Genome-wide association studies have identified several type 2 diabetes (T2D) risk loci linked to impaired β-cell function. The identity and function of the causal genes in these susceptibility loci remain, however, elusive. The HHEX/IDE T2D locus is associated with decreased insulin secretion in response to oral glucose stimulation in humans. Here we have assessed β-cell function in Ide knockout (KO) mice. We find that glucose-stimulated insulin secretion (GSIS) is decreased in Ide KO mice due to impaired replenishment of the releasable pool of granules and that the Ide gene is haploinsufficient. We also show that autophagic flux and microtubule content are reduced in β-cells of Ide KO mice. One important cellular role for IDE involves the neutralization of amyloidogenic proteins, and we find that α-synuclein and IDE levels are inversely correlated in β-cells of Ide KO mice and T2D patients. Moreover, we provide evidence that both gain and loss of function of α-synuclein in β-cells in vivo impair not only GSIS but also autophagy. Together, these data identify the Ide gene as a regulator of GSIS, suggest a molecular mechanism for β-cell degeneration as a consequence of Ide deficiency, and corroborate and extend a previously established important role for α-synuclein in β-cell function.
Genome-wide association studies have emphasized a key role of β-cell function in type 2 diabetes (T2D) (1–3). The identity and function of the causal genes in these susceptibility loci remain, however, largely unknown. One such locus is positioned in the HHEX/IDE gene region (2–4), which in humans is associated with lower 30-min oral glucose-stimulated insulin levels (5,6). The Ide gene encodes a multifunctional protein implicated in a diverse set of processes, including development, cell growth and differentiation, proteasome activity, steroid-mediated signaling, and protein degradation (7). An important protective cellular function for IDE is to, via degradation or the formation of stable, irreversible complexes, limit intracellular levels of aggregate prone, amyloidogenic proteins and peptides, thereby preventing formation of toxic oligomers (7,8). Ide knockout (KO) mice have been reported to be glucose intolerant and hyperinsulinemic, which was suggested to be the result of reduced insulin degradation and peripheral insulin resistance (9,10). In contrast, Miller et al. (11) reported that blood glucose and insulin levels were normal in fasted-refed Ide KO mice.
Here, we have assessed β-cell function in mice lacking the Ide gene. We show that insulin secretion is impaired and that islet autophagic flux and microtubule content are reduced in Ide KO mice. We also find that IDE can form stable complexes with α-synuclein and that levels of α-synuclein are increased in Ide KO islets. Moreover, we find that decreased IDE levels are associated with increased levels of α-synuclein in human T2D islets and that increased expression of α-synuclein in β-cells of normal mice impairs glucose-stimulated insulin secretion (GSIS) and autophagic flux. Thus, our data suggest that IDE is a T2D risk gene, causing decreased GSIS and reduced autophagy via impaired suppression of α-synuclein levels in β-cells. These findings suggest that increasing IDE activity and/or expression may represent a therapeutic strategy for treatment of human T2D.
RESEARCH DESIGN AND METHODS
Strains.
The animal studies were approved by the Institutional Animal Care and Use Committee of Umeå University and conducted in accordance with Guidelines for the Care and Use of Laboratory Animals. The Ide Hz mice (C57BL/6albino:129SvEvBrd mixed background) were generated by Lexicon Genetics (12) and bred to generate Ide wild-type (WT) and KO mice, which were then bred independently as previously described (9). The Snca KO mice were obtained from the Jackson laboratory (stock 003692). The Rip/Snca transgenic construct was generated by cloning a 546-bp, PCR-generated, full-length mouse Snca cDNA fragment (primers used: 5′-ctagcccgggatcggagttcttcagaagcctagg-3′ and 5′-ctagcccgggaactgagcacttgtacgccat-3′) behind the Rip2-hsp68 hybrid promoter (subcloned from Addgene DM#265). Transgenic mice were generated at the Umeå transgene core facility (Umeå University).
Glucose, insulin, and islet amyloid polypeptide/amylin measurements.
