Designed peptides as nanomolar cross-amyloid inhibitors acting via supramolecular nanofiber co-assembly

The study addresses the challenge of inhibiting amyloid self-assembly, a process implicated in neurodegeneration and β-cell loss in AD and T2D. Epidemiological links and molecular evidence suggest cross-interactions between Aβ and IAPP that can either cross-seed or cross-suppress amyloid formation. Aβ and IAPP co-localize in human and mouse amyloid deposits, and sequence and structural similarities between IAPP(8–28) and Aβ(15–40/42) suggest shared interaction hotspots. Building on these insights and prior IAPP-derived interaction surface mimics (ISMs) that redirected aggregation, the authors hypothesized that peptides mimicking the Aβ amyloid core (Aβ(15–40)) could inhibit both IAPP and Aβ42 self-assembly and reciprocal cross-seeding by stabilizing alternative, non-amyloidogenic folds and redirecting aggregation pathways into non-toxic states.

Prior work established polymorphic cross-interactions between Aβ and IAPP capable of reciprocal cross-seeding and cross-suppression depending on assembly states. Aβ and IAPP share ~25% identity and ~50% similarity, especially within amyloid core segments, and previous SAR studies identified common “hot segments” mediating self- and cross-interactions. Structural similarities between fibrils have been proposed, including models of hetero-amyloids. Previously, IAPP-derived ISM peptides sequestered Aβ and/or IAPP into non-toxic amorphous aggregates. N-methylation in amyloidogenic sequences has been shown to reduce amyloidogenicity while preserving binding. Despite these advances, few cross-amyloid inhibitors target both polypeptides effectively and none previously suppressed reciprocal Aβ/IAPP cross-seeding.

Design: The Aβ(15–40) segment, modeled in a β-strand–loop–β-strand U-shaped fold (fAβ40), was modified to yield Aβ amyloid core mimics (ACMs). Modifications included: (a) loop substitution Aβ(24–26) (Val-Gly-Ser) with β-sheet-promoting hydrophobic tripeptides ((Nle)3, (Leu)3, (Phe)3); (b) selective N-methylation of two alternating residues within Aβ(17–20) (either Leu17/Phe19 or Val18/Phe20) to suppress intrinsic amyloidogenicity; and (c) Met35→norleucine substitution. A series of 13 analogs was synthesized; six ACMs with combinations of loop tripeptides and N-methylation patterns were advanced: Nle3-VF, Nle3-LF, L3-VF, L3-LF, F3-VF, F3-LF. Synthesis and characterization: Peptides were synthesized by Fmoc-SPPS, purified by RP-HPLC, and verified by MS. Labeled variants (fluorescein/FITC, TAMRA, Atto647N, HiLyte647) were prepared for imaging and binding assays. Biophysical assays: Amyloid fibrillogenesis monitored via Thioflavin T (ThT) fluorescence for IAPP and Aβ42 under defined buffer conditions, including self- and cross-seeding with preformed fibrils (fIAPP or fAβ42). Secondary structure assessed by far-UV circular dichroism (CD). Oligomer/assembly states probed by glutaraldehyde cross-linking followed by NuPAGE and Western blotting, size-exclusion chromatography (SEC), and ESI-IMS-MS. Morphology analyzed by TEM, immunogold-TEM, and X-ray fiber diffraction. Microscopy: Confocal laser-scanning microscopy (CLSM), stimulated emission depletion (STED) nanoscopy, two-photon microscopy (2PM), and FLIM-FRET were used to visualize co-assemblies and determine nanoscale proximity between components. Binding and potency: Apparent binding affinities (Kd, app.) of ACMs to IAPP and Aβ42 determined by fluorescence titrations using 5 nM labeled monomeric peptides. Cellular toxicity assessed by MTT reduction assays in RIN5fm (IAPP) and PC12 (Aβ42) cells to determine IC50 values. Functional assays: Seeding competence of hetero-nanofibers tested by adding hf-IAPP/ACM or hf-Aβ42/ACM to fresh IAPP or Aβ42. Thermostability evaluated by boiling followed by TEM and ThT. Proteolytic susceptibility tested by proteinase K (PK) digestion with dot blot quantification. Phagocytosis quantified in primary murine bone marrow-derived macrophages (BMDMs) and BV2 microglia using TAMRA-labeled tracers. Ex vivo electrophysiology measured Aβ42-mediated hippocampal LTP impairment and its rescue by ACMs in mouse slices.

