Aberrant fatty acid metabolism in skeletal muscle contributes to insulin resistance in zinc transporter 7 (znt7)-knockout mice

ZnT7 (Slc30a7) is a widely expressed zinc transporter involved in sequestration of zinc into the Golgi apparatus and vesicular compartments. znt7-knockout (KO) mice are mildly zinc-deficient and lean. Despite their lean phenotype, adult male znt7-KO mice are prone to insulin resistance. We hypothesized that fat partitioning from adipose to nonadipose tissues causes insulin resistance in znt7-KO mice. Here, we used biological and biochemical methods, including fatty acid and oxylipin profiling, EM, immunohistochemistry, quantitative RT-PCR, and Western blot analysis, to identify the underlying mechanism of insulin resistance in znt7-KO mice. We found that insulin resistance in this model was primarily associated with increased intracellular fatty acid levels in the skeletal muscle, which promoted intracellular lipid accumulation and production of bioactive lipid mediators, such as 12,13-dihydroxyoctadecanoic acid (12,13-DiHOME) and 12-hydroxyeicosatetraenoic acid (12-HETE). The expression of fatty acid–binding protein 3 (Fabp3) was dramatically up-regulated in the znt7-KO muscle cells accompanied by increased expression of Cd36, Slc27a1, and Slc27a4, the three major fatty acid transporters in the skeletal muscle. We also demonstrated that znt7-KO muscle cells had increased fatty acid oxidative capacity, indicated by enlarged mitochondria and increased mRNA or protein expression of key enzymes involved in the fatty acid mitochondrial shuttle and β-oxidation. We conclude that increased fatty acid uptake in the znt7-KO skeletal muscle is a key factor that contributes to the excessive intracellular lipid deposit and elevated production of bioactive lipid mediators. These mediators may play pivotal roles in oxidative stress and inflammation, leading to insulin resistance.

Cellular zinc homeostasis is tightly regulated through two families of zinc transporters (Slc30a and Slc39a) 3 (1,2) and other small proteins, such as metallothioneins (3). In general, Slc30a family members (Slc30a1-10) are involved in sequestration of zinc into intracellular compartments or export of zinc out of the cell when cellular zinc is in excess. In contrast, Slc39a family members (Slc39a1-14) play roles in either zinc uptake from the extracellular space or zinc release from the intracellular stores into the cytoplasm when cellular zinc concentration is low (1,2). The expression of zinc transporters can be ubiquitous or tissue-or cell-specific. For example, Slc30a1 is expressed ubiquitously, whereas Slc30a8 is abundantly and almost exclusively expressed in the pancreatic islet. Between the two extremes, expression of many other zinc transporters can be widespread at various levels. For example, ZnT7 is ubiquitously expressed among tissues/cell types; however, the expression level differs greatly. As a result, the impact of a null mutation in an Slc30a or Slc39a gene on zinc metabolism varies due to differences in expression levels in the body.
We reported previously that ZnT7 was expressed in mouse skeletal muscle (4), adipose tissue (5), ␤-cells in the islet of the pancreas (6), epithelium of the small intestine (7), and prostate (8). Consequently, the znt7-knockout (KO) mouse displays phenotypes that are closely associated with its protein expression level in a given tissue or cell type (9). A reduced capability of maintaining cellular zinc level at a normal range contributes to zinc deficiency in znt7-KO mice, which results from decreased dietary zinc absorption in the gut and consequently reduced zinc accumulation in organs, such as the bone, liver, and kidney (10). We have also demonstrated that male adult znt7-KO mice are susceptible to diet-induced insulin resistance. We showed that the abnormality in glucose metabolism was attributed to the disruption of insulin metabolism in pancreatic ␤-cells in znt7-KO mice (4,6) and reduced insulin sensitivity in skeletal muscle, a tissue responsible for 70 -80% of glucose disposal following a carbohydrate load (4). The underlying molecular mechanism by which the znt7-null mutation leads to peripheral insulin resistance in znt7-knockout mice is still unknown. We reported previously that the znt7-null mutation decreased lipid accumulation in mouse adipose tissue (5). Hence, we hypothesized that insulin resistance in the muscle of znt7-KO mice might be the result of fat partitioning from adipose to nonadipose tissues. According to the ZnT7 protein expression pattern in the skeletal muscle (abundant) and liver (undetectable in hepatocytes), we further hypothesized that the lipid partitioning in the znt7-KO mouse mainly occurred in the skeletal muscle but not the liver.
In this study, we used fatty acid and oxylipin profiling, EM, immunohistochemistry, quantitative RT-PCR, and Western blot analysis to probe the underlying mechanism of insulin resistance in the skeletal muscle of znt7-KO mice. Here, we provide evidence that insulin resistance in znt7-KO mice is associated with increased uptake and binding of free fatty acids in the muscle cell. Bioactive lipid mediators were elevated in the skeletal muscle of znt7-KO mice and showed the potential to drive insulin resistance. In addition, we demonstrated that the fatty acid oxidative capacity of znt7-KO muscle cells was increased, evident by the enlarged mitochondria, increased mRNA and protein expression of key enzymes involved in fatty acid mitochondrial shuttle and ␤-oxidation, and reactive oxygen species (ROS) production.