Glucose tolerance, insulin secretion, and insulin tolerance tests are detailed in the Supplementary Experimental Procedures. Total pancreatic insulin, proinsulin, and islet amyloid polypeptide (IAPP)/amylin from 20-week-old mice were measured using a sensitive rat insulin radioimmunoassay (Linco), rat/mouse proinsulin ELISA kit (Mercodia), and rat/mouse amylin enzyme immunosorbent assay kit (Phoenix Pharmaceuticals), respectively.
Immunohistochemistry, quantification of mRNA, and cell counting.
Immunohistochemical staining, quantification of mRNA expression levels, and cell counting were performed essentially as previously described (13,14) and as in the Supplementary Experimental Procedures. The primary and secondary antibodies used for immunohistochemistry and oligonucleotide sequences used for real-time RT-PCR are listed in Supplementary Tables 1 and 2.
In vitro culture of islets.
For islet culture experiments, islets were isolated by collagenase perfusion and recovered and equilibrated as previously described (14). In vitro analyses on isolated islets are detailed in the Supplementary Experimental Procedures.
Western blot analyses.
Western blot (WB) analyses of isolated mouse and human islets are detailed in the Supplementary Experimental Procedures.
Protein aggregation and degradation assays.
Aggregation and degradation assays were performed using recombinant rat-IDE (Calbiochem), human α-synuclein (Prospec), and human insulin (Novo Nordisk) as described in the Supplementary Experimental Procedures.
Human islets.
Human islet studies were approved by the Human Research Ethics Committee of the Karolinska Institute. Human islets were isolated from brain-dead, heart-beating, T2D (n = 4) and nondiabetic multiorgan donors (n = 4) as previously described (15). All donors were in the age range of 48–72 years. The donors with diabetes were on treatment with metformin or insulin. The use of pancreata for scientific purpose was approved in all cases.
Quantification and statistical analyses.
Quantification of WB experiments was performed using Image-J software, and all the statistical analyses were performed by heteroscedastic two-tailed Student t test. We considered a value of P < 0.05 to be statistically significant.
RESULTS
Glucose intolerance and impaired insulin secretion in Ide KO mice.
To assess the role of Ide in β-cells, we determined GSIS in Ide-null mutant mice, denoted Ide KO, and Ide WT mice. First-phase insulin secretion was modestly reduced and second-phase insulin secretion was severely blunted in Ide KO mice (Fig. 1A). Consequently, Ide KO mice were hyperglycemic (nonfasted glucose levels = 9.9 ± 0.3 mmol/L in Ide WT mice and 12.7 ± 0.8 mmol/L in Ide KO mice, P = 0.004) and glucose intolerant (Fig. 1B). IDE levels are decreased in liver and brain of Ide Hz mice (11), and we found a ∼30% decrease in IDE levels in Ide Hz islets (Fig. 1C and D). In addition, Ide Hz mice showed blunted insulin secretion (Fig. 1E) and were glucose intolerant (Fig. 1F), providing evidence that the Ide gene is haploinsufficient. Ide KO mice were not, however, hyperinsulinemic (fasted insulin = 0.39 ± 0.06 ng/mL in Ide WT mice and 0.49 ± 0.10 ng/mL in Ide KO mice, P = 0.40) and responded to exogenous insulin by reducing glucose levels at a rate identical to that of Ide WT mice (Supplementary Fig. 1A and B).
To further explore the role for Ide in β-cell function and to mimic conditions linked to the development of overt diabetes in humans, i.e., insulin resistance, we next analyzed high-fat diet (HFD)–treated and old Ide WT and KO mice. Predictably, Ide WT mice developed insulin resistance (Supplementary Fig. 1C and D) and glucose intolerance (Fig. 1G and H) on HFD and with age, but responded to the developed insulin resistance by enhancing insulin secretion (Fig. 1I and J). Ide KO mice also developed insulin resistance (Supplementary Fig. 1C and D) on the HFD and with age, but the insulin secretion defect became further pronounced with significantly impaired first- and second-phase insulin secretion (Fig. 1I and J), manifested as a further deterioration of glucose tolerance where glucose levels actually exceeded the detection limit of the gluconometer, i.e., 33.3 mmol/L, for several of the HFD mice, which therefore had to be counted as 33.3 mmol/L (Fig. 1G and H). Thus, β-cells of Ide KO mice fail to adequately adapt to two of the major triggers of insulin resistance and T2D in humans: obesity and ageing. These findings provide evidence for a role for Ide in ensuring GSIS and are consistent with the association of the HHEX/IDE susceptibility allele to decreased insulin secretion in humans (5,6,16).
