Mechanisms of Islet Amyloidosis Toxicity in Type 2 Diabetes

Abstract

Amyloid formation by the neuropancreatic hormone, islet amyloid polypeptide (IAPP or amylin), one of the most amyloidogenic sequences known, leads to islet amyloidosis in type 2 diabetes and to islet transplant failure. Under normal conditions, IAPP plays a role in the maintenance of energy homeostasis by regulating several metabolic parameters, such as satiety, blood glucose levels, adiposity and body weight. The mechanisms of IAPP amyloid formation, the nature of IAPP toxic species and the cellular pathways that lead to pancreatic β-cell toxicity are not well characterized. Several mechanisms of toxicity, including receptor and non-receptor-mediated events, have been proposed. Analogs of IAPP have been approved for the treatment of diabetes and are under investigation for the treatment of obesity.

1. Introduction

Amyloids are partially ordered, fibrillar, protein aggregates that are rich in β-sheet structure. Amyloid formation has been implicated in more than 30 different human disorders including such debilitating diseases as Alzheimer’s disease (AD), Parkinson's disease (PD) and type 2 diabetes (T2D) (Table-1). Amyloid formation is not restricted to in vivo pathological conditions; a large number of proteins that do not form amyloid in vivo can be induced to do so in vitro under non-physiological conditions [1–4]. Amyloid can, in some cases, be functional and beneficial [4]. In this review, we focus on amyloid formation by islet amyloid polypeptide (IAPP, amylin), a neuropancreatic hormone that forms pancreatic islet amyloid in T2D and contributes to β-cell dysfunction and death. We first outline the biosynthesis of IAPP and describe its normal physiological roles. We then discuss IAPP amyloid formation, with emphasis on potential mechanisms of toxicity, drawing analogy to proteins and physiological consequences documented in other amyloidosis diseases, which have not yet been characterized for islet amyloid. We conclude with a brief description of the clinical applications of IAPP analogs.

Table 1.

Common pathological amyloidoses and their major protein components.

Amyloidosis Disease	Amyloidogenic Protein	Deposition Type
Alzheimer’s Disease; Inclusion-body myositis; Down's syndrome; Cerebral β-amyloid angiopathy	Amyloid-β peptides (1–40 and 1–42)	Neurodegenerative
Hereditary cerebral haemorrhage with amyloidosis	Mutants of Amyloid-β peptides	Neurodegenerative
Huntington’s disease (HD)	Autosomal dominant mutation of human huntingtin leading to expanded polyglutamine inserts	Neurodegenerative
Parkinson’s Disease and other synucleinopathies	α-Synuclein	Neurodegenerative
Familial amyotrophic lateral sclerosis (ALS); also known as motor neuron disease or Lou Gehrig’s disease	Mutations in Superoxide dismutase (SOD1); TDP-43 and FUS/TLS	Neurodegenerative
Serpinopathies	Mutations in members of the serine protease inhibitor or serpin superfamily of proteins (Serpins)	Neurodegenerative or Local
Bovine spongiform encephalopathies (BSE or Mad Cow disease); Creutzfeldt-Jakob disease	Prions in the Scrapie form (PrpSc)	Neurodegenerative
Intracytoplasmic neurofibrillary tangles; Tauopathies; Alzheimer’s Disease	Tau protein	Neurodegenerative
Icelandic hereditary cerebral amyloid angiopathy (CAA); also known as hereditary cyctatin C amyloid angiopathy	Mutant of cystatin C	Neurodegenerative
Familial British dementia	ABri polypeptide (ABriPP)	Neurodegenerative
Familial Danish dementia	ADan polypeptide (ADanPP)	Neurodegenerative
Type 2 diabetes, Pancreatic islet amyloidosis	Amylin, also known as Islet Amyloid Polypeptide (IAPP)	Local
Aortic medial amyloidosis	Medin (a fragment of lactadherin)	Local
Atrial amyloidosis	Atrial natriuretic factor	Local
Medullary carcinoma of the thyroid (MTC)	Pro-calcitonin	Local
Injection-localized amyloidosis	Insulin	Local
Critical illness myopathy (CIM)	Hyperproteolytic state of myosin ubiquitination	Local
Lichen amyloidosis	Keratins	Local
Restrictive amyloid heart; also known as cardiac amyloidosis, amyloid cardiomyophathy or ApoA-I amyloidosis	Apolipoprotein A-I (Apo-A1)	Local or Systemic
Cataract	Crystallin family of proteins	Local
Pituitary prolactinoma	Prolactin	Local
Pulmonary alveolar proteinosis (PAP)	Pulmonary surfactant protein C	Local
Familial amyloid polyneuropathy (FAP); Familial amyloid cardiopathy (FAC); Senile systemic amyloidosis (SAA)	Transthyretin (TTR)	Systemic
Familial amyloidosis of Finnish type (FAF)	Fragments of gelsolin mutants	Systemic
Amyloid light chain(AL) amyloidosis; also known as Primary systemic amyloidosis (PSA)	Immunoglobulin light chains	Systemic
Amyloid heavy chain(AH) amyloidosis	Immunoglobulin heavy chains	Systemic
Dialysis-related amyloidosis	β2-microglobulin (β2m)	Systemic
Corneal amyloidosis associated with trichiasis	Variation of lactoferrin (LF)	Local
Hereditary lattice corneal dystrophy	Mainly C-terminal fragments of kerato-epithelin	Local
AA amyloidosis or Secondary amyloidosis (associated with inflammatory disorders such as tuberculosis, rheumatoid arthritis, bronchiectasis, ulcerative colitis, Crohn’s disease, renal cell carcinoma, ankylosing spondylitis, nephritic syndrome, chronic osteomyelitis, Hodgkin disease, familial Mediterranean fever)	Serum amyloid A (SAA) protein	Systemic
Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL)	Mutations in the Notch3 gene	Neurodegenerative or Systemic
ApoA-II amyloidosis	Apolipoprotein A-II (ApoA2)	Systemic
ApoA-IV amyloidosis	N-terminal fragment of apolipoprotein A-IV (ApoA4)	Systemic
Fibrinogen amyloidosis	Variants of fibrinogen α-chain	Systemic
Lysozyme amyloidosis	Mutants of lysozyme	Systemic

