Potential of peptides and phytochemicals in attenuating different phases of islet amyloid polypeptide fibrillation for type 2 diabetes management

Rlit O. Abioye, Chibuike C. Udenigwe,b,*

a Department of Chemistry and Biomolecular Sciences, Faculty of Science, University of Ottawa, Ottawa, Ontario, K1N 6N5, Canada

b School of Nutrition Sciences, Faculty of Health Sciences, University of Ottawa, Ottawa, Ontario, K1H 8M5, Canada

Keywords:

Islet amyloid polypeptide

Fibril formation

Disaggregation

Bioactive peptides

Phytochemicals

ABSTRACT

Islet amyloid polypeptide (IAPP), or amylin, has been identified as a key factor in the development of type 2 diabetes (T2D). IAPP aggregates, which form amyloid fibrils, contribute to cytotoxicity of the pancreatic β-cells, resulting in loss of function and subsequent reduction in insulin production. As a result, surviving β-cells overcompensate for this reduction of insulin production, further contributing to the loss of function because of increased stress, thus leading to insulin resistance. Endogenously, IAPP monomers function in a variety of roles; however, aggregation renders them non-functional. The use of naturally occurring compounds, including peptides and phytochemicals, has been explored as a way to mitigate or inhibit IAPP fibril formation. This review discusses the structure, endogenous roles and mechanism of IAPP fibril formation, recent advances on inhibitors of IAPP fibril formation, and new insights on the future development and application of foodderived inhibitors towards T2D management.

1. Overview of the structure and role of islet amyloid polypeptide in type 2 diabetes

1.1 Structural information

Islet amyloid polypeptide (IAPP), also known as amylin, is a 37-residue basic polypeptide [1-4]. As a result of the amidated C-terminus, this polypeptide lacks any negatively charged groups. In addition, the presence of four positively charged groups (N-terminus,Lys1, Arg11, and His18) contributes to the cationic nature of IAPP,with net charge ranging from +2 to +4 depending on the pH [1]. The structure of IAPP includes a rigid disulfide bridge between Cys2 and Cys7, embedded within the membrane-binding domain spanning from residues 1-19 of the N-terminus [2,3]. Residues 8-19 form the highly transient random coil structure that interchanges with an α-helical conformation in aqueous solution, while the central 20-29 residues govern the amyloidogenic region, which plays an instrumental role in fibril formation [3,4]. Lastly, the C-terminal region is involved in peptide self-association [3].

The aggregation of monomeric IAPP, forming protofibrils and subsequent mature fibrils, is a spontaneous and favorable reaction in vivo. This transformation consists of a change from the intrinsically disordered IAPP monomer to the highly ordered β-sheets, forming amyloids of varying sizes [1,3]. Each monomer adopts a U-shaped confirmation with residues 18-27 embedded within the turn region between the two β-strands of a monomer [1,5]. Initial protofibrils are composed of two C2-symmetric stacks of monomeric IAPP that lack interchain hydrogen bonding or any other charge-charge interactions across the axis between the proto fibrils. However, hydrogen bonding occurs within monomers of a single stack running parallel to the fibrillary axis [1].

1.2 Endogenous roles

IAPP is a small, neuroendocrine hormonal peptide that is copackaged, co-localized and co-secreted with insulin at an IAPP/insulin ratio of 1:100 by the pancreatic β-cells in the islets of Langerhans[2,3,6-8]. It is stored in secretory granules along with insulin and released to facilitate glucose metabolism in events of elevated blood glucose levels by inhibiting glucagon secretion [7,9]. IAPP plays several regulatory roles for maintaining post-prandial glucose levels by suppressing gastric emptying, and modulating food intake and glucagon secretion [2,10]. Its effects in centrally stimulating satiety suggests the presence of potential IAPP binding sites within the brain cells [10].

1.3 IAPP turnover

Under normal circumstances, IAPP routinely undergoes proteolytic degradation via the ubiquitin proteasome system(UPS) where the 26S proteasome complex plays a pivotal role in IAPP turnover [11]. UPS recognizes and degrades damaged and dysfunctional proteins as well as fully functional proteins that are no longer needed [12]. Routine turnover of IAPP helps to prevent intracellular accumulation and cytotoxicity to pancreatic β-cells.During type 2 diabetes (T2D), however, there is a marked decrease in proteasome activity where the polyubiquitinylated proteins become more abundant, indicating functional impairment of the proteolytic complex [11].

