Our study confirms this hypothesis. We decided to deliver factors secreted by BMSCs in the form of concentrated conditioned medium, to optimize potential effects. We also chose to use a fraction of this BMSC-CM that contains factors of a molecular weight.10 kDa, as this fraction was earlier described as beneficial on myocardial infarct size, oxidative stress and apoptosis. Moreover, we eliminated serum from the medium because of its poorly defined composition. Two complementary behavioural tests were used to assess recovery of motor functions: BBB as a global recovery test, and grid navigation to evaluate more precisely fine motor movements. In both tests, control animals show spontaneous recovery after spinal cord contusion. The great spontaneous recovery obtained here can also be explained by the type of the lesion, which, despite the use of a 250 kdyn force, doesn’t severely affect the ventral part of the spinal cord, thus preserving white matter tracts. According to the literature, 200 kdyn already corresponds to a severe injury, but this parameter is also known to vary upon the animal weight or strain. Despite this fact, BMSC-CM treated rats obtained significantly Diacerein higher scores in both tests compared to control animals. Difference between groups started from days 4 to 7, indicating a rapid effect after administration, as also described after BMSC transplantation in a model of spinal cord contusion injury. Later on, BBB scores continue to improve, with treated rats displaying at each delay higher scores compared to controls. In their study, Namiki et al. showed an improvement of inclined plane scores one week after BDNF infusion by osmotic pump in a SCI clip model. However, this evolution stopped very early after the end of the BDNF treatment. The prolonged effect that we obtain may be due to the combined action of several factors present in our CM, with maybe redundant molecules with variable bioavailabilities. It is also possible that the cocktail of administered factors favours the survival of endogenous cells preserving the tissue from further lesion extension. In other studies, after BMSC transplantation in contusion or compression models of rat SCI, locomotor improvements were also observed during weeks following the graft, with BBB scores reaching similar levels. For example, 4 weeks after spinal contusion and BMSC graft, Karaoz and colleagues obtain a BBB score of 15 while our treated rats reach the score of 16 at the same delay. Also, 3 weeks after BMSC graft in a compression injury model, Quertainmont et al. show that rats reach a BBB score of 12 which is even lower to the score of 15 reached in the present study. This reinforces our hypothesis that BMSC-secreted factors are beneficial after SCI and that they improve functional recovery as do BMSC transplants. Improved recovery can be a consequence of neural tissue protection, which is favoured by angiogenesis. Angiogenesis is indeed known to induce axonal regrowth via the supply of oxygen and nutritional factors that helps preserving injured spinal tissue from further degradation. Moreover, it has been shown that the Dimesna pro-angiogenic VEGF molecule possesses neuroprotective properties. This factor is present in BMSC-CM and our in vitro data on aortic rings confirm that it is involved in the proangiogenic effect. Other factors which are known to promote angiogenesis were also detected in BMSC-CM: osteopontin, matrix metalloproteinase-13 and fibroblast growth factor-binding protein. In vivo, blood vessels observed at the lesion epicentre of treated animals, even if not higher in number, have larger diameters compared to controls, and could thus provide, via increased blood flow, more nutritional substances and oxygen to damaged tissue.