The kinetics of amyloid formation is complex and displays a sigmoidal profile with three observable phases. The initial steps of aggregation, which lead to formation of an active seed, occur in the lag phase and represent the rate limiting process. In this phase, monomers oligomerize and convert into species that nucleate an exponential fibril growth phase. Fibrils elongate by addition of peptide to their ends. Secondary nucleation that involves the catalyst of fibril formation from existing fibrils also occurs. This may involve breakage of existing fibrils to increase the concentration of free ends, or the templating of new fibrils on the surface of existing ones. Finally, a steady state is reached where soluble peptide is at equilibrium with amyloid fibrils. Off-pathway steps leading to amorphous aggregates also occur. Amyloid formation can be accelerated by the addition of small amounts of preformed fibrils in a process known as “seeding” (Figure 1A).

Figure 1. Amyloid formation by IAPP.

(A) Schematic diagram of amyloid formation (solid blue curve). During the lag phase monomers associate to form oligomeric species which then assemble to nucleate an exponential growth phase. Secondary nucleation and off-pathway steps such as formation of amorphous aggregates are omitted for clarity. Amyloid formation can be accelerated by the addition of small amounts of preformed fibrils (dashed red curve). (B) The primary sequence of human IAPP. The peptide has a free N-terminus, an amidated C-terminus and an intramolecular disulfide bond between residues 2 and 7.

Although there is no sequence homology or structural similarity between the proteins that form amyloid, all amyloid deposits share common characteristics. Amyloid fibrils are typically unbranched, 5–10 nm in width, variable in length, polymorphic, and form a cross β-sheet structure. This structure is defined by perpendicular orientation of the individual polypeptide chains to the long axis of the fibril; with the interchain hydrogen bonds aligned parallel to the long axis [4]. Amyloids can form from proteins that fold to a compact tertiary structure in their unaggregated state, or they can originate from ‘intrinsically disordered polypeptides’ that fail to adopt compact tertiary structures in their soluble native state [4]. Amyloid formation from a folded precursor normally involves global or local unfolding to an aggregation prone state, while intrinsically disordered proteins can aggregate directly from their native ensemble. Well studied examples of globular proteins that form amyloid include β2-microglobulin, transthyretin (TTR) and mutants of human lysozyme. Stabilization of the native fold can prevent amyloid by this class of proteins; this is the basis of the first clinically approved, rationally designed, small molecule anti-amyloid agent [5]. Aβ, α-synuclein, and IAPP are important examples of intrinsically disordered polypeptides that form amyloid in vivo [2, 6].