Results
We previously demonstrated that adult male znt7-KO mice were more sensitive to high fat diet-induced abnormalities in glucose and insulin metabolism than the WT control mice (4). We have also shown that znt7-KO mice gain less body weight than the controls mainly due to reduced body fat accumulation (10,11). Hence, we hypothesized that znt7-KO mice had dysregulation of lipid metabolism and that this genetically predisposed factor might render the znt7-KO mouse sensitive to high fat diet-induced glucose intolerance. We therefore studied a cohort of 18.5-week-old male znt7-KO and age-matched male WT littermates to determine the underlying mechanism of insulin resistance in metabolically relevant tissues, such as the skeletal muscle and liver. znt7-KO mice used in this study were congenic mice with a C57BL/6J (B6) genetic background in which the znt7 allele (129P2) was backcrossed to the B6 strain for at least 12 generations. The KO and WT mice were littermates; therefore, both the znt7-KO and WT mice were genetically identical except for the znt7 allele and its flanking regions on chromosome 3.
The study design is laid out in Fig. 1A. Both genotypes of mice were fed a semipurified diet containing 30 mg of zinc/kg of diet starting at 5 weeks of age. Difference in body weight was seen around 9 -10 weeks of age and became significant at 13 weeks between the two genotype groups (Fig. 1B). We observed that mice in both znt7-KO and WT groups gained limited weight 1 week after they were singly housed at 11 weeks of age. But the body weight gain resumed from 13 weeks until euthanization at 18.5 weeks (Fig. 1B). Table 1 lists the measurements of body weights, fat pad weights, and organ weights at necrosis. As we expected, when compared with the WT controls, znt7-KO mice accumulated less body fat as measured by individual or total fat pad weights (epididymal, subcutaneous, and retroperitoneal fat pads). When adjusting the body fat (total measured fat pad weights) to the body weight, znt7-KO mice appeared leaner than the WT controls at 18.5 weeks (Table 1), consistent with our previous findings in znt7-KO mice with a mixed B6 and 129P2 genetic background (10,11). The weights of other organs, including the liver, heart, and brain, were not apparently different between the two genotypes ( Table 1). Food intake was decreased in the KO group (Fig. 1C). However, when the food intake was adjusted for the body weight gain, the food intake per gram of body weight gain was comparable between the two genotype groups (Fig. 1D), suggesting a similar growth efficiency.
Overnight fasting plasma triglycerides and cholesterol levels were slightly low in znt7-KO mice compared with the WT controls, whereas overnight fasting glucose and total fatty acid levels were similar between the two genotypes ( Table 2). In contrast, an elevated blood glucose level (17%, p Ͻ 0.01) was observed in znt7-KO mice compared with the controls after short-term fasting (6 h), suggesting abnormal glucose metabolism in znt7-KO mice. Nevertheless, the increased 6-h fasting glucose level in znt7-KO mice was not accompanied by an elevated blood insulin level ( Table 2). The effect of the znt7-KO on the whole-body glucose metabolism was further confirmed by glucose and insulin tolerance tests. As shown in Fig. 2A, znt7-KO mice had a higher blood glucose level than the controls at 120 min (p Ͻ 0.05) after a glucose load ( Fig. 2A). Plasma insulin levels in znt7-KO mice during glucose tolerance tests were not significantly changed compared with the controls (Fig.  2B). Furthermore, blood glucose levels in znt7-KO mice during insulin tolerance tests were higher than those of the controls, suggesting a resistance to the blood glucose-lowering effect of exogenously administered insulin in znt7-KO mice (Fig. 2C). This resistance was also reflected by the area under the curve glucose measurements of the insulin tolerance test (Fig. 2D). Taken together, our data suggest that znt7-KO mice had impaired glucose tolerance, likely due to insulin resistance.
The homeostatic control of blood glucose levels is largely accomplished by the skeletal muscle, liver, and adipose tissue in which the muscle takes up and utilizes about 70 -80% of daily absorbed glucose (12). We have shown that the activity of the insulin signaling pathway and glucose uptake are compromised by znt7 knockdown in adipocytes, leading to a reduction in fat accumulation (5). Thus, we hypothesized that fat partitioning from adipose to nonadipose tissues might be the underlying mechanism of insulin resistance observed in znt7-KO mice. Hence, we performed metabolomic profiling of fatty acids in the skeletal muscle and liver from znt7-KO and the control mice.
As shown in Fig. 3A, the total fatty acid content in the skeletal muscle of znt7-KO mice was 35% higher than that of the control mice (p Ͻ 0.05). The sum of concentrations of saturated (SFA), Muscle insulin resistance in znt7-KO mice monounsaturated (MUFA), and polyunsaturated fatty acids (PUFA) was elevated in the skeletal muscle of znt7-KO mice compared with that of the controls. The ratio of -3 PUFA to -6 PUFA was decreased (KO, 0.56 versus WT, 0.66), suggesting an increased risk for insulin resistance in the skeletal muscle of znt7-KO mice. In addition, fatty acid composition analysis indicated that elevated fatty acids were mostly concentrated in the long-chain fatty acid group (Fig. 3C). The relative abundance of C18:1(n-9) to C18:0 was increased in the muscle of znt7-KO mice (p ϭ 0.038), consistent with elevated stearoyl-CoA desaturase 1 activity, which has been reported to contribute to abnormal lipid partitioning and insulin resistance (13).
Arachidonic acid (C20:4(n-6)), a key inflammatory intermediate in skeletal muscle (14), was also increased by ϳ25% in the znt7-KO muscle compared with the controls (p ϭ 0.029) (Fig.  3C). Meanwhile, the anti-inflammatory docosahexaenoic acid (C22:6(n-3)) (15) was increased by ϳ30% in the znt7-KO muscle compared with the controls (p ϭ 0.043). Regardless, the balance of the pro-and anti-inflammatory long-chain fatty acids appeared to be unchanged in the znt7-KO muscle (p ϭ 0.5). In contrast to the changes observed in the znt7-KO muscle, little to no change in the total fatty acid content or composition was observed in the liver of znt7-KO compared with the control mice (Fig. 3, B and D). This discrepancy may be associ-  A, a schematic diagram of the study. Mice were weaned at 3 weeks old. An AIN96-based rodent diet containing 30 mg of zinc/kg of diet was introduced to both znt7-KO and control mice at 5 weeks old. All mice were euthanized at 18.5 weeks old. B, growth curves during the food intake measuring period. Body weights were monitored for 3 weeks after the special diet was introduced until before the insulin resistance test was performed. C, total food (grams) consumed by znt7-KO and control mice during the 3-week data collection period. znt7-KO and control mice were housed individually at 11 weeks old. Food intake was measured 2 weeks after mice were singly housed. D, food intake per gram of body weight gain. BW, body weight; wk, week; IPGTT, intraperitoneal glucose tolerance test; IPITT, intraperitoneal insulin tolerance test. Data are means, and error bars represent S.D. *, p Ͻ 0.05.