Increased pancreatic insulin and IAPP content in Ide KO mice.
To examine the underlying cause for the observed insulin secretion defect, we analyzed islet cell number and organization as well as the patterns of expression of the islet hormones insulin, glucagon, somatostatin, and pancreatic polypeptide, glucose transporter type 2 (Glut2), and the transcription factors Ipf1/Pdx1 and Isl1, all of which were normal (Fig. 2A–C). qRT-PCR analyses revealed that the expression of key β-cell and endocrine transcription factors such as Isl1, Ipf1/Pdx1, Nkx6.1, NeuroD, MafA, and Hhex, as well as that of the kinesin motor protein Kif11, which is also part of the HHEX/IDE locus (5), was similar in islets of Ide WT and Ide KO mice (Supplementary Fig. 2A). Ide KO mice showed a significant increase in pancreatic insulin (∼85%) (Fig. 2D) and IAPP (∼30%) (Fig. 2E) content, and confocal microscopy analyses provided evidence of increased density of insulin vesicles (Supplementary Fig. 2B). No significant change in glucagon content was observed (data not shown), and the proinsulin/insulin ratio was normal in Ide KO islets (Supplementary Fig. 2C). Thus, the reduced GSIS observed in Ide KO mice does not appear to result from impaired β-cell differentiation, a decreased amount of β-cells, or insufficient insulin content.
Under conditions involving reduced insulin secretion, pancreatic insulin content homeostasis is maintained via induction of autophagy (17,18). The increased insulin and IAPP content of Ide KO β-cells suggested that autophagy-mediated maintenance of pancreatic insulin content homeostasis may be reduced. In agreement with this notion, the autophagic marker LC3, the autophagic substrate p62, and polyubiquitinated proteins were accumulated in islets of fasted and fed Ide KO mice as compared with those of Ide WT mice (Fig. 2F, Supplementary Fig. 2D, and data not shown). Moreover, LC3 turnover assays (19) showed that, in contrast to Ide WT islets, the lysosomal inhibitor chloroquine failed to increase LC3-II levels in starved Ide KO islets, and the rate of autophagic flux was reduced by 28% in Ide KO islets (Fig. 2G and H). The expression of key autophagy genes such as Atg5, Atg6/Beclin, Atg7, and Atg8/LC3 was normal in β-cells of Ide KO mice (Supplementary Fig. 2E), arguing against a reduced expression of these genes as the underlying cause for the β-cell autophagy defect. Together, these findings indicate that the increased pancreatic insulin and IAPP content of Ide KO mice, at least in part, is the consequence of reduced autophagic flux in β-cells of Ide KO mice.
Replenishment of the releasable pool of insulin granules is impaired in Ide KO mice.
To elucidate at what level Ide ascertains insulin secretion, we next analyzed the effect of secretagogues on insulin secretion in Ide KO and Ide WT mice. Glibenclamide, arginine, and carbachol stimulated insulin secretion to the same extent in Ide KO and Ide WT mice (Fig. 3A–C). These secretagogues act, however, primarily by stimulating exocytosis of insulin from the releasable pool of granules (20–22), suggesting that docking, priming, and triggering of exocytosis of granules are unimpaired in Ide KO mice.
The expression levels of genes involved in glucose uptake, glucose metabolism, and membrane depolarization, i.e., Glut2, glucokinase (GK), SUR1, and Kir6.2, were similar between control and Ide KO mice (Supplementary Fig. 3A). Control and Ide KO islets also showed a similar increase in ATP/ADP ratio in response to glucose (Supplementary Fig. 3D), providing evidence that glucose uptake and metabolism are unaffected in β-cells of Ide KO mice. The expression of genes involved in exocytosis, such as Rab3c, Rab27a, Vamp2, Snap25, Stx1a, Stx4, Syt7, and Syt9, was indistinguishable between Ide WT and Ide KO mice, and there was a nonsignificant tendency to decreased expression of Rab3b and increased expression of Rab3a in islets of Ide KO mice (Supplementary Fig. 3B and C). Thus, expression of key components of the insulin exocytosis machinery is largely unimpaired in islets of Ide KO mice.