Amyloidosis diseases can be divided into three broad classes: neurodegenerative, systemic and local amyloidosis [7]. In neurodegenerative diseases, amyloids are deposited in the brain and spinal cord. Important examples of this class include AD, PD and Huntington’s disease (HD). In systemic amyloidoses, such as in the case of amyloid light-chain (AL) amyloidosis, aggregation occurs in multiple organs and tissues. A subset of systemic amyloidoses includes lysozyme amyloidosis; senile systemic amyloidosis; familial transthyretin-associated amyloidosis, which arises from deposition of wild-type or one of more than 50 mutated forms of TTR; and diseases of chronic inflammation, in which an N-terminal fragment of the acute phase protein serum amyloid A (SAA) forms amyloid deposits. In the non-neurological, localized amyloidoses, deposition of amyloid occurs in one target organ, usually proximal to the production site of the amyloidogenic peptide. Important examples of this third class include amyloid formation by the crystallins associated with cataract; atrial amyloid, caused by atrial natriuretic factor; and islet amyloidosis in T2D. The association between amyloid formation and disease pathogenesis is common for all amyloidoses and the cytotoxic properties of amyloidogenic peptides may be similar [8].

2. IAPP is one of the most amyloidogenic sequences known

2.1 Biosynthesis of IAPP

IAPP is synthesized as an 89 residue pre-prohormone. Removal of the 22 residue signal sequence leads to the 67 residue pro-IAPP, which is further processed in the Golgi and in the insulin secretory granule to the 37 residue mature hormone [9, 10]. Additional post translational modifications include formation of an intramolecular disulfide bridge between residues 2 and 7, and amidation of the C-terminus (Figure 1B). Mature IAPP is stored in the insulin secretory granule at a ratio of 1:50 to 1:100 relative to insulin and is co-secreted with insulin [11].

IAPP is also subject to spontaneous, non-enzymatic, post-translational modifications that may impact its function and its tendency to aggregate. For example, the human polypeptide contains six chemically liable Asn residues which can undergo spontaneous deamidation to yield Asp or iso-Asp residues. This modification results in the replacement of a neutral amide side chain with a carboxyl group and will lower the net charge of human IAPP at physiological pH, and presumably decrease its solubility. Deamidation of human IAPP accelerates amyloid formation in vitro, but its role in vivo is not understood [12].

2.2 The primary sequence of IAPP correlates with in vivo amyloidogenicity

IAPP is a member of the calcitonin like family of polypeptides and has been found in all animals studied, although not all species form amyloid [13, 14]. Mice and rats do not develop islet amyloid, but cats, non-human primates and humans do. The human polypeptide is extremely amyloidogenic in vitro, while rat IAPP is not, even though the two polypeptides differ at only six positions. Notably, rat IAPP contains three proline residues within the 20 to 29 sequence and these are believed to be responsible for its inability to form amyloid. Much attention has been focused on the sequence within the 20 to 29 region and the role it plays in controlling amyloid formation. This portion of the polypeptide chain is considered to be one of the major determinants of the ability of IAPP variants to form amyloid. The characterization of designed variants of IAPP support this conjecture; analogs of IAPP which contain prolines or N-methylated residues at other positions within the 20 to 29 segment are considerably less amyloidogenic than wild type human IAPP [15–17]. Proline and N-methyl amino acids are well known to disrupt β-sheets. There is one known mutation in vivo within this region, a Ser20Gly substitution, which enhances amyloid formation [18, 19]. While there is a strong correlation between the primary sequence of the 20 to 29 segment and in vitro amyloidogenicity, mutations outside of this region can abolish amyloid formation, indicating that it cannot be the sole factor controlling IAPP’s amyloidogenicity [20–22].

2.3 Physiological function of IAPP

The physiological roles of IAPP are receptor mediated. The IAPP receptor is formed from a complex of the calcitonin receptor with a receptor activity modifying protein (RAMP) [23, 24]. IAPP binds the calcitonin receptor (CTR) in the absence of RAMPs, but the affinity is low. The affinity of IAPP for the CTR-RAMP complex is higher, with an IC₅₀ reported to be on the order of 8 nanomolar for the CTR-RAMP-1 complex [25]. Six different subtypes of the IAPP receptor are generated by different combinations of the two splice variants of the calcitonin receptor with different RAMPs, but the distribution of the subtypes is not fully characterized [24, 25].

Circulating concentrations of IAPP have been reported to be between 3–5 picomolar in rats, rising to 15–20 picomolar with increased blood glucose levels [26]. However these values are unlikely to be relevant to amyloid formation since IAPP is stored at a much higher level in the insulin secretory granule, between 500 micromolar to several millimolar. This implies that the local concentration of IAPP after release from the granule will be temporarily much higher than the circulating concentration.

