Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Spatiotemporal Dynamics of Dissemination of Non-Pandemic HIV-1 Subtype B Clades in the Caribbean Region

  • Marina Cabello,

    Affiliation Laboratório de AIDS e Imunologia Molecular, Instituto Oswaldo Cruz, FIOCRUZ, Rio de Janeiro, Brazil

  • Yaxelis Mendoza,

    Affiliations Department of Genomics and Proteomics, Gorgas Memorial Institute for Health Studies, Panama City, Panama, Department of Biotechnology, Acharya Nagarjuna University, Guntur City, Andhra Pradesh, India, Department of Genetics and Molecular Biology, University of Panama, Panama City, Panama, INDICASAT-AIP, City of Knowledge, Clayton, Panama City, Panama

  • Gonzalo Bello

    gbellobr@gmail.com

    Affiliation Laboratório de AIDS e Imunologia Molecular, Instituto Oswaldo Cruz, FIOCRUZ, Rio de Janeiro, Brazil

Spatiotemporal Dynamics of Dissemination of Non-Pandemic HIV-1 Subtype B Clades in the Caribbean Region

  • Marina Cabello, 
  • Yaxelis Mendoza, 
  • Gonzalo Bello
PLOS
x

Abstract

The Human immunodeficiency virus type-1 (HIV-1) epidemic in the Caribbean region is mostly driven by subtype B; but information about the pattern of viral spread in this geographic region is scarce and different studies point to quite divergent models of viral dissemination. In this study, we reconstructed the spatiotemporal and population dynamics of the HIV-1 subtype B epidemic in the Caribbean. A total of 1,806 HIV-1 subtype B pol sequences collected from 17 different Caribbean islands between 1996 and 2011 were analyzed together with sequences from the United States (n = 525) and France (n = 340) included as control. Maximum Likelihood phylogenetic analyses revealed that HIV-1 subtype B infections in the Caribbean are driven by dissemination of the pandemic clade (BPANDEMIC) responsible for most subtype B infections across the world, and older non-pandemic lineages (BCAR) characteristics of the Caribbean region. The non-pandemic BCAR strains account for >40% of HIV-1 infections in most Caribbean islands; with exception of Cuba and Puerto Rico. Bayesian phylogeographic analyses indicate that BCAR strains probably arose in the island of Hispaniola (Haiti/Dominican Republic) around the middle 1960s and were later disseminated to Trinidad and Tobago and to Jamaica between the late 1960s and the early 1970s. In the following years, the BCAR strains were also disseminated from Hispaniola and Trinidad and Tobago to other Lesser Antilles islands at multiple times. The BCAR clades circulating in Hispaniola, Jamaica and Trinidad and Tobago appear to have experienced an initial phase of exponential growth, with mean estimated growth rates of 0.35–0.45 year−1, followed by a more recent stabilization since the middle 1990s. These results demonstrate that non-pandemic subtype B lineages have been widely disseminated through the Caribbean since the late 1960s and account for an important fraction of current HIV-1 infections in the region.

Introduction

Globally, an estimated 34 million people were infected with the human immunodeficiency virus-type 1 (HIV-1), the aetiologic agent of acquired immunodeficiency syndrome (AIDS), at the end of 2012 [1]. The Caribbean is one of the most severely affected regions in the world after Sub-Saharan Africa. About 250,000 people (1.0% of the adult population) were living with HIV-1 in the Caribbean in 2012, 78% of whom were reported in Hispaniola, the island shared by the Dominican Republic and Haiti [1]. HIV prevalence greatly varies among countries ranging from <0.1% in Cuba to over 2% in the Bahamas and Haiti [1]. The first AIDS cases were recognized in Haiti in 1978–1979 [2] and in other Caribbean countries in 1982–1984 [3], [4], [5]. The main mode of HIV transmission in the region is heterosexual sex [6].

The subtype B dominates the HIV-1 epidemic in most Caribbean islands [7], [8], [9], [10], [11], [12], [13], [14], with exception of Cuba where several non-B genetic forms are collectively more prevalent [15], [16], [17], [18], [19], [20]. Genetic evidence suggests that the HIV-1 subtype B was introduced from Central Africa into America through Haiti around the middle 1960s, coinciding with the return of many Haitian professionals who worked in the Democratic Republic of Congo [21]. According to that study, one subtype B strain was disseminated from Haiti to the United States (US) around 1969 and from the US to the rest of the world, establishing a “BPANDEMIC” clade. Other subtype B lineages, here called “BCAR” clades, remain mostly restricted to Haiti and neighboring Caribbean islands. The study of Gilbert et al (2007), however, analyzed a very low number of HIV-1 subtype B Caribbean sequences (Haiti = 11 and Trinidad and Tobago = 11) and the relative prevalence of the BPANDEMIC and BCAR clades across different Caribbean islands remains largely unknown.