Autophagy-lysosomal system (ALS) is the second most abundant degradation system regulating IAPP turnover [12]. While UPS degrades proteins exclusively, the dynamic membrane recognition system of ALS allows for the degradation of a broader range of macromolecules, in addition to proteins and their aggregates, such as lipids and DNA [12]. With respect to IAPP degradation, the dynamic membrane recognition allows for the identification of IAPP macromolecules in need of degradation. Upon identification of IAPP,uptake via autophagy is induced, followed by intercellular transport to specialized lysosomes called autophagosomes [12]. Autophagosome accumulation occurs because of rapid encapsulation offibrils by the aforementioned autophagosomes in event of overwhelmingly high levels offibril formation. In order for degradation to occur, lysosomes must fuse with autophagosomes to form autophagolysosomes.These newly formed specialized lysosomes are able to degrade ubiquitinylated fibrils through a process known as aggrephagy [13].However, under stress conditions, such as rapid fibril formation, the rate of autophagosome maturation is overburdened. This negatively impacts the rate of amylin degradation in the process and further increases the stress experienced by the pancreatic β-cells [12,13].

1.4 Role in T2D development

Overwhelming evidence points to IAPP cytotoxicity in T2D development, making it an attractive target for T2D mitigation. While IAPP fibrillation is not considered to be the primary cause of T2D, it is still found to be a contributing factor responsible for the decreased level of insulin production by pancreatic β-cells [2]. IAPP amyloid deposits have been observed in over 90%-95% of T2D patients,and its presence has been linked to membrane damage of β-cells,causing cytotoxicity [2,7,8,14-16]. The damages caused by IAPP amyloid deposits have been correlated with a loss in β-cell mass and function, resulting in a marked decrease in insulin production[2,4]. This subsequent loss of β-cell mass and function serves as the initiator of chronic insulin resistance [4,17]. Pancreatic β-cell function has also been shown to be spatially correlated with amylin protein deposition where β-cell mass is strongly reduced in islets containing IAPP deposits while nearby islets lacking amylin remain unaffected [15,18].

2. IAPP fibril formation

2.1 Mechanisms of formation

IAPP fibril formation follows the classical aggregation process consisting of a lag/nucleation, elongation/growth, and stationary phases (Fig. 1). However, prior to fibril formation, misfolding or unfolding must occur, making the nucleation phase the rate determining step in IAPP fibril formation [4,19]. IAPP is intrinsically transient in its native α-helical state, making unfolding or misfolding highly favorable. The rate-determining nature of this stage presents in the association of two IAPP monomers. At physiological pH, IAPP maintains a positive charge due to the basic residues Lys1, Arg11,and His18 [1]. Of the residues, only Lys1 and Arg11 remain exposed to solvent while His18 is oriented towards the core [1,3]. Due to the orientation of IAPP monomers, nucleation includes the association of two cationic residues needing to overcome repulsive forces, thus explaining the rate determining nature of the nucleation phase [1,20].Formation of dimers from IAPP monomers indicates the beginning of the nucleation phase while subsequent aggregation into oligomers mark the transition into the elongation phase. Proto fibrils are intermediate molecules that form immediately afterwards, leading to mature fibril formation and the stationary phase where fibrils are in equilibrium with monomers [4,19]. As shown in Fig. 1, the reaction curve of IAPP fibril assembly has been described by many as a sigmoidal curve, similar to that of an allosteric enzyme kinetics plot [4,19].

Fig. 1 IAPP fibrillation kinetics plot. Blue curve represents the classical fibrillation kinetics of monomeric IAPP while the red curve represents the fibrillation kinetics of seeded fibrillation. Schematic outlines the classical fibril formation process separated into the three phases: lag,exponential, and stationary.

Fragments of mature fibrils are often used as “seeds” or templates to form additional fibrils. This alternative template pathway to fibril formation is more favorable and less energy demanding compared to nucleation initiated by the aggregation of two monomers. Consequently,the process results in a much shorter lag phase followed by a rapid elongation phase, and finally the expected stationary phase observed with the template pathway (Fig. 1) [4,20]. This favorability is explained by the initial nucleation and elongation phases in normal fibril formation.Dimerization and oligomerization of IAPP monomers are unfavorable steps in fibrillation that result in the formation of highly unstable fibril intermediates, which fragment in order to achieve a more stable arrangement. The fragment in turn serves as a template for binding additional monomers, forgoing the nucleation phase almost completely and increasing the elongation phase in the process [20].

