Introduction
Type 2 diabetes is characterised by hyperglucagonaemia and a decrease in the pancreatic beta/alpha cell ratio [
1,
2]. Alpha cell hyperplasia and/or hypertrophy has been postulated to be a principal cause of decreased beta/alpha cell ratio, yet the operative cellular and molecular mechanisms remain poorly understood. Although some evidence points to beta-to-alpha cell transdifferentiation and/or increased alpha cell proliferation as possible causes of hyperglucagonaemia [
3,
4], these results remain controversial and need further study.
Genome-wide association studies have identified a locus on chromosome 10 linked to risk for type 2 diabetes. This chromosomal region includes the
Ide gene that codes for the insulin-degrading enzyme (IDE) [
5]
. Supporting a causal role for IDE, several
Ide polymorphisms have been identified that are linked to type 2 diabetes risk [
6]. Moreover, the Goto–Kakizaki (GK) rat was shown to contain two coding mutations within the
Ide gene [
7]. As a protease that avidly degrades insulin in vitro, IDE had been expected to mediate insulin clearance in vivo; however, emerging evidence has identified several non-proteolytic functions that might be perturbed in type 2 diabetes, implying a complex functional role for IDE [
8]. For instance, although pancellular genetic deletion of
Ide in mice consistently produces a pronounced diabetic phenotype [
9,
10], this phenotype has variously been associated with hyperinsulinaemia [
9] or hypoinsulinaemia [
10]. Liver-specific deletion of
Ide results in decreased insulin sensitivity but does not impair insulin clearance [
11,
12]. Of special interest, in studies of
Ide-null mice, a key role for IDE in the regulation of insulin secretion from beta cells was identified [
10]. Genetic deletion of
Ide exclusively in beta cells produced impaired glucose-stimulated insulin secretion, along with elevated basal constitutive insulin secretion, decreased cell surface-associated levels of GLUT2, and a phenotype of beta cell functional immaturity [
13]. These findings, among many others [
8], suggest that IDE is a pleiotropic protein whose involvement in the pathogenesis of type 2 diabetes is likely to be multifaceted.
IDE is present within all islet cells but it is expressed at higher levels in alpha cells from both mice and humans [
2]. This, combined with the fact that IDE can degrade glucagon as well as insulin [
14], suggests that IDE may play an important role in alpha cell function and/or glucagon homeostasis in vivo. To investigate this topic, we generated a novel mouse line with alpha cell-specific deletion of
Ide, the A-IDE-KO mouse model, and we performed in vivo and in vitro studies to understand the implication of IDE in alpha cell function and proliferation.
Discussion
IDE has been implicated in the pathogenesis of type 2 diabetes but its role in different tissues involved in glucose homeostasis has only recently begun to be elucidated [
11‐
13,
24]. To clarify the role of IDE in alpha cell function, we developed a novel alpha cell-specific
Ide-knockout mouse model (A-IDE-KO). Deletion of
Ide in alpha cells resulted in a metabolic phenotype consisting of hyperglucagonaemia and hyperinsulinaemia but with normal glucose tolerance. The hyperinsulinaemia is likely due to an exacerbated paracrine effect wherein excess glucagon release by alpha cells stimulates the glucagon receptors on beta cells, leading to activation of the cAMP– protein kinase A (PKA)–exchange protein directly activated by cAMP (EPAC) pathway and thereby stimulating insulin secretion [
25‐
28]. Alternatively, or in addition, this phenotype could be attributed to dysregulated paracrine control as a consequence of impaired alpha cell ciliogenesis, as previously reported for beta cell paracrine control in the beta cell-specific IFT88-KO model [
22]. In this model it was shown that the response of beta cells to somatostatin is dependent on the cilia, consistent with the fact that the somatostatin receptor 3 (SSTR3) localises to beta cell cilia. It was also observed that global changes occurred in modulation of pathways governing paracrine signalling, hormone secretion, islet cell connectivity and calcium activation [
22].
