The absence of ABCD2 sensitizes mice to disruptions in lipid metabolism by dietary erucic acid.

ABCD2 (D2) is a peroxisomal transporter that is highly abundant in adipose tissue and promotes the oxidation of long-chain MUFA. Erucic acid (EA, 22:1ω9) reduces very long chain saturated fatty acids in patients with X-linked adrenoleukodystrophy but promotes dyslipidemia and dilated cardiomyopathy in rats. To determine the role of D2 in the metabolism of EA, we challenged wild-type and D2 deficient mice (D2 KO) with an enriched EA diet. In D2 KO mice, dietary EA resulted in the rapid expansion of adipose tissue, adipocyte hypertrophy, hepatic steatosis, and the loss of glycemic control. However, D2 had no impact on the development of obesity phenotypes in two models of diet-induced obesity. Although there was a significant increase in EA in liver of D2 KO mice, it constituted less than 2% of all fatty acids. Metabolites of EA (20:1, 18:1, and 16:1) were elevated, particularly 18:1, which accounted for 50% of all fatty acids. These data indicate that the failure to metabolize EA in adipose results in hepatic metabolism of EA, disruption of the fatty acid profile, and the development of obesity and reveal an essential role for D2 in the protection from dietary EA.

ever, the absence of D2 had no impact on the development of obesity phenotypes when challenged with a high-fat (HF) diet or when crossed with leptin-defi cient (ob/ob) mice. Although adipose tissue is greatly expanded after EA feeding, the lipid profi le is only marginally affected with little accumulation of EA. Although there is a signifi cant increase in EA in liver, it constitutes less than 2% of the total fatty acid pool. Conversely, hepatic SFA and MUFA content are elevated by 2-and 3-fold, respectively. These data suggest that in the absence of D2, peroxisomal metabolism of EA proceeds effi ciently in the liver but that EA metabolites disrupt hepatic lipid metabolism, resulting in obesity, insulin resistance, and steatosis.

Animal husbandry
Mice harboring the mutant Abcd2 allele are maintained on the C57BL/6J background as heterozygotes. Strain refreshing is conducted every fi ve generations using C57BL6/J females obtained from The Jackson Laboratory (Bar Harbor, ME). Genotyping experiments to differentiate Abcd2-defi cient (D2 KO) from heterozygous and wild-type mice were conducted as previously described ( 18 ). Animals were housed in individually ventilated cages in a temperature-controlled room with 14:10 light:dark cycle and provided with enrichment in the form of acrylic huts and nesting material. All mice were maintained on standard rodent chow (Harlan Teklad 2014S) until initiation of experiments. All animal procedures conformed to PHS policies for humane care and use of laboratory animals and were approved by the institutional animal care and use committee at the University of Kentucky.
The EA-enriched diet was custom formulated to contain levels of EA found in native rapeseed oil and contained 21.6% total energy from fat (Research Diets, New Brunswick, NJ) (Supplementary Tables I and II). EA accounted for 55.6% of total fatty acids as determined by GC-MS. The diet was stored at 4°C, provided ad libitum, and replaced twice weekly to limit oxidation. Diets were initiated at 8 weeks of age and continued until termination of the experiment after 8 weeks of feeding (n = 7). A second cohort of wild-type and D2 KO littermates were analyzed for the development of obesity phenotypes after low fat (LF) (10% kCal, Research Diets #D12450B, n = 10) or HF (45% kCal, Research Diets #D12451, n = 13 wt, 8 D2 KO) diets. Diets were initiated at 8 weeks of age. Mice were monitored weekly for weight gain and monthly for body composition. Mice were euthanized at 24 weeks (16 weeks on diet). D2 KO mice were also crossed to a leptin-defi cient strain ( Lep ob , Strain #000632; The Jackson Laboratory), maintained on standard rodent chow, and analyzed at 8, 12, and 16 weeks of age (n = 11 ob/ob; n = 8 ob/ob D2 KO).
Body composition was determined by EchoMRI at the initiation of the feeding period and the day before the termination of the study. During week 7, fasting blood glucose levels were measured using a standard glucometer from a drop of blood obtained by tail-vein prick after a 4 h fast beginning at lights-on. To determine glucose tolerance, mice were injected with sterilized 20% glucose solution (10 l/g of body weight, i.p.). Blood glucose levels were measured before and 30, 60, 90 and 120 min after glucose injection.