To further probe how Ide regulates second-phase insulin secretion, we next performed dual arginine stimulation of insulin secretion in vivo with a 10-min interval (23). The first arginine stimulus will empty the releasable pool of granules, and hence the second arginine stimulus will challenge the ability to replenish this pool (23). The insulin secretion response to the first arginine injection was similar in Ide WT and Ide KO mice, whereas insulin secretion in response to the second arginine stimulus was severely blunted in Ide KO mice (Fig. 3D). Additionally, in vitro perifusion of isolated islets showed that insulin secretion in response to high (16.8 mmol/L) glucose levels was reduced in Ide KO islets (Supplementary Fig. 3E). After an intermittent 3-min period of low (2.8 mmol/L) glucose levels, sequential addition of high glucose levels together with glibenclamide was unable to correct the insulin secretion defect in Ide KO islets (Supplementary Fig. 3E). These findings indicate that recruitment of granules from the reserve pool and/or priming of granules in the releasable pool are decreased in Ide KO mice.
The glucose-amplifying pathway contributes to the second phase of insulin secretion by accelerating the priming process of granules (24–26). To investigate the glucose amplifying pathway in Ide KO mice, islets were depolarized by high K+ in the presence of the KATP channel activator diazoxide. Under these conditions, high glucose (16.8 mmol/L) concentrations generated a robust increase in insulin secretion from both Ide WT and Ide KO islets, although the relative amount of insulin secreted from Ide KO islets was reduced (Fig. 3E). Thus, Ide KO islets respond to the amplifying pathway of insulin secretion by glucose, indicating that priming of granules is largely unaffected in Ide KO β-cells.
Reduced microtubule levels in islets of Ide KO mice.
The perturbation of insulin granule recruitment as well as autophagy in β-cells of Ide KO mice urged us to look for a potential common denominator that might explain these phenotypes. The cytoskeleton and, in particular, microtubules are implicated in efficient insulin granule transport (27), autophagosome formation, transport, and fusion with lysosomes (28,29). Total tubulin was reduced by ∼27% in starved Ide KO islets compared with Ide WT islets (Fig. 4A and B), and the amount of polymerized tubulin, i.e., microtubules, in starved Ide KO islets was ∼57% of that observed in Ide WT islets (Fig. 4A and B). Glucose stimulation increases the amount of polymerized tubulin (30), and although both control and Ide KO islets increased their levels of polymerized tubulin upon glucose stimulation, total and polymerized tubulin content were still reduced, by ∼30%, in glucose-stimulated Ide KO islets compared with Ide WT islets (Fig. 4A and B). Glucose stimulation ensures mobilization and trafficking of granules also by inducing the reorganization of F-actin (27). Consequently, F-actin–depolymerizing agents, such as latrunculin B (Supplementary Fig. 4A and B), have been shown to enhance GSIS (27). Phalloidin staining of Ide WT and Ide KO islets did not, however, reveal any apparent difference in actin organization (Supplementary Fig. 4C), and, in contrast to Ide WT islets, Ide KO islets failed to increase insulin secretion in response to latrunculin B (Fig. 4C). These data indicate that the insulin secretion defect of Ide KO β-cells is upstream of F-actin reorganization, suggesting that the insulin secretion defect observed in Ide KO mice, at least in part, is a consequence of reduced, microtubule-dependent recruitment of granules.
α-Synuclein levels are inversely correlated with IDE levels in islets of Ide KO mice and T2D patients.
