Uranium kidney toxicity is well documented after acute uranium intake with a threshold of around 1-3 µg/g of kidney. Results of chronic exposure studies are controversial even after long-term exposure via drinking water during 9 months reaching 6 µg/g of kidney. During nuclear fuel processing, workers can potentially be exposed to repeated inhalations of uranium compounds. A biokinetic model after uranium intake has been recently updated by the International Commission on Radiological Protection to include recent data. However most of experimental data consider either acute inhalation, chronic ingestion via drinking water, or wound contamination (shrapnels). To evaluate uranium kidney concentration after protracted inhalations, assumptions of repeated acute intake are made, but no experimental data are available to support them. The objectives of this work is therefore to verify 1) if the uranium biokinetic model developed from the data obtained for acute inhalation is consistent with the data obtained for repeated contaminations of animals; 2) if nephrotoxicity modifies uranium retention and excretion; 3) if nephrotoxicity threshold can be predicted by the models. Mice (C57BL6/J) were exposed to different uranyl nitrate hexahydrate concentrations (0.03-3 mg/kg/day) via intranasal instillation four times a week during two weeks. Concentrations of uranium in urine and in tissues were measured at several time points during exposure and up to 42 days post exposure. In parallel, kidney toxicity was evaluated in urine using Clusterin and Kim-1 kidney biomarkers and confirmed by in situ kidney mRNA levels of early and late nephrotoxicity markers. For the lower uranium concentrations, the experimental retention in kidney was well predicted by the standard rat specific biokinetic model when accounting for the exposure profile. For every concentration tested, the amount of activity retained in the kidney reflected the two consecutive weeks of instillation. Uranium biokinetics was modified for higher doses, probably due to tissue alteration and organ dysfunction not considered in biokinetic models. A clear nephrotoxicity and renal impairment was detected for the highest concentration of 3 mg/kg/day and moderate nephrotoxicity was observed at 1 mg/kg/day. Our results suggest that specified biokinetic models should be developed to take into account the alteration of excretion and retention due to organ toxicity. The threshold of uranium nephrotoxicity should be adapted according to the contamination mode and its duration (acute, chronic, and repeated).