Muscle insulin resistance in znt7-KO mice
ated with the difference in ZnT7 protein expression levels in the two tissues (Fig. 3, C and D, insets). We showed that ZnT7 was readily detectable in the sarcoplasm of the myofibril, whereas ZnT7 was absent from the hepatocyte. Only the sinusoid endothelial cell in the liver stained positive for ZnT7 (Fig. 3, C and D, insets).
To examine whether increased fatty acid content leads to an increased lipid accumulation in the skeletal muscle of znt7-KO mice, we stained the femoral skeletal muscle prepared from znt7-KO and WT mice with Oil Red O, a dye that detects neutral lipids. As shown in Fig. 4A, lipid accumulation was noticeably increased in the myofibrils of znt7-KO mice compared with the controls. This observation was further confirmed by triglyceride quantification in the skeletal muscle of znt7-KO and WT mice. We showed that znt7-KO skeletal muscle contained ϳ50% higher triglycerides than the control (Fig. 4B). Together, we conclude that znt7-KO impairs lipid metabolism in the skeletal muscle.
Next, we examined whether increased fatty acid concentrations in the skeletal muscle of znt7-KO mice would result in an increased capacity for fatty acid utilization. First, we performed EM of the femoral skeletal muscle isolated from znt7-KO and WT mice. As shown in Fig. 5A, in the WT, the mitochondria (dark and ovoid structures) were localized at the level of the I band and centered around the z-line between the myofibrils (16). In contrast, in the znt7-KO muscle, the sarcoplasmic space of the myofibril was inflated with nonuniformly distributed mitochondria. In addition, the size of the mitochondria in the znt7-KO muscle was expanded, covering two or more z-lines along the myofibril (Fig. 5A). High magnification images revealed a high level of lipid accumulation between the mito-chondria in the znt7-KO muscle (Fig. 5B). Nevertheless, we did not observe a noticeable internal structure change in the mitochondria of the znt7-KO muscle (Fig. 5B). In addition, mitochondrial DNA contents in the muscle cells between znt7-KO and the WT controls were not different (data not shown).
Dysregulation of several factors, such as fatty acid transport, intracellular fatty acid binding, and/or fatty acid oxidation, could increase fatty acid concentrations in the znt7-KO muscle. Thus, we decided to investigate the mRNA expression of genes involved in the fatty acid uptake and oxidation in the skeletal muscle, including Cd36 (fatty acid translocase), Slc27a1 (longchain fatty acid transport protein 1), Slc27a4, Fabp3 (fatty acid-binding protein 3), Acsl1 (long-chain fatty acid CoA ligase 1), Acacb (acetyl-CoA carboxylase 2), Cpt1b (carnitine palmitoyltransferase 1b), Acadl (long-chain acyl-CoA dehydrogenase), and Hadhb (3-ketoacyl-CoA thiolase, ␤ subunit), in differentiated primary myoblasts isolated from skeletal muscles of znt7-KO and WT mice (4). Although primary isolated skeletal muscle cells might not reflect skeletal muscle in vivo, we used these cells to achieve improved results for quantitative PCR, Western blot assay, fatty acid uptake, and ROS measurements. As shown in Fig. 6A, among the genes examined, the mRNA expression of Fabp3, a gene associated with intracellular fatty acid binding, was significantly higher (25-fold) in the znt7-KO primary myotubes than in the WT controls. Up-regulation of Fabp3 in znt7-KO myotubes was also confirmed at the protein level (Fig. 6B). Four genes that are implicated in fatty acid uptake, Cd36, Slc27a1, Slc27a4, and Acsl1, were moderately up-regulated (2.5-5-fold higher in the znt7-KO muscle than in the controls). We further confirmed that the up-regulation of expression of these fatty acid uptake genes in znt7-KO myotubes led to an increase in fatty acid uptake by ϳ27% compared with the controls (Fig. 6C). Interestingly, the mRNA expression of Acacb, an enzyme that converts acetyl-CoA to malonyl-CoA, was also up-regulated in the znt7-KO primary myotubes. Malonyl-CoA is a potent inhibitor of mitochondrial fatty acid uptake through allosteric inhibition of Cpt1 enzymatic activity, a rate-limiting step in fatty acid uptake and oxidation by the mitochondria. Increased mRNA expression of Acacb suggests a negative impact of the znt7-KO on mitochondrial fatty acid oxidation in myotubes even though Acadl and Hadhb were upregulated in the znt7-KO primary myotubes compared with the controls (Fig. 7, A and B). Taken together, our results suggest that the znt7-null mutation may increase fatty acid uptake and expand the Fabp3-bound fatty acid pool in the skeletal muscle. znt7-KO may also impair fatty acid oxidation in the skeletal muscle. As a result, lipids accumulate in the skeletal muscle of znt7-KO mice.
To determine whether high intracellular fatty acid levels in znt7-KO muscle cells would result in oxidative stress, we examined mRNA expression of genes associated with oxidative stress and inflammation, such as Alox5 (arachidonate 5-lipoxygenase), Alox12 (arachidonate 12-lipoxygenase), Alox15 (arachidonate 15-lipoxygenase), and Ephx1-4 (epoxide hydrolases 1-4) in primary znt7-KO and WT myotubes. We also examined oxylipin contents in skeletal muscles isolated from znt7-KO and WT mice. As shown in Fig. 7C, quantitative RT-PCR analyses showed that znt7-KO myotubes had a 25-fold