To further assess how loss of IDE activity impairs insulin secretion, autophagy, and tubulin polymerization, we looked for a candidate IDE substrate that would negatively affect these processes. α-Synuclein, encoded by Snca, is an amyloidogenic protein negatively correlated with secretion, autophagy, microtubule polymerization, and trafficking (31–34). Moreover, α-synuclein contains the four–amino acid GAXX amyloidogenic region, where X corresponds to amino acids with an aliphatic side chain, which is also present in the IDE substrates amyloidß and IAPP (35). Analyses of α-synuclein protein levels revealed an ∼30–40% increase of α-synuclein monomers in Ide KO islets compared with Ide WT islets and an intermediate (∼15%) increase of α-synuclein in Ide Hz islets (Fig. 5A). Increased levels of α-synuclein were also observed by immunohistochemistry in β-cells of Ide KO mice (Fig. 5B). These findings suggest that Ide, directly or indirectly, regulates α-synuclein levels in β-cells. To further evaluate the impact of increased α-synuclein levels on insulin secretion and autophagic flux, we generated Rip/Snca transgenic mice overexpressing α-synuclein in β-cells. Rip/Snca mice had blunted GSIS and reduced glucose tolerance (Fig. 5C), decreased autophagic flux (∼25%) (Fig. 5D), and decreased microtubules (∼30–40%) (Supplementary Fig. 5A and B) but normal IDE levels, β-cell numbers, and β-cell turnover (Supplementary Fig. 5C and data not shown). Thus, increased levels of α-synuclein in β-cells negatively affect insulin secretion, autophagic flux, and microtubule content.
To explore whether the increased levels of α-synuclein in β-cells of Ide KO mice reflect a role for IDE in α-synuclein degradation, recombinant α-synuclein was incubated with recombinant IDE. Although IDE efficiently degraded insulin already within 1 h of incubation, IDE was not capable of degrading α-synuclein even if incubated for 48 h (Fig. 5E and data not shown). IDE has also, however, been suggested to function as a “dead-end chaperone” that neutralizes the aggregation of amyloidogenic proteins by forming irreversible, SDS-resistant complexes with monomeric forms of the protein (8). Incubation of α-synuclein alone resulted in the formation of high-molecular-weight complexes corresponding to α-synuclein pentamers and decamers already after 30 min (Fig. 5E and Supplementary Fig. 5D). Incubation of α-synuclein with IDE not only reduced the formation of these oligomeric complexes but also generated, in a time- and dose-dependent manner, an SDS-resistant, ∼130-kDa complex that was recognized by α-synuclein antibodies (Fig. 5E and F and Supplementary Fig. 5D). The ∼130-kDa complex corresponds to the combined molecular weight of IDE (∼110 kDa) and α-synuclein (∼17 kDa on an SDS-PAGE) and was recognized also by IDE antibody (Supplementary Fig. 5E). Moreover, IDE antibodies dose-dependently competed with the formation of the ∼130-kDa IDE/α-synuclein complex (Supplementary Fig. 5F). IDE was not, however, capable of affecting preformed high-molecular-weight α-synuclein oligomers, although the ∼130-kDa complex was still generated (Supplementary Fig. 5G). These findings indicate that IDE, by complexing with α-synuclein monomers, reduces the formation of α-synuclein oligomers.
We next assessed IDE and α-synuclein protein levels in islets isolated from T2D patients and nondiabetic controls. IDE levels were decreased (by ∼40%) in human T2D islets compared with nondiabetic controls (Fig. 5G). α-Synuclein monomeric levels were very low and in some cases barely detectable in islets from nondiabetic controls (Fig. 5H). Levels of α-synuclein monomers were, however, increased (by ∼100%) in islets from T2D patients (Fig. 5H). In addition, an ∼50–60-kD protein, or protein complex, recognized by α-synuclein antibodies was increased (by ∼130%) in T2D islets compared with nondiabetic islets (Fig. 5H). Taken together, these findings provide evidence that levels of α-synuclein are inversely correlated with IDE levels in islets of Ide-deficient mice and human T2D patients.
