Elovl6 Deficiency Improves Glycemic Control in Diabetic db/db Mice by Expanding β-Cell Mass and Increasing Insulin Secretory Capacity

The increasing prevalence of obesity worldwide has become an alarming public health concern because of dramatic increases in the incidence of obesity-associated diseases, including type 2 diabetes (T2D) (1,2). Obesity leads to progressive deterioration of the insulin secretory function of pancreatic β-cells and reduced capacity to compensate for increased peripheral insulin resistance (3,4). β-Cell dysfunction and actual decrease in β-cell mass were both implicated in the deterioration of functional β-cell capacity.

Lipid accumulation in nonadipose tissues, a phenomenon known as lipotoxicity, was implicated in both pathologies as a molecular link between obesity and dysregulated glucose homeostasis (5,6). Pancreatic β-cells are highly susceptible to lipotoxicity, and numerous studies showed that saturated fatty acids (SFAs), such as palmitate (C16:0), suppressed insulin gene expression and secretion and ultimately induced β-cell apoptosis and/or dedifferentiation through multiple processes, including generation of ceramides and reactive oxygen species, endoplasmic reticulum (ER) stress, and inflammation (7–12). Monounsaturated FAs (MUFAs), such as palmitoleate (C16:1n-7) and oleate (C18:1n-9), and the polyunsaturated FA (PUFA) eicosapentaenoate (C20:5n-3), can protect β-cells from apoptosis and insulin secretory defects induced by SFAs (9,13,14). In addition to exogenous FA composition, evidence indicates that the intracellular capacity to modulate FA composition of lipid species might be another determinant of β-cell lipotoxicity. FA desaturase and elongase enzymes modify FAs by adding a cis-double bond or two carbons, respectively, to fatty acyl-CoA (15,16). Unique roles for these enzymes in pancreatic β-cells remain to be defined.

Elovl6 is a microsomal enzyme involved in the elongation of SFA and MUFA with 12, 14, and 16 carbons (17,18). Loss of Elovl6 function reduces stearate (C18:0) and oleate levels and increases palmitate and palmitoleate levels (19,20). In our previous study, we reported that mice with the targeted disruption of Elovl6 (Elovl6^−/−) were protected against the development of hepatic insulin resistance and deterioration of insulin secretory function of pancreatic β-cells in animals fed a high-fat and high-sucrose diet, despite similar levels of hepatosteatosis and obesity between Elovl6-deficient and wild-type mice (19,21). These findings suggested that the vital role of alterations in FA composition by Elovl6 deficiency extended beyond lipid accumulation and affected insulin sensitivity and β-cell function. Therefore, Elovl6 inhibition could be a potential therapeutic approach in T2D treatment. A critical question that needs to be answered is whether inhibition of this elongase confers reduced susceptibility to T2D. In the current study, we investigated the effects of Elovl6 deletion in leptin receptor–deficient Lepr^db/db (db/db) mice in a model of T2D (22).

All animal husbandry and animal experiments complied with the guidelines of the University of Tsukuba’s regulations of animal experiments and were approved by the University of Tsukuba Animal Experiment Committee. Elovl6^−/− mice were generated as described previously (19). We obtained Lepr^db/+ (db/+) mice on a C57BL/KsJ background from Charles River Japan and crossed with Elovl6^−/− mice to obtain db/+;Elovl6^+/− mice. db/+;Elovl6^+/− mice were then crossed more than seven generations into the C57BL/KsJ background. Finally, double-heterozygous male and female mice were bred to generate mice with the double mutation of Lepr and Elovl6 (db/db;Elovl6^−/−). All animals were housed in a pathogen-free barrier facility with a 12-h light/dark cycle and were given free access to normal chow and water. Age- and sex-matched littermates were used for all experiments. Mice were sacrificed during the light phase after food deprivation for 4 h.

Plasma glucose, insulin, triglyceride (TG), total cholesterol (T-Cho), free FA (FFA), and glycosylated hemoglobin A_1c (HbA_1c) levels were determined as described previously (19,23).