IAPP-advanced glycation end products (AGE) conjugates (AGEIAPP) can also play the role of “seeds” for new fibril formation,resulting in a dramatically reduced nucleation phase, and in causing more lethal cytotoxicity compared to native IAPP [4,21]. AGE-IAPP can also be used to seed fibrillation with as low a ratio as 10% total AGE-IAPP present [4]. This non-enzymatic conversion to AGEIAPP renders the modified polypeptide resistant to proteolytic degradation, thus promoting its accumulation in vivo [4]. AGEIAPP plays a role in activating fibril formation compared to native IAPP, even at low pH where aggregation is largely unfavorable.It is hypothesized that the AGE modification drastically changes the aggregation rate by promoting conformational change from the intrinsically disordered random coil of IAPP to the highly ordered β-sheet structure offibrils [4,22]. This conversion significantly reduces the high-energy threshold governed by the nucleation step,thus allowing rapid transition to the formation of high-molecular weight cytotoxic IAPP oligomers [22]. Other non-enzymatic selfmodifications such as deamidation is known to encourage IAPP aggregation in a similar mechanism as glycation [23,24].

2.2 Cytotoxicity

Cytotoxicity associated with IAPP occurs primarily because of the formation of pores within the pancreatic β-cell membrane, which reduces membrane integrity and permeability, leading to disrupted ionic homeostasis and signal transduction and triggering cell death(Fig. 2) [3]. This hypothesis is plausible given the presence of the 19-residue membrane-binding domain of human IAPP (hIAPP),which modulates fibril formation and incites membrane leakage upon aggregation [25]. While this is the widely accepted mechanism, there is no consensus on the species causing the membrane permeating pores. Some studies point to oligomeric species and other prefibrillary soluble intermediates whereas others suggest the mature fibrils as the mediators [3,4,19,25,26].

Oligomeric species play a principal role in cytotoxicity and disease development [27]. Related amyloidogenic peptides previously implicated in diseases, such as the β-amyloids, and pre-fibrillary assemblies, particularly soluble oligomers, have significant cytotoxic effects in vivo [19]. Oligomerization causes the initial membrane leakage through pore formation or aspecific leakage, which may contribute to cytotoxicity [25]. Challenges in elucidating the true cytotoxic abilities of these oligomeric species are mainly due to the extremely short-lived nature of oligomeric aggregates and, as such, structural characteristics and other cues have yet to be identified [3,4,19,20].

In contrast, insoluble amyloid fibrils play an indirect role in β-cell cytotoxicity [13,21]. Alone, these insoluble fibrils are relatively inert and not cytotoxic [13], but their negative effect may be facilitated through the formation offibrils. Monomeric IAPP interacts with the lipid membrane of the β-cell through the membrane-binding domain[28]. This insertion leaves the amyloidogenic region of IAPP free to aggregate. Cytotoxicity comes with the growth of the fibrils to form larger pores, which eventually disrupts the membrane altogether[13]. In addition, another indirect model suggests that cytotoxicity is also triggered by a number of changes in the cellular environment because offibril formation [29]. These changes include an increase in the formation of reactive oxygen and nitrogen species, abnormalities in cellular redox systems, loss of protein function post-aggregation,and hyper-phosphorylation of proteins that accumulate in aggregative protein deposits [29]. Specific contributors to aggregation-induced cytotoxicity including the roles of monomeric, oligomeric, and protofibrillar species in inducing β-cell death have been reviewed elsewhere [30].

Furthermore, intracellular overproduction of IAPP followed by delays or errors in processing may result in the accumulation of IAPP due to the overwhelmed IAPP turnover mechanisms, thus causing fibril formation intracellularly [31]. Subsequent pores formed within the cell membrane by extracellular IAPP fibrils can function as calcium-permeable ion channels, resulting in the increased in flux of calcium ions into the cell [31]. Thus, the prolonged increase in intracellular calcium concentrations results in cell damage and apoptosis.Further exacerbation of this process may occur with the release of intracellular pre- fibrillar intermediates into the extracellular environment followed by subsequent formation of IAPP fibrils (Fig. 2).

2.3 Role of microenvironment in fibril formation

The local environment of IAPP also influences the switch from stable monomeric IAPP to fibrils. Factors such as salt concentration,ionic strength and glucose concentration play a major role in fibril formation [14].

Salt concentration and ionic strength have a complex effect on amyloidogenesis. These conditions are hypothesized to play a major role in establishing favorable conditions for IAPP amyloid formation[1]. In general, amyloidogenesis increases significantly with increasing ionic strength; however, the interplay between ionic strength and fibril formation is more complex than a simple linear correlation because pH, ion species, and concentration also affect the behavior of IAPP [1]. At high salt concentration, the effect of ion strength in IAPP fibrillation is directed by the Hofmeister effect, which associates protein solubility to the concentration of salt in solution[32]. Thus, IAPP fibrillation will be more favorable in conditions of high salt concentration, which explains the increased rate offibril formation as a result [1,33]. Furthermore, a higher salt concentration in the presence of anions yields an increased dependence on anion binding between salt and polypeptide, which favors fibril formation[1]. More research on the effects of salt alone on fibril formation would provide independent insights on the interplay between the two,as well as potential targets for IAPP fibrillation inhibition. Similar to the effect of monovalent anionic composition on fibrillation, IAPP aggregation accelerates in the presence of anionic lipids such as phosphatidylglycerol and phosphatidylserine [3].