At termination of the experiments, mice were euthanized by exsanguination under ketamine/xylazine anesthesia after a 4 h fast beginning shortly after lights-on. Blood was collected from the right ventricle with a 1 ml syringe fi tted with a 20 gauge hypodermic X-linked adrenoleukodystrophy (ALD). ALD is a pleiotropic disorder characterized by the accumulation of VLCFAs in plasma and tissues that presents with varying neurological, adrenocortical, and Leydig cell defi ciencies ( 6,7 ). The closest paralog of D1 is D2, which shares greater than 60% amino acid identity in humans and mice and at least some overlapping function in vitro and in vivo (8)(9)(10)(11)(12). Although direct evidence for transmembrane transport of fatty acyl-CoA substrates is lacking, human D1 and D2 transgenes have been shown to rescue growth in oleatecontaining media and to promote the oxidation of distinct subsets of fatty acids in yeast lacking endogenous peroxisomal fatty acyl-CoA transporters ( 13,14 ). In this system, D2 was most active toward the 22-carbon saturated fatty acid and 22-and 24-carbon PUFAs ( 14 ).
Whereas D1 is constitutively expressed in many tissues, D2 is more restricted and dynamically regulated by transcription factors involved in lipogenesis, fatty acid oxidation, and cholesterol elimination. D2 mRNA and protein have been detected in brain, liver, lung, adrenal gland, and skeletal muscle but are most abundant in adipose tissue ( 9,15 ). Within adipose, D2 is restricted to adipocytes and is upregulated during adipogenesis, presumably by the lipogenic transcription factor sterol receptor element binding protein 1 (SREBP1) ( 15,16 ). Conversely, the expression of D2 is relatively low in liver, but is up-regulated by fasting and peroxisome proliferator activated receptor (PPAR) ␣ agonists ( 9 ). Hepatic expression of D2 is suppressed by the liver X receptor (LXR), suggesting a role for D2 in cholesterol metabolism ( 16 ). However, the functional signifi cance of this observation is unclear and has not been investigated.
Despite its regulation by transcriptional factors essential for lipid homeostasis in liver and adipose, the physiological role for D2 in lipid metabolism remains largely unknown. The absence of D2 results in a modest accumulation of VLCFAs in liver after feeding with a diet enriched in saturated fat and after prolonged fasting, but these differences did not result in overt metabolic phenotypes, nor did they affect the expression of key metabolic enzymes ( 12 ). Similarly, there is no apparent adipose dysfunction, and fatty acid profi les in adipose tissue were largely unaffected in D2 defi cient mice ( 15 ). As in liver, adrenal, and neural tissues, there was a modest accumulation of 22-and 24-carbon saturated fatty acids (SFAs) and MUFAs. When acutely challenged with a diet enriched in erucic acid (EA) (22:1 -9), there was a gene-dosage dependent increase in the levels in adipose and plasma ( 15 ).
EA is abundant in native cultivars of rapeseed ( Brassica napus ) and is used clinically in combination with oleic acid to reduce plasma levels of VLCFA in patients with ALD due to its ability to inhibit the elongation of SFAs ( 6 ). However, it is also associated with a transient lipidosis, dyslipidemia, and the development of dilated cardiomyopathy in rats, effects not observed in mice and opposed by treatment with the PPAR ␣ agonist clofi brate ( 17 ).
We hypothesized that the absence of D2 would sensitize mice to disturbances in lipid metabolism by dietary EA.
Our results indicate that in the absence of D2, dietary EA promotes obesity, insulin resistance, and steatosis. How-Radiolabeled Chemicals, Inc., St. Louis, MO) and 20 M of the indicated fatty acids in the presence of 100 M CoA, 1 mM ATP, and 1 mM NADPH at 37°C for 10 min. Fatty acids were extracted from the reaction mixture, and the total incorporation of [ 14 C] malonyl-CoA to fatty acid substrates was measured by scintillation counting.

Protein and RNA analysis
Membrane proteins were prepared and analyzed by SDS-PAGE and immunoblot analysis as previously described ( 25 ). The isolation of total RNA and the determination of relative transcript abundance by quantitative real-time PCR for both tissues and cells was conducted as previously described ( 25 ).

