Evolución del VIH-1 durante el proceso de recuperación de la eficacia biológica in vitro

Resumen

Human immunodeficiency virus type 1 (HIV-1) constitutes one of the major medical challenges due to infectious diseases the world is facing, affecting more than 30 million people. Due to its high mutation rate, enormous population sizes and rapid turnover, HIV-1, as the other RNA viruses, generates a swarm of mutants known as viral quasispecies. The quasispecies and not single genomes are the object of evolution, thus studying its properties will lead to a better understanding of viral evolution. For that purpose, ¿in vitro¿ experiments are usually employed to simulate the population size changes the virus suffers during natural infections. In the present thesis, we have subjected 6 biological clones, that had suffered drastic fitness losses due to plaque-to-plaque passages, to 30 large population passages in cell culture to recover fitness. After the passages, we performed a complete analysis of the viruses including fitness, phenotypic properties, nucleotide sequences at the consensus level and quasispecies heterogeneity. We also studied individually some of the mutations that appeared during the passages performed. Finally, we used all the nucleotide sequences and obtained along the passages and their fitness values to generate a real fitness landscape of the recovery process of the viruses. All viral populations recovered fitness after the 30 passages with a continuing overall grow (from a mean fitness of 0.52 in passage 1 to 1.58 in passage 31), but with significant differences between viruses. When the potential causes of fitness recovery were analysed, a high positive correlation between heterogeneity and fitness (p-value = 0.0057) was obtained. The quasispecies heterogeneity together with the evolutionary rate were the variables most associated with the fitness recovery using a multiple regression analysis. We also observed that this increase in quasispecies complexity allowed the virus to fix new beneficial mutations, of which almost 90% were non-synonymous, with an average of 8.3 mutations per virus at passage 31. The mutational pattern, although presenting some convergent mutations and deletions, was very different between viruses and did not present hotspots. Interestingly, some mutations associated with important phenotypic effects, such as tropism change or antiviral resistance, appeared in these ¿in vitro¿ conditions without any external selective pressure. In addition, we detected that not only the consensus sequences were positively selected but also, at the quasispecies level, the mutant spectra was shaped by positive selection. Using all quasispecies sequences within each virus, we generated phylogenetic networks to study the appearance of mutations, to detect the imposition of beneficial changes in viral populations and to describe the quasispecies dynamics throughout the passages. Ultimately, we generated a real fitness landscape with all the sequences obtained during these analyses, using phylogenetic methods that produced descriptive landscapes. Moreover, we used artificial neural networks (ANN) to gain projective capacity, allowing the construction of predictive fitness landscapes that could be used to estimate unknown fitness values. Using all the analyses performed in this thesis, we have disclosed different pathways of ¿in vitro¿ fitness recovery at the consensus level and analysed the dynamics of variability generation and mutant selection within viral quasispecies. To conclude, in the process of HIV-1 ¿in vitro¿ biological fitness recovery the quasispecies heterogeneity is one of the most important factors driving viral evolution. The generation of complexity leads to the fixation of beneficial mutations and determines the properties of the viral populations.