Aggregation kinetics define the temporal window of IAPP proteotoxicity.

Aggregation of islet amyloid polypeptide (IAPP) to form amyloid contributes to β-cell dysfunction in type 2 diabetes, yet the identity and temporal persistence of the toxic species is unresolved. Competing models attribute toxicity to mature fibrils, fibril growth, or transient prefibrillar intermediates formed during the lag phase or via secondary nucleation. Here, we directly test these models by combining time-resolved β-cell functional assays with concurrent biophysical measurements of IAPP aggregation across multiple perturbations and sequence variants. Across 22 independent experiments spanning more than a 450-fold range in lag times, we find that maximal toxicity occurs during the lag phase and declines as fibrils accumulate. The duration of β-cell dysfunction scales linearly with lag phase length, establishing aggregation kinetics as a quantitative predictor of the onset, peak, and termination of toxicity. Perturbations that alter aggregation kinetics, including concentration, temperature, and sequence, predictably shift the temporal window of toxicity. The diabetes-associated S20G variant produces higher peak toxicity over a compressed time window, whereas the slower-aggregating variants examined prolong toxicity without increasing its magnitude. These results resolve competing models by demonstrating that transient lag-phase intermediates, rather than growth phase processes or mature fibrils, dominate β-cell dysfunction, and establish aggregation kinetics as a predictor of the timing and duration of cellular exposure to toxic intermediates.