Zinc ions play an indirect role in fibril formation through the modulation of insulin concentration. Insulin, along with IAPP,is stored in granular forms within pancreatic β-cells where the intracellular environment consists of a high Zn2+concentration and ionic strength [1,14]. The Zn2+transporter in pancreatic β-cells,ZnT8, is responsible for transporting zinc against its concentration gradient into the β-cells, thus indirectly modulating IAPP aggregation through insulin binding [14]. Insulin monomers and dimers preferentially bind IAPP monomers, whereas insulin oligomers are insoluble and do not bind IAPP. Instead, zinc ions bind and stabilize insulin hexamers, retaining their crystal forms within the β-cell granules. As a result, zinc ions modulate insulin granulation, thus altering the concentration of insulin monomers and dimers in the process, and impacting the formation of insulin-IAPP complexes and IAPP aggregation [14,34]. The binding of insulin monomers and dimers to IAPP occurs at the amyloidogenic region of IAPP,thus preventing IAPP self-assembly and inhibiting aggregation in the process [14]. Zinc ions also bind to the monomeric IAPP at His18, destabilizing the α-helical region and favoring the β-hairpin conformation instead [35]. While the β-hairpin is an intermediate structure in the formation of the amyloidogenic β-sheet, Zn2+blocks the process by stabilizing the much lesser amyloidogenic hairpin conformation [34,35]. A decrease in extracellular Zn2+concentration allows for an increase in oligomerization of insulin and, as a result,the concentration of monomeric and dimeric IAPP increases. Since both insulin conformations bind competitively to IAPP monomers,an increase in insulin-IAPP complexes is observed with increased Zn2+concentration, thus blocking the amyloidogenic region of IAPP and decreasing the propensity of IAPP aggregation [14]. Insulin-IAPP complex formation provides another route for IAPP shuttling in environments of excess peptide as well.

A number of other changes in the micro-environment can affect IAPP aggregation, either encouraging or discouraging fibrillation.Additional factors that play a role in fibril formation of IAPP include the presence of other metal ions such as Cu2+[36,37].

2.4 Role of IAPP sequence in fibril formation

Like humans, IAPP analogues are present in many different mammalian species, including baboon, cat, dog, rat, mouse, and cow [13]. However, unlike hIAPP, some analogues do not exhibit fibril formation and, more importantly, these mammals do not develop hyperglycemia or T2D [14]. Rat IAPP (rIAPP) analogues have been extensively studied to understand the importance of the 6-residue difference in the sequence that seem to elimination of the amyloidogenic propensity [6,13,18]. A high sequence conservation exists in rats and human’s IAPP, except for the amyloidogenic region,half of which consist of proline substitutions. Amyloidogenesis studies suggest the importance of Pro25 of rIAPP in fibril formation [6,38].This is likely because proline plays the role of a β-breaker, preventing the formation of β-sheet at the C-terminal and disrupting the U-shape topology seen in fibril formation, thus reducing the ability of IAPP to form aggregates [6,39,40]. The Pro25 residue introduces a kink in the middle of the peptide fragment of the mature fibrils, resulting in the loss of β-sheet topology in favor of aggregated coils [6].

Although the driving forces in IAPP self-assembly are not completely elucidated, many hypotheses have been proposed. Originally,it was thought that aromatic residues, specifically phenylalanine, played a fundamental role in facilitating the π-π interactions that encourage aggregation [41]. Mutational studies revealed that phenylalanine, and other aromatic residues alone, were not absolutely required for the formation offibrils. This led to the rationale that hydrophobic interactions,amongst other intermolecular forces, were the driving force behind fibril formation, while aromacity simply provided structural stabilization[42,43]. While no consensus exists on what plays the decisive factor on fibril formation, Pro fit et al. [44] suggested that both π-π and hydrophobic interactions play roles in fibril formation and when either is disrupted,disaggregation is expected to occur.