Muscle insulin resistance in znt7-KO mice
increase in Alox12 mRNA expression, whereas no significant difference in the expression of Alox5 and Alox15 was seen (data not shown). Moreover, in znt7-KO primary myotubes, the mRNA expression of all Ephx genes except Ephx3 was up-regulated. The oxidative status of znt7-KO in muscle cells was further illustrated by ROS levels in living cells. As shown in Fig.  7D, a higher ROS level (2.2-fold, p Ͻ 0.01) was detected in znt7-KO myotubes than in the WT control cells.
Oxylipins are bioactive lipids, and elevations in these compounds are often associated with oxidative stress and inflammation (17). We showed that znt7-KO skeletal muscle had increased levels of linoleic acid, ␣-linolenic acid, and arachidonic acid (Fig. 3C). The mRNA expression of genes that encode enzymes involved in metabolizing these fatty acids to oxylipins, including lipoxygenases and epoxide hydrolases, were also up-regulated (Fig. 7C). We next investigated the nonesterified oxylipin pool of femoral skeletal muscle isolated from znt7-KO and WT mice. We found that 14 oxylipins among metabolites examined, including both n-6 -and n-3-derived compounds, were significantly higher in the znt7-KO skeletal muscle than those of the control (Fig. 8A). Although their relative potencies can be argued, 10 of these oxylipins have been associated with oxidative stress or inflammation (Table 3). In addition, four metabolites were found to be slightly lower in concentration in the muscle of znt7-KO mice than that in the control (Fig. 8A).
To better understand the relationship between increased oxylipins and insulin resistance, we investigated metabolites from two distinct biosynthesis cascades with plausible links to glucose regulation, specifically, modulation of insulin signaling and glucose uptake by 12,13-dihydroxyoctadecanoic acid (12,13-DiHOME; a product of the soluble epoxide hydrolase) and 12-hydroxyeicosatetraenoic acid (12-HETE; a product of 12-lipoxygenase) in rat L6 myotubes. A decrease in the activity of the insulin signaling pathway is a signature of insulin resistance in skeletal muscle cells. Thus, we examined phosphorylation levels of Akts, a key step in modulation of insulin signaling and translocation of Glut4 from an intracellular storage pool to the cytoplasmic membrane for glucose uptake. L6 myotubes were treated with either 12,13-DiHOME (2 M) or 12-HETE (2 M) for 18 h. As shown in Fig. 8, B and C, the insulin-dependent phosphorylation of Akts was decreased by 40 and 35% in L6 myotubes treated with 12,13-DiHOME and 12-HETE, respectively, compared