A more recent study that analyzed 836 HIV-1 subtype B pol gene sequences from 13 different Caribbean countries, suggests a very different model of viral dispersion [22]. The study indicates that most of the current subtype B variability in the Caribbean islands was generated since the early 1980s and onwards and that the virus was mainly disseminated from Antigua and Puerto Rico following two distinguishable routes. During the 1980s the virus would have jumped from Antigua to other Lesser Antilles (Barbados, Dominica, Grenada, Trinidad and Tobago, St. Lucia and St. Vincent), Bahamas, Haiti and Jamaica. In the same period, the virus would have spread from Puerto Rico to Cuba, Jamaica, Haiti and Dominican Republic. The authors proposed that BCAR clades early disseminated from Haiti resulted in dead-end infections and that a second introduction of the BPANDEMIC clade from the US through Puerto Rico during the 1980s generated the actual epidemic in the Caribbean region. Other phylogenetic studies, however, reveal that some subtype B pol Caribbean sequences recently isolated branched within the BPANDEMIC clade while other branched at more basal positions within the subtype B phylogeny [23], [24], [25], thus suggesting a continuous circulation of non-pandemic BCAR clades in the Caribbean region.

The objective of this study was to estimate the current prevalence of the BPANDEMIC and BCAR clades in the Caribbean islands and to reconstruct the spatiotemporal dynamics of dissemination of the HIV-1 subtype B in the region. For this, we used a comprehensive dataset of HIV-1 subtype B pol sequences (n = 1,806) isolated from 17 different Caribbean islands between 1996 and 2011. These Caribbean sequences were combined with subtype B sequences from the US (n = 525) and France (n = 340) and subjected to Maximum Likelihood and Bayesian phylogeographic analyses.

Materials and Methods

HIV-1 subtype B pol sequence dataset

We retrieved all HIV-1 subtype B pol sequences with known sampling date from the Caribbean, US and France that covered the entire protease and partial reverse transcriptase (PR/RT) regions (nt 2253–3260 relative to the HXB2 clone) that were available at the Los Alamos HIV Database (http://www.hiv.lanl.gov) by June 2013. Additional HIV-1 subtype B pol sequences available in the same database, but only covering part of the RT (nt 2673–3203 relative to the HXB2 clone) were also downloaded from Barbados, Guadeloupe, Haiti, Martinique, Puerto Rico, and US Virgin Islands. Only one sequence per subject was selected and those sequences containing frameshift mutations were removed from the alignment. For the analyses to run in a reasonable time, some sequences from the US, the most overrepresented country in our dataset with about 10,000 sequences, were removed. In order to generate a “non-redundant” subset representative of the HIV-1 subtype B diversity in the US, highly similar (identity ≥95%) sequences from this country were clustered with the CD-HIT program [26] using an online web server [27] and only one sequence per cluster was selected. This resulted in a final data set of 2,671 subtype B pol sequences isolated from the Caribbean (n = 1,806), US (n = 525), and France (n = 340) between 1982 and 2011 (Table 1). The number and geographic representation of subtype B Caribbean sequences available from other genomic regions was very limited, thus we decided to focus on the pol gene only. The subtype assignment of all sequences included was confirmed using the REGA HIV subtyping tool v.2 [28] and by performing Maximum Likelihood (ML) phylogenetic analyses (see below) with HIV-1 group M subtype reference sequences. Sequences were aligned and all sites associated with major antiretroviral drug resistance in PR (30, 32, 46, 47, 48, 50, 54, 76, 82, 84, 88 and 90) and RT (41, 65, 67, 69, 70, 74, 100, 101, 103, 106, 115, 138, 151, 181, 184, 188, 190, 210, 215, 219 and 230) detected in at least two sequences were excluded. All alignments are available from the authors upon request.

Phylogenetic analysis

Maximum Likelihood (ML) phylogenetic trees were inferred under the GTR+I+Г4 nucleotide substitution model selected using the jModeltest program [29]. The ML trees were reconstructed with the PhyML program [30] using an online web server [31]. Heuristic tree search was performed using the SPR branch-swapping algorithm and the reliability of the obtained topology was estimated with the approximate likelihood-ratio test (aLRT) [32] based on the Shimodaira-Hasegawa-like procedure. The ML trees were visualized using the FigTree v1.4.0 program [33].

Analysis of the spatiotemporal dispersion pattern

The evolutionary rate, the age of the most recent common ancestor (TMRCA) and the spatial diffusion pattern of HIV-1 BCAR clades were jointly estimated using a Bayesian Markov Chain Monte Carlo (MCMC) approach implemented in BEAST v1.8 [34], [35] with BEAGLE to improve run-time [36]. Analyses were performed using the GTR+I+Г4 nucleotide substitution model and a relaxed uncorrelated lognormal molecular clock model [37]. The mean evolutionary rates previously estimated for the subtype B pol gene (1.7–3.0×10−3 subst./site/year) [24], [38], [39], [40] were incorporated as an informative prior interval. Migration events were reconstructed using a reversible discrete phylogeography model [41] with a CTMC rate reference prior [42]. The most relevant migration pathways and the number of viral migrations between locations were estimated by using the Bayesian stochastic search variable selection (BSSVS) [39] and the ‘Markov jump’ counts [43], [44], [45] approaches, respectively. Three MCMC chains were run for 4×108 generations and then combined. Effective Sample Size (ESS) and 95% Highest Probability Density (HPD) values were inspected using Tracer v1.6 [46] to asses convergence and uncertainty of parameter estimates. The maximum clade credibility (MCC) tree was summarized with TreeAnnotator v1.8 and visualized with FigTree v1.4.0. Migratory events were summarized using the SPREAD application [47].