Statistical analysis
Body composition data were analyzed by two-way repeated measures ANOVA. Bonferroni posts tests were used where indicated. Comparisons between genotypes for other measures were conducted with two-tailed t -tests. All statistical analyses were conducted using GraphPad Prism statistical analysis software.

D2 defi cient mice are susceptible to EA-induced obesity, dyslipidemia, and loss of glycemic control
To determine the role of D2 in the metabolism of dietary EA, mice were challenged with an EA-enriched diet for 8 weeks beginning at 8 weeks of age. As previously reported, differences in body weight, body composition, and other measures of obesity and lipid metabolism were not observed at the initiation of the study (Fig. 1, Reference 15). After EA feeding, we observed a modest but significant increase in body weight in D2 KO mice compared with wild-type controls. Analysis of body composition by MRI indicated a signifi cant increase in fat mass in D2 KO mice after 8 weeks on diet. The accumulation of excess adipose tissue occurred primarily in epididymal and retroperitoneal depots ( Fig. 1B ). Whereas lean mass increased in wild-type mice during EA feeding, it remained relatively constant in D2 KO and was signifi cantly lower than in wildtype controls at the termination of the study ( Fig. 1A ). The needle. Serum was separated by centrifugation and stored at Ϫ 20°C. Tissues were excised, rinsed with PBS to remove blood, and snap frozen in liquid nitrogen. Liver, heart, and individual fat pats were dissected and weighed. Tissue samples were stored at Ϫ 80°C. Additional tissue samples were immediately embedded in OCT, frozen on dry ice, and stored at Ϫ 20°C or formalin fi xed overnight and stored in 70% ethanol at 4°C until processing and histological analysis.

Histology
Tissue was processed in a dehydrating ethanol gradient followed by xylene incubation and paraffi n embedding. Paraffi n blocks of tissue were cut into sections of 1 to 3 m thickness and stained for hematoxylin and eosin (H&E). Frozen sections were cut at 10 m thickness and subjected to Oil-Red-O staining.
The acyl-carnitine extraction method is a modifi cation of the method previously described ( 21 ). Tissues were weighed and homogenized in water and extracted with 80% acetonitrile with a series of vortexing, sonication, and centrifugation at 4000 rpm for 10 min. The supernatant was transferred into a vial, dried under nitrogen, and reconstituted with 100 l of methanol. D3-C2, D3-C4, and D3-C16 acylcarnitines were used as internal standards for short-and long-chain acyl carnitines, respectively. Analyses of acylcarnitines was carried out using a Shimadzu UFLC coupled with an AB Sciex 4000-Qtrap hybrid linear ion trap triple quadrupole mass spectrometer in multiple reaction monitoring mode. Acyl carnitines were analyzed using an XTerra MS C8 3.5 m, 3.0 × 100 mm column. The mobile phase consisted of 75/25 methanol/water, 0.5% formic acid + 0.1% ammonium formate as solvent A and 80% of (99/1 methanol/water, 0.5% formic acid + 0.1% ammonium formate): 20% chloroform as solvent B. For the analyses of acyl carnitines, the separation was achieved using a gradient of 0 to 100% solvent B in 10 min and maintaining at 100% B for the next 8 min. The column equilibrated back to the initial conditions in 3 min. The fl ow rate was 0.5 ml/min with a column temperature of 30°C. The sample injection volume was 10 uL. The mass spectrometer was operated in the positive electrospray ionization mode with optimal ion source settings determined by synthetic standards of acylcarnitines with a declustering potential of 86 V, entrance potential of 10 V, collision energy of 53 V, collision cell exit potential of 6 V, curtain gas of 20 psi, ion spray voltage of 5500 V, ion source gas1/gas2 of 40 psi, and temperature of 550°C. Multiple species of acyl carnitines wewre quantitated by monitoring species specifi c precursor product ion pairs.

Enzyme assays
Fatty acid elongation was determined as described previously ( 22 ). Briefl y, microsomes (30 g total protein) were incubated in a reaction mixture containing 150 M [ 14 C] malonyl-CoA (American absence of infl ammation in D2 KO adipose tissue after EA feeding despite the increase in adipose mass and adipocyte hypertrophy.