The normal physiological roles of soluble IAPP are not completely understood in humans, but studies in rodent models show that IAPP is involved in the suppression of satiety and adiposity, as well as the regulation of glucose homeostasis via inhibition of glucose-stimulated insulin secretion (GSIS), gastric emptying, suppression of glucagon release, vasodilatation, and the excretion of calcium, potassium and sodium [26–35]. IAPP’s anorectic effect appears to be mediated mainly at the area postrema (AP) of the CNS. Several recent reviews provide a critical and in depth examination of the physiological roles of IAPP [26–29, 34, 35].

3. IAPP impacts T1D, T2D and islet cell transplantation

Amyloid accumulates in the pancreatic Islets of Langerhans in the majority of individuals with T2D. Pancreatic amyloid deposits were first described more than 100 years ago, but the protein component was not identified until much later when, in 1987 two groups independently isolated a 37 residue polypeptide from ex vivo samples of pancreatic amyloid [36–38]. Interest in islet amyloid has undergone a resurgence in the last ten years with the realization that β-cell dysfunction and the loss of β-cell mass are key features of T2D [39]. The decline in β-cell mass and function is attributed to several factors, including islet inflammation, glucolipotoxicity, accumulation of cholesterol and islet amyloid formation [40–43].

Amyloid deposition is also an important factor in the failure of islet cell transplants, and correlates with graft failure [44–46]. IAPP amyloid forms rapidly upon transplantation of human islets into nude mice and occurs before the recurrence of hyperglycaemia; this is correlated with the loss of β-cells [44, 45]. Islet amyloid has been detected in transplanted human islets in a patient that suffered islet graft failure [46]. Conversely, prevention of amyloid formation by transplantation of porcine islets prolongs islet graft survival [47].

In T2D, formation of islet amyloid by IAPP is a significant problem and the clinical goal is to inhibit amyloidosis-induced toxicity. The situation is different in type 1 diabetes (T1D); here the issue is a lack of production of adaptive IAPP. IAPP is produced and released with insulin by pancreatic β-cells, thus it is absent in T1D, as is insulin. The absence of physiologic concentrations of IAPP in T1D and late stage T2D may have deleterious effects, motivating the development of non-toxic bioactive analogs of human IAPP for the clinical goal of hormone replacement therapy, as is done with insulin. One analog has been approved as an adjunct for insulin therapy for the treatment of diabetes and its administration is reported to improve glycemic control. In contrast, there are no clinically approved inhibitors of pathologic IAPP amyloid formation indicated for T2D [48]. The nature of toxic species produced during IAPP amyloidosis is not known, making it difficult to design therapeutics that target the pathological form of the polypeptide. Currently, drug design aimed at preventing amyloid formation is an active area of research.

3.1 Islet amyloid includes other components which may influence the kinetics of amyloid formation and fibril stability

In vivo, amyloid formation takes place in a heterogeneous environment with the potential for interactions with molecules of the extracellular matrix and with membranes, as well as with other factors. Like other amyloid deposits, islet amyloid contains serum amyloid P component (SAP), apolipoprotein E (apoE), and the heparan sulfate proteoglycan (HSPG) perlecan [49–52]. There is a well documented correlation between the e4 allele of the apolipoprotein E gene (apoe4) and AD [53]. The e4 allele plays an important role in AD and has more widespread effects than any other genetic factor that has been implicated in sporadic, late-onset AD. Studies with apoE knockout mice have shown that this is not the case for T2D [51]. Interactions between SAP and IAPP also do not appear to play a role in amyloid deposition, but interactions with the glycosaminoglycan (GAG) component of HSPGs might. GAG chains of HSPGs have been shown to significantly accelerate amyloid formation by IAPP and partially processed forms of pro-IAPP in vitro [54].

One model for IAPP amyloid formation in vivo assigns an important role to the interaction of incorrectly processed proIAPP with perlecan [55]. Perlecan is a component of the extracellular matrix and is associated with islet amyloid. A fraction of the IAPP that is secreted in T2D is incompletely processed and includes the N-terminal flanking peptide. Impaired processing leads to increased amyloid [56, 57], and the processing intermediate, denoted here as IAPP-Npro, is found in islet amyloid. IAPP-Npro is less amyloidogenic in solution than mature IAPP, but it interacts more effectively with GAGs. Interactions with model GAGs significantly enhance the rate of in vitro amyloid formation by IAPP-Npro and the resulting fibrils are able to seed amyloid formation by fully processed IAPP [58]. Hence release of increased amounts of incorrectly processed IAPP-Npro which bind to GAGs could generate a high local concentration of the polypeptide and initiate amyloid formation. These deposits might recruit mature IAPP and additional IAPP-Npro. Are interactions with GAGs important in vivo? The answer is not known, but over-expression of heparanase inhibits amyloid formation in a transgenic mouse that over-expresses IAPP, and the inhibition of GAG synthesis in cultured islets has been demonstrated to reduce amyloid deposition [59, 60].