Reconstruction of demographic history

The mode and rate of population growth of some country-specific HIV-1 BCAR clades was also estimated using the BEAST v1.8 software. Changes in effective population size through time were initially estimated using a Bayesian Skyline coalescent tree prior [48] and estimates of the population growth rate were subsequently obtained using the parametric model (logistic, exponential or expansion) that provided the best fit to the demographic signal contained in datasets. Comparison between demographic models was performed using the log marginal likelihood (ML) estimation based on path sampling (PS) and stepping-stone sampling (SS) methods [49]. The MCMC analyses were run for 100–200 million generations. Convergence of parameters and uncertainty in parameter estimates were assessed as described in the previous paragraph.

Results

Evidence of co-circulation of BPANDEMIC and BCAR clades in the Caribbean

In order to estimate the relative prevalence of pandemic (BPANDEMIC) and non-pandemic (BCAR) subtype lineages in the Caribbean, subtype B pol (PR/RT) sequences from different Caribbean islands (n = 743) were combined with subtype B sequences from the US (n = 525) and France (n = 340), two countries with epidemics dominated by the BPANDEMIC clade. The ML analysis revealed that, as expected, most HIV-1 subtype B sequences from the US (89.7%) and France (99.7%) branched in a well supported (aLRT = 0.85) monophyletic subgroup (the BPANDEMIC clade) nested within basal non-pandemic lineages (the BCAR clades) that branched closer to the subtype B root (Fig. 1A). Most of the deepest BCAR lineages were traced to the Caribbean region (83.9%), although their prevalence greatly varies across islands ranging from 96.1% in Trinidad and Tobago to 0.9% in Cuba (Table S1). This analysis also revealed that BCAR sequences from the Dominican Republic and Haiti, two nations located in the island of Hispaniola, were highly intermixed with each other and occupied the deepest branches in the subtype B phylogeny; whereas most BCAR sequences from Jamaica and Trinidad and Tobago branched in two country-specific subclades, BCAR-JM (aLRT = 0.92) and BCAR-TT (aLRT = 0.84), that were nested among basal sequences from Hispaniola (Fig. 1A). The number of BCAR sequences from other Caribbean islands was too small to evaluate their clustering pattern.

thumbnail
Figure 1. ML phylogenetic tree.

A) HIV-1 subtype B pol PR/RT sequences (∼1,000 nt) circulating in the Caribbean (n = 743), US (n = 525), and France (n = 340); B) HIV-1 subtype B pol RT (∼600 nt) sequences from Barbados (n = 14), Guadeloupe (n = 243), Haiti (n = 15), Martinique (n = 452), Puerto Rico (n = 285), US Virgin Islands (n = 54) and representative sequences of the BPANDEMIC (US = 165, France = 135) and the BCAR (Caribbean = 200) clades. Branches are colored according to the geographic origin of each sequence as indicated at the legend (bottom right). Arcs indicate the BPANDEMIC and BCAR clades. Purple and green shaded boxes highlight the position of the Caribbean clades BCAR-TT and BCAR-JM. The aLRT support values are indicated at key nodes: *(0.81–0.90), **(0.91–1). Trees were rooted using HIV-1 subtype D reference sequences. The branch lengths are drawn to scale with the bar at the bottom indicating nucleotide substitutions per site.

https://doi.org/10.1371/journal.pone.0106045.g001

In order to obtain more accurate estimates of the prevalence of BPANDEMIC and BCAR clades in the Caribbean region, we performed a second ML analysis that combined shorter subtype B pol (RT) Caribbean sequences (n = 1,063) from countries poorly represented in the PR/RT dataset (Barbados, Guadeloupe, Haiti, Martinique, Puerto Rico and US Virgin Islands) with reference sequences representative of the BPANDEMIC (US/France = 300) and BCAR (Caribbean = 200) clades selected from the previous analysis. The new ML analysis confirmed the complete segregation of the BPANDEMIC and BCAR reference sequences and the co-circulation of both pandemic and non-pandemic lineages in all Caribbean islands (Fig. 1B); although their relative prevalence greatly vary across countries (Table S1). These results support the existence of three major HIV-1 molecular epidemiologic scenarios in the Caribbean region (Fig. 2). The first one, represented by Haiti, Dominican Republic, Trinidad and Tobago and some other Lesser Antilles, is characterized by the predominance (>70%) of non-pandemic BCAR clades. The second one, represented by Jamaica, Guadeloupe, Martinique, US Virgin Islands and probably the Bahamas, is characterized by roughly similar frequency of both BPANDEMIC and BCAR clades. The third one is represented by Cuba and Puerto Rico where the vast majority (>97%) of subtype B sequences belong to the BPANDEMIC clade.

thumbnail
Figure 2. Estimated proportion of BCAR and BPANDEMIC clades among HIV-1 subtype B infected individuals from different Caribbean countries according to the ML analyses.

The total number of sequences analyzed in each locality is indicated.

https://doi.org/10.1371/journal.pone.0106045.g002

Spatiotemporal dispersal pattern of the HIV-1 BCAR clades in the Caribbean

The origin and spatiotemporal dynamics of non-pandemic subtype B Caribbean lineages was then reconstructed using a Bayesian phylogeographic analysis. The BCAR pol (PR/RT) sequences from the Dominican Republic, Haiti, Jamaica, and the Lesser Antilles were classified into nine discrete locations: Hispaniola (n = 136), Jamaica (n = 73), Trinidad and Tobago (n = 52), Antigua and Barbuda (n = 4), Dominica (n = 2), Grenada (n = 3), Montserrat (n = 1), Saint Lucia (n = 4) and Saint Vincent and the Grenadines (n = 6), and combined with subtype D sequences from the Democratic Republic of Congo (DRC). The estimated evolutionary rate of the HIV-1 B/D pol dataset was 1.74×10−3 (95% HPD: 1.70×10−3–1.80×10−3) substitutions/site per year and the corresponding coefficient of rate variation was 0.27 (95% HPD: 0.22–0.32), thus supporting the selection of a relaxed molecular clock model.