D2 does not infl uence the development of diet-induced obesity
D2 is up-regulated in liver of obese and insulin-resistant mice, suggesting that D2 might modulate the development of obesity phenotypes in multiple mouse models ( 26 ) (Supplementary Figure I). First, mice were challenged with a HF (45% kCal) diet for 12 weeks beginning at 8 weeks of age ( Fig. 2 ). Although there was a modest increase in fat mass, we observed no differences in body weight, adiposity, fasting glucose, or glucose tolerance. There were also no differences in food intake or blood lipids (Supplementary  Table III). We next crossed D2 defi cient mice with the leptin-defi cient (ob/ob) strain, maintained them on standard rodent chow, and evaluated them at 8, 12, and 16 weeks of age. As in HF feeding, there was a tendency toward increased fat mass at 12 weeks, but there were no differences in other measures of obesity phenotypes (Supplementary  Figure II, Supplementary Table III).

Dietary EA alters adipose fatty acid profi le and gene expression
GC-MS analysis of the fatty acid profi le in epididymal adipose showed that the relative abundance of EA was similar in wild-type and D2 KO mice and comprised less than 7% of the total fatty acids within adipose tissue ( Fig. 3 ). The relative abundance of C18 and C20 MUFA was lower in D2 defi cient mice and offset by an increase in C14 and C16 SFAs. The reduction in 20:1 and 18:1 in D2 KO mice is consistent with a reduction in ␤ -oxidation and shortening of dietary EA before storage in adipose tissue triglycerides (TGs). However, given the increase in adipose tissue mass, the total amount of all fatty acids within adipose is greater in D2 KO than in wild-type mice. On the whole, there was an increase in SFAs and a decrease in MUFAs in adipose tissue, whereas the levels of PUFA were unaffected ( Fig. 3 , inset). These data are consistent with previous reports indicating that EA is not a preferred substrate for TG synthesis in adipose tissue ( 27 ).
To explore potential mechanisms for increased adiposity, we analyzed mRNA levels of key genes involved in the regulation of adipose function and lipid and carbohydrate metabolism. Insulin-induced gene-1 (INSIG1), but not INSIG2, was down-regulated in D2 KO mice. Although the expression of SREBP1 was not altered, several SREBP1 target expansion of adipose mass in the absence of an increase in lean mass resulted in a substantial increase in adiposity in D2 KO mice ( Fig. 1C ). Histological analysis of epididymal fat indicates that the expansion of adipose tissue was associated with the enlargement of adipocytes ( Fig. 1D ). Fasting glucose levels were increased in D2 KO mice compared with wild-type controls ( Fig. 1E ). A glucose tolerance test conducted after 7 weeks on diet revealed a decrease in glucose disposal consistent with the development of insulin resistance ( Fig. 1F, G ). Whereas fasting levels of triglycerides were unaffected, serum cholesterol was signifi cantly elevated ( Table 1 ).
The expansion of adipose and the loss of glycemic control are generally associated with adipose tissue infl ammation. Consistent with the expansion of adipose tissue, leptin levels were elevated in D2 KO mice ( Table 1 ). However, we failed to detect any differences in other adipocytokines typically associated with infl ammation. We also stained adipose tissue sections for CD68 and did not observe an accumulation of macrophages or the presence of crownlike structures (data not shown). These data suggest the EA is an inhibitor of fatty acid elongation, an essential step in PUFA synthesis ( 28 ). The deletion of fatty acid elongase 5 (ELOVL5) reduces hepatic levels of DHA and AA and promotes steatosis in mice ( 22 ). Therefore, we measured elongation of several ELOVL5 substrates to determine if EA reduced the activity of this enzyme in D2 KO mice ( Fig. 5 ). Rather than causing a reduction in ELOVL5 activity, the elongation of intermediates of PUFA synthesis was signifi cantly elevated. We next measured the abundance of mRNAs for lipogenic genes ( Fig. 5B ). Although mRNA for SREBP1 was elevated, we did not observe signifi cant increases in key lipogenic genes, including ACC1 and FAS. There was a modest but signifi cant increase in SCD1 mRNA, but there was no increase in enzyme activity ( Fig. 5C ). These data indicate that hepatic steatosis is unrelated to the relative reduction in PUFAs and disinhibition of hepatic SREBP1 processing and activity. Unlike adipose, there was no increase in hepatic expression of D1 in D2 KO liver.