Binding of HSPGs to amyloid deposits could influence their stability and their clearance. Salts accelerate IAPP amyloid formation and also stabilize IAPP fibrils by reducing electrostatic repulsion [61, 62]. The polyanionic GAG chains of HSPGs could be even more effective than simple salts since the spacing of the negative charges in GAGs can match the spacing of the positively charged sites in amyloid fibrils [54]. The effect of GAGS on IAPP amyloid stability and on the clearance of islet amyloid is a largely unexplored area.

3.2 The initiation site of islet amyloid formation is controversial

Determining if islet amyloid originates intracellularly or extracellularly is important because it will directly impact therapeutic approaches. The location of the initial site of islet amyloid deposition in vivo is currently controversial [27, 63, 64]. Amyloid deposits found in T2D appear to be extracellular and initial histological studies with rodent models argue in favor of an extracellular origin. Transgenic rodent models that over-express human IAPP are consistent with an intracellular origin [63]. In contrast, work with a cultured islet model is consistent with an extracellular origin of islet amyloid [64]. That study showed that the secretion of IAPP is an important factor in islet amyloid formation and β-cell toxicity. The effect of reagents that increased IAPP secretion, but did not increase the amount of IAPP produced, were used together with reagents that inhibited IAPP secretion, but maintained the level of production of IAPP. Increasing secretion increased amyloid formation and toxicity, while inhibiting secretion reduced amyloid formation and toxicity [64]. The conflicting results may be related to the level of IAPP produced and to the techniques used in the detection of amyloid [27, 63–65].

4. Mechanisms of toxicity in other amyloidosis–potential lessons for islet amyloidosis

A range of mechanisms has been proposed for the general toxic effects of amyloidosis, however, the exact mechanisms of cell death are still not completely clear. In some cases, amyloid fibril deposits physically disrupt tissue architecture and lead to organ dysfunction, however, in most cases, activation of multiple overlapping cellular mechanisms and downstream signaling pathways have been proposed to lead to disease pathogenesis. These include receptor-mediated interactions and non-receptor mediated phenomena.

Non-receptor based mechanisms have primarily focused on membrane disruption and permeabilization by soluble oligomers. Membrane disruption leads to an increase in intracellular Ca²⁺ and has been shown to activate several pathogenic pathways, including production of reactive oxygen species (ROS) [66], altered signaling pathways [67, 68] and mitochondrial dysfunction [69]. Membrane permeabilization by amyloid oligomers may also induce an oxidative stress response in cells [66, 70]. Permeabilization of mitochondrial membranes by amyloid oligomers and the accumulation of Ca²⁺ in the matrix of mitochondria lead to an increase in ROS production, cytochrome C release and apoptosis [71, 72]. ER stress has also been proposed to contribute to cytotoxicity [73].

Receptor-mediated mechanisms of toxicity include FAS (also known as APO-1, APT or CD95), p75NTR (p75 neurotrophin receptor) and RAGE (receptor for advanced glycation end products). In particular, FAS and p75NTR have been investigated in IAPP and Aβ toxicity [74, 75], while RAGE has been shown to engage amyloidogenic species of Aβ1–40 and Aβ1–42, serum amylin A (SAA) and prion-derived peptide, among others [76]. RAGE was originally named for its ability to bind advanced glycation end products (AGEs), but is now recognized as a multi-ligand pattern recognition receptor with several classes of ligands including amyloid forming polypeptides and proteins. RAGE not only elicits signaling pathways that lead to inflammation and apoptosis, but is also involved in internalization of bound amyloidogenic ligands. RAGE-Aβ interactions have received prominence [76–78]; however, the role of RAGE in islet amyloidosis has not been investigated.

Upregulation of autophagy has been recognized as a common protective response to the accumulation of toxic amyloidogenic aggregates in degenerative diseases [79–82]. However, autophagocytosis and lysosomal degradation of misfolded amyloidogenic polypeptides and proteins is not entirely successful as accumulation of amyloidogenic aggregates also leads to autophagy-mediated lysosomal dysfunction and cell death [80, 81].