Supporting this, A-IDE-KO mice have normal beta cell mass and display no beta cell hypertrophy or hyperplasia. The hyperglucagonaemia in A-IDE-KO mice appears to be a primary phenotype produced by a constellation of underlying causes, including augmented alpha cell mass (attributable to alpha cell hyperplasia and hypertrophy) and dysregulation of glucagon secretion. In particular, high glucose failed to inhibit glucagon secretion in A-IDE-KO mouse islets. Similarly, A-IDE-KO mouse islets were unresponsive to the normal inhibitory effect of insulin. Together, these two effects result in a phenotype of constitutively elevated glucagon secretion, closely paralleling the phenotype of constitutive insulin secretion produced by deletion of
Ide from beta cells [
2]. The fact that the common consequence of
Ide deletion in both alpha and beta cells is dysregulation of hormone secretion strongly supports the idea that IDE plays an important functional role in secretory processes [
10,
13]. It is noteworthy that the proteolytic function of IDE does not seem to play a significant role in the regulation of these peptide hormones in vivo as was once assumed.
Surprisingly, A-IDE-KO mice display normal fasting glucose levels and physiological glucose tolerance, apparently attributable to liver glucagon resistance in the form of reduced glucagon signalling. This may represent a compensatory response to chronic exposure of hepatocytes to hyperglucagonaemia that prompts a reduction in glucagon receptor levels and impaired p-Creb/Creb signalling. Glucagon resistance would be circumventing hepatic glucose production and hyperglycaemia in this preclinical model [
29].
There are differences in the effects of IDE deletion on glucagon secretion in vivo vs ex vivo. Although glucagon secretion was increased in both paradigms, isolated islets showed constitutive glucagon secretion that was not inhibited by high glucose or insulin. By contrast, in vivo glucagon levels were significantly attenuated at 5 min after glucose challenge. Two factors may help to account for this discrepancy. First, plasma glucagon levels reflect the balance between glucagon secretion and its clearance in vivo [
30] but clearance mechanisms are not present ex vivo. Second, an important mechanism for in vivo glucagon secretion is hypoglycaemia-induced activity of the pancreatic innervation [
31], which is not operative in isolated islets.
At the level of gene expression, islets isolated from A-IDE-KO mice expressed increased levels of
Arx and
Mafb relative to control islets.
Arx is required for alpha cell development, promoting their specialisation and differentiation, and overexpression of this transcription factor has been strongly implicated in alpha cell hyperplasia [
32].
Mafb is expressed in both alpha and beta cells during endocrine pancreas development [
33,
34] but becomes specific to the alpha cell lineage 2 weeks after birth [
34]. This transcriptional factor has been described as a key regulator of glucagon gene expression [
35]. In alignment with the increase in these two transcription factors,
Gcg transcripts were non-significantly elevated in A-IDE-KO mouse islets. Deletion of
Ide also resulted in increased expression of genes coding for several members of the SNARE protein complex, including
Snap25,
Stx1a and
Vamp2. Because the SNARE complex plays a key role in facilitating the fusion of glucagon granules to the plasma membrane, regulating cellular exocytosis, it is reasonable that these genes would be upregulated to meet the demand of continuous glucagon secretion [
36].
Histomorphometric studies also revealed that alpha cell number and size were increased in A-IDE-KO mouse pancreases, resulting in greater alpha cell mass. Evidence supporting this phenotype being attributable to augmented cell proliferation was provided by independent studies in cultured alpha cell lines with or without siRNA-mediated knockdown of IDE. It is surprising that ~40% reduction in IDE resulted in a ~50% activation of cell proliferation. Based on published studies, one could argue that this effect might be mediated by an interaction between IDE and the retinoblastoma protein (pRb), a tumour suppressor that inhibits cell-cycle progression at the G
1/S transition when interacting with E2F transcription factors [
37]. IDE co-purifies with pRb on proteasomal preparations of breast cancer and hepatoma cells [
38]. Similarly, IDE has been shown to co-immunoprecipitate with the tumour suppressor phosphatase and tensin homologue (PTEN), accelerating its degradation by sirtuin-4 (SIRT4) in response to nutritional starvation stresses [
39]. Although the functional significance of these protein–protein interactions remains to be fully elucidated, these findings are consistent with a functional role for IDE in regulating cell proliferation and, possibly, oncogenesis. Interestingly, we observed that proliferating alpha cells exhibit a significant diminution in the abundance of cilia, an important hallmark of alpha cell differentiation with important functions in beta cells and in paracrine islet signals [
22]. These findings raise several questions. For example, how precisely do cilia contribute to alpha cell proliferation? Are the effects of IDE deficiency on ciliogenesis specific or perhaps symptomatic of a more general effect on cytoskeletal homeostasis? Could the reduction in cilia abundance contribute to the dysregulation of glucagon and insulin secretion? What precisely is the role of IDE in this connection?
