Molecular epidemiology of HIV-1 subtypes in India: origin and evolutionary history of the predominant subtype C

Abstract

Background: India has the third largest HIV-1 epidemic with 2.4 million infected individuals. Molecular epidemiological analysis has identified the predominance of HIV-1 subtype C (HIV-1C). However, the previous reports have been limited by sample size, and uneven geographical distribution. The introduction of HIV-1C in India remains uncertain due to this lack of structured studies. To fill the gap, we characterised the distribution pattern of HIV-1 subtypes in India based on data collection from nationwide clinical cohorts between 2007 and 2011. We also reconstructed the time to the most recent common ancestor (tMRCA) of the predominant HIV-1C strains.

Methodology/principal findings: Blood samples were collected from 168 HIV-1 seropositive subjects from 7 different states. HIV-1 subtypes were determined using two or three genes, gag, pol, and env using several methods. Bayesian coalescent-based approach was used to reconstruct the time of introduction and population growth patterns of the Indian HIV-1C. For the first time, a high prevalence (10%) of unique recombinant forms (BC and A1C) was observed when two or three genes were used instead of one gene (p<0.01; p = 0.02, respectively). The tMRCA of Indian HIV-1C was estimated using the three viral genes, ranged from 1967 (gag) to 1974 (env). Pol-gene analysis was considered to provide the most reliable estimate [1971, (95% CI: 1965-1976)]. The population growth pattern revealed an initial slow growth phase in the mid-1970s, an exponential phase through the 1980s, and a stationary phase since the early 1990s.

Conclusions/significance: The Indian HIV-1C epidemic originated around 40 years ago from a single or few genetically related African lineages, and since then largely evolved independently. The effective population size in the country has been broadly stable since the 1990s. The evolving viral epidemic, as indicated by the increase of recombinant strains, warrants a need for continued molecular surveillance to guide efficient disease intervention strategies.

Figure 1. Prevalence of HIV-1 subtypes and recombinant forms in India based on one gene, two and three genes.

The p values are presented along with the values for the predominant HIV-1C and recombinants. Subtyping was carried out as described in the methods section. For sequences that indicated an event of recombination, bootscan analysis was performed in Simplot version 3.5.1 using 100 nucleotide window size and 20 nucleotide step size to map the precise breakpoint. The recombination event was further confirmed by region-specific phylogenetic analysis using ML tree in MEGA 5.

Figure 2. Distribution of HIV-1 subtypes and recombinants in the clinical cohorts based on two genes.

The regions [southern (Karnataka, Tamil Nadu and Andhra Pradesh), northern (Punjab and Haryana), north-eastern (Manipur) and central (Madhya Pradesh)] from where the samples were collected are shown in colors. The pie chart depicts the percentage of subtypes and recombinant strains in respective regions.

Figure 3. Annotated phylogenetic trees for the gag- (Panel A), env (Panel B) and pol- (Panel C) genes.

Phylogenetic trees were constructed using the Indian HIV-1 subtype C from clinical samples taken from different parts of India (shown in green), including southern (Karnataka, Tamil Nadu and Andhra Pradesh), northern (Punjab and Haryana), north-eastern (Manipur) and central (Madhya Pradesh) regions of India plus data base retrieved reference sequences from western India and other countries, and two out-group subtype B strains. The Indian tMRCA included in this study were shown with an arrow. Pol-gene analysis was considered to provide the most reliable estimate of tMRCA (shown with a red arrow) for Indian HIV-1C, indicating around 1971 (95% CI: 1965–1976)]. Posterior probability plots showing the tMRCA estimates of Indian HIV-1C (Panel D): The posterior probability density plots for the three genes (env, pol and gag) point to a tMRCA for Indian HIV-1C from 36.9 (env), 40.5 (pol) and 44.3 (gag) years ago, with an overall mean for the combined graph at 40.6 years ago before 2011, that is in 1970.4 (95% CI: 1960.3–1978.1).

Figure 4. Bayesian skyline plot representing estimates of the effective number of infections through time.

The demographic history from the pol BSP identified three epidemic growth phases divided by the vertical lines; an initial slow growth followed by an exponential phase which slowed down to a stationary phase around 1990. The identification of first HIV case in India is indicated by an arrow.