Chronic inflammation may be an important contributing factor to amyloidosis protein toxicity as it is frequently observed in local and systemic amyloidosis diseases [83–85]. In neurodegenerative diseases, such as AD, PD and HD, activation of microglia [86] and pro-inflammatory processes (e.g. production of cytokines, chemokines, nitric oxide (NO), ROS, and arachidonic acid metabolites) can lead to oxidative stress, mitochondrial dysfunction and neural toxicity [87–90]; however, the exact mechanisms of cell death are controversial. Aβ-induced activation of NF-kappaB, a potent immediate-early transcriptional regulator of numerous proinflammatory genes, has been shown to be stimulated by ROS formation in primary neurons derived from AD patients [91]. Activation of the inflammasome protein complex by misfolded protein aggregates has also been implicated in the pathogenesis of several amyloidosis diseases such as AD and ALS, providing a common mechanism of IL-1β cytokine production in these diseases [92].

Amyloid oligomers, which precede mature fibrils, have been implicated, albeit indirectly, in the pathogenesis of several diseases, including AD, AL and TTR amyloidosis in vitro and in vivo; and have been shown to trigger oxidative stress and activation of apoptotic pathways [93–96]. The identity and nature of the specific toxic species has not been addressed for islet amyloidosis.

Disease-specific events and pathways characterize amyloidoses. Mutations that lead to the destabilization, unfolding and/or misfolding of proteins, are commonly associated with inherited diseases. Mutations in proteins can also enhance the amyloidogenicity of polypeptides and proteins resulting in acceleration or increase in the formation of toxic amyloidogenic species. Other disease-specific mutations can occur in genes that do not encode proteins that accumulate as amyloids, but rather impact the production of amyloidogenic proteins. Examples include the presenilins in AD and parkin in PD. These mutations can have the effect of increasing the concentration of aggregation prone proteins and polypeptides. Presenilin mutations alter the proteolytic processing of APP, leading to an increase in the more aggregation prone Aβ-42 isoform [97]. Covalent modification of proteins, such as oxidation, glycation and racemization can also promote unfolding or mis-folding and may contribute to amyloid formation [95].

These common pathological mechanisms in amyloidogenic diseases may stem from common properties that are shared among toxic amyloidogenic species [8, 95, 96]. Ongoing efforts to elucidate the mechanisms of cytotoxicity and tissue damage by other amyloid proteins may ultimately redirect therapeutic efforts in islet amyloidosis.

5. Mechanisms of IAPP-induced β-cell toxicity in islet amyloidosis

A subset of the mechanisms described above has been investigated in the context of islet amyloidosis. Cell membrane permeabilization or disruption by IAPP aggregates has been suggested to be a mechanism of toxicity. Other mechanisms of IAPP toxicity include localized islet inflammation, defects in autophagy, ER stress, as well as receptor-mediated mechanisms involving FAS and the activation of downstream signaling pathways such as cJUN N-terminal kinase (JNK). The available data suggest that IAPP exerts its toxic effects on β-cells by multiple mechanisms. Several of these overlap and share common signaling pathways [Figure 2].

Figure 2. Physiological and pathophysiological effects of IAPP.

According to this scheme, amyloidosis by adaptive IAPP leads to the production of cytotoxic species. The toxic species of IAPP may form either intracellularly or extracellularly, leading to a series of parallel, overlapping and potentially additive or synergistic pathogenic pathways.

Amyloid formation by IAPP has been shown to induce apoptosis in cell culture and in isolated human islets [65, 98–101]. The pathways that lead to IAPP induced β-cell apoptosis are not yet fully elucidated, although a growing body of literature and progress is being made [102, 103]. The JNK pathway, a stress activated pro-apoptotic pathway in β-cells, mediates apoptosis in cultured β-cells and in islets that have been exposed to exogenous IAPP. Importantly, it has also been shown that the JNK pathway is activated in response to amyloid formation from endogenous IAPP [103].

It has been proposed that IAPP might exert nonspecific cytotoxicity by permeabilizing cell surface membranes [104, 105]. Numerous studies have demonstrated clustering of IAPP amyloid fibrils on or near membranes in vivo, and exogenous IAPP disrupts cell membranes in vitro. The ability of IAPP to induce ion leakage depends on the lipid composition of the membrane and the ratio of lipid to peptide examined. An important caveat may be that most of the in vitro membrane model systems employed in biophysical studies of membrane leakage utilize much higher fractions of anionic lipids, and very different types of lipids than those found innately in β-cells [106]. These membrane mimetics lack cholesterol and gangliosides, which have been shown to play a role in mediating IAPP clearance and membrane interactions [107]. Studies have shown that IAPP variants that are not toxic to β-cells in vivo have the ability to disrupt some of these in vitro model membrane systems. Permeabilization and loss of membrane integrity by IAPP may indeed be one mechanism of toxicity, particularly at high peptide concentrations, however, caution should be taken when extrapolating from studies that employ non-physiological model membranes and peptide concentrations to the more complicated situation in vivo. More multifarious model membranes are now being employed and are expected to provide new mechanistic insights under more physiologically relevant conditions [106].