3. Mechanisms of inhibition of IAPP fibril formation

Aromatic and π-π interactions have been hypothesized to play an instrumental role in IAPP fibrillation [41-44]. As a result, potential agents that can compete with the major interactions governing fibril formation have been extensively studied. Preventing these interactions can lead to the inhibition offibril formation and mitigation of the deleterious effects of IAPP aggregation on pancreatic β-cells.Consequently, inhibitors of IAPP fibrillation can be used in preventing and managing T2D.

3.1 Peptides

This group of inhibitors are produced specifically to target IAPP, or can also be cross-reactive peptides that are repurposed for fibrillation inhibition. Nonetheless, the peptides interact with some conformation of IAPP and induce local changes, promoting the monomeric and other less harmful IAPP conformations in the process.

The use of hairpin peptides provides a novel approach to peptidebased fibrillation inhibition which, until then, had been explored using hIAPP-derived peptide truncations with additional modifications[9,45]. Through these studies, tyrosine and tryptophan-rich peptides were observed to be more effective at fibrillation inhibition, likely as a result of the aromatic moieties allowing for the formation of π-π interactions with the aromatic residues of hIAPP [9]. In this approach,the cyclized β-hairpin, cyclo-WW2, mimics the secondary structure that commonly occurs in early stages of aggregation in the fibrillation pathway, subsequently competing with the self-association propensity of IAPP. Aryl-rings of the β-hairpin allows for its more favorable binding to the hydrophobic hot spots of IAPP compared to the selfassociation with other IAPP monomers, thus inhibiting subsequent aggregation [45]. The added stability of the inhibitor by way of cyclization further increases its potent inhibitory effects down to substiochiometric levels of inhibitor concentration [9]. Cyclo-WW2 appears to be a general fibrillation inhibitor as favorable inhibition was also observed with other amyloidogenic peptides, such as α-synuclein [9]. The study by Sivanesam et al. [9] provided a crucial perspective on maximizing inhibitor stability and its potential role in potency, which is an important factor that should always be addressed when considering feasibility.

Fig. 2 Role of intracellular and extracellular IAPP fibrillation in the induction of apoptosis in β-cells; ROS, reactive oxygen species.

The pentapeptide inhibitor FLPNF was originally designed to mimic the RLANF sequence of the membrane-binding domain of hIAPP, where residue substitutions were made in order to increase hydrophobicity of the peptide. The alanine residue of RLANF was substituted with a proline to prevent the pentapeptide from forming β-strands and subsequent self-assembly [46]. Based on a BLAST search, peptide FLPNF is present in several plant-based food proteins,such as soy, pigeon pea, cowpea and rice proteins. Anchoring of this peptide with monomeric hIAPP was facilitated via π-π and cation-π interactions of phenylalanine and proline, as well as hydrogen bonding afforded by the asparagine and phenylalanine residues.While it has some inhibitory effects, its potency was not enough to completely block IAPP fibrillation [46]. The mechanism of action of FLPNF was likely through the binding and stabilization of monomeric IAPP, thus temporarily arresting IAPP in the lag phase for as long as possible [46]. The favorable interaction of FLPNF with IAPP was hypothesized to prevent IAPP self-association. In terms of cell viability, FLPNF successfully increased the lifespan of INS-1 rat cells in a dose-dependent manner. This effect was likely due to FLPNF associating with monomeric IAPP, thus preventing the formation of toxic oligomers and other pre-fibrillar species. Radical scavenging studies should be considered to elucidate potential antioxidative activities of the peptide as an inhibitory mechanism.

A different approach with synthetic peptides by Xuan et al. [47]resulted in the creation of the peptide, LA12, designed to specifically bind the amyloidogenic core of IAPP. This peptide binds both residues 11-16 and 19-28 of the membrane-binding domain and α-helical region of IAPP (Fig. 3). LA12 functions by inserting into the key amyloidogenic sites, destabilizing parallel β-sheets and initiating disaggregation of existing fibrils. In addition, LA12 associated with monomeric IAPP at the same loci, preventing further self-aggregation by stabilizing the random coil region [47]. Consequently, a 78%reduction offibrils was reported after 7-day incubation of IAPP fibrils with LA12 at a molar concentration ratio range of 30:1 to 50:1, LA15/amylin [47]. Based on these findings, a comprehensive mining of the food proteome stands to provide a structurally diverse array of natural peptide motifs with the potential to bind IAPP and inhibit its fibrillation.

3.2 Phenolic compounds

Polyphenolic compounds contain aromatic rings that competitively interact with the aromatic residues of IAPP, disrupting π-π interactions and impacting IAPP self-assembly [38]. As a result,aromatic interactions are considered to be very important to the amyloidogenic process as it is used to stabilize the anti-parallel β-sheet structure [29]. Many natural and food-derived compounds containing multiple aromatic rings possess disaggregative properties that can disrupt the π-π interactions and break down of IAPP fibrils. Table 1 outlines all the aromatic ring-containing compounds discussed in this section.