Muscle insulin resistance in znt7-KO mice
with vehicle-treated cells. Accordingly, insulin-stimulated glucose uptake was down in the L6 myotubes treated with 12,13-DiHOME and 12-HETE (Fig. 8D). Interestingly, 12,13-DiHOME treatment also lowered the basal glucose uptake in L6 myotubes (Fig. 8D). It is known that the proinflammatory cytokine Ccl2 (also called Mcp1) is a prominent inducer of insulin resistance in skeletal muscle through inhibition of insulin-stimulated Akt phosphorylation and subsequent glucose uptake at a dose above 200 pg/ml in cells (18). We next examined whether L6 myotubes treated with either 12,13-DiHOME or 12-HETE would increase Ccl2 secretion into the culture medium. As shown in Fig. 8E, treatment of L6 myotubes with either 12,13-DiHOME or 12-HETE was able to induce Ccl2 secretion to a level much higher than 200 pg/ml, suggesting the presence of a negative impact of 12,13-DiHOME or 12-HETE on glucose uptake in muscle cells via Ccl2. Taken together, we conclude that ZnT7 and thus intracellular zinc play an important role in lipid metabolism of the skeletal muscle. The znt7 knockdown increased intracellular fatty acid levels in the skeletal muscle; led to up-regulation of Alox12, Ephx1, Ephx2, and Ephx4 gene expression; and promoted downstream fatty acid metabolite production, modulating insulin signaling in the skeletal muscle of znt7-KO mice (Fig. 9).

Discussion
It is known that reduced storage of fat in adipose tissue results in greater partitioning of dietary lipids to nonadipose tissues, such as skeletal muscle and liver, causing fat deposition in these tissues (19). We previously demonstrated that znt7-KO disrupted lipid accumulation in adipose tissue (5). Thus, we hypothesized that lipid partitioning from adipose tissue to nonadipose tissues triggered insulin resistance in peripheral tissues of znt7-KO mice (4). We investigated fatty acid profiles in the skeletal muscle, liver, and plasma isolated from male 18.5-week-old znt7-KO and WT littermates. We found significant abnormality in fatty acid metabolism in the skeletal muscle of znt7-KO mice, whereas little alteration was found in the liver (Fig. 3D) and plasma (data not shown) of znt7-KO mice. These abnormalities were characterized by a general increase in tissue lipids with a 2-fold increase in the stearoyl-CoA desaturase activity index (20). This suggests an adaptive shift in metabolism toward lipid storage (21). We showed that the abnormality in fatty acid metabolism in the skeletal muscle of znt7-KO mice was in accordance with the expression of ZnT7 protein, which was abundantly expressed in the skeletal muscle with little to no expression in hepatocytes of the liver (Fig. 3, C  and D). Thus, our data support not only that ZnT7 promotes fat storage in adipose tissue but that it is also involved in regulating fatty acid metabolism in the skeletal muscle.
In the heart and skeletal muscle, long-chain fatty acids are taken up by Cd36 and/or Slc27a1/Slc27a4 and subsequently bound to Fabp3 (22). The binding of fatty acids to Fabp3 provides a positive feedback loop in which more long-chain fatty acids are transported into the muscle fibrils (23). Fabp3 functions as an intracellular chaperone to transport fatty acids to appropriate intracellular compartments for ␤-oxidation, lipid synthesis, and lipid bioactive mediator production (Fig. 9). It has been demonstrated that Fabp3 is required for oxidation of exogenous fatty acids in the muscle, and the Fabp3 expression level in adipocytes was strongly correlated with obesity development (24). In the skeletal muscle, Fabp3 mRNA expression is induced in response to an increased dietary fat load in mice leading to lipid accumulation in the skeletal muscle (24 -26). Because Fabp3 mRNA and protein expression is markedly increased in the znt7-KO skeletal muscle, accompanied by increased expression of fatty acid transporters and long-chain fatty acid CoA ligase 1, we postulate that elevated concentrations of intracellular fatty acid; intracellular lipid deposit; and linoleic acid-, ␣-linolenic acid-, and arachidonic acid-derived bioactive lipid mediators are elicited by the Fabp3-dependent influx of long-chain fatty acids in the znt7-KO muscle fibril. Increased intracellular fatty acid concentrations also up-regulated catabolic/fasting-type metabolism in the znt7-KO muscle cells, evident by the extended and enlarged mitochondria and increased gene/protein expression of key enzymes involved in the long-chain fatty acid shuttle between the sarcoplasm and mitochondria and in mitochondrial ␤-oxidation. This reprogramming in the znt7-KO skeletal muscle is suspected to increase the rate of ␤-oxidation to reduce lipotoxicity, but we have not directly measured this effect.
In addition to the robust up-regulation of Fabp3 mRNA expression in the znt7-KO muscle cell (25-fold higher than the control), we detected a similar rigorous expression pattern for Alox12 transcripts in the znt7-KO muscle cell (25-fold increase compared with the controls). Alox12 is a nonheme iron-containing dioxygenase involved in converting polyunsaturated fatty acids to specific hydroperoxide analogs (e.g. arachidonic acid to 12S-hydroperoxyeicosatetraenoic acid), which leads to the downstream production of a spectrum of bioactive lipid mediators involved in inflammatory regulation (27). Although

