Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disease characterized by a loss of motor neurons in the brain (brainstem and cortex) and the spinal cord that leads to a motor neurological symptomatology. Approximately 10% of ALS cases have a familial form of ALS. Among the familial cases, 20% are caused by dominantly inherited mutations (around 100 different mutations) in the protein Cu/Zn superoxide dismutase (SOD1). To date, genetic ALS models on mutated SOD1 (SOD1G93A, SOD1G85R and SOD1G37R) transgenesis in rodents have been successful in recapitulating the main features of the human pathology, especially the loss of spinal or facial motoneurons, the increased astrogliosis, the activation of microglia and the cellular and molecular disruptions in motoneurons. The therapies toward a treatment or cure for ALS have been met with limited success. In the present study, SOD1G93A transgenic mice have been used to test new gene therapies to slow disease progression and pharmacological approach to obtain a better comprehension of the pathological process, more precisely, to define the involvement of microglial activation in the spread of the disease. For the approach by gene therapy, injection of viral vectors in a localized region of central nervous system (CNS) constitutes an excellent tool for the development of new therapies. In the present study, the viral-mediated expression of potential neuroprotective factors was explored in the lumbar spinal cord or facial nucleus of SOD1G93A transgenic mice. As HIV-1-derived lentiviral vectors can efficiently transduce neurons in CNS, these retroviral vectors were used to over-express the neurotrophic factor GDNF (glial cell line-derived neurotrophic factor) to supply a trophic support to lumbar and facial motoneurons, or RNAi molecules specifically targeting the SOD1G93A gene to inhibit the mutant SOD1 expression, the cause of the disease, in the lumbar spinal cord of SOD1G93A transgenic mice. For the study of the level of microglial involvement in the pathology, SOD1G93A transgenic mice had a daily pharmacological treatment with an inhibitor of microglial activation, the minocycline. The present thesis demonstrates that, on the one hand, at the difference of laboratories that delayed the disease progression by intramuscular injection of adenoviral vectors or adenoassociated vectors encoding for GDNF, the intraspinal injection of lentivirus encoding for GDNF did not lead to a protection of lumbar motoneurons and a delay of the disease. But this study showed a significantly rescue of facial motoneurons by the injection of lentivirus encoding for GDNF directly in facial nucleus. A selective vulnerability between motoneurons in facial nucleus and these in lumbar spinal cord has been therefore underlined. On the other hand, we showed that the therapeutic approach to inhibit the mutant SOD1 expression look as the most effective to delay the progression of the ALS pathology. Both studies of gene therapy, compared to others studies using the same therapeutic factors, demonstrated clearly that to optimize a therapeutic treatment, it seems better to act at muscular junction level with a retrograde transport form muscle to motoneuronal bodies of the therapeutic molecule or viral vector. Nevertheless, motoneuron-restricted silencing of mutant SOD1 would be useful for better understanding of neuronal and glial cellular mechanism mediating ALS-linked mutated SOD1 toxicity. Regarding the involvement of microglia in the pathological process, the inhibition of microglial activation by the minocycline treatment, suggests, on the one hand, that microglial activation do not initiate the pathological process and that this microgliosis might be a consequence and triggered by multiple disruptions. On the other hand, that minocycline could protect completely facial motoneurons and partially spinal motoneurons by its anti-apoptotic property. Therapeutic approaches aiming to act directly on the cause of the disease as well as a better comprehension of the role of non-neuronal cells in the pathological process may therefore open new perspectives for the treatment of ALS pathology.