Other proposed mechanisms of IAPP-induced β-cell death include defects in autophagy, local inflammation, mitochondrial dysfunction and receptor-mediated mechanisms linked to oxidative stress, cytokine production and activation of signaling cascades leading to apoptosis [73, 102, 103, 108–118]. These toxic mechanisms have been shown to be activated by either intracellular or extracellular aggregates. The pro-apoptotic JNK pathway mediates β-cell apoptosis in cultured cells and in islets exposed to high concentrations of IAPP, and has recently been shown to become upregulated in response to endogenous IAPP amyloid formation [103]. In β-cells, JNK becomes activated by a range of events leading to cellular stress, including ER stress, ROS formation and oxidative stress, increases in glucose concentration and the production of pro-inflammatory cytokines. Similar downstream signaling pathways have also been found to become activated in both the intrinsic (Bim) and extrinsic (Fas, Fadd) pathways. Interaction of either endogenous or exogenous IAPP aggregates with FAS, also known as the death receptor, leads to caspase 3 activation, while deletion of Fas protects β-cells from IAPP toxicity [74]. These studies are supported by in vivo experiments that demonstrate that inhibition of caspase 3 protects β-cells from IAPP-induced β-cell apoptosis [119]. The downstream signaling pathways that regulate β-cell death in response to IAPP amyloidosis are not fully characterized, and more work is needed to understand how many of these pathways intertwine and overlap, and whether these intracellular consequences are triggered by interactions with cell surface receptors and/or membrane disruption.

Defects in autophagy have been shown to play a role in amyloidosis diseases and have been proposed to be a factor in IAPP toxicity. Chaperone-mediated autophagy and macroautophagy ensure synthesis of cellular components, recycle damaged or dysfunctional organelles, and clear ubiquitinated proteins. Studies have shown that activation of autophagy protects β-cells from IAPP-induced apoptosis, while inhibition of autophagy-lysosomal digestion promotes IAPP toxicity [110]. Over expression of IAPP in β-cells, before the development of hyperglycemia, has been shown to impair autophagy [110, 115]. Thus, impairment in autophagy could lead to the build up of toxic aggregates and promote β-cell death.

Local islet inflammation induced by toxic forms of IAPP may play a role in β-cell dysfunction and death via activation of pro-inflammatory responses [112, 114]. IAPP has been shown to activate inflammasomes, which are multi-protein caspase activating complexes that have been implicated in metabolic disease. Inflammasome activation triggers signaling cascades leading to the production of pro-inflammatory cytokines such as Interleukin-1β (IL-1β) [112]. IL-1β has been reported to be a mediator of IAPP-induced β-cell toxicity, however the exact source(s) of IL-1β production associated with islet amyloidosis in vivo is still an open question.

Defects in endoplasmic reticulum associated protein degradation (ERAD), unfolded protein response (UPR) and ER stress have been reported to induce β-cell death by IAPP aggregates produced both intracellularly and extracellularly [108, 109, 116, 117]. In the case where toxicity arises from intracellular aggregate formation, ProIAPP and not mature IAPP may be the deleterious species, as proIAPP miss-processing has been shown to occur in diabetes and post-translational modification is completed in the Golgi and insulin secretory granules [56, 57]. The exact role of ER stress in IAPP-induced β-cell dysfunction in vivo is currently not well understood. Studies using transgenic animal models that significantly overexpress IAPP support a role for ER stress, while no ER stress was detected in cultured islets expressing lower levels of IAPP [111].

One of the major unresolved issues in the field of IAPP biology is the difficulty in differentiating between the functional and toxic forms of the polypeptide, making it difficult to determine whether the outcome of a particular experiment is relevant to the physiological or pathophysiological situation. This problem is compounded by the large variation in the methods used to prepare IAPP by different workers. Small variations in peptide concentration, residual buffers and co-solvents can potentially alter peptide secondary structure, affecting stability, aggregation kinetics and potentially activation of off-target cellular stress responses. Caution should be applied when interpreting studies that use uncharacterized IAPP for in vitro or ex vivo investigations. Structural and biochemical characterization of amyloidogenic polypeptides and proteins is traditionally carried out by specialized spectroscopic techniques primarily utilized by biophysicists and chemists. The current gap between the disciplines of biology and biophysical chemistry hinder research efforts, not only in the field of islet amyloidosis, but in the study of amyloidosis diseases in general.