Muscle insulin resistance in znt7-KO mice
further studies are warranted, our findings suggest that insulin resistance in the skeletal muscle of znt7-KO mice is due to the up-regulation of Fabp3 and subsequently increased levels of bioactive lipid mediators involved in oxidative stress and inflammation. Our results also provide an intriguing clue with respect to a potential molecular marker for intracellular dysregulation of zinc homeostasis. It appears that increased fatty acid concentrations in the znt7-KO skeletal muscle cell provoke a compensatory mechanism to effectively metabolize fatty acids and to avoid lipotoxicity in the muscle. We observed that the mitochondria were enlarged and expanded in the znt7-KO skeletal muscle cells, and mRNA or protein expression of key enzymes involved in mitochondrial transport of fatty acids and subsequent ␤-oxida-tion, including Acadl, Cpt1b, and Hadhb, were up-regulated. However, the increased capacity was outpaced by fatty acid uptake. Thus, the excessive fatty acids were deposited as triglycerides in the znt7-KO skeletal muscle, leading to high intracellular lipid accumulation. From our oxylipin profiling result, it is evident that znt7-KO muscle cells had elevated levels of mediators associated with either inflammation or oxidative stress (Table 3). Our results in mRNA expression of Ephx1, Ephx2, and Ephx4; ROS production in znt7-KO primary myotubes; and Ccl2 secretion from oxylipin-treated L6 myotubes support the notion that znt7-KO skeletal muscle was under

Muscle insulin resistance in znt7-KO mice
oxidative stress, which activates inflammatory pathways, leading to insulin resistance in znt7-KO mice. It is noteworthy that, in the znt7-KO skeletal muscle, up-regulation of Fabp3 mRNA expression was highly and positively correlated with the change in Alox12 mRNA expression but not Alox5 or Alox15 mRNA expression (data not shown). One may speculate that the promoters of Fabp3 and Alox12 genes share one or more regulatory elements for transcription factors that are sensitive to change in intracellular zinc concentrations in the skeletal muscle. By searching the database for transcriptional factor-binding sites in promoters provided by MAPPER2 (multigenome analysis of positions and patterns of elements of regulation), we found that the promoter regions of Fabp3 and Alox12 (2000 bp upstream of the transcriptional start sites) share many potential transcriptional factor-binding sites (28). Among them, paired box protein (Pax6)-and peroxisome proliferator-activated receptor-␣ (Ppar␣)-binding sequences are commonly present in the promoters of Fabp3 and Alox12 but not in those of Alox5 and Alox15. Pax6 is a transcriptional factor that plays key roles in the development of neural tissues. The null mutation in the Pax6 gene causes severe defects in eye development in mice. Pax6 also regulates glucagon gene expression in the pancreatic islet implicated in glucose metabolism. Thus, it is not surprising that Fabp3 and Alox12 are target genes for Pax6. Ppar␣ is a nuclear hormone receptor transcription factor that is involved in regulating fatty acid oxidation in the heart and skeletal muscle (29). Ppar␣ expression is positively correlated with gene expression of crucial enzymes

Muscle insulin resistance in znt7-KO mice
involved in fatty acid transport, fatty acid mitochondrial shuttle, and mitochondrial energy production (30). Further exploration of the effect of Pax6 and Ppar␣ on the transcription of Fabp3 and Alox12 is required.
In conclusion, the znt7-null mutation resulted in altered fatty acid metabolism in skeletal muscle. Increased fatty acid accu-mulation in the znt7-KO muscle cell triggered compensatory mechanisms to oxidize excessive fatty acids in the mitochondria, leading to an expansion of the mitochondria and induction of key enzymes required for mitochondrial ␤-oxidation. Excessive intracellular fatty acids in the znt7-KO muscle cell also increased production of ROS and PUFA-derived mediators of

Muscle insulin resistance in znt7-KO mice
oxidative stress and inflammation, leading to muscle insulin resistance in znt7-KO mice.

Animals and diets
The generation and characterization of znt7-KO mice were described previously (4,10). Both znt7-KO and WT littermates were generated from heterozygous mating. Male mice with desired genotypes were weaned at 3 weeks old and fed a semipurified diet containing 30 mg of zinc/kg of diet (Research Diets, New Brunswick, NJ) ad libitum from 5 to 18.5 weeks. At 11 weeks old, mice were singly housed, and food consumption was measured every other day from 13 to16 weeks. At the end of the study, mice were fasted (16 -18 h) and euthanized. Blood and tissues, including skeletal muscle and liver, were collected for subsequent experiments. All mice were housed in a temperature-controlled room at 22-24°C with a 12-h light:dark cycle. Breeding mice were fed a standard laboratory chow diet (Laboratory Rodent Diet 5001, LabDiet, Brentwood, MO) and dou-ble-distilled water ad libitum. All animal experiments were conducted in accordance with National Institutes of Health Guidelines for the Care and Use of Experimental Animals and were approved by the Institutional Animal Care and Use Committee of the University of California, Davis.

Genotyping
Genotyping was performed as described previously using tail clips (10).

Muscle insulin resistance in znt7-KO mice Intraperitoneal insulin tolerance test
Before the test, mice were fasted for 4 h and weighed. Insulin (5.5 units of Humulin R (U-100)/kg of body weight; Lilly) was given intraperitoneally. Blood glucose was determined at 0, 15, 30, 60, and 120 min after insulin injection using a One-Touch UltraMini meter.