To functionally assess whether the increase in levels of α-synuclein alone accounts for the impaired β-cell function of Ide-deficient mice, we bred the Snca KO allele onto the Ide KO background. Glucose tolerance tests (GTTs) showed that lack of Snca not only failed to rescue the impaired β-cell function of Ide mutant mice but insulin secretion was further blunted in Ide KO; Snca KO double mutant mice as compared with Ide KO mice (Fig. 6A and B). These data imply a role for Snca in GSIS, and GTTs revealed that Snca KO mice were glucose intolerant and had a severely impaired GSIS (Fig. 6C and D). Similarly, autophagic flux was reduced in islets of Snca KO mice (Fig. 6F) and worsened in Ide;Snca double KO islets as compared with Ide KO islets (Fig. 6E). Together these data not only show that loss of Snca in Ide KO mutant mice fails to rescue both the GSIS and autophagic flux phenotype in these mice but also suggest that Snca, in a dose-dependent manner, is required for normal GSIS and glucose tolerance as well as β-cell autophagy.
DISCUSSION
Our study shows that Ide controls β-cell function; mice lacking a functional Ide gene are hyperglycemic, glucose intolerant, and insulin secretion deficient due to reduced replenishment of the releasable pool of insulin granules. Farris et al. (9) and Abdul-Hay et al. (10) found that Ide KO mice were glucose intolerant, had increased fasting insulin levels, and showed evidence of insulin resistance and impaired insulin degradation, whereas Miller et al. (11) reported that Ide KO mice had normal glucose and insulin levels 4 h after refeeding of overnight-fasted mice. We do not find fasting insulin levels to be significantly elevated in Ide KO mice, nor do we find evidence of insulin resistance. Whether the differences in phenotype reflect differences in genetic background and/or methodologies used is unclear but underscores the still unresolved question regarding the role for IDE in insulin degradation (36). Nonetheless, we do find that Ide KO mice have significantly impaired GSIS, which was not explored in any of the other studies (9–11). We also find that the Ide gene is haploinsufficient and that IDE levels are inversely correlated with α-synuclein levels in mouse and human islets. Moreover, we show that increased levels of α-synuclein in β-cells impair GSIS and autophagy in vivo, and that IDE can form stable complexes with α-synuclein. These findings are 1) consistent with the decreased insulin secretion associated with the HHEX/IDE T2D locus in humans (5,6,16), 2) identify Ide as a key β-cell gene in vivo, thus implicating IDE as a potential T2D susceptibility allele, and 3) suggest a mechanism for this risk allele in insulin secretion and diabetes in humans. These findings do not, however, rule out other role(s) for Ide in β-cells. Likewise, although our in vitro analyses are supportive of a cell-autonomous role for Ide in β-cells, the Ide KO mice analyzed in this study are global KOs; thus, we cannot exclude that the β-cell defect displayed by these mutants, at least in part, may be influenced by Ide-related deficiencies in other cells and tissues.
Our findings that GSIS is reduced in mice overexpressing α-synuclein in β-cells are supportive of previous data showing that overexpression of α-synuclein in an insulinoma cell line reduces basal, and to a lesser extent glucose-stimulated, insulin secretion (31). Together, these findings suggest that α-synuclein suppresses insulin secretion. Increased α-synuclein levels have been shown to negatively influence microtubule polymerization and trafficking (33,34), and a role for microtubules is implicated both in GSIS (27) and autophagy (28,29). Thus, the decreased microtubule content in both Ide and Rip/Snca islets suggests that the reduced GSIS and autophagy observed in these mice, at least in part, may result from negative effects of α-synuclein on microtubule polymerization and function. The exact role, if any, for microtubules in GSIS is however unclear; numerous studies implicate a role for microtubules in GSIS, in particular for the recruitment of granules for sustained insulin secretion (reviewed in 27), whereas others provide evidence against a role for microtubules in GSIS (26). α-Synuclein has been shown to interact with Kir6.2 at insulin granules (31) and overexpression of α-synuclein in vitro, possibly by negatively affecting Rab1a function, impairs growth hormone secretion and autophagy in mammalian cells (32), providing additional, microtubule-independent, mechanistic implications of increased levels of α-synuclein (Fig. 7).