Hematoxylin and eosin staining and immunohistochemistry for insulin- and glucagon-containing cells were performed as described previously (21,23,24). Pancreatic sections were stained with an anti-BrdU antibody (Abcam, Cambridge, U.K.). TUNEL staining was performed using the DeadEnd Fluorometric TUNEL system (Promega, Madison, WI). All images were acquired using a BZ-X710 microscope (Keyence, Osaka, Japan), and data were analyzed using a BZ-H3 analyzer (Keyence) and Adobe Photoshop software (Adobe Systems, San Jose, CA).

Isolation of islets from mice was performed by Ficoll-Conray density-gradient centrifugation (21,23,24). Insulin secretion, insulin content, TG content, cholesterol (Cho) content, and FA composition of islets were measured as previously described (21,23,24).

Islets were treated with FA as described previously (14) with some modifications. Briefly, palmitate and oleate were dissolved in 100 mmol/L in ethanol to prepare stock solutions, which were then diluted in RPMI 1640 supplemented with 1% FBS and 0.5% FA-free BSA to a final concentration of 500 μmol/L.

Apoptotic β-cells were detected by immunofluorescence staining with antibody to the cleaved form of caspase-3 (Cell Signaling Technology Japan, Tokyo) and insulin (Abcam), followed by detection with secondary antibodies. The cells were counterstained with DAPI. Cell proliferation was assessed by 5-ethynyl-2’-deoxyuridine (EdU) incorporation using the Click-iT EdU Alexa Fluor 488 Imaging Kit (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s protocol. EdU-labeling reagent was added to islet culture medium during the last 24 h of culture. All images were acquired using a TCS SP5 confocal microscope (Leica, Wetzlar, Germany), and data were analyzed using ImageJ software.

Total RNA extraction from isolated islets, cDNA synthesis, and quantitative real-time PCR were performed as previously described (21,23–27). Primer sequences for Ins1, Ins2, Gcg, Pdgfra, Reg3a, Reg3b, Reg2, Aldh1a3, Ccr2, Ccl21a, Glycam1, Saa3, Xbp1s, Ddit3, Casp1, Atf4, Ppp1r15a, and Txnip are presented in Supplementary Table 1. mRNA expression levels were normalized to that of cyclophilin mRNA.

Values are expressed as means ± SEM and were analyzed using ANOVA. Differences were considered significant for P of <0.05.

To investigate the role of Elovl6 in T2D development and progression, we first determined the effect of endogenous Elovl6 in leptin receptor–deficient db/db mice, which exhibit obesity, insulin resistance, and resultant β-cell failure as a model of T2D, generating Elovl6-deficient db/db mice (db/db;Elovl6^−/−). During 6–16 weeks of age the db/db;Elovl6^+/+ and db/db;Elovl6^−/− mice both gradually gained body weight (BW) without any significant differences (Fig. 1A). At 40 weeks of age, when emergence of diabetes caused BW loss in db/db;Elovl6^+/+ mice, db/db;Elovl6^−/− mice still sustained a gain, with significant difference compared with db/+;Elovl6^+/+ controls. There was no significant difference in food intake between the two genotypes (4.5 ± 0.5 g/day in db/db;Elovl6^+/+ mice vs. 4.7 ± 0.1 g/day in db/db;Elovl6^−/− mice). Surprisingly, although obesity in db/db mice was not improved, Elovl6 deficiency markedly improved hyperglycemia (Fig. 1B) and showed an adaptive increase in insulin (Fig. 1C) in db/db mice. These metabolic changes also manifested in ameliorated polydipsia, polyuria (Fig. 1D), and elevated HbA_1c levels in db/db;Elovl6^−/− mice (Fig. 1E). Furthermore, improved plasma glucose clearance (Fig. 2A) and increased insulin secretion (Fig. 2B) were observed in db/db;Elovl6^−/− mice by oral glucose tolerance test, whereas insulin sensitivity, as evidenced by insulin tolerance test, did not differ between the two genotypes (Fig. 2C and D). These results suggested that Elovl6 deficiency prevented T2D progression by increasing insulin secretory capacity of pancreatic β-cells in db/db mice.