Oil Red O staining of skeletal muscle
Femoral skeletal muscle was dissected and rinsed in ice-cold 1ϫ PBS, blotted dry, and placed in ice-cold 4% paraformalde-hyde overnight at 4°C. Tissue was washed with 10 ml of 1ϫ PBS buffer (pH 7.4) for 2, 10, and 30 min; blotted dry; and placed in ice-cold 30% sucrose overnight at 4°C. Muscle was then embedded in optimal cutting temperature compound (OCT) in a liquid nitrogen-chilled isopentane bath. Muscle was then sectioned in 8-m thicknesses and allowed to dry overnight at room temperature. Muscle was subsequently stained with Oil Red O and counterstained with Harris modified hematoxylin and Light Green.

Immunohistochemistry
Femoral skeletal muscle and liver were dissected, fixed, paraffin-embedded, and sectioned as described previously (8). Tissues were stained with an antibody against mouse ZnT7 (33), and sections were counterstained with hematoxylin.

Electron microscopy
Muscle samples were fixed using Karnovsky's fixative in 0.1 M sodium phosphate buffer (Sorenson's) (34). A 1-mm 3 piece was cut and washed with the same buffer. Dehydration was accomplished using ascending concentrations of acetone, and then the piece was infiltrated and embedded in an epoxy resin mixture. Ultrathin sections were cut with a diamond knife (Diatome, Hatfield, PA) and picked up on copper grids. The sections were stained with uranyl acetate and lead citrate before viewing on a Phillips CM120 Biotwin (Hillsboro, OR). Digital images were taken with a Gatan BioScan model digital camera (Pleasanton, CA).

Isolation and culture of primary myoblasts and myotubes
Myoblast isolation from skeletal muscle of znt7-KO mice was conducted as described previously (4). Briefly, myoblasts were maintained in a growth medium consisting of 40% Ham's F-10, 40% DMEM, 20% FBS (bovine fetal serum), 2.5 ng/ml basic fibroblast growth factor (Gemini Bio-Products, West Sacramento, CA), 100 units/ml penicillin G, and 100 g/ml streptomycin (Thermo Fisher Scientific, Carlsbad, CA). Myoblasts were differentiated into myotubes in 4 -6 days in DMEM con- In the blood Albumin Fabp3 fatty acid Insulin resistance Figure 9. A schematic diagram of fatty acid metabolism and pathways leading to insulin resistance in the znt7-KO muscle cell. In the skeletal muscle cell, long-chain fatty acids are taken up by Cd36, Slc27a1, and Slc27a4. Once fatty acids are brought into the sarcoplasm, they are immediately bound to Fabp3. This binding also provides a positive feedback for more fatty acid uptake. Bound fatty acids can then be transported intracellularly for ␤-oxidation in the mitochondria to produce ATPs, for triglyceride synthesis to avoid lipid toxicity from elevated fatty acid concentrations, and for bioactive lipid mediator production to regulate physiological function of the muscle. In znt7-KO muscle cells, fatty acid uptake and Fabp3 binding are up-regulated, leading to increased ␤-oxidation in the mitochondria. However, it seems that the increased ␤-oxidation cannot keep up with the increased fatty acid levels in the znt7-KO muscle cell. Excessive levels of fatty acids are converted into lipids and lipid mediators, resulting in inflammation and oxidative stress, two key processes that drive insulin resistance.

Muscle insulin resistance in znt7-KO mice
taining 2% house serum before harvest for quantitative RT-PCR and Western blot analysis.

2-[ 3 H]Deoxyglucose uptake assay
L6 myoblasts (6 ϫ 10 4 ) were seeded in 12-well plates, and differentiation was induced as described previously (4). After 7-day differentiation, cells were starved in MEM-␣ in the presence of 2% fatty acid-free BSA (Gold Biotechnology, St. Louis, MO). Then 2 M 12-HETE or 12,13-DiHOME (Cayman Chemical) was added and incubated for 18 h. The cells were washed and preincubated in KRH buffer (4) at 37°C for 30 min followed by a 15-min incubation in fresh KRH buffer containing 0.33 M 2-[ 3 H]deoxyglucose (PerkinElmer Life Sciences) and 150 M D-glucose in the presence of 0 or 100 nM insulin (Sigma-Aldrich). After insulin stimulation, the cells were washed with ice-cold 1ϫ PBS containing 20 mM D-glucose four times and lysed in 300 l of 0.1 N NaOH. Radioactivity was determined by liquid scintillation counting. Protein concentrations were determined using a BCA protein assay kit (Thermo Fisher Scientific).

ELISA analysis for Ccl2
L6 myotubes were differentiated as described above. After 7-day differentiation, cells were treated with 10 M 12-HETE or 12,13-DiHOME in MEM-␣ containing 2% fatty acid-free BSA at 37°C for 48 h. The culture medium was collected, and cell debris was removed by centrifugation. The supernatant was collected and stored at Ϫ80°C until use. Ccl2 concentration in the culture medium was quantified using a rat Mcp-1 ELISA kit according to the manufacturer's instructions (Thermo Fisher Scientific).