The accumulation of diacylglycerol and ceramide is associated with the development of steatosis and the progression of NAFLD ( 29 ). Total diacylglycerols and ceramides appear to be elevated in hepatic lipid extracts from D2 KO mice. However, the levels were quite variable, and these differences did not reach statistical signifi cance (Supplementary Figure V).
In contrast to adipose, where tissue mass is comprised almost exclusively of TGs, liver fatty acid profi les differ substantially from fatty acid content. When reported on a per gram tissue basis, the fatty acids that accumulate in liver are limited to EA and its chain-shortened products of peroxisomal metabolism (20:1, 18:1, and 16:1) ( Fig. 6 ).
genes were down-regulated in adipose tissue of D2 KO mice fed EA, including acetyl-CoA carboxylase (ACC1), FAS, and stearoyl CoA-desaturase (SCD1) ( Fig. 3B ). There was also a 2.5-fold increase in the expression of D1 in adipose of D2 KO mice fed EA, but D1 expression in adipose is near the limit of detection in adipocytes ( 15 ). Transcription factors that regulate cholesterol, fatty acid, and glucose homeostasis were down-regulated in D2 KO mice (LXR, ChREBP) (Supplementary Figure III). These data are largely consistent with the development of insulin resistance in adipose tissue associated with adipocyte hypertrophy, suggesting these changes are refl ective of the development of the obesity and insulin resistance rather than mechanisms for EAdependent phenotypes in D2 KO mice.

EA promotes hepatic steatosis in D2 defi cient mice
The development of obesity and adipose insulin resistance is associated with ectopic lipid accumulation. H&E staining of liver sections revealed vacuolar structures that stained positively with Oil-Red-O in D2 KO mice ( Fig. 4 ). Hepatic triglycerides and cholesterol were elevated in the D2 KO mice compared with controls. Whereas the alteration in the fatty acid profi le in adipose tissue was limited to a few specifi c fatty acids, the hepatic fatty acid profi le revealed signifi cant differences for many fatty acid species. There is dramatic shift from wild-type to D2 KO mice toward more MUFAs (C18-C22) and fewer PUFAs (C18-C22) ( Fig. 4E ). The relative levels of DHA and AA, key inhibitors of hepatic lipogenesis that suppress SREBP1 processing and transcriptional activity, are signifi cantly reduced. Fig. 2. D2 defi ciency does not infl uence the development of diet-induced obesity. Wild-type and D2 KO mice were challenged with LF and HF diets for 12 weeks beginning at 8 weeks of age. Body weight (A), lean mass (B), fat mass (C), adiposity (D), fasting glucose (E), and glucose tolerance (F) were determined at the termination of the study. Data are mean ± SD (n = 10-13). Two-way ANOVAs were used to determine differences due to genotype and diet. Bonferroni multiple comparisons between genotypes were conducted within a diet to determine effects of genotype within each time point. * P < 0.01. resulted in a robust disruption of lipid metabolism leading to an expansion of adipose tissue, the loss of glycemic control, and steatosis. This suggests that mice have suffi cient D2 present at all times or can rapidly up-regulate D2 to compensate for the presence of EA in the diet.
D2 is an SREBP1 target gene and is up-regulated during adipogenesis, suggesting that D2 dependent peroxisomal metabolism may play an essential role in lipogenesis and TG storage. Indeed, an alternative pathway involving peroxisomal dihydroxyacetone phosphate acyltransferase accounts for approximately 40% of TG synthesis in 3T3-L1 cells ( 31 ). Although D2 may supply this pathway with fatty acyl-CoAs, the absence of D2 did not reduce lipid accumulation in mouse embryonic fi broblasts differentiated into adipocytes ( 15 ). Likewise, we observed no reduction in adipose mass in mouse models of diet-induced obesity regardless of whether adipose expansion was driven by excess dietary fat or by hyperphagia and lipogenesis. Consequently, the functional signifi cance of SREBP1 regulated D2 expression remains unclear.