Because of the development of hepatosteatosis, adipocyte inflammation, and hyperlipidemia accompanied by the obesity model with pancreatic β-cell dysfunction and glucose intolerance, we weighed liver and epididymal white adipose tissue (eWAT) and measured metabolic parameters. Liver weight and hepatic TG and T-Cho content were higher, whereas plasma levels of TG and FFA were lower in db/db;Elovl6^−/− mice than in db/db;Elovl6^+/+ mice (Table 1). Analysis of liver histology further confirmed the dramatic increase in lipid droplets in db/db;Elovl6^−/− mice compared with db/db;Elovl6^+/+ mice (Supplementary Fig. 1A). Hepatic gene expression analysis showed a consistent increase in glucokinase (Gck) and stearoyl-CoA desaturase 1 (Scd1) expression in db/db;Elovl6^−/− mice, indicating increased glycolysis and lipogenesis (Supplementary Fig. 1B). In addition, Elovl6 deficiency reduced glucose-6-phosphatase (G6pc) mRNA levels in both db/+ and db/db mice, consistent with a reduction in blood glucose levels (Supplementary Fig. 1B). Histological examination of eWAT sections revealed that Elovl6 deficiency in db/db mice did not improve adipose tissue inflammation (Supplementary Fig. 2A). Real-time PCR analysis showed that expression levels of lipogenic and proinflammatory genes in the eWAT of db/db;Elovl6^+/+ and db/db;Elovl6^−/− mice were similar (Supplementary Fig. 2B). These results demonstrated that amelioration of T2D progression by Elovl6 deficiency in db/db mice was not caused by reduced lipid accumulation and inflammation in liver and WAT. However, as a consequence of the greater steatosis observed in db/db;Elovl6^−/− mice, a more efficient FA storage could improve FA sequestration and reduce β-cell lipotoxicity.

Pancreatic islets isolated from db/db;Elovl6^−/− mice were morphologically compared with those from db/+;Elovl6^+/+, db/+;Elovl6^−/−, and db/db;Elovl6^+/+ mice. The number (Fig. 3A) and size (Fig. 3B) of islets were both greater in db/db;Elovl6^−/− mice than in db/db;Elovl6^+/+ mice. Histological examination indicated that the islets in db/db;Elovl6^−/− mice were markedly enlarged compared with those in db/db;Elovl6^+/+ mice (Fig. 3C). Immunostaining of pancreatic sections from 16-week-old animals with antibodies to insulin and glucagon revealed that Elovl6 deficiency in db/db mice resulted in a significant increase in β-cell mass, whereas β-cell mass was smaller in db/db mice than in db/+;Elovl6^+/+ mice (Fig. 3D). In contrast, we did not detect any apparent differences in scattered α-cell distribution between db/db;Elovl6^+/+ and db/db;Elovl6^−/− mice.

To elucidate how Elovl6 deletion may enlarge β-cell mass, we studied β-cell proliferation and apoptosis. To evaluate β-cell proliferation, mice were administered BrdU. The number of BrdU-positive cells was ∼2.5-fold greater in islets from db/db;Elovl6^−/− mice than in those from db/db;Elovl6^+/+ mice (Fig. 3E). TUNEL staining demonstrated that β-cell apoptosis was ∼4.3-fold greater in db/db;Elovl6^+/+ mice than in db/+;Elovl6^+/+ mice (Fig. 3F). Elovl6 deletion, however, dramatically reduced β-cell apoptosis in db/db mice to levels observed in db/+;Elovl6^+/+ and db/+; Elovl6^−/− mice. Therefore, islet hyperplasia in db/db;Elovl6^−/− mice was a consequence of both increased proliferation and reduced apoptosis of β-cells.

To evaluate the effects of Elovl6 deficiency on insulin secretion ex vivo, pancreatic islets were isolated from db/+;Elovl6^+/+, db/+;Elovl6^−/−, db/db;Elovl6^+/+, and db/db;Elovl6^−/− mice. Insulin secretion from db/db;Elovl6^+/+ islets in response to high glucose was blunted (Fig. 4A). Elovl6 deficiency enhanced glucose-stimulated insulin secretion (GSIS) in db/db islets. Elovl6 deficiency also showed the tendency to enhance KCl-stimulated insulin secretion in both db/+ and db/db islets. The insulin content of db/db;Elovl6^+/+ islets was significantly lower than that of db/+;Elovl6^+/+ islets, whereas the reduction in insulin was prevented by Elovl6 deficiency (Fig. 4B). These results suggested that Elovl6 deficiency improved impaired GSIS from db/db;Elovl6^−/− islets, which contributed to improvement of glucose intolerance observed in db/db mice.

