its receptor remains to be identified (Makide et al., 2009; Jo et al., 2008). Lysophospholipids have previously been implicated in promoting metabolic disease via their ability to dictate membrane dynamics and initiate cell signaling, particularly in immune cells (Wymann et al., 2008; Skoura and Hla, 2009). We assume that defective LPG degradation favors the reesterification of LPG to PG, which may explain the elevated PG levels in the ABHD6-knockdown liver. PG is also an important precursor for other complex lipids including cardiolipin and bis(monoacylglycerol)phosphate (Hullin-Matsuda et al., 2007). Therefore, changes in ABHD6-driven LPG metabolism could alter many mitochondrial or lysosomal metabolic processes. Interestingly, a recent genome-wide association study (GWAS) in Pima Indians identified the lysophosphatidylglycerol acyltransferase 1 (LPGAT1) loci as being predictive of body mass index (BMI), providing additional genetic evidence that LPG metabolism may influence BMI and adiposity (Traurig et al., 2012).
Projecting forward, there are several important factors to consider for developing ABHD6 as a drug target. For example, it will be important to determine how ABHD6 expression and subcellular distribution are regulated in diverse cell and tissue contexts. Furthermore, it will be essential to evaluate whether ABHD6 has other physiological substrates in vivo, and if tissue-selective inhibitors will be necessary for safe and effective prevention of metabolic disease. Another important consideration will be whether chronic ABHD6 inhibition results in CB1 receptor desensitization in the brain or other receptor signaling abnormalities. Our data suggest that chronic ABHD6 inhibition with ASOs does not cause CB1 receptor desensitization in mouse liver (Figure 5D). However, whether CB1 receptor desensitization occurred in other tissues was not examined here. As previously documented (Bachovchin et al., 2010), we show that ABHD6 is ubiquitously expressed in mice (Figure 1B). Additional studies will be required to assess the risk versus benefits of inhibiting ABHD6 in specific tissues. Our studies provide key information in regards to ABHD6's biochemical and physiological role in certain peripheral tissues (liver, adipose, and kidney) given the tissue-selective inhibition observed with ASOs (Figure S1). However, our studies do not address ABHD6's primary role in other tissues where it is abundantly expressed (brain, small intestine, pancreas, and skeletal muscle), all of which are key sites regulatory sites for energy metabolism.
It is important to compare and contrast the results of inhibiting ABHD6 with systemic small molecule inhibitors versus a more selective approach using ASO technology. Although both ASO-mediated and WWL-70-mediated inhibition of ABHD6 protected against high fat diet-induced obesity and glucose intolerance, only ABHD6 ASO treatment improved hepatic steatosis. The mechanism underlying this difference is unclear at this point, but this most likely stems from systemic versus tissue-restricted inhibition. Interestingly, ASO-mediated inhibition of ABHD6 was not associated with alteration in food intake at any time point examined (Figure 2H). However, WWL-70 treatment caused a 20% reduction in food intake (data not shown), indicating that the mechanism by which WWL-70 protected against obesity and glucose intolerance is likely driven by hypophagia, while ASO-mediated inhibition more specifically dampened hepatic lipogenesis and increases in energy expenditure.
In addition to further studies with selective ABHD6 inhibitors, the generation of global and conditional knockout mouse models is necessary to define the tissue-specific role for ABHD6 in chronic diseases of altered lipid metabolism. In summary, our data identify ABHD6 as a key determinant in the pathogenesis of HFD-induced obesity, insulin resistance, and hepatic steatosis. Furthermore, we have uncovered novel lipid substrates of ABHD6 by coupling in vivo lipidomic and in vitro biochemical approaches. Looking forward, we believe that utilizing a similar in vivo loss-of-function approach may prove useful for mapping natural enzyme-substrate relationships for the other uncharacterized enzymes in the ABHD family. It is interesting to note that during the preparation of this manuscript that another ABHD enzyme ABHD12 was likewise identified as a dual monoacylglycerol lipase and lysophospholipase with preference towards lysophosphatidylserine species using a similar in vivo substrate identification approach (Blankman et al., 2013). Taken together, our findings suggest that ABHD6 is a key lipase involved in monoacylglycerol and lysophospholipid hydrolysis, and that ABHD6 inhibitors hold promise as therapeutics for obesity, non-alcoholic fatty liver disease, and type II diabetes.