Surprisingly, our analyses of Snca KO and Ide KO;Snca KO double mutant mice are supportive of a positive role for α-synuclein in β-cells; Snca KO mice showed severely reduced GSIS and attenuated autophagic flux in islets, and lack of Snca on an Ide KO background resulted in a further deterioration of GSIS and autophagy. In contrast, previous studies of in vitro glucose stimulation of α-synuclein–deficient islets showed evidence of an increased insulin secretion rate at low and medium, but not high, glucose levels (31). Whether the discrepancy between these findings reflects the in vitro versus in vivo experimental design is unclear but underscores the need for additional studies to address the mechanism by which Snca regulates insulin secretion. Notably, the exact physiological function of α-synuclein remains largely unknown but a role for α-synuclein in diverse processes like SNARE complex formation and mitochondrial calcium homeostasis has been suggested as has the idea that the toxic activities of α-synuclein may in fact reflect its normal function (37,38).
IDE has been shown to form stable, irreversible complexes with amyloidogenic proteins such as amyloidß (8), and our studies provide evidence that α-synuclein oligomerization is antagonized by this proposed “dead-end chaperone” activity of IDE in β-cells (Fig. 7). Human IAPP is also an IDE substrate, and in contrast to rodent IAPP, human IAPP can develop toxic oligomers and amyloid deposits, which are commonly observed in β-cells of T2D patients (39,40). In humans, reduced IDE expression and/or activity may thus lead to increased levels of IAPP oligomers and, with time, the formation of IAPP amyloid deposits and β-cell destruction. Autophagy is important for selective degradation of accumulated and aggregated proteins, sometimes referred to as aggrephagy (41). Thus, independent of IDE expression or activity, conditions leading to reduced autophagy in β-cells likely lead to increased intracellular levels of not only α-synuclein but also IAPP oligomers and an enhanced likelihood of IAPP aggregation and consequently amyloid formation in human β-cells.
T2D is associated with both age and obesity, and IDE activity has been shown to decline with age (42) and be inhibited by long-chain free fatty acids (43). Moreover, autophagy is inhibited by insulin, and conditions leading to increased insulin levels, e.g., ageing and obesity, are associated with impaired autophagy (44). Hence, increasing age and obesity likely result in reduced IDE activity as well as autophagy and consequently decreased turnover and/or neutralization of amyloidogenic proteins such as α-synuclein and IAPP, which in turn may lead to β-cell degeneration. Thus, stimulation of IDE expression or activity could represent a valuable therapeutic strategy for the prevention and/or restoration of β-cell function in T2D patients.
ACKNOWLEDGMENTS
This work was supported by grants from the Swedish Research Council, the Strategic Research Program in Diabetes at Umeå University, the Wallenberg Foundation, the European Union (Integrated Project EuroDia LSHM-CT-2006-518153 in the Framework Program 6 of the European Community), the Kempe Foundation, the Swedish Diabetes Association, and the Novo Nordisk Foundation (to H.E.).
H.E. is a cofounder, shareholder, and consultant of the unlisted biotech company Betagenon. C.-G.Ö. is on the advisory boards of Novo Nordisk and Eli Lilly and is a shareholder of the Swedish biotech company Orexo. No other potential conflicts of interest relevant to this article were reported.
P.S. designed and performed in vivo analyses of the control, HFD-fed, aged, and transgenic mice, as well as in vitro islet culture; interpreted results; helped write the paper; and determined insulin, amylin, and glucagon pancreatic content. L.B. performed WB analyses, tubulin and actin fractionations, autophagic flux analyses, and protein aggregation and degradation assays and helped write the paper. S.E. assisted with immunofluorescence studies, cell counting, and mRNA expression analyses. L.L. performed WB analyses, tubulin and actin fractionations, autophagic flux analyses, and protein aggregation and degradation assays. F.B. prepared mouse islets and performed islet perifusion analyses. C.-G.Ö. provided human islets. H.E. designed and supervised the study, analyzed and interpreted the data, and wrote the paper. H.E. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Parts of this study were presented as an oral presentation at the Korea-Sweden Symposium on Frontier Sciences, Seoul, Korea, 30–31 May 2012.
The authors thank members of their laboratory for technical instructions, suggestions, and helpful discussions; Dr. Malcolm A. Leissring (Mayo Clinic, Jacksonville, Florida) for the generous gift of IDE 6A1 mAb; and Drs. Thomas Edlund and Sara Wilson (Umeå University) for critical reading and help in the preparation of the manuscript.