Spatio-Temporal History of HIV-1 CRF35_AD in Afghanistan and Iran

Abstract

HIV-1 Circulating Recombinant Form 35_AD (CRF35_AD) has an important position in the epidemiological profile of Afghanistan and Iran. Despite the presence of this clade in Afghanistan and Iran for over a decade, our understanding of its origin and dissemination patterns is limited. In this study, we performed a Bayesian phylogeographic analysis to reconstruct the spatio-temporal dispersion pattern of this clade using eligible CRF35_AD gag and pol sequences available in the Los Alamos HIV database (432 sequences available from Iran, 16 sequences available from Afghanistan, and a single CRF35_AD-like pol sequence available from USA). Bayesian Markov Chain Monte Carlo algorithm was implemented in BEAST v1.8.1. Between-country dispersion rates were tested with Bayesian stochastic search variable selection method and were considered significant where Bayes factor values were greater than three. The findings suggested that CRF35_AD sequences were genetically similar to parental sequences from Kenya and Uganda, and to a set of subtype A1 sequences available from Afghan refugees living in Pakistan. Our results also showed that across all phylogenies, Afghan and Iranian CRF35_AD sequences formed a monophyletic cluster (posterior clade credibility> 0.7). The divergence date of this cluster was estimated to be between 1990 and 1992. Within this cluster, a bidirectional dispersion of the virus was observed across Afghanistan and Iran. We could not clearly identify if Afghanistan or Iran first established or received this epidemic, as the root location of this cluster could not be robustly estimated. Three CRF35_AD sequences from Afghan refugees living in Pakistan nested among Afghan and Iranian CRF35_AD branches. However, the CRF35_AD-like sequence available from USA diverged independently from Kenyan subtype A1 sequences, suggesting it not to be a true CRF35_AD lineage. Potential factors contributing to viral exchange between Afghanistan and Iran could be injection drug networks and mass migration of Afghan refugees and labours to Iran, which calls for extensive preventive efforts.

Fig 1. Genome structure of HIV-1 CRF35_AD.

In this recombinant lineage, subtype D fragments are found within the gag, pol, and env genomic regions, in a genomic background that matches subtype A1. We created four non-overlapping alignments, each of which belonged to one parental lineage of the CRF35_AD only. These alignments include gag_1, gag_2, pol_1, and pol_2. From the pol_1 alignment, which mainly covers parental D fragments, the A1 sections were removed, and the remaining D sections were concatenated. A similar approach was taken to remove parental D section from the pol_2 alignment, which mainly covers parental A1 fragments. Details of nucleotide positions related to these alignments can be found in the Methods.

Fig 2. Time-scaled Bayesian MCC trees of the HIV-1 CRF35_AD.

(a) gag_1; (b) gag_2; (c) pol_1; (d) pol_2. Key nodes with a posterior clade credibility of 0.7 or greater are marked with black rectangles. Most probable location state and its Posterior State Probability (PSP) are indicated in front of the nodes, corresponding to between-country migration events of CRF35_AD. These nodes are highlighted in Yellow (for Iran → Afghanistan transmission), green (for Afghanistan → Iran transmission), and gray (for Kenya → USA transmission). A highly supported subtype A1 cluster indicating transmission from Afghan refugees to Native Pakistanis is highlighted in blue (Part b). Other parental A1 or D sequences are collapsed for visual clarity. To the upper left of all phylogenies is a bar chart indicating posterior probabilities for the location states of the root of the CRF35_AD_{Afghan-Iranian} cluster (node A). The color code is indicated in the legend, on the lower left corner. PSP: Posterior State Probability; A1: HIV-1 subtype A1; D: HIV-1 subtype D; AF: Afghanistan; IR: Iran; AF refugees _PK: Afghan refugees living in Pakistan; BJD: Boroujerd; HE: Herat; KB: Kabul; KE: Kenya; KSH: Kermanshah; MSH: Mashhad; PK: Pakistan; SDJ: Sanandaj; SHZ: Shiraz; TEH: Tehran; UG: Uganda.

Fig 3. Spatio-temporal dynamics of HIV-1 CRF35_AD dispersion among Afghanistan, Iran, and Afghan refugees living in Pakistan, between 1990 and 2005.

Arrows between locations represent branches in the time-scaled Bayesian MCC tree for the gag_1 (a), gag_2 (b), pol_1 (c), and pol_2 (d) datasets. Only migration events supported by a posterior clade credibility of 0.7 or greater are indicated. Arrow colors reflect the timing of location transitions according to the legend at upper right corner. Lighter colors represent more recent transitions, and darker colors represent older history. AF: Afghanistan; IR: Iran; PK: Pakistan; BJD: Borujerd; HE: Herat; KB: Kabul; KSH: Kermanshah; MSH: Mashhad; SDJ: Sanandaj; SHZ: Shiraz; TEH: Tehran.