Although D2 opposes the accumulation of EA in adipose tissue and plasma after an acute dietary challenge ( 15 ), EA levels in wild-type and D2 KO mice did not differ after 8 weeks of EA feeding. Despite the fact that EA constituted 50% of the dietary fatty acid species, it accounts for only 5% of the fatty acids in adipose and did not differ between genotypes. This is in keeping with previous reports indicating There is also an increase in 14-18 carbon SFA. Consequently, there is an accumulation of SFA and MUFA in the absence of any change in PUFA ( Fig. 6 , inset). Additional analysis of gene expression in liver revealed reductions in UCP2 and malic enzyme but no alterations in the expression of transcriptional regulators of sterol and carbohydrate metabolism or fatty acid modifying enzymes (Supplementary Figure IV). Collectively, the data suggest that the increase in hepatic MUFA is a consequence of peroxisomal shortening of EA. The increase in SFA does not appear to be a consequence of elevated lipogenesis and is more likely due to substrate competition for incorporation into secreted lipoproteins and mitochondrial oxidation.

DISCUSSION
The present study reveals a novel role for D2 in opposing disturbances in lipid metabolism associated with dietary EA. EA causes a transient lipidosis and dilated cardiomyopathy in several species and is associated with an elevation in hepatic triglycerides after long-term feeding in rats ( 30 ). However, mice are resistant to these effects. Although the absence of D2 sensitizes mice to EA feeding, the phenotype differs substantially from rats. We did not observe hypertriglyceridemia in the present study or in previous studies with an acute challenge ( 15 ). We also saw no evidence of cardiomyopathy. Nonetheless, dietary EA Fig. 3. Dietary EA alters adipose fatty acid profi le and gene expression in D2 KO mice. A: Fatty acid profi les in total lipid extracts from epididymal fat pads were determined by GC-MS. Relative signal intensities were normalized to an internal standard (C17:0) and expressed as percentage of total fatty acids. Inset: Sum percentage of fatty acids by degree of saturation. B: The relative abundance of mRNAs for lipogenic regulators and enzymes was measured by qRT-PCR. Data are shown as mean ± SD (n = 7). Comparisons between genotypes for individual fatty acids and mRNAs were made using an unpaired t -test. * P < 0.05; ** P < 0.01. the accumulation of EA, it fails to prevent its disruptive effects on lipid metabolism.
The synthesis of 20-and 22-carbon PUFAs requires the activity of fatty acid elongases. In the absence of ELOVL5, mice develop steatosis due to PUFA defi ciency ( 22 ). EA inhibits fatty acid elongation and is used clinically to reduce VLCFAs in ALD patients, suggesting this could be a mechanism for EA-mediated disruptions in lipid metabolism in the absence of D2. Indeed, there is a relative reduction in hepatic PUFAs, including DHA and AA in D2 KO mice ( Fig. 4 ). However, fatty acid elongase activity was elevated in microsomes isolated from D2 KO mice. This increase in activity also appears to be the case in vivo based on the increase in 20:3 3, the product of 18:3 3 elongation. Consequently, the relative decline in PUFAs in the fatty acid profi le may be misleading. There was no reduction in the absolute levels of PUFAs in the liver of D2 KO mice ( Fig. 6 ), but whether the absolute or relative levels of PUFAs is critical to feedback inhibition of lipogenesis is not known. The expression of lipogenic genes was largely unaffected, and there was no increase in the activity of SCD1, an essential enzyme for TG storage. These data suggest that hepatic PUFA content rather than relative levels are important for regulation of SREBP1 activity in vivo. that EA is a poor substrate for esterifi cation and storage as TG and suggests that EA incorporation into adipose TGs reaches its maximum irrespective of the presence or absence of D2 ( 27 ). In the presence of D2, EA is likely shortened to 16-and 18-carbon MUFAs in peroxisomes. Indeed, these are elevated in the adipose tissue of wild-type compared with D2 KO mice. If not stored as EA in the adipose TG pool and not metabolized to shorter chain MUFAs, dietary EA must be metabolized outside of adipose tissue.
