Supervisors info:
Παρασκευής Δημήτριος, Επίκουρος Καθηγητής, Ιατρική, ΕΚΠΑ
Σύψα Βάνα, Επίκουρη Καθηγήτρια, Ιατρική, ΕΚΠΑ
Μαγιορκίνης Γκίκας, Λέκτορας, Ιατρική, ΕΚΠΑ
Summary:
Introduction: Previous studies have shown that HIV-1 subtypes A1 and B are the predominant clades in Greece.
Aim: To investigate the patters of subtypes A1 and B dispersal in Greece, to identify the proportion of infections occurred locally, and to estimate the factors that are associated with local transmissions, using molecular epidemiology methods.
Material: We studied 1,230 A1 and 2,156 B sequences isolated from HIV-1 diagnosed patients during 1996-06/2015 in Greece. Sequences were available in the protease (PR) and partial reverse transcriptase (RT) regions.
Methods: Maximum-likelihood (ML) phylogeny reconstruction with bootstrap evaluation was conducted in RAxML8, using GTR+G as nucleotide substitution model. Phylogenetic analysis was performed separately on the 1,230 A1 and 2,156 B sequences from Greece along with a random set of globally sampled sequences (A1: N=1,500; B: N=2,000), used as references. Phylogenetic analysis was performed for each subtype in five replicates, using a different reference dataset each time. Local transmission networks (LTNs) were defined as phylogenetic clusters including sequences from Greece at proportions >70% (geographic criterion), found in all five replicates (phylogenetic confidence). Local dispersal was estimated by dividing the number of sequences found within LTNs with the total number of sequences sampled in Greece for each subtype. The statistical analysis was carried out using multivariate logistic regression models. Presence in LTNs was the binary outcome variable, while age, period of sampling, risk group, nationality and gender were chosen as possible explanatory variables.
Results: Phylogenetic analysis revealed that 91.5% (1,125 out of 1,230) of the A1 sequences sampled in Greece formed 23 LTNs. The size of the LTNs ranged between 2 and 1,059 sequences. The 86.1% of the A1 sequences fell within a single LTN consisted of 1,059 sequences. The rest of the A1 sequences (N=105, 8.5%) clustered at different points in the ML tree. The majority of the unclustered sequences had been isolated mainly from heterosexuals with non-Greek nationality. Phylogenetic analysis also revealed that 72.8% (1,570 out of 2,156) of the B sequences sampled in Greece formed 133 LTNs, with a range of 2 to 157 sequences. Multivariate logistic regression analysis showed that gender (male; OR=1.8; p=0.034), risk group (MSM; OR=7.2; p<0.001) and nationality (Greek; OR=7.2; p<0.001) were associated with regional clustering of subtype A1, while period of sampling (2011-2015; OR=7.21; p<0.001) with local transmissions of subtype B.
62
Conclusions: Our analysis suggests considerable differences in the levels of regional clustering of HIV-1 subtype A1 and B in Greece. For subtype A1 we found a lower number of introductions than for subtype B. This probably due to that subtype B is more frequently found in developed countries, suggesting that cross-border transmissions are more likely than non-B clades. It was also shown that the majority of A1 unclustered sequences had been isolated mainly from heterosexuals with non-Greek nationality. This finding suggests that this population group was infected with A1 strains with non-Greek origin and it is not related to further dispersal. Moreover, we found that local dispersal of subtype B transmissions is mostly associated with recent diagnosis.
Keywords:
HIV, Molecular epidemiology, Transmission networks, Phylogenetic analysis, Subtypes