Because excessive lipid accumulation in pancreatic β-cells is known to be associated with lipotoxicity and reduces insulin secretion, we further determined islet TG and Cho contents. The TG content of db/+;Elovl6^−/− islets was slightly but significantly lower than that of the db/+;Elovl6^+/+ islets (Fig. 4C). The TG content of db/db;Elovl6^+/+ islets was markedly higher than that of db/+;Elovl6^+/+ islets. Conversely, the TG content of db/db;Elovl6^−/− islets was markedly lower than that of db/db;Elovl6^+/+ islets and similar to that of db/+;Elovl6^+/+ and db/+;Elovl6^−/− islets. No significant difference was found for Cho contents between the four genotypes (Supplementary Fig. 4).

To further estimate the effect of the diabetic state and Elovl6 deficiency on FA profile in pancreatic islets, we analyzed the FA composition of islets isolated from db/+;Elovl6^+/+, db/+;Elovl6^−/−, db/db;Elovl6^+/+, and db/db;Elovl6^−/− mice (Fig. 4D). The difference in FA composition between db/+;Elovl6^+/+ and db/+;Elovl6^−/− mice was not significant. Compared with db/+;Elovl6^+/+ mice, the islet FA composition in db/db;Elovl6^+/+ mice showed a significant decrease in stearate and an increase in palmitoleate and oleate composition. Compared with db/db;Elovl6^+/+ islets, there was a marked reduction in oleate composition of db/db;Elovl6^−/− islets, which was similar to that observed in db/+;Elovl6^+/+ and db/+;Elovl6^−/− controls. The ratio of MUFAs in db/db;Elovl6^−/− islets was consistently lower than that in db/db;Elovl6^+/+ islets (Fig. 4D, inset). We also analyzed the plasma FA composition of db/+;Elovl6^+/+, db/+;Elovl6^−/−, db/db;Elovl6^+/+, and db/db;Elovl6^−/− mice (Supplementary Fig. 5). We found that the changes in plasma FA composition resulting from Elovl6 deletion in db/db mice were different from those in islets: db/db;Elovl6^−/− mice had decreased levels of stearate and oleate and increased levels of palmitate and palmitoleate in plasma. These results suggested that the modulation of intracellular FA composition in β-cells by Elovl6 was associated with β-cell dysfunction and that Elovl6 deficiency could play a positive role in β-cell mass and function through a reduction of oleate and TG content under diabetic conditions.

To provide insight into how Elovl6 deficiency might alter the gene expression profile to preserve β-cell function and mass, we analyzed isolated islets for genes encoding proteins with functions in FA and glucose metabolism, insulin production, cell proliferation, apoptosis, inflammation, and ER stress (Fig. 5). Expression levels of lipogenic genes, including Elovl6, Scd1, and Scd2, were significantly lower in db/db;Elovl6^+/+ islets than in control db/+;Elovl6^+/+ islets (Fig. 5A). Expression levels of Scd1 and Scd2 were similarly decreased in db/db;Elovl6^−/− islets, whereas the expression of sterol regulatory element binding protein 1c (Srebf1c) and its target gene uncoupling protein-2 (Ucp2) tended to decrease in db/db;Elovl6^−/− islets compared with db/db;Elovl6^+/+ islets. Expression of peroxisome proliferator activated receptor-α (Ppara), a key regulator of FA oxidation, was elevated similarly in both db/db;Elovl6^+/+ and db/db;Elovl6^−/− islets (Fig. 5B). However, expression levels of PPARα target genes involved in β-oxidation, such as carnitine palmitoyltransferase 1a (Cpt1a) and acyl-CoA dehydrogenase, medium chain (Acadm), were significantly decreased in db/db;Elovl6^−/− islets compared with db/db;Elovl6^+/+ islets, suggesting the difference in TG accumulation between db/db;Elovl6^+/+ islet and db/db;Elovl6^−/− islet was not caused by the difference in the capacity for islet FA oxidation.

