Food amyloid fibrils are safe nutrition ingredients based on in-vitro and in-vivo assessment

The study addresses whether food-derived protein amyloid fibrils, despite their advantageous properties, are safe for human consumption. Safety concerns stem from the association of amyloids with human diseases and the possibility of amyloid cores resisting digestion, crossing the intestinal epithelium, and cross-seeding pathological amyloids. The authors contextualize amyloids as fibrillar protein aggregates with cross-β architecture, noting the existence of non-pathological functional amyloids across organisms and the growing applications of artificial amyloids from food proteins. The central questions are: Do amyloid nuclei resist gastrointestinal digestion? Are digested amyloids cytotoxic? Can digested counterparts cross the epithelium into the bloodstream and induce cross-seeded pathological aggregates? The purpose is to comprehensively assess the digestive fate and safety of β-lactoglobulin and lysozyme amyloid fibrils via in-vitro digestion and in-vivo models, comparing outcomes to native monomers.

The paper reviews that amyloid fibrils, once considered primarily pathological, can also be functional and non-disease-related in various organisms, including humans. Examples include spider silk, curli fibrils in E. coli, and human receptor-interacting protein kinase, with physiological roles such as hormone storage in endocrine granules. Artificial amyloids from natural proteins and synthetic peptides have been applied in nanotechnology, with food proteins (β-lactoglobulin, lysozyme, ovalbumin, oat proteins) serving as ideal sources due to abundance and non-toxicity. These materials exhibit superior functionalities in biomedicine, environmental science, and materials applications, including nutrient/drug delivery and therapeutic uses. Despite this, dietary use has been hindered by safety concerns extrapolated from pathological amyloids, including potential nanotoxicity and cross-seeding of pathological aggregates. β-lactoglobulin’s structural properties and allergenicity motifs are discussed in relation to pepsin resistance at low pH.

• Proteins and fibril preparation: β-lactoglobulin (β-lg) and hen egg white lysozyme monomers prepared; amyloid fibrils formed by heating 2 wt% monomer solutions at 90 °C (β-lg 6 h; lysozyme 24 h) with stirring; fibril solutions dialyzed (100 kDa MWCO) at pH 2.
• Simulated gastrointestinal digestion: INFOGEST static in vitro protocol used with simulated gastric fluid (pH 3, pepsin 280 U/mL) for 120 min and simulated intestinal digestion with pancreatin (trypsin activity ≥0.3 U/mg protein; explored up to 20–50 U/mL trypsin equivalents) for 120 min at 37 °C. Enzyme concentrations varied to determine minimal amount for full digestion; stirring vs shaking effects assessed.
• Analytical techniques during/after digestion:
  • SDS-PAGE to monitor proteolysis of monomers and fibrils.
  • ELISA (AgraQuant kits) to assess antigenicity/allergenicity epitopes.
  • Fluorescence assays: Thiazole orange (TO) and Thioflavin T (ThT) to track β-sheet content reduction.
  • Circular dichroism (CD) spectroscopy for secondary structure transitions.
  • MALDI-TOF/TOF MS (linear and reflector modes) to profile peptide mass distributions and sequence candidates; low enzyme to limit autolysis; additional 10 kDa filtration.
  • Microfluidic polarized light microscopy to monitor birefringence decay kinetics of shear-aligned fibrils during digestion (imaging every 0.5 s, intensity quantification).
  • Small-angle X-ray scattering (SAXS) to compare scattering profiles before and after digestion vs buffer/water; STORM fluorescence imaging for fibril visualization pre/post gastric digestion.
  • Atomic force microscopy (AFM) large-field and random-scan imaging to detect any residual fibril fragments and define minimal enzyme concentrations for complete digestion.
• In-vitro cytotoxicity and ROS:
  • Human intestinal cell lines: Caco2 and HCEC-1CT exposed 4 h to freeze-dried, reconstituted digested β-lg/lysozyme monomers and fibrils across 50–1250 µg/mL; ATP-based viability (CellTiter-Glo); ROS assays; controls included digestion matrix and TBHP.
  • Also tested digested Aβ42 fibrils as pathological control.
• In-vivo models:
  • C. elegans: Mobility/healthspan assays with supplementation at day 1 of adulthood (0.1667 mg/mL) for monomers, fibrils, and digested fibrils; polyQ35-YFP aggregation model (AM140) fed with β-lg or lysozyme fibrils at 0.25 and 1.5 mg/mL; aggregates quantified by fluorescence microscopy and ImageJ; statistical analyses via one-way ANOVA with Dunnett’s tests.
  • Mice (Kunming males): Oral gavage with 200 µL of 1 mg/mL β-lg/lysozyme monomers or fibrils; after 4 h, small intestine and colon contents analyzed by LC-MS/MS (Q Exactive) for peptide identification; serum sampled at 1 h for peptide detection. 5-DTAF-labeled β-lg monomers/fibrils used to track in-vivo distribution over 6–24 h; organs and blood imaged by IVIS. Long-term administration: daily gavage for 30–60 days at low (1 mg/mL) or high (5 mg/mL) doses; major organs processed for Congo Red staining to detect amyloid plaques; AD FAD+ mice as positive control.
• Data availability and statistics as per Nature Communications standards.