Fig. 3 Domain outline of IAPP as well as inhibitor binding regions grouped by compound class. Bottom: amino acid sequence of IAPP showing the transient-helical region and amyloidogenic FGAIL region.

Table 1Structure of natural IAPP fibrillation inhibiting compounds from different sources.

Table 1 (Continued)

Epigallocatechin gallate (EGCG) is a well-studied polyphenol that has been repurposed as a broad inhibitor for amyloidogenic proteins,and is especially functional for disaggregating IAPP fibrils [29,48].Due to its disaggregative properties, EGCG is also able to protect rat insulinoma (INS-1) β-cells from IAPP-induced cytotoxicity. EGCG potency assays revealed that a working ratio of 2:1 to 5:1 (IAPP/EGCG) yields significant effects on lowering IAPP aggregation[48]. A multitude of interactions facilitates the binding of EGCG to hIAPP monomers, including the π-π, Van der Waals, alkyl, π-alkyl,conventional hydrogen bonds, and carbon-hydrogen bond interactions[49]. The preferential binding site of EGCG is between the coil and helix of hIAPP, specifically at Arg11, Leu12, Ser19, Ala25, Ile26,Leu27, and Tyr37 residues (Fig. 3) [49]. Due to the strong affinity of EGCG to IAPP monomers, it is likely to successfully inhibit the formation of IAPP dimers, thus preventing aggregation at the earlier steps of self-aggregation [49,50].

Genistein (4,5,7-trihydroxyisoflavone), a phytoestrogen from soybeans, plays a dual role in the inhibition offibril formation of both hIAPP and Aβ [51]. Genistein exclusively binds IAPP monomers to prevent fibril formation and, upon seeding, its inhibitory abilities are effectively minimized [51]. Interaction of genistein with monomeric IAPP is facilitated through π-π interactions afforded by the two phenolic rings of this natural isoflavone compound (Table 1). In addition, genistein binds preferentially to the β-turn and N-terminal region of IAPP [50,51]. Molecular interaction simulations revealed that genistein binds to Lys1, Asn3, Arg11, Phe15, Val17, His18,Phe23, Ala25, Asn31, and Tyr37 residues of IAPP (Fig. 3) [49]. While genistein is a potent inhibitor offibril formation at the early stages of self-assembly, it is rendered ineffective once nucleation has occurred[51]. Despite the inability of genistein to disaggregate formed fibrils,the polyphenol is still able to mitigate the cytotoxic effects of IAPP fibrils. This is accomplished through genistein-induced remodeling of IAPP fibrils into unstructured aggregates, thus reducing the cytotoxic effects on rat insulinoma (RIN-m5F) cells [51]. More research should be done to identify if other isoflavone compounds have more potent and multifaceted inhibitory mechanisms in fibril formation beyond IAPP self-aggregation inhibition.

Lycopus lucidus, from the family Lamiaceae, is commonly consumed for its role in traditional medicine, and also as food. Many compounds such as flavonoids and phenolic acids have been isolated from the rhizome and studied for their health applications [52].Schizoteniun A, lycopic acid A, and lycopic acid B are polyphenols with catechol moieties isolated from Lycopus lucidus and suggested to exhibit anti-aggregative properties on hIAPP [52]. The three compounds showed extremely strong inhibitory effects on IAPP fibrillation with IC50values of 0.58, ＜ 0.01, and 0.023 μmol/L,respectively. The presence of multiple catechol moieties within their structures supports the suggestion that π-π interactions between the phenolic compounds and monomeric IAPP are more favorable than hIAPP self-association, thus resulting in potent inhibition of the lag phase [52]. Moreover, reduction of cytotoxicity was suggested given the antioxidative activities afforded by the phenolic compounds,which act as reactive oxygen species (ROS) scavengers. However,the true protective abilities of these compounds against IAPP fibril toxicity still need to be verified using cell cultures.

Rosmarinic acid, isolated from the plant Isodon japonicus,has been studied for its anti-aggregative properties against hIAPP fibrillation [53]. The aerial part of the plant are used as functional food and in traditional Chinese and Japanese medicine. Rosmarinic acid and its derivative, caffeic acid, were observed to possess inhibitory abilities on IAPP fibrillation, with IC50values of 3.1 and 57.6 μmol/L, respectively. As previously mentioned, polyphenols are known for their ability to inhibit aggregation via the disruption of the π-π interactions between the aromatic residues of monomeric IAPP. Hence, the stronger inhibitor potency observed for rosmarinic acid in the lag phase may because of its additional catechol moieties compared to caffeic acid. The increased aromaticity suggests the presence of a stronger affinity of rosmarinic acid for hIAPP,facilitated by an increased hydrophobic and π-π interactions. In terms of protective abilities, the compounds also acted as antioxidants, with rosmarinic acid exhibiting stronger effects than caffeic acid, thus minimizing the presence and subsequent damages and cytotoxicity caused by ROS, which become more prevalent during IAPP fibrillation [53].