Consistent with the previous ML analysis, the overall topology of the Bayesian MCC tree showed that most sequences from Jamaica and Trinidad and Tobago branched in two country-specific subclades that were nested within the basal sequences from Hispaniola (Fig. 3). The mean estimated TMRCA for the major HIV-1 lineages were as follows: subtypes B/D = 1952, subtype D = 1965, subtype B = 1964, BCAR-TT = 1969, and BCAR-JM = 1971; very similar to that previously described by Gilbert et al (2007) (Table 2). The most probable root location of the HIV-1 subtype B ancestor was placed in Hispaniola (posterior state probability [PSP] = 0.92) (Figs. 3 and 4A). Because most of the HIV-1 BCAR sequences included in our dataset were from Hispaniola (48.7%), we generated five “balanced” subsets containing up to 25 sequences from each location (Table S2). Hispaniola was pointed out as the most probable root location of the B clade in all “balanced” subsets, although the support was lower than that obtained for the complete dataset (PSP = 0.44–0.81) (Figs. 4B to 4F).

thumbnail
Figure 3. Time-scaled Bayesian MCMC tree of the HIV-1 BCAR lineages from the Caribbean and subtype D reference sequences from the Democratic Republic of Congo (DRC).

Branches are colored according to the most probable location state of their descendent nodes. Color code is indicated at the legend at bottom right. Colors circles indicate the positions of nodes corresponding to the most recent common ancestors of major clades, as indicated at the legend at bottom left. Branch lengths are depicted in units of time (years). The tree was automatically rooted under the assumption of a relaxed molecular clock.

https://doi.org/10.1371/journal.pone.0106045.g003

thumbnail
Figure 4. Root location of the HIV-1 subtype B.

Graphics depict the location state posterior probability distributions at the root of the subtype BCAR clades at the Bayesian MCC trees obtained from complete dataset (A) and country “balanced” sub-sets (B to G) from the Caribbean region. Color code is indicated at the legend at bottom left.

https://doi.org/10.1371/journal.pone.0106045.g004

thumbnail
Table 2. Bayesian time-scale estimates for the origin of HIV-1 subtypes B and D.

https://doi.org/10.1371/journal.pone.0106045.t002

Reconstruction of viral migrations across time from the complete dataset revealed a rapid dissemination of BCAR clades across the Caribbean region (Figs. 5A–5D). After the introduction of HIV-1 subtype B into Hispaniola around the middle 1960s, non-pandemic BCAR lineages were independently disseminated to Trinidad and Tobago and Jamaica around the late 1960s and the early 1970s, respectively. Those early introductions seeded secondary outbreaks in Trinidad and Tobago and Jamaica that resulted in the origin of the BCAR-TT and BCAR-JM clades. Several independent transmissions of non-pandemic BCAR clades from Hispaniola to Jamaica (n = 7), Trinidad and Tobago (n = 1) and the other Lesser Antilles (n = 8) were detected from the late 1970s onwards. In the same time period, our data indicates that the BCAR-TT clade was also independently disseminated from Trinidad and Tobago to other Lesser Antilles (n = 6), Jamaica (n = 2) and Hispaniola (n = 1). In contrast, we found no evidence of dissemination of the BCAR-JM clade out of Jamaica.

thumbnail
Figure 5. Spatiotemporal dynamics of dissemination of non-pandemic HIV-1 BCAR clades in the Caribbean region.

A–D) Viral migration events occurred between 1964 and 2005 are indicated. Lines between locations represent branches in the Bayesian MCC tree along which location transitions occurred. The line’s color informs the source location and only the earliest transitions between each location pair were represented. E) Most significant epidemiological links of the dissemination process of BCAR clades. Only epidemiological links supported by Bayes factor rates >3 are displayed. The values of Bayes factor rates are indicated by a color gradient that ranges from light to dark grey as indicated at the legend (up left). F) Viral migrations among locations as measured using ‘Markov jump’ counts. The width of the arrows is proportional to the corresponding mean estimated number of viral transitions between locations. No arrows were displayed when the mean estimated number of transitions was below one.

https://doi.org/10.1371/journal.pone.0106045.g005

The Bayes factor tests for significant nonzero rates, supports epidemiological linkage between DRC and Hispaniola; between Hispaniola and Jamaica/Trinidad and Tobago/Antigua and Barbuda/Dominica/St. Lucia/St. Vincent and the Grenadines; and between Trinidad and Tobago and Jamaica/Grenada/Montserrat/St. Vincent and the Grenadines (Fig. 5E and Table S3). The Markov jump counts analysis indicates that the most viral transitions between epidemiologically linked locations were from Hispaniola to Jamaica and the Lesser Antilles (except Trinidad and Tobago), and from Trinidad and Tobago to the other Lesser Antilles (Fig. 5F and Table S4). Lower numbers of viral migrations were detected between Jamaica/Trinidad and Tobago and between Hispaniola/Trinidad and Tobago. The highest net viral migration flux (efflux minus influx) was for Hispaniola (11.45), followed by Trinidad and Tobago (7.30), Jamaica (−4.05), and other Lesser Antilles (−14.70).