• Digestion behavior (β-lg):
  • SDS-PAGE showed β-lg monomer largely intact during gastric phase (18 kDa), with limited fragmentation; β-lg fibrils rapidly degraded into <10 kDa polypeptides within minutes of pepsin exposure, further to <5 kDa in intestinal phase.
  • ELISA indicated high allergenicity for gastric-phase β-lg monomer due to resistant epitopes; β-lg fibrils showed significant allergenicity drop during gastric digestion and further reduction in intestinal phase.
  • TO fluorescence: β-lg monomer decreased ~20% in gastric phase, with steep decay in early intestinal phase to background; β-lg fibrils decreased ~70% in gastric phase and to background in intestinal phase. ThT assays corroborated.
  • CD: β-lg monomer shifted from mixed β/α to random coil across digestion; β-lg fibrils’ β-sheet minimum at 218 nm shifted to 200 nm (gastric) and 198 nm (intestinal), indicating random coil—no residual β-sheet cores post-digestion.
  • SAXS: Post-intestinal digestion, scattering profiles of β-lg monomer, β-lg fibrils, and buffer were identical and indistinguishable from deionized water; STORM imaging showed disappearance of fibrillar structures after gastric digestion.
  • Microfluidics polarized light: Birefringence intensity of aligned fibrils dropped immediately upon pepsin addition, decayed exponentially in first 30 min, and reached water baseline after intestinal phase, indicating absence of fibrils.
  • MALDI-MS: β-lg monomer before digestion showed full-length species (Mavg ~18,366 Da) with singly/doubly/triply charged peaks. After intestinal digestion, both monomer and fibril samples predominantly contained oligopeptides <5 kDa centered around 1–2 kDa (~10–20 aa). Reflector mode: 100 peaks in digested monomer vs 55 in digested fibrils; 38 peaks common; only 2 unique to fibrils (815.45 Da and 1354.71 Da) with sequence candidates also present within larger monomer fragments, indicating more extensive amyloid digestion; many peaks attributed to enzyme autolysis.
  • AFM single-molecule analysis showed complete digestion of fibrils at trypsin 20–50 U/mL (2–5x lower than INFOGEST); stirring reduced enzyme needs by ~5–10x; no fibril fragments detectable at these conditions.
• Lysozyme findings:
  • Lysozyme fibrils showed improved digestibility vs monomers: large signal drop in gastric ThT/TO (~4-fold reduction) and more extensive digestion to <1 kDa peptides (MALDI), while monomers were less digested even after intestinal phase.
• In-vitro cell assays:
  • No loss of viability in Caco2 or HCEC across 50–1250 µg/mL for digested fibrils vs digestion matrix; exposure often promoted growth relative to matrix alone, consistent with nutritive peptides. ROS levels comparable between digested fibrils and monomer controls. Digested Aβ42 fibrils also did not reduce cell viability.
• C. elegans:
  • No mobility impairment; β-lg and lysozyme fibrils prolonged healthspan (increased active movement) compared to controls; equal protein amounts fed (0.1667 mg/mL). PolyQ35 aggregation unaffected by β-lg fibrils; mild increase in aggregates only at high lysozyme fibril dose (1.5 mg/mL) but not at 0.25 mg/mL; digested fibrils did not promote aggregation.
• Mice:
  • LC-MS/MS of small intestine/colon after 4 h gavage detected numerous low-MW peptides; amyloid-derived peptides were of lower or comparable MW vs monomer-derived, confirming superior or equal digestibility; for β-lg, almost all peptides overlapped between fibrils and monomers except LNENKV and NGECAQK (non-amyloidogenic) in fibrils; for lysozyme, all amyloid peptides also found in monomers.
  • Serum at 1 h post-gavage: no β-lg or lysozyme peptides detected for either monomers or fibrils, indicating no absorption into bloodstream under tested conditions.
  • 5-DTAF-labeled tracking: strong intestinal fluorescence over 6 h, with excretion via feces; no fluorescence detected in blood serum up to 24 h or in organs, corroborating lack of systemic absorption.
  • Long-term administration (30–60 days, low/high doses): Congo Red staining showed no amyloid plaques in brain or major organs in any fibril-fed group; AD FAD+ positive controls displayed plaques. Histology indistinguishable from non-AD controls. Overall: Food amyloid fibrils (β-lg, lysozyme) are digested at least as well as monomers to small, random-coil oligopeptides without residual amyloid cores; no cytotoxicity, no ROS increase, no health detriments in Caco2/HCEC, C. elegans, or mice; no evidence of epithelial crossing or cross-seeding pathology.

