The precise role of inflammation in the constitution of late injuries is less clear. Similarly to others studies, a chronic release of proinflammatory mediators was observed in irradiated tissues in our model. The TNF-a levels of expression were increased during all the observation period. This was consistent with other PR-171 studies reporting that in the lungs, an increase of TNF-a expression was detected during the first 10 days after irradiation and that values of TNF-a mRNA remained elevated 25 weeks after irradiation. IL-2 levels were also elevated during all the study, with the same profile of evolution. On the contrary, IL-1a increased only later in our model. This could be surprising, as in vitro and in vivo studies highlighted that thoracic irradiation induced prolonged release of IL-1a in serum and in lung samples. But Rube et al. already found in irradiated lung samples that after an initial increase of IL-1 during the first hours after irradiation, IL-1 returned to a baseline threshold and elevated markedly only after 8 and 16 weeks. In the present rat model, 6 weeks might correspond to the period of baseline threshold of Rube et al. study. In fact, histological findings and scores were similar to these molecular changes: following an initial inflammatory phase, there was a decrease in inflammation at 6 months and then a rebound of inflammatory reaction was found at 1 year. This late intensification of inflammatory reaction was particularly marked with a very significant increase of the IL1a expression between 6 and 12 months. Similar late intensification was also observed for muscles in our model. Chronic inflammatory reaction is usually considered as the trigger for radiation-induced fibrosis and some authors proposed that pro- and anti-inflammatory cytokines might be used as markers for predicting late injuries: Chen and coworkers found that elevated pre-irradiation levels in IL-1a and IL-6 were predictive of radiation-induced symptomatic pneumopathy. Similarly, Arpin and coworkers found that high levels of IL-10 were protective against radiation induced symptomatic pulmonary fibrosis after thoracic radiotherapy, while elevation of IL-6 was associated with more severe lesions. This might be explained by the known direct antagonism between IL-10 and TGF-b1, and by the inhibition of the production of IL-10 by TGF-b1 associated with IL-6 in the T-reg pathway. Nevertheless, the precise role of these cytokines in the constitution of late lesions remains controversial and some studies were in contradiction with Chen’s and Arpin’s results. Barthelemy-Brichant and coworkers found no relation between IL-6 and radiation pneumonitis and a multicytokine analysis of plasma of patients showed that only low pre-treatment levels of IL-8 were predictive of radiation pneumonitis. Our model confirms that IL-10 might play a protective role against fibrosis, as elevation of IL-10 was slightly related with lower levels of TGF-b, but this role is not confirmed by histological scores. Similarly, the increase of IL-1a expression was not associated with more severe lesions. Finally, the relation between levels of expression of TNF-a, and the fibrosis, the inflammation or the vascular alterations scores was only moderate in our model. On the contrary, the inflammation, cellular alterations and fibrosis scores were related to variations of IL-2.