MELAS (mitochondrial encephalomyopathy, lactic acidosis and stroke-like episodes) is a mitochondrial disorder caused mainly by the m.3243A>G mutation in mitochondrial DNA. In this thesis, we report on how the severity of pathophysiological alterations is differently expressed in fibroblasts derived from patients with MELAS disease. We evaluated mitophagy activation and mitochondrial biogenesis which are the main mechanisms regulating the degradation and genesis of mitochondrial mass in transmitochondrial cybrids and fibroblasts derived form MELAS patients. Our results suggest a critical balance between mitophagy and mitochondrial biogenesis which leads to the expression of different degrees of pathological severity among MELAS fibroblast cell lines according to their heteroplasmy load and the activation of AMP-activated protein kinase (AMPK). AMPK-activators such as 5-aminoimidazole-4-carboxamide 1-ß-D-ribofuranoside (AICAR) or coenzyme Q10 (CoQ) increased peroxisome proliferator-activated receptor alpha (PGC-1¿) nuclear translocation, mitochondrial biogenesis, antioxidant enzyme system response, autophagic flux and ultimately improved pathophysiological alterations in MELAS fibroblasts with the most severe phenotype. Our findings support the hypothesis that mitochondrial biogenesis, increased antioxidant response and autophagy clearance serve as compensatory mechanisms in response to mitophagic degradation of dysfunctional mitochondria and point out that AMPK is an important player in this balance. These results are particularly important since currently no efficient treatments are available for this chronic progressive disorder. Furthermore, in this thesis we propose the evaluation of the effectiveness of putative beneficial pharmacological agents in the treatment of MELAS by using cellular models such as transmitochondrial cybrids and fibroblasts with high mutational load. According to our results, supplementation with riboflavin or coenzyme Q10 effectively reversed the pathologic alterations in MELAS cybrid and fibroblast cell models. Our results indicate that cell models manifesting severe pathophysiological alterations and high heteroplasmy load have great potential as a screening and validation assays of novel drug candidates for MELAS treatment and presumably also for other diseases with mitochondrial impairment