Expression levels of master transcription factors regulating insulin gene expression and GSIS in mature β-cells, such as pancreatic and duodenal homeobox 1 (Pdx1), musculoaponeurotic fibrosarcoma oncogene family A (MafA), and neurogenic differentiation 1 (Neurod1), as well as insulin 1 (Ins1), Ins2, and synaptotagmin-like 4 (Sytl4), a crucial component of the insulin secretion, were coordinately repressed in db/db;Elovl6^+/+ islets compared with those in db/+;Elovl6^+/+ islets (Fig. 5C). Conversely, expression of these genes was not completely but was significantly restored in db/db;Elovl6^−/− islets compared with db/db;Elovl6^+/+ islets, although the expression level of neurogenin 3 (Neurog3), the islet-defining factor, is similar in all groups. Expression levels of glucose transporter 2 (Slc2a2), Gck, and insulin receptor substrate 2 (Irs2), genes important for glucose metabolism and β-cell mass expansion, were similar between db/db;Elovl6^+/+ and db/db;Elovl6^−/− islets (Fig. 5C).

Expression levels of genes involved in β-cell proliferation and islet regeneration, such as platelet-derived growth factor receptor, α-polypeptide (Pdgfra), regenerating islet-derived 3-α (Reg3a), Reg3b, and Reg2, were markedly higher in db/db;Elovl6^+/+ islets than in db/+;Elovl6^+/+ islets (Fig. 5D). These increases in db/db;Elovl6^+/+ islets were cancelled in db/db;Elovl6^−/− islets. Expression levels of apoptosis-related genes, such as B-cell leukemia/lymphoma 2 (Bcl2) and BCL2-associated X protein (Bax), were similar in all groups (Fig. 5E). Recent animal and human studies show that β-cells become dedifferentiated in diabetes, reverting to a progenitor-like stage and partly converting to other endocrine cell types (28,29), and that aldehyde dehydrogenase 1A3 (Aldh1a3) is a newly identified biomarker of dysfunctional β-cells (30). We tested the expression of Aldh1a3 and found that it was elevated similarly in both db/db;Elovl6^+/+ and db/db;Elovl6^−/− islets (Fig. 5F), indicating Elovl6 deficiency did not protect from β-cell dedifferentiation in db/db mice.

Expression levels of the macrophage marker CD68 antigen (Cd68), inflammatory cytokines and their receptors, such as tumor necrosis factor-α (Tnfα), Ccl2, chemokine (C-C motif) receptor 2 (Ccr2), interleukin 1β (Il1β), chemokine (C-C motif) ligand 21A (Ccl21a), the L-selectin ligand glycosylation–dependent cell adhesion molecule 1 (Glycam1), and the acute-phase protein serum amyloid A 3 (Saa3), were higher in db/db;Elovl6^+/+ islets than in control db/+;Elovl6^+/+ islets (Fig. 5G). However, Elovl6 deficiency significantly decreased the expression of these proinflammatory genes in both db/+ and db/db islets. Moreover, Elovl6 deficiency significantly decreased the expression of caspase-1 (Casp1), a gene involved in inflammation and cell death, in both db/+ and db/db islets (Fig. 5G). Decreased pancreatic islet inflammation in db/db;Elovl6^−/− mice was further examined by immunohistochemistry (Supplementary Fig. 6). The data indicated that the number of infiltrating Mac3-positive macrophages was markedly greater in db/db;Elovl6^+/+ mice than in db/+;Elovl6^+/+ mice, whereas macrophage infiltration was protected in pancreatic islets of db/db;Elovl6^−/− mice.

ER stress has been proposed as a mechanism for β-cell dysfunction and death in T2D (3,31–34). Chan et al. (35) reported that the expression of adaptive unfolded protein response (UPR) genes progressively declined in db/db islets and that the maintenance of the adaptive UPR is associated with β-cell compensation in obese mice. Consistent with this report, the expression levels of genes for ER stress, such as activating transcription factor 4 (Atf4) and DNA-damage inducible transcript 3 (Ddit3), were significantly decreased in both db/db;Elovl6^+/+ and db/db;Elovl6^−/− islets compared with db/+;Elovl6^+/+ and db/+;Elovl6^−/− islets (Fig. 5H). Expression of the spliced form of X-box binding protein 1 (Xbp1s), which controls the development and maintenance of ER, was also significantly decreased in db/db;Elovl6^+/+ islets yet was maintained in islets from db/db;Elovl6^−/− mice. In contrast, the expression levels of genes for proapoptotic and terminal UPR, such as protein phosphatase 1 regulatory subunit 15a (Ppp1r15a), thioredoxin interacting protein (Txnip), and NLR family pyrin domain containing 3 (Nlrp3), and activating transcription factor 3 (Atf3), a stress-inducible transcription factor involved in modulating proinflammatory responses and apoptosis, were higher in db/db;Elovl6^+/+ islets than in control db/+;Elovl6^+/+ islets. Strikingly, the expression of these genes was markedly suppressed in db/db;Elovl6^−/− islets compared with db/db;Elovl6^+/+ islets. These results demonstrate that Elovl6 deficiency partially recovers the adaptive UPR and suppresses the terminal UPR in db/db islets. Together, these results indicated that Elovl6 regulated both β-cell mass and insulin production potentially thorough modulation of the islet inflammation and ER stress in obese diabetic mice.