Demographic history of the HIV-1 BCAR clades in the Caribbean

We reconstructed the population dynamic pattern of the HIV-1 BCAR clades from Hispaniola (n = 136), the BCAR-TT clade from Trinidad and Tobago (n = 49) and the BCAR-JM clade from Jamaica (n = 50). Substitution rate and TMRCA estimates obtained in the previous Bayesian analysis were used as prior intervals for demographic reconstructions. The Bayesian skyline plot (BSP) coalescent analysis suggests that all Caribbean clades experienced an initial phase of fast exponential growth followed by a more recent decline in growth rate since the middle 1990s, consistent with a model of logistic growth (Fig. 6). The log ML for the logistic, exponential, and expansion growth models were then calculated using both PS and SS methods. The model of logistic population growth was strongly supported over the other for all HIV-1 Caribbean clades (log BF >3) (Table 3). According to the logistic growth coalescent model, the Caribbean clades exhibited mean initial growth rates that range from 0.36 year−1 to 0.46 year−1, with great overlap of the 95% HPD intervals (Fig. 6).

thumbnail
Figure 6. Demographic history of the HIV-1 BCAR (A, B), BCAR-JM (C, D), and BCAR-TT (E, F) clades circulating in Hispaniola, Jamaica and Trinidad and Tobago, respectively.

Effective number of infections (y-axis; log10 scale) through time (x-axis; calendar years) estimated using Bayesian skyline (A, C, E) and logistic (B, D, F) growth coalescent model. Median estimates of the effective number of infections (solid line) and 95% HPD intervals of the estimates (dashed lines) are shown in each graphic. The median growth rate (with the corresponding 95% credibility interval in parenthesis) of each clade estimated under logistic growth model is indicated in the upper left corner.

https://doi.org/10.1371/journal.pone.0106045.g006

thumbnail
Table 3. Best fit demographic model for different HIV-1 subtype BCAR clades.

https://doi.org/10.1371/journal.pone.0106045.t003

Discussion

This study demonstrates that the HIV-1 subtype B epidemic in the Caribbean is driven by dissemination of the pandemic clade (BPANDEMIC), responsible for most subtype B infections across the world, as well as non-pandemic lineages (BCAR). The relative prevalence of the different subtype B clades greatly varies among countries, giving rise to three major epidemiologic scenarios: 1) islands where BCAR lineages are predominant (Haiti, Dominican Republic and some Lesser Antilles); 2) islands where epidemic is mainly driven by the BPANDEMIC clade (Cuba and Puerto Rico); and 3) islands where both BPANDEMIC and BCAR clades circulates at roughly similar proportions (Jamaica, Guadeloupe, Martinique, US Virgin Islands and the Bahamas).

The differential spreading of BPANDEMIC and BCAR clades across Caribbean islands probably resulted from the combination of chance effects and socio-ecological factors. The HIV-1 epidemic in most Caribbean countries is mainly driven by heterosexual sex [6], with exception of Cuba and Puerto Rico where the epidemic is driven primarily by populations of men having sex with men (MSM) [20], [50] and injection drug users (IDUs) [6], [51], respectively. As a commonwealth of the United States, an extensive migration/travel of Puerto Rican IDUs between Puerto Rico and New York has been reported [52], [53]. It is also interesting to note that the origin of the major HIV-1 subtype B Cuban clades, estimated around the early 1990s [24], coincides with an abrupt increase in the number of tourists from North America and Europe that visited Cuba [54]. These singular socio-ecological factors may have fueled the chance introduction of the BPANDEMIC clade in the Cuban MSM and Puerto Rican IDUs, which explains the preferential dissemination of this subtype B variant in those islands.

Our study indicates that HIV-1 subtype B likely entered in the Americas through the island of Hispaniola around the middle 1960s, in agreement with the model proposed by Gilbert et al (2007). HIV sequences from Haiti and the Dominican Republic were phylogenetically intermixed among each other, consistent with the geographic proximity and intense human mobility between those countries [55], making it difficult to determine which country of Hispaniola was the entrance point for this subtype. Notably, the HIV prevalence in the Dominican Republic is particularly high (5%) among the Haitian-origin communities of the bateyes [56]; suggesting that the Dominican Republic epidemic has been partially driven by links to Haiti. This epidemiological context combined with historical data that also links Haiti to the DRC in the 1960s [21], supports the notion that subtype B was most likely first introduced into Haiti and later disseminated to the Dominican Republic at multiple times.

Non-pandemic BCAR strains probably started to spread from Haiti and/or Dominican Republic to Trinidad and Tobago and Jamaica between the late 1960s and the early 1970s, seeding secondary outbreaks that gave rise to local non-pandemic Caribbean clades here called BCAR-TT and BCAR-JM. The BCAR-TT clade was previously described by Gilbert et al (2007) and comprises 94% of subtype B sequences from Trinidad and Tobago here included, supporting that this epidemic mainly resulted from dissemination of a single founder non-pandemic strain [7], [21], [57]. The BCAR-JM clade comprises 34% of subtype B sequences from Jamaica, indicating a polyphyletic origin of the Jamaican epidemic. Indeed, the second largest Jamaican-specific clade that comprises 19% of subtype B sequences from this country was nested among BPANDEMIC strains. The BCAR strains may have also seeded secondary HIV epidemics in other Lesser Antilles islands, although more sequences should be analyzed to confirm this observation.