Dihydrocaffeic acid isolated from the Lycii Cortex, a dried root bark of Lycium chinense, commonly used as a Chinese medicinal herb,has also been reported to exhibit dose-dependent anti-aggregative properties towards IAPP [54]. The inhibition occurred through the disruption of the π-π interactions between the amyloidogenic regions of IAPP, possibly during the lag phase where the interactions most frequently occur. The IC50value reported for dihydrocaffeic acid was 9.3 μmol/L, which is higher than those reported for compounds with multiple aromatic moieties, suggesting a correlation between the number of phenolic moieties present within a structure and its potency in IAPP fibrillation inhibition [54]. Interestingly, caffeic acid,a structural analogue to dihydrocaffeic acid exhibited a significantly higher IC50value compared to dihydrocaffeic acid. This suggests that other moieties apart from the aromatic ring may be playing important roles in inhibitor potency. In the case of caffeic acid, the presence of α,β unsaturation seems to reduce the potency of the inhibitor, compared to dihydrocaffeic acid, possibly by introducing structural flexibility that potentially influence the stability of the IAPP-phenolic complex and thus reducing binding favorability. As a result, a decreased ability for an inhibitor to form a stable complex with IAPP would potentially diminish its sustained inhibitory role in mitigating fibrillation.

Flavonoids isolated from the halophyte Tamarix gallica L. have also been extensively studied for its numerous biochemical and pharmacological functions [55]. Hmidene et al. [55] studied the effects of some of the flavonoids on mitigating IAPP fibril formation,and also the effect of flavonoids containing glucuronide moieties to provide insight on the seemingly absolute requirement of aromatic compounds in a potent inhibitor. From the studies, the strongest inhibitory activities were observed for quercetin and its glucuronidesubstituted derivative (QGlcA), with IC50values of 1.8 and 1.7 μmol/L,respectively. Notably, antioxidative activities used to mitigate the effects of ROS and reactive nitrogen species caused by the formation of oligomeric species should be taken into consideration when evaluating inhibitor suitability and potency. QGlcA showed a much higher antioxidant activity than quercetin [55]. While the sugar substitution itself may have enhanced the activity, it is also possible that the increased interaction with IAPP was due to the significant increase in hydroxyl groups present on the sugar moiety,or the solubility and accessibility of QGlcA compared to quercetin.As a result, the glucyronide derivative has the potential to provide a stronger protective effect on cells against damages caused by IAPP fibrillation. It is also possible that inhibition of IAPP fibril formation occurred earlier at the lag phase where the inhibitors are able to bind more favorably with monomeric IAPP, thus decreasing the favorability for self-assembly and mitigating fibril formation [50,51].

3.3 Terpenoids

Extracted from Salvia miltiorrhiza Bge or red sage, for use as functional food and medicine, the tanshinone class of compounds has been explored for their pharmacological uses in treatment of a myriad of diseases [56]. Notably, tanshinone I (TS1) and tanshinone IIA (TS2)were reported to inhibit IAPP fibrillation, disaggregate preformed fibrils, and mitigate IAPP fibrillation-induced cytotoxicity in cultured rat insulinoma cells [56]. Incubation of the tanshinone derivatives with IAPP resulted in a dose-dependent reduction in the exponential/growth phase but no effect on the lag phase, indicating that binding of the compounds does not favor the monomeric conformation [56].While TS1 and TS2 bind to hydrophobic and aromatic residues of IAPP, the preferred binding sites are located on opposing ends of IAPP [56]. TS1 binds the N-terminal β-strand on the interior face,towards the U-bent cavity, interacting with Leu12, Ala13, Asn14,Phe23, Gly24, Ala25, and Ile26, whereas TS2 binds the exterior face of the N-terminal β-sheet at Ala13, Asn14, Phe15, and Leu16 (Fig. 3).TS2 also exhibits favorable binding to the C-terminal β-sheet residues of Leu27, Ser28, Ser29, Val32, Gly33, Ser34, and Asn35 (Fig. 3).The mechanisms of inhibition of TS1 and TS2 are similar in that they both prevent self-association binding of IAPP monomers by binding to the β-strands of IAPP sites, thus blocking elongation as well as subsequent transformation to pleated β-sheet structure, a common hallmark of IAPP fibrillation [56]. In addition to the contribution of hydrophobic interactions, charge-transfer complex interactions mainly in the form of parallel (TS1) and T-shaped (TS2) π-π interactions were observed between the terpenoids and the aromatic residues of IAPP [56].