- Six designed ACM peptides (Nle3-VF, Nle3-LF, L3-VF, L3-LF, F3-VF, F3-LF) potently inhibit IAPP fibrillogenesis and cytotoxicity with nanomolar IC50 and bind IAPP with nanomolar affinity. IC50 (IAPP toxicity, nM): Nle3-VF 65.0±5.2; Nle3-LF 82.1±10.2; L3-VF 112.5±8.1; L3-LF 133.2±29.0; F3-VF 78.5±13.6; F3-LF 41.7±4.1. Apparent Kd to IAPP (nM): 69.5±1.4; 55.4±5.9; 77.3±2.9; 143.2±5.0; 15.0±1.9; 37.6±2.9, respectively. - ACMs inhibit Aβ42 fibrillogenesis and cytotoxicity and bind Aβ42 with nanomolar-to-submicromolar affinity. IC50 (Aβ42 toxicity, nM): Nle3-VF 367±79; L3-VF 261±140; F3-VF 1032±297; F3-LF 262±115 (n.d. for Nle3-LF, L3-LF). Apparent Kd to Aβ42 (nM): Nle3-VF 14.5±8.0; Nle3-LF 11.1±6.0; L3-VF 38.0±2.0; L3-LF 2.6±1.4; F3-VF 160.8±12.9; F3-LF 430.6±7.1. - Mechanism: ACMs co-assemble with IAPP or Aβ42 into ThT-invisible, non-cytotoxic, heteromeric nanofibers (hf-IAPP/ACM; hf-Aβ42/ACM) and highly ordered superstructures (bundles, loops, ribbons, nanotubes). Evidence includes immunogold-TEM (dual labeling), pull-down/WB, SEC, cross-linking/NuPAGE, CD, X-ray fiber diffraction (cross-β signature), STED/2PM imaging, and FLIM-FRET showing close proximity (<5.5 nm; e.g., donor lifetime ~0.8 ns; FRET efficiency ~55% for IAPP/ACM; node regions ~60% for Aβ42/ACM). - Morphology: IAPP/ACM fibrils are TEM-indistinguishable from fIAPP (6–10 nm width; 100–200 nm length) but ThT-invisible and non-toxic. Aβ42/ACM fibrils are significantly longer (2–4× fAβ42 length) with similar widths (7–8 nm). - Co-assembly pathway: Large amorphous hetero-aggregates form early and evolve into nanofibers; IAPP monomers/prefibrillar species can template co-assembly. No fibrils form with non-amyloidogenic rat IAPP or N-methylated IAPP-GI. - Seeding: hf-IAPP/ACM and hf-Aβ42/ACM are seeding-incompetent, in contrast to fIAPP and fAβ42. - Stability and degradability: hf-IAPP/ACM and hf-Aβ42/ACM are thermolabile (converted to amorphous after 95°C, 5 min) and protease-sensitive. PK degradation: hf-IAPP/ACM fully degraded <6 h (ACM component within minutes) vs fIAPP stable ≥30 h; hf-Aβ42/Nle3-VF degraded ~30 min vs fAβ42 ~2 h. - Cellular clearance: Phagocytosis of hetero-nanofibers by BMDMs and BV2 microglia increased 3–10-fold compared with homofibrils (IAPP and Aβ42 cases). - Functional rescue: ACMs fully ameliorate Aβ42-induced impairment of hippocampal LTP ex vivo when added to monomers or pre-oligomerized Aβ42. - Cross-seeding suppression: ACMs inhibit fIAPP-mediated cross-seeding of Aβ42 and fAβ42-mediated cross-seeding of IAPP; formation of non-toxic, cross-seeding-incompetent ternary co-assemblies (Aβ42/ACM/IAPP) observed by 2PM.

The findings demonstrate a distinct inhibition mechanism wherein designed Aβ(15–40)-based peptides co-assemble with IAPP or Aβ42 into heteromeric, amyloid-like but non-toxic and ThT-invisible nanofibers. By combining N-methylations (to block β-sheet propagation) with hydrophobic loop substitutions (to stabilize alternative interaction surfaces), ACMs maintain high-affinity interactions with amyloidogenic segments while suppressing pathological fibril propagation. This redirection yields labile, protease-sensitive, and readily phagocytosed assemblies, contrasting sharply with the stability and seeding competence of pathogenic fibrils. The approach successfully suppresses reciprocal cross-seeding between IAPP and Aβ42, a proposed mechanistic link between T2D and AD. Ex vivo LTP rescue supports physiological relevance. The data also suggest that potentially beneficial hetero-amyloids with labile properties may exist, and that harnessing amyloid fold polymorphism can generate functional materials and therapeutic strategies.

Six Aβ amyloid core-mimicking peptides were identified as potent cross-amyloid inhibitors of IAPP and Aβ42, binding with nanomolar affinities and blocking toxic self-assembly and reciprocal cross-seeding. ACMs act via co-assembly into non-toxic, ThT-invisible hetero-nanofibers that are seeding-incompetent, thermolabile, protease-sensitive, and efficiently phagocytosed. These properties make ACMs promising leads for anti-amyloid therapeutics relevant to both T2D and AD and provide design principles for functional hetero-amyloid-based nanomaterials. Future work should resolve atomic structures of hetero-assemblies, elucidate detailed mechanisms of co-assembly and polymorphism, evaluate in vivo efficacy and safety, and optimize pharmacological properties.

The atomic structures of Aβ/IAPP hetero-amyloids and ACM-containing hetero-nanofibers remain unresolved; proposed models are hypothetical. Mechanistic steps of co-assembly are not fully elucidated. Binding affinity estimates use simplified 1:1 models despite potential complexity due to self-assembly. Most data are from in vitro and ex vivo assays; in vivo validation is not reported.