For ABHD6 knockdown studies, at 6-8 weeks of age male C57BL/6N mice (Harlan) were either maintained on standard rodent chow or switched to a high fat diet for a period of 4-12 weeks, and were simultaneously injected with murine specific ABHD6 antisense oligonucleotides biweekly (25 mg/kg BW) as previously described (Lord et al., 2011; Brown et al., 2010; Brown et al., 2008a; Brown et al., 2008b). For small molecule inhibitor studies, mice were fed a high fat diet and simultaneously treated with either a vehicle or 10 mg/kg WWL-70 for 8 weeks. Intraperitoneal glucose tolerance tests and insulin tolerance tests were performed essentially as previously described (Brown et al., 2008a, Brown et al., 2010) in mice treated with diet and ASO for 10-11 weeks. Metabolic measurements were conducted in the comprehensive lab animal monitoring systems (CLAMS) from Columbus Instruments. Body composition was determined by magnetic resonance imaging (MRI). A detailed description of all mouse experiments is provided in the online supplementary methods section.
Mice were injected with control ASO or ABHD6 ASOβ and maintained on a HFD for a period of 8 weeks prior to experiment. After an overnight fast (9:00 p.m. - 9:00 a.m.), mice were anesthetized with isoflurane (4% for induction, 2% for maintenance), and were maintained on a 37°C heating pad to control body temperature. A minimal midline laparotomy was performed and the portal vein was visualized. Thereafter, mice received either vehicle (cremophor EL:ethanol:saline at a 1:1:18 ratio) or the CB1 agonist (CP-55,490, Cayman Chemical # 13241) at a dose of 0.1 mg/kg body weight directly into the portal vein. Exactly 5 minutes later, the liver was excised and immediately snap frozen in liquid nitrogen. Protein extracts from tissues were analyzed by Western blotting to examine CB1 signaling as described below.
VLDL secretion rates were determined using the detergent block method (Li et al., 1996). Dietary fat absorption was measured using the Olestra® method (Jandacek et al., 2004), and de novo fatty acid synthesis was measured using the 3H-water method (Shimano et al., 1996). These methods are described in detail in the online supplementary methods section.
Hepatocytes were isolated from chow-fed ASO-treated mice using the collagenase perfusion method, and the rate of de novo lipogenesis was determined following the conversion of 14C-acetate and 3H-oleate into newly synthesized triacylglycerol in the presence of lipase inhibitors to block lipolysis/re-esterification. A detailed method is provided in the online supplementary methods.
A maltose binding protein ABHD6 fusion protein construct was generated to create affinity-purified rabbit polyclonal antibodies against murine ABHD6. A detailed description of cloning, expression, and purification are included in the online supplementary methods section.
Whole tissue homogenates were made from multiple tissues in a modified RIPA buffer, and Western blotting was conducted as previously described (Brown et al., 2004).
Tissue RNA extraction was performed as previously described for all mRNA analyses (Lord et al., 2011; Brown et al., 2010; Brown et al., 2008a; Brown et al., 2008b). Microarray analyses were performed by the Wake Forest School of Medicine Microarray Shared Resource Core using standard operating procedures, and quantitative real time PCR (qPCR) analyses were conducted as previously described (Lord et al., 2011; Brown et al., 2010; Brown et al., 2008a; Brown et al., 2008b). A detailed description of RNA methods is available in the online supplementary methods section.
Extraction of liver lipids and quantification of molecular species by mass spectrometry was performed as previously described (Lord et al., 2011; Ivanova et al., 2007; Myers et al., 2011; Callender et al., 2007; Saghatelian et al., 2004).
The coding sequence of murine ABHD6 was cloned into the yeast expression vector pYEX4T-1. The resulting protein was purified using glutathione-sepharose beads, and used for enzymology studies as described in the online supplementary methods section.
All data are expressed as the mean ± S.E.M. or S.D., and were analyzed using either a one-way or two-way analysis of variance (ANOVA) followed by Student's t tests for post hoc analysis using JMP version 5.0.12 software (SAS Institute, Cary, NC).
We thank Larry Rudel, Paul Dawson, and Ryan Temel (Wake Forest School of Medicine) for insightful comments and suggestions. We also sincerely thank Marc Prentki and Murthy Madiraju (Montreal Diabetes Research Center) for critical review of this work. This work was supported by the Department of Pathology at Wake Forest School of Medicine, a pilot grant from the Wake Forest School of Medicine Venture Fund, and a pilot grant awarded under the Wake Forest and Harvard Center for Botanical Lipids (P50-AT002782). These studies also received generous funding by the National Institute of General Medical Sciences LIPID MAPS (U54-GM069338 to H.A.B.), and National Institute of Diabetes and Digestive and Kidney Diseases (F32-DK084582 to J.L.B.).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Supplemental information including Extended Experimental Procedures, 7 Supplemental Figures, and 1 Supplemental Table can be found with this article online.
Other than Richard Lee, Rosanne Crooke, and Mark Graham, who are employees at ISIS pharmaceuticals, Inc. (Carlsbad, CA), all other authors report that they have no conflicts of interest.