To test whether islet and β-cell function could be altered by changes in cellular FAs regulated by Elovl6, we evaluated the effects of palmitate and oleate on islets isolated from Elovl6^+/+ and Elovl6^−/− mice. Consistent with many previous studies, Elovl6^+/+ islets treated with palmitate had increased cleaved caspase-3–positive cells, a marker of apoptosis (Fig. 6A and Supplementary Fig. 7A). However, this effect was significantly suppressed in Elovl6^−/− islets. Meanwhile, oleate led to a mild increase in apoptosis in both Elovl6^+/+ and Elovl6^−/− islets. This result is different from previous studies indicating that oleate did not induce β-cell apoptosis (9,13). The discrepancy is presumably because we used a much lower serum concentration (1%, in contrast to 10% in previous studies) in culture medium. β-Cell proliferation was not affected by palmitate but was strongly stimulated by oleate (Fig. 6B and Supplementary Fig. 7B). Consistent with this result, Moullé et al. (36) recently reported that lipid highly enriched in oleate stimulates β-cell proliferation both in vivo and ex vivo. Moreover, oleate-induced β-cell proliferation was accelerated by Elovl6 deficiency. This effect could somehow mask the lipotoxic side of oleate in previous studies. Further studies will be required to identify the mechanism whereby oleate promotes β-cell proliferation.

We also evaluated the effects of palmitate and oleate on the insulin secretion of islets isolated from Elovl6^+/+ and Elovl6^−/− mice. Basal insulin secretion was not affected by FAs and Elovl6 deficiency; however, GSIS was inhibited by the addition of palmitate or oleate in Elovl6^+/+ islets (Fig. 6C). Although the GSIS of Elovl6^−/− islets tended to be lower than that of Elovl6^+/+ islets, the suppressed GSIS by palmitate or oleate was not observed in Elovl6^−/− islets. Insulin contents were also decreased by the addition of palmitate or oleate in Elovl6^+/+ islets (Fig. 6D). Insulin content of Elovl6^−/− islets was significantly lower than that of Elovl6^+/+ islets. Insulin content of Elovl6^−/− islets treated with palmitate or oleate was similar to that of Elovl6^+/+ islets.

Next, the effects of palmitate and oleate on mRNA expression in islets isolated from Elovl6^+/+ and Elovl6^−/− mice were determined. Elovl6^+/+ islets treated with palmitate exhibited downregulation of Ins1, Ins2, Pdx1, and Mafa compared with control islets (Fig. 6E). Oleate treatment also led to a significant decrease in the expression of these β-cell–specific genes, indicating β-cell dysfunction was induced by both palmitate and oleate. Elovl6^−/− islets treated with palmitate or oleate showed a significant decrease in the expression of these β-cell–specific genes to a similar extent as Elovl6^+/+ islets, indicating that protection of β-cell dysfunction by Elovl6 absence is not mediated through expression of these genes.