6. Therapeutic applications of IAPP for the treatment of diabetes and obesity

IAPP is deficient in individuals suffering from T1D and advanced T2D. Co-administration of IAPP with insulin helps to normalize fluctuating glucose levels to a greater degree than is possible with insulin alone [48, 120]. However, the extreme amyloidogenicity of human IAPP prevents its direct use as an adjunct to insulin therapy. Consequently, a non-amyloidogenic analog of human IAPP, denoted as Pramlintide, which contains proline substitutions at the same positions found in rat IAPP, was developed and has been approved by the FDA for use in the treatment of diabetes [48].

IAPP and IAPP analogs, such as Pramlintide, are also being explored for the treatment of obesity. Particularly exciting is the potential of combining leptin and IAPP [121–123]. Leptin is an adipokine that plays a major role in maintaining energy homeostasis. The protein acts on the CNS to suppress appetite and alters metabolism in peripheral tissue. Treatment of leptin deficient (ob/ob) mice with leptin has been shown to reduce food intake and lead to weight loss. Individuals that lack functional leptin suffer from extreme obesity, which can be reversed by treatment with leptin [121]. Unfortunately, administration of exogenous leptin does not lead to weight loss or reduction of food intake for obese individuals who generate normal leptin, and the restoration of leptin responsiveness in obese individuals is difficult. This has led to the concept of leptin resistant obesity, where obese individuals are insensitive to high circulating concentrations of leptin. The molecular mechanism of leptin signaling has been extensively studied, although the origins of leptin resistance are not fully characterized. Binding of the protein to its receptor stimulates Janus kinase 2 (JAK2) and activates several pathways, including the activator of transcription and signal transducer pathway (STAT3, STAT5) which are important for the effects of leptin on body weight [122–124]. Leptin-induced signaling is negatively controlled by feedback systems that involve protein tyrosine phosphatase1B and suppression of cytokine signaling-3. The negative feedback loop serves to prevent prolonged leptin receptor activation. Leptin-induced signaling has been recently reviewed [124].

Weight-lowering effects mediated by IAPP have been documented in obese rats and humans. IAPP and IAPP analogs have been shown to reduce food intake without producing signs of conditioned taste aversion or visceral illness [121, 125]. Animal studies with food-matched controls led to the notion that IAPP-induced weight loss occurs via mechanisms similar to those found with enhanced leptin sensitivity [122, 125]. Leptin and IAPP have been proposed to affect energy homeostasis synergistically [121]. This suggests that co-administration of leptin and IAPP might be beneficial. Administration of leptin to leptin resistant, diet-induced obese rats showed that leptin by itself is ineffective at inducing weight loss. Treatment with IAPP alone led to modest loss in body weight. However, the sustained use of the two polypeptides led to significant effects [121, 123]. A recent report suggests that some of the effects of IAPP and leptin might be additive rather than cooperative [126]. In that study, administration of IAPP and leptin was shown to activate extracellular-regulated kinase (ERK), STAT3, AMP-activated protein kinase (AMPK) and the AKT signaling pathways in an additive, but not synergistic fashion, and the effects were abolished by ER stress. The co-administration of IAPP and leptin is still potentially an attractive therapeutic strategy in the absence of strong synergy since even additive effects could be beneficial. Other gastric satiety signals have also been shown to have beneficial effects when administered with leptin, and this is an active area of research.

7. Perspectives, puzzles and future directions

Increasing evidence supports an adaptive role for IAPP in the regulation of adiposity and energy homeostasis. Under normal circumstances, this neuropancreatic hormone is co-produced, co-stored and co-secreted with insulin from β-cell secretory granules into the circulation. In T1D and late stage T2D, lack of production of adaptive IAPP may have deleterious consequences, while in early T2D, overproduction and misprocessing of IAPP, along with other factors, lead to islet amyloidosis and β-cell dysfunction and cell death. Progress has been made in understanding the pathological effects of IAPP amyloid formation in vivo, however many unanswered questions still remain. These include the mechanisms of islet amyloid formation and cytotoxicity in vivo and in vitro; the initial site of amyloid deposition in vivo; the role of membranes and HPSGs in vivo, the properties of the toxic species; and mechanisms of clearance. Major hurdles that hinder progress include a lack of physiologically relevant model systems for biophysical studies, and the lack of detailed biophysical characterization of IAPP prior to testing in biological assays. The latter is essential in order to differentiate between adaptive and toxic forms of IAPP. Bridging the gap between the disciplines will be important for accurate interpretation of ex vivo biological and pharmacological studies of IAPP using cultured cells and islets, as well as for identification of therapeutic targets and the design of pharmacological agents for the treatment of islet amyloidosis.