As in adipose tissue, the levels of EA are quite low in the liver and constitute less than 2% of the fatty acid profi le. However, there is a signifi cant increase in EA in D2 KO mice, as well as 16-, 18-, and 20-carbon MUFAs. The relatively small amount of hepatic EA and the accumulation of its chain-shortened metabolites suggest that EA is metabolized by peroxisomes in the liver. In contrast to adipose tissue, the liver expresses relatively low levels of D2 and high levels of D1 ( 9,15 ). D1 is the closest family member to D2, and, although the transporters have distinct substrate spectra, there is signifi cant overlap, including EA. D1 expression was increased in adipose of D2 KO mice fed EA but was unchanged in liver. This transporter may therefore compensate for the absence of D2 with respect to EA metabolism. Although the compensatory pathway prevents Total triglycerides and cholesterol in hepatic tissue were determined using colorimetric assays. E: Fatty acid profi les of total hepatic lipids were measured by GC-MS. Fatty acid levels were normalized to an internal standard (C17) and expressed as percentage of total fatty acids. Inset: Sum percentage of fatty acids by degree of saturation. Data are shown as mean ± SD (n = 7). Comparisons between genotypes were made using an unpaired t -test. * P < 0.05; ** P < 0.01. tissue. One potential explanation is that the feeding period for the present experiment was only 8 weeks. Perhaps with increased time, these hallmarks of obesity and insulin resistance would have presented. Regardless, the data indicate that adipose dysfunction and infl ammation was not causative in the EA-dependent phenotype in D2 KO mice.
The role of peroxisomal metabolism in the development of obesity phenotypes is poorly understood but may play a signifi cant role. A recent study indicates that D2 expression and hepatic peroxisomal metabolism are increased during the development of obesity and insulin resistance in HF fed mice ( 26 ). The products of peroxisomal metabolism include acylcarnitines that are capable of bypassing carnitine palmitoyltransferase 1 for entry into mitochondria irrespective of cellular levels of malonyl-CoA. HF feeding and excess caloric intake result in mitochondrial overload, incomplete ␤ -oxidation, and the accumulation of medium-chain acylcarnitines in the plasma of rodents ( 32 ). Therefore, peroxisomal ␤ -oxidation could potentially contribute to mitochondrial overload and promote obesity phenotypes after HF feeding. We measured acylcarnitines in liver of wild-type and D2 KO mice but did not observe differences in individual acylcarnitines or the sums of short-, medium-, or long-chain species (Supplementary Figure VI).
Conversely, there is an increase in peroxisomal lipid metabolism in mice resistant (A/J) to obesity and steatosis after HF feeding, suggesting that peroxisomes may play a protective role with respect to the development of obesity phenotypes ( 33 ). We have confi rmed an increase in hepatic expression of D2 after HF feeding and have also observed an increase in ob/ob mice (Supplementary Figure I). However, the absence of D2 played neither a protective nor a deleterious role in the development of obesity phenotypes in mouse models of obesity and insulin resistance. Although there may be a role for peroxisomes in the development of obesity and steatosis, it is independent of the presence or absence of D2 unless mice are challenged with a diet containing EA and perhaps other fatty acids that require peroxisomal metabolism before storage in TG pools.
A surprising observation is that expansion of the adipose tissue and loss of glycemic control was not associated with the accumulation of infl ammatory cytokines in plasma. Similarly, there was no evidence for macrophage infi ltration and the formation of crown-like structures in adipose Fig. 5. Dietary EA increases elongase activity but has no effect on lipogenic genes or SDC1 activity. A: Elongation of fatty acids in liver microsomes was determined by incorporation of malonyl CoA into the indicated fatty acid substrates. B: mRNA levels of liver lipogenic genes were measured by qRT-PCR. C: Activity of stearoyl-CoA desaturase was determined by conversion of labeled stearate to oleate in isolated liver microsomes. Data are shown as mean ± SD (n = 4). Comparisons between genotypes were made using an unpaired t -test. * P < 0.05; ** P < 0.01; *** P < 0.001. Fig. 6. Hepatic fatty acid content after EA feeding in wild-type and D2 KO mice. Fatty acid content (mg/g tissue) in total hepatic lipids was measured by GC-MS. Inset: Sum of fatty acid content by degree of saturation. Data are shown as mean ± SD (n = 7). Comparisons between genotypes were made using an unpaired t -test. * P < 0.05; ** P < 0.01.