Because loss of β-cells was accompanied with increased inflammation and terminal UPR, we analyzed the expression of genes for inflammation and UPR (Fig. 6F). In Elovl6^+/+ islets, palmitate treatment significantly increased the expression of genes for inflammation (Tnfa, Ccl2) and UPR (Ddit3, Ppp1r15a). Oleate treatment, however, led only to a modest increase in Ddit3 gene expression in Elovl6^+/+ islets, suggesting that exogenous oleate does not cause islet inflammation and cell death. These effects of palmitate were significantly attenuated in Elovl6^−/− islets except for Ccl2. Oleate-treated Elovl6^−/− islets showed a slight but significant decrease of Tnfa and Ppp1r15a expression compared with Elovl6^+/+ islets. Administration with palmitate or oleate also impaired β-cell functions of MIN6 cells, accompanied by decreased Pdx1, Mafa, and Ins1 expression and increased Ccl2, Ddit3, and Atf3 expression (Supplementary Fig. 8), indicating other evidence that oleate was more harmful than palmitate in β-cells. Collectively, these findings demonstrate that oleate contributes to impaired GSIS and palmitate contributes to β-cell dysfunction, inflammation, and cellular stress responses and that palmitate lipotoxicity can be partly modified by Elovl6 activity. These results suggest that susceptibility to FA-induced inflammation and ER stress is reduced in islets with a reduced Elovl6 activity.

In the current study, we demonstrated that Elovl6 deletion in db/db mice significantly reduced circulating blood glucose levels without affecting obesity or insulin resistance. We observed that compared with db/db;Elovl6^+/+ mice, db/db;Elovl6^−/− mice had improved glucose tolerance, enhanced GSIS, an adaptive increase in insulin, and markedly increased β-cell mass associated with increased proliferation and decreased apoptosis. Phenotypic changes in db/db;Elovl6^−/− mice involve attenuated islet inflammation accompanied with decreased islet oleate and TG levels. As far as we know, this is the first study to test the effect of Elovl6 deficiency of function in a T2D model of islet failure, and our findings indicated that Elovl6-mediated modulation of intracellular FA metabolism in β-cells was essential in preventing the toxic effects of FAs and preserving proper β-cell function, which might contribute to the expansion of β-cell mass in obesity.

Elucidation of the factors triggering islet inflammation in T2D is of emerging importance. Previous studies demonstrated that macrophage recruitment and inflammation in pancreatic islets played a key role in β-cell dysfunction in patients with T2D and animal T2D models (37–39). β-Cells respond to endogenous stimuli, including FAs, by initiating macrophage recruitment via production of chemokines, and subsequent interactions between β-cells and macrophages perpetuate inflammatory processes within islets that lead to β-cell dysfunction (3,39,40). Importantly, db/db;Elovl6^−/− islets displayed a marked reduction in oleate and MUFA content without increasing SFA compared with db/db;Elovl6^+/+ islets. Oleate-induced β-cell dysfunction was observed in β-cell lines and pancreatic islets that exhibited impaired GSIS after oleate exposure (41–43). Busch et al. (40) performed a microarray analysis of palmitate- and oleate-exposed MIN6 cells and found that the expression of proinflammatory genes encoding chemokines and mediators of the acute phase response were increased by both FAs. In line with these previous reports, we confirmed that oleate contributed to β-cell dysfunction within islets and MIN6 cells, suggesting that the combination of oleate with various stresses, such as high glucose, oxidative stress, and mitochondrial dysfunction, worsens β-cell function and death by cytokine production and promotion of inflammation.

Whether the effect of oleate on β-cell dysfunction is mediated by oleate itself or via intermediate products of oleate metabolism remains to be determined. The underlying mechanism of oleate-associated toxic effects in β-cells remains unknown, but Elovl6-mediated compositional changes in specific FAs may affect cellular functions important for T2D development. Previous studies showed that TG content in pancreatic islets was associated with impaired β-cell secretory capacity and dysfunction (5,6). A potential mechanism of reduction in islet TG in db/db;Elovl6^−/− mice is suppression of TG synthesis secondary to decreased oleate levels in the islets. Because oleate is a good substrate for diacylglycerol acyltransferase, the rate-limiting step in TG synthesis, excessive TG accumulation with oleate might contribute to β-cell dysfunction. The inhibition of conversion from palmitate to stearate resulting in a reduction in oleate levels in β-cells is likely a major contributor to decreased islet TG content and resolved inflammation in the current study.