While non-pandemic BCAR clades started to spread some years earlier or around the same time as the BPANDEMIC clade, the final outcome of each subtype B lineage was very different. Whereas the BPANDEMIC lineage was disseminated across the world, the BCAR clades seems mainly confined to the Caribbean region. HIV-1 BCAR clades circulate in some Caribbean countries (Haiti, Bahamas, Barbados, Trinidad and Tobago and Jamaica) with high HIV prevalence rates (>1%), thus arguing against the hypothesis of a low epidemic potential of non-pandemic B clades. Our results also revealed that BCAR clades moved from the Caribbean into the US at multiple occasions (Fig. 1A); but most introductions failed to ignite significant outbreaks. This suggests that socio-ecological factors have probably played a key role in the final outcome of the different subtype B lineages. The chance introduction of the BPANDEMIC ancestor into a group of individuals with high rates of partner exchanges and living in a globally interconnected country like the US may explain the successful worldwide dissemination of that viral clade.

The first AIDS cases reported in Jamaica and Trinidad and Tobago in the early 1980s were among MSM who engaged in sex with North American MSM [58], [59]. This observation reinforced the hypothesis that HIV-1 was first introduced into the Caribbean islands via homosexual contact with North American foreigners in the late 1970s or early 1980s [7], [10], [58], [59]. Our results confirm that the BPANDEMIC clade has been introduced and disseminated in most Caribbean islands. The firsts AIDS cases described in Jamaica and Trinidad and Tobago; however, were most probably linked to the transmission of BCAR viruses disseminated out from Hispaniola since the late 1960s. The mean estimated TMRCA of the BCAR-TT (1969) and BCAR-JM (1973) clades coincide with that estimated for the BPANDEMIC clade (1969) [21]; suggesting that HIV epidemics in Jamaica, Trinidad and Tobago and the US started around the same time which is fully consistent with the nearly simultaneous description of the firsts AIDS cases in the those countries at the early 1980s.

The results presented here also clearly contrast with the hypothesis that BCAR clades early disseminated from Haiti resulted in dead-end infections in the Caribbean and that a reintroduction of the BPANDEMIC clade from the US generated the actual epidemic in the Caribbean region [22]. The authors recognize the existence of two major clusters in the Caribbean: cluster I (which grouped most of the sequences from Haiti, the Lesser Antilles, plus half of the Jamaican sequences) and cluster II (which included most of sequences from Cuba, Puerto Rico, and about half of the Jamaican sequences); but both lineages were associated to the BPANDEMIC clade [22]. In our opinion, clusters I and II matched with the BCAR and BPANDEMIC clades here described, respectively. The clear distinction between pandemic and non-pandemic subtype B lineages was probably hampered in the previous study because the absence of reference subtype B sequences from US/Europe and the use of an unrooted phylogenetic tree.

The study of Holguin and Pagan (2013) proposes that the earliest subtype B Caribbean epidemics arose in the islands of Puerto Rico and Antigua around 1980 and that epidemics in Haiti, Dominican Republic and Jamaica only arose around the middle 1980s (95% HPD: 1980–1990). Our study and the study of Gilbert et al (2007), however, indicate that the subtype B epidemics in Hispaniola, Jamaica and Trinidad and Tobago probably arose before 1975. Of note, the mean evolutionary rate estimated for the HIV-1 pol gene in the previous study (3.6×10−3 subs/site/year) [22] was two times higher than the corresponding mean rate estimated here (1.7×10−3 subs/site/year). That rate was also higher than that usually estimated for the pol gene of HIV-1 subtype B (1.0–3.0×10−3 subs/site/year) [24], [38], [39], [40], [60], [61]. and other HIV-1 group M clades (1.0–2.5×10−3 subs/site/year) [24], [61], [62], [63], [64], [65], [66], [67], [68], [69], [70], [71], [72], [73]. That extremely fast calibration clock rate for the HIV-1 pol gene may have pushed TMRCAs estimates of Caribbean epidemics toward misleading young ages.

The study of Holguin and Pagan (2013) also suggests that subtype B was mainly disseminated through the Caribbean following two routes: clade I would have jumped from Antigua to other Lesser Antilles, the Bahamas, Haiti and Jamaica; and clade II would have spread from Puerto Rico to Cuba, Jamaica, Haiti and Dominican Republic. Our study suggests a very different scenario in which the BCAR clades (clade I) were disseminated from both Hispaniola and Trinidad and Tobago to the other Caribbean islands. We have not determined the most important hubs of dissemination of the BPANDEMIC clade (clade II); but it is highly improbable to trace the origin of all BPANDEMIC Caribbean sequences to Puerto Rico. The real scenario is probably more complex and other countries that maintain intensive migration/travel with the Caribbean like the US, England, Netherlands, France and Spain may have also acted as important hubs of dissemination of the BPANDEMIC clade into the region.

