Read-Across from D-Allulose Supports an Animal-Free Safety Assessment of 5-Keto-D-Fructose (5-KDF)

Abstract Background: Modern toxicological science recognizes that structurally similar food ingredients can be assessed through validated alternative methods rather than animal testing. 5-Keto-D-fructose (5-KDF), a rare ketohexose, shares substantial structural similarity with the approved rare sugar D-allulose. Objective: To determine whether 5-KDF and D-allulose exhibit equivalent safety profiles across computational, physicochemical, and experimental endpoints, and whether this structural analogy obviates the need for animal studies. Methods: Structural similarity was quantified using Morgan fingerprints (RDKit v2024.09.6). Computational toxicology employed Derek Nexus v6.1.0, VEGA QSAR, ToxTree v3.1.0, and ProTox-II. Experimental genotoxicity assessment followed OECD TG 471 (bacterial reverse mutation, five strains ±S9) and OECD TG 473 (chromosomal aberration in human lymphocytes ±S9). Mechanistic safety was evaluated in differentiated Caco-2 intestinal and HepG2 hepatic cells with critical osmotic controls (mannitol). Results: Structural analysis demonstrated 60% similarity (Tanimoto coefficient 0.60; Dice 0.75) with equivalent physicochemical properties (cLogP: −3.17 vs −3.38; TPSA: 115 vs 118 Ų). All computational platforms unanimously predicted identical non-mutagenic, non-clastogenic, non-carcinogenic profiles. Experimental validation confirmed: bacterial reverse mutation assay negative across all five strains ±S9 (up to 5,000 μg/plate); chromosomal aberration test negative in human lymphocytes ±S9 (up to 1,000 μg/mL). Mechanistic assessment demonstrated statistical equivalence to D-allulose: intestinal barrier integrity preserved (TEER 84±7% vs 86±5%, p=0.58), hepatic ATP maintained (29.8±2.8 vs 30.1±2.5 nmol/mg, p=0.92), no lipogenic gene activation (FASN 0.94±0.15 vs 1.08±0.18-fold, p>0.05), no oxidative stress (109±14% vs 106±12%, p=0.81). Conclusions: 5-KDF exhibits an identical safety profile to approved D-allulose across all measured endpoints. Both compounds are simple monosaccharides lacking structural toxicophores, with equivalent physicochemical properties and metabolic inertness. The comprehensive negative genotoxicity data (Tier 1 battery) combined with structural analogy demonstrates that animal testing (Tier 2) would provide no additional discriminating information. This weight-of-evidence approach satisfies modern safety assessment standards while advancing the 3Rs principles (Replacement, Reduction, Refinement) in toxicological science.