Chronic exposure of palmitate to islets or β-cell lines has long been known to cause lipotoxicity leading to β-cell apoptosis (14,31,44). Our current studies clearly demonstrate that Elovl6 deficiency did not cause accumulation of palmitate in islets, but significantly decreased the sensitivity and extent to which palmitate induced inflammation and ER stress, suggesting that FA synthesized by Elovl6, rather than palmitate itself, exerts inflammation and terminal UPR. Consistent with these results, Green et al. (45) reported that increased Elovl6 activity in INS-1 cells significantly increased palmitate-induced ER stress and apoptosis. The underlying mechanism whereby long-chain FAs more than C18 or longer could exert a more toxic effect than C16 is currently unknown. The data from exogenous treatment with oleate in isolated islets and MIN6 cells suggest that lipotoxicity by oleate is not very simple and that the intracellular trafficking of FA or the process by Elovl6 might be important. Because previous studies have shown that increased ceramide formation plays a key role in lipotoxicity-induced apoptosis and loss of function of pancreatic β-cells (7,46,47), Elovl6 may lead to inflammation and terminal UPR through the ceramide signaling pathway.

The current study implicates multiple pathways in long-chain FA–mediated β-cell lipotoxicity. β-Cell dysfunction is mediated by palmitate or oleate through impaired gene expression involving insulin secretion, which highlights underrecognized oleate and TGs lipotoxicity, and is partly ameliorated by Elovl6 absence as estimated by decreases in islet oleate and TGs. Meanwhile, inflammation and ER stress leading to β-cell mass loss involving apoptosis and cell proliferation is essentially cancelled by Elovl6 absence (Fig. 7). These results at least partly explain previously described complex effects of long-chain FAs on β-cells. It can be concluded that endogenous conversion of palmitate to stearate by Elovl6 and presumably subsequent reaction is critical for palmitate-mediated lipotoxic action.

Our results support the hypothesis that Elovl6 deletion protects against β-cell failure in db/db mice as a genetic model of T2D. However, because Elovl6 deletion was not targeted to the pancreas and was achieved in all tissues, the possibility remains that Elovl6 deletion might affect the FA profile in other tissues to affect overall organismal metabolism and, therefore, β-cell function. Indeed, we previously showed that Elovl6 deficiency altered FA composition in serum and various organs and affected a wide range of cellular functions and the disease progression (19,21,48–50). We assume the changes of plasma FA composition might influence the FA composition of islets; however, changes in FA composition of db/db;Elovl6^−/− islets were different from changes observed in plasma: a marked reduction in oleate component without a reduction in stearate and an increase in palmitate, compared with db/db;Elovl6^+/+ islets. Thus, these results suggested that the Elovl6-regulated FA chain elongation from C16 to C18 and subsequent SCD-mediated C18 desaturation in β-cells plays a pivotal role for maintaining proper function and preventing the toxic effects of FAs. Future studies in mice with conditional deletion of Elovl6 in pancreatic β-cells and other cell types are under way and necessary to further delineate its role in T2D.

In summary, we provide evidence showing the beneficial effect of Elovl6 loss in db/db mice supporting previously not fully appreciated β-cell toxic effect of oleate. The results of the current study, combined with our previous findings demonstrating that Elovl6 deficiency improved obesity-induced peripheral insulin resistance, suggest that limiting Elovl6 expression or activity in individuals during early diabetes or in those with metabolic syndrome might be beneficial for T2D prevention and treatment.

Acknowledgments. The authors thank Katsuko Okubo, Yuko Tamai, and Chizuko Fukui (Department of Endocrinology and Metabolism, University of Tsukuba) for technical assistance and the members of our laboratories for discussion and helpful comments on the manuscript. The authors thank Enago (www.enago.jp) for the English language review.

Funding. This work was supported by Grants-in-Aid for Scientific Research 22117502 (to H.Sh.) and 23659464 (to T.M.) and Program to Disseminate Tenure Tracking System (to T.M.) from the Ministry of Science, Education, Culture, and Technology of Japan; Front Runner of Future Diabetes Research from Japan Foundation for Applied Enzymology (to T.M.); and Ono Medical Research Foundation (to T.M.).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. T.M. and H.Sh. designed the project. H.Z., T.M., Y.Nakan., K.M., and N.T. performed experiments. H.Z., T.M., Y.Nakan., T.Y., Y.O., S.-i.H., Y.T., Y.A., H.I., S.Y., H.Su., M.S., N.Yah., Y.Nakag., H.So., N.Yam., and H.Sh. analyzed and interpreted data. H.Z., T.M., and H.Sh. prepared the manuscript. All authors reviewed the manuscript. T.M. and H.Sh. are guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.