Our demographic reconstruction suggests that BCAR clades circulating in Hispaniola, Jamaica and Trinidad and Tobago experienced an initial phase of exponential growth followed by a more recent stabilization since the middle 1990s. This reconstructed demographic pattern fully agrees with the epidemiological profile of the Caribbean region, where the number of people living with HIV has remained relatively stable since the late 1990s [1], [6], and resemble the patterns previously described for subtype B epidemics in other American countries including Brazil [74] the US [75] and Panama [25]. Interestingly, the mean growth rates estimated for the BCAR clades from Hispaniola, Jamaica and Trinidad and Tobago (0.35–0.45 year−1) were similar to those estimated for BPANDEMIC clades mainly circulating among heterosexual populations from Panama (0.20–0.40 year−1) [25]; but lower than those estimated for BPANDEMIC clades mainly transmitted among MSM populations from Cuba, Italy, Hong Kong and the United Kingdom (0.5–1.6 year−1) [24], [38], [39], [40]. This suggests that ecological factors, rather than viral lineage characteristics, are the major determinants of the HIV-1 subtype B epidemic growth rate across different countries.

In summary, this study demonstrates that non-pandemic HIV-1 subtype B viral strains have been widely disseminated through the Caribbean since the late 1960s and account for an important fraction (≥50%) of current HIV-1 infections in Haiti, Dominican Republic, Jamaica, the Bahamas and the Lesser Antilles. This study also indicates that Haiti, Dominican Republic and Trinidad and Tobago were probably the major hubs of dissemination of BCAR in the region. Although this is the most comprehensive study of HIV-1 spread in the Caribbean performed to date, future studies would be improved by the use of more geographically balanced datasets as well as longer (ideally full-length) viral genomic sequences.

Supporting Information

Table S1.

Distribution of HIV-1 subtype pol sequences from different countries within the BPANDEMIC and the BCAR clades.

https://doi.org/10.1371/journal.pone.0106045.s001

(PDF)

Table S2.

Number of sequence per location included in the complete and in the country “balanced” HIV-1 BCAR datasets.

https://doi.org/10.1371/journal.pone.0106045.s002

(PDF)

Table S3.

Bayes factor (BF) rates of epidemiological links between locations for dispersal of non-pandemic BCAR lineages in the Caribbean region.

https://doi.org/10.1371/journal.pone.0106045.s003

(PDF)

Table S4.

Mean estimated number of viral migrations between locations for dispersal of non-pandemic BCAR lineages in the Caribbean region.

https://doi.org/10.1371/journal.pone.0106045.s004

(PDF)

Acknowledgments

We thank Vera Bongertz for critical review of the manuscript and comments on English language usage and Ana Carolina Paulo Vicente for logistic support. We also thank the article reviewers for helpful comments.

Author Contributions

Conceived and designed the experiments: GB. Performed the experiments: MC YM GB. Analyzed the data: MC GB. Contributed reagents/materials/analysis tools: MC GB. Contributed to the writing of the manuscript: GB MC YM.