Total RNA isolation, cDNA synthesis, and quantitative RT-PCR analysis
Total RNA was isolated according to the manufacturer's instructions (Thermo Fisher Scientific). Total RNA (1 g) was converted into cDNA using an iScript cDNA synthesis kit (Bio-Rad). For quantitative RT-PCR analysis, synthesized cDNA was diluted 10-fold with double-distilled water, and 2 l was used in a SYBR Green-based PCR using SsoAdvanced TM SYBR Green Supermix (Bio-Rad). Quantitative PCR was performed on a CFX384 Real-Time System (Bio-Rad). Primers were designed using Primer-Blast (https://www.ncbi.nlm.nih.gov/tools/primerblast/) or adapted from the PrimerBank (https://pga.mgh. harvard.edu/primerbank/index.html). 4 All primers were confirmed with melting curve analysis. Primer sequences are listed in Table S1. Quantitative PCR was run in triplicate, and the average crossing-point value was used for relative expression calculations (Ϫ2 ⌬⌬CT method) (35). The expression of the target genes was normalized to the expression of Actb.

Antibody
Antibody against ZnT7 was described previously (33). Antibodies against Fabp3 and Actb were purchased from Proteintech (Chicago, IL) and Sigma-Aldrich, respectively. A mouse monoclonal Hadhb antibody (C-6) was purchased from Santa Cruz Biotechnology (sc-271496; Dallas, TX). Horseradish peroxidase-conjugated secondary antibodies were purchased from Cell Signaling Technology. [1-14 C]Palmitic acid (100 Ci/ml; 55 mCi/mmol) was purchased from MP Biomedicals (Solon, OH). Primary myoblasts (WT and znt7-KO) (2.5 ϫ 10 5 cells/well) were plated in 48-well plates and differentiated into myotubes in 4 days in DMEM containing 2% house serum. Myotubes were preincubated with 0.1% (w/v) fatty acid-free BSA in DMEM for 2 h at 37°C in 5% CO 2 . Preincubation medium was replaced with 250 l of 80 M [1-14 C]palmitic acid (1.2 Ci), 40 M fatty acid-free BSA in DMEM and incubated for 0 or 3 min at 37°C in 5% CO 2 . Uptake was stopped by aspiration of uptake buffer and addition of icecold stop buffer (0.1% BSA, 0.2 mM phloretin in DMEM) for 2 min on ice. Cells were washed three times with 300 ml of icecold DMEM containing 0.1% fatty acid-free BSA. Cells were lysed by addition of 150 l of lysis buffer (0.1 N NaOH, 0.2% SDS) and incubated for 30 min on ice. Radioactivity was determined by liquid scintillation counting (Tri-Carb 2500 TR liquid scintillation counter, Packard). Protein concentrations were determined using a BCA protein assay kit (Bio-Rad). Results are mean Ϯ S.D. (three independent experiments with three to six replicate).

Western blot analysis
Primary or L6 myotubes were lysed in M-PER buffer (Thermo Fisher Scientific) containing phosphatase and kinase inhibitors as described previously (4). Protein concentrations were determined using a BCA protein assay (Bio-Rad or Thermo Fisher Scientific). Western blots were performed as described previously (5).

ROS detection and quantitation
Primary myoblasts (WT and znt7-KO) (ϳ20,000/well) were plated in 96-well black wall/clear bottomed microplates and differentiated into myotubes in 4 days in DMEM containing 2% house serum. To detect and quantify ROS accumulation levels in myotubes, we used a Cellular Reactive Oxygen Species Detection Assay kit (orange fluorescence; Abcam, Cambridge, MA). Cells were preincubated in 1ϫ PBS for 15 min at 37°C in 5% CO 2 . Cells were then incubated in ROS orange stain working solution for 60 min in the dark at 37°C in 5% CO 2 . The orange fluorescence was read using a Spectra Max GEMINI XS microplate reader (excitation, 540 nm; emission, 570 nm; Molecular Devices, Sunnyvale, CA). Protein concentrations were determined using a BCA protein assay kit (Thermo Fisher Scientific). Results are mean Ϯ S.E. (three independent experiments with three to six replicates, n ϭ 12).

Data and statistical analysis
Results are presented as the mean Ϯ S.E. or mean Ϯ S.D. Student's t test was used in comparisons of two genotype groups. Differences were considered to be significant at p Ͻ 0.05.
Author contributions-L. H. planned and designed experiments, established primary myoblasts, analyzed the results, and wrote the manuscript. S. T. conducted the animal study; performed blood chemistry measurements, muscle triglyceride concentrations, and RT-PCR; and analyzed the roles of oxylipins in Akt phosphorylation and glucose uptake. C. P. K. conducted immunohistochemistry, Oil Red O staining in muscle, electron microscopy (with help from Patricia Kysar in the Electron Microscopy Laboratory at UC Davis), RT-PCR, and Western blotting for Fabp3. W. R. K. and T. L. P. conducted fatty acid and oxylipin analyses, respectively. J. W. N. analyzed fatty acid and oxylipin results. J. D. conducted cytokine secretion in L6 myotubes.