Deletion of
Ide in alpha cells (this study) and beta cells [
10,
13] produces dysregulation of glucagon and insulin secretion, respectively, and in both cases also results in increases in the accumulation of oligomeric α-synuclein. IDE binds avidly to monomeric α-synuclein, leading to the formation of stable and irreversible complexes, thereby slowing the formation of higher-n aggregates of α-synuclein. Steneberg et al postulated that
Ide deletion impairs insulin secretion from beta cells by promoting aggregation of α-synuclein, which in turn disrupts microtubule function and impairs secretion processes dependent on the integrity of the cytoskeleton [
10]. We show here that pancreatic alpha cells also express α-synuclein and that levels of oligomeric α-synuclein species are increased in both A-IDE-KO mouse pancreases and IDE-knockdown cells. The hypothesis that α-synuclein oligomer formation leads to cytoskeletal disorders is corroborated by the present study, which found that alpha cells lacking IDE harbour increased α-synuclein aggregates together with decreased levels of acetylated α-tubulin, which is required for microtubule stabilisation and the assembly of primary cilia [
21]. Cilia are microtubule-based structures that protrude from the cell surface and function as sensors for mechanical and chemical ecological cues that regulate cellular differentiation and division [
40]. Beta cell cilia are required for normal insulin secretion [
20] and it has been reported that beta cell cilia loss affects paracrine interactions in the islet and causes altered glucagon and somatostatin secretion [
22]. Notably, Gerdes et al established a link between primary cilia and diabetes in GK rats, finding impaired glucose-stimulated insulin secretion and fewer ciliated beta cells in these animals relative to controls [
20]. Interestingly, GK rats harbour loss-of-function mutations in the
Ide gene [
7], resulting in inhibition of IDE’s ability to degrade amyloid peptides [
41]. In addition, the dynamics of the microtubule network play an important role in pancreatic beta cell secretion. Microtubule depolymerisation by glucose or pharmacological agents enhances insulin secretion by increasing the incorporation of granules at exocytotic sites [
42].
If cilia provide information that serves to retain cells in their functional, differentiated G
0 state, then defects in this pathway could cause proliferative disorders such as cancer [
23]. Furthermore, the absence of cilia has been associated with increased proliferation in several cell types, including beta cells [
43]. Thus, impaired ciliogenesis may underlie the increased proliferation in IDE-deficient alpha cells. In another connection, α-synuclein also interacts with the cytoplasmic terminus of Kir6.2, a major subunit of the of ATP-sensitive potassium channel (K
ATP), common in both beta and alpha cells, inducing impaired insulin secretion [
44]. Whether this interaction between α-synuclein and Kir6.2 occurs in alpha cells and contributes to the observed glucagon secretion dysregulation warrants further study.
There is another important question with respect to elevated α-synuclein levels, since it has been previously shown that maintenance of continuous presynaptic SNARE complex assembly requires a nonclassical chaperone activity mediated by synucleins in neurons. More specifically, α-synuclein directly binds to the SNARE protein VAMP-2 and promotes SNARE complex assembly [
45]. This evidence may explain elevated levels of SNARE complex proteins in A-IDE-KO mice. Indeed, it has been recently reported that treatment of isolated islets with α-synuclein monomers increases glucose-stimulated insulin secretion [
46], suggesting that the effect of α-synuclein on exocytosis also occurs in islet cells.
Our findings indicate that decreased IDE expression in A-IDE-KO mouse islets mainly affected glucagon secretion at steps downstream of Ca
2+ signals [
47‐
49]. Indeed, several molecules involved in the secretory process, such as SNARE proteins, α-synuclein and microtubules, were altered in the A-IDE-KO model. These changes have been associated with facilitated secretion in the pancreatic beta cell. Comparably, the overall effect of these alterations on the alpha cell secretory process could account for the lack of suppression of glucagon secretion at 16 mmol/l glucose in A-IDE-KO mouse islets, despite decreased alpha cell Ca
2+ signalling, since the secretory output may depend on the balance of multiple regulatory factors.
In conclusion, selective deletion of Ide in alpha cells triggers hyperglucagonaemia and alpha cell hyperplasia, resulting in elevated constitutive glucagon secretion. We propose that loss of IDE expression in alpha cells may contribute to hyperglucagonaemia in type 2 diabetes.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.