The comprehensive multi-modal in-vitro and in-vivo analyses address the core safety concerns of dietary amyloid fibrils. Amyloid fibrils from β-lg and lysozyme disassemble under gastric pepsin and are fully degraded during intestinal digestion into random-coil peptides indistinguishable from monomer digestion products, as evidenced by fluorescence, CD, SAXS, microfluidics imaging, MALDI-MS, and AFM. Lysozyme fibrils, in particular, exhibit greater digestibility than monomers, mitigating concerns about persistent amyloid cores. First-tier cytotoxicity testing in human intestinal cell lines shows no adverse effects and suggests nutritive benefits. In vivo, C. elegans healthspan improvements upon fibril feeding suggest benign or beneficial metabolic impacts, with no cross-seeding except a slight effect only at very high doses of undigested lysozyme fibrils in a sensitized polyQ model. In mice, intestinal peptide profiles confirm extensive digestion of fibrils; the absence of serum peptides and organ fluorescence indicates no systemic absorption. Long-term feeding yields no amyloid plaque formation, refuting cross-seeding risk in tested conditions. These findings collectively support that digested food amyloids are at least as safe as digested monomers and potentially safer for lysozyme, addressing key risk hypotheses and supporting their consideration as safe nutritional ingredients.

This study demonstrates that β-lactoglobulin and lysozyme amyloid fibrils are efficiently digested by simulated and in-vivo gastrointestinal processes into small, random-coil oligopeptides similar to or smaller than those from monomers, with no detectable residual amyloid cores. Across in-vitro human cell assays and in-vivo C. elegans and mouse models, digested fibrils showed no cytotoxicity, no oxidative stress elevation, no impairment of healthspan, no systemic absorption of peptides, and no induction of amyloid plaques, with some indications of improved digestibility and healthspan benefits. These results support the safety of food protein amyloid fibrils as potential nutritional ingredients and suggest they could be considered, pending regulatory assessment, as Generally Regarded as Safe (GRAS). Future research directions:

• Clinical or human-relevant gastrointestinal models to confirm digestibility, absorption, and safety in humans.
• Broader panel of food proteins and processing conditions to generalize findings.
• Long-term dietary intervention studies assessing metabolic and immunological outcomes, including allergenicity in sensitive populations.
• Detailed peptidomics and bioactivity profiling of digestates to evaluate potential functional benefits.

• Enzyme concentration for some in-vitro analyses (e.g., MALDI-MS) was intentionally kept low to limit enzyme autolysis; although follow-up AFM and bulk methods established full digestion at INFOGEST-compliant or lower enzyme levels, this difference may affect direct comparability across assays.
• The study focuses primarily on two proteins (β-lactoglobulin and lysozyme); while representative, generalization to all food protein amyloids requires further validation.
• In-vitro digestion models and cell assays, while informative, do not fully replicate human gastrointestinal complexity and interindividual variability (e.g., microbiome, transit times).
• A mild increase in polyQ aggregation in C. elegans occurred only at very high concentrations of undigested lysozyme fibrils, highlighting that extreme exposures to intact fibrils could have distinct effects not observed with digested fibrils.
• No serum peptides were detected at 1 h post-gavage in mice; different time points, doses, or sensitive targeted assays could further confirm systemic non-absorption.