3.4 Alkaloids

Kukoamines A and B, also isolated from Lycii Cortex, are polyamine spermidine alkaloids that exhibited dose-dependent inhibitory abilities against IAPP fibrillation [54]. Potency was also likely due to the aromatic rings, which enhance binding of the alkaloids to monomeric IAPP, thus leading to a more favorable interaction than self-association with another IAPP monomer.The IC50values for kukoamine A and B were 8.7 and 3.3 μmol/L,respectively [54]. Despite their effectiveness in inhibiting IAPP fibrillation, there is a dearth of information on the aggregation phases targeted by the alkaloids. However, it is reasonable to hypothesize that the compounds inhibited aggregation at the lag phase where association with monomeric IAPP is more favorable, but may not be as potent in disaggregation of preformed fibrils. As both alkaloids contain the same number of catechols, the distance between the catechol moieties and the alkylamine substitution needs to be investigated to understand the interaction with IAPP hydrophobic pocket and in making the IAPP binding more favorable. Notably, the alkylamine chain might enhance binding of kukoamine B with IAPP instead of introducing stearic hindrance, especially through additional hydrogen bonding with the amine group, thus further stabilizing its interaction with IAPP and increasing its potency.

4. Polyphenols in pancreatic β-cell protection

Polyphenols have been largely identified to possess dosedependent protective effects on pancreatic β-cells, especially in minimizing oxidative stress [57,58]. EGCG, cinnamon and red wine polyphenols were previously reported to exert hypoglycemic effects through the modulation of oxidative stress [57,58]. While the mechanisms of these polyphenols in relation to β-cell protection have been studied primarily in the context of their antioxidative abilities,further investigations linking the role of polyphenols in mitigating antioxidative stress and IAPP fibrillation should be considered. This is important given that the phenolic compounds have demonstrated IAPP binding activities in vitro and beneficial roles in glucose metabolism, suggesting a potential link between the two properties[59]. Furthermore, IAPP fibril formation induces cellular oxidative stress [60], which suggests a potential antioxidative mechanism of phenolic compounds in inhibiting IAPP fibrillation and β-cell toxicity.

5. Conclusions and future directions

The shift to identifying natural or food-derived compounds as a treatment for T2D via the inhibition of IAPP fibril formation has several benefits, especially considering the relative safety of these compounds compared to synthetic drugs. However, there are potential drawbacks. Polyphenolic compounds, for example, are generally not very soluble, posing difficulties in choosing their delivery vehicles. In addition, they may have low bioavailability, thus higher concentrations would be ingested in order to observe beneficial effects. For the most part, these issues have not been specifically studied for IAPP-induced toxicity and will need to be taken into consideration when elucidating the practicality of using natural compounds for controlling the prevalence of IAPP fibrils. Future investigations should also be considered in identifying the governing mechanisms behind IAPP cytotoxicity as it can allow for deliberate targeted treatments.Understanding the governing interactions controlling fibril assembly,particularly in elucidating the importance of hydrophobic and π-π stacking in fibril formation, would be beneficial in identifying the modes of inhibition. Furthermore, a majority of the proposed natural and food-derived inhibitors need to be evaluated in vivo to validate their appropriateness as candidates for controlling IAPP fibril formation and T2D treatment. In addition to aromatic moieties that are thought to play an important role in the disruption of π-π interactions between monomeric IAPP, the contribution of other structural factors in inhibitor potency should be investigated through rational design and structure-function relationship studies. Lastly,there is a disproportionate focus of research on the inhibition of IAPP fibrils and little focus on preventative measures. Particularly, there is a need to understand the mechanisms that trigger IAPP misfolding and dimerization with other misfolded IAPP monomers, and the role of the natural and food-derived compounds at these early stages.A combination of the preventative and treatment approaches with natural and food-derived compounds promises to provide a powerful,safer and effective approach for controlling IAPP fibril formation towards the prevention, management and treatment of T2D.

Conflict of interest statement

Authors declare that there are no conflicts of interest associated with this manuscript.

Acknowledgements

Authors acknowledge the financial support from the Natural Sciences and Engineering Research Council of Canada (NSERC)(grant number RGPIN-2018-06839), and the University Research Chairs Program of the University of Ottawa, Canada.