References

  1. 1. UNAIDS (2013) Report on the global AIDS epidemic. Accessed from: http://www.unaids.org/en/media/unaids/contentassets/documents/epidemiology/2013/gr2013/UNAIDS_Global_Report_2013_en.pdf. Accessed 2014 June 20.
  2. 2. Pape JW, Liautaud B, Thomas F, Mathurin JR, St Amand MM, et al. (1983) Characteristics of the acquired immunodeficiency syndrome (AIDS) in Haiti. N Engl J Med 309: 945–950.
  3. 3. Bartholomew C, Raju CC, Jankey N (1983) The acquired immune deficiency syndrome in Trinidad. A report on two cases. West Indian Med J 32: 177–180.
  4. 4. Koenig RE, Pittaluga J, Bogart M, Castro M, Nunez F, et al. (1987) Prevalence of antibodies to the human immunodeficiency virus in Dominicans and Haitians in the Dominican Republic. JAMA 257: 631–634.
  5. 5. Narain JP, Hull B, Hospedales CJ, Mahabir S, Bassett DC (1989) Epidemiology of AIDS and HIV infection in the Caribbean. Bull Pan Am Health Organ 23: 42–49.
  6. 6. Figueroa JP (2008) The HIV epidemic in the Caribbean: meeting the challenges of achieving universal access to prevention, treatment and care. West Indian Med J 57: 195–203.
  7. 7. Cleghorn FR, Jack N, Carr JK, Edwards J, Mahabir B, et al. (2000) A distinctive clade B HIV type 1 is heterosexually transmitted in Trinidad and Tobago. Proc Natl Acad Sci U S A 97: 10532–10537.
  8. 8. Gittens MV, Roth WW, Roach T, Stringer HG Jr, Pieniazek D, et al. (2003) The molecular epidemiology and drug resistance determination of HIV type 1 subtype B infection in Barbados. AIDS Res Hum Retroviruses 19: 313–319.
  9. 9. Vaughan HE, Cane P, Pillay D, Tedder RS (2003) Characterization of HIV type 1 clades in the Caribbean using pol gene sequences. AIDS Res Hum Retroviruses 19: 929–932.
  10. 10. Noel RJ Jr, Chaudhary S, Rodriguez N, Kumar A, Yamamura Y (2003) Phylogenetic relationships between Puerto Rico and continental USA HIV-1 pol sequences: a shared HIV-1 infection. Cell Mol Biol (Noisy-le-grand) 49: 1193–1198.
  11. 11. Heslop OD, Smikle MF, Vickers IE, Christian NA, Harvey KM, et al. (2009) High genetic diversity in human immunodeficiency virus-type 1 in Jamaica. West Indian Med J 58: 195–200.
  12. 12. Nadai Y, Eyzaguirre LM, Sill A, Cleghorn F, Nolte C, et al. (2009) HIV-1 epidemic in the Caribbean is dominated by subtype B. PLoS ONE. 4: e4814.
  13. 13. Roye ME, Amarakoon, II, Hamilton CL, Eyzaguirre LM, Figueroa P, et al. (2011) Genotypic characterization of HIV type 1 in Jamaica. AIDS Res Hum Retroviruses 27: 91–95.
  14. 14. Hamilton CL, Eyzaguirre LM, Amarakoon, II, Figueroa P, Duncan J, et al. (2012) Analysis of protease and reverse transcriptase genes of HIV for antiretroviral drug resistance in Jamaican adults. AIDS Res Hum Retroviruses 28: 923–927.
  15. 15. Cuevas MT, Ruibal I, Villahermosa ML, Diaz H, Delgado E, et al. (2002) High HIV-1 genetic diversity in Cuba. Aids 16: 1643–1653.
  16. 16. Kouri V, Aleman Y, Perez L, Perez J, Fonseca C, et al. (2012) High frequency of antiviral drug resistance and non-B subtypes in HIV-1 patients failing antiviral therapy in Cuba. J Clin Virol 55: 348–355.
  17. 17. Machado LY, Blanco M, Dubed M, Diaz HM, Ruiz NM, et al. (2012) HIV type 1 genetic diversity in newly diagnosed Cuban patients. AIDS Res Hum Retroviruses 28: 956–960.
  18. 18. Machado LY, Dubed M, Diaz H, Ruiz N, Romay D, et al. (2013) Transmitted HIV type 1 drug resistance in newly diagnosed Cuban patients. AIDS Res Hum Retroviruses 29: 411–414.
  19. 19. Perez L, Kouri V, Aleman Y, Abrahantes Y, Correa C, et al. (2013) Antiretroviral drug resistance in HIV-1 therapy-naive patients in Cuba. Infect Genet Evol 16C: 144–150.
  20. 20. Perez L, Thomson MM, Bleda MJ, Aragones C, Gonzalez Z, et al. (2006) HIV Type 1 molecular epidemiology in cuba: high genetic diversity, frequent mosaicism, and recent expansion of BG intersubtype recombinant forms. AIDS Res Hum Retroviruses 22: 724–733.
  21. 21. Gilbert MT, Rambaut A, Wlasiuk G, Spira TJ, Pitchenik AE, et al. (2007) The emergence of HIV/AIDS in the Americas and beyond. Proc Natl Acad Sci U S A 104: 18566–18570.
  22. 22. Pagan I, Holguin A (2013) Reconstructing the Timing and Dispersion Routes of HIV-1 Subtype B Epidemics in The Caribbean and Central America: A Phylogenetic Story. PLoS ONE 8: e69218.
  23. 23. Junqueira DM, de Medeiros RM, Matte MC, Araujo LA, Chies JA, et al. (2011) Reviewing the history of HIV-1: spread of subtype B in the Americas. PLoS ONE 6: e27489.
  24. 24. Delatorre E, Bello G (2013) Phylodynamics of the HIV-1 epidemic in Cuba. PLoS ONE 8: e72448.
  25. 25. Mendoza Y, Martinez AA, Castillo Mewa J, Gonzalez C, Garcia-Morales C, et al. (2014) Human Immunodeficiency Virus Type 1 (HIV-1) Subtype B Epidemic in Panama Is Mainly Driven by Dissemination of Country-Specific Clades. PLoS ONE 9: e95360.
  26. 26. Li W, Godzik A (2006) Cd-hit: a fast program for clustering and comparing large sets of protein or nucleotide sequences. Bioinformatics 22: 1658–1659.
  27. 27. Huang Y, Niu B, Gao Y, Fu L, Li W (2010) CD-HIT Suite: a web server for clustering and comparing biological sequences. Bioinformatics 26: 680–682.
  28. 28. de Oliveira T, Deforche K, Cassol S, Salminen M, Paraskevis D, et al. (2005) An automated genotyping system for analysis of HIV-1 and other microbial sequences. Bioinformatics 21: 3797–3800.
  29. 29. Posada D (2008) jModelTest: phylogenetic model averaging. Mol Biol Evol 25: 1253–1256.
  30. 30. Guindon S, Gascuel O (2003) A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst Biol 52: 696–704.
  31. 31. Guindon S, Lethiec F, Duroux P, Gascuel O (2005) PHYML Online–a web server for fast maximum likelihood-based phylogenetic inference. Nucleic Acids Res 33: W557–559.
  32. 32. Anisimova M, Gascuel O (2006) Approximate likelihood-ratio test for branches: A fast, accurate, and powerful alternative. Syst Biol 55: 539–552.
  33. 33. Rambaut A (2009) FigTree v1.4: Tree Figure Drawing Tool. Available from: http://treebioedacuk/software/figtree/. Accessed 2014 June 20.
  34. 34. Drummond AJ, Nicholls GK, Rodrigo AG, Solomon W (2002) Estimating mutation parameters, population history and genealogy simultaneously from temporally spaced sequence data. Genetics 161: 1307–1320.
  35. 35. Drummond AJ, Rambaut A (2007) BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol Biol 7: 214.
  36. 36. Suchard MA, Rambaut A (2009) Many-core algorithms for statistical phylogenetics. Bioinformatics 25: 1370–1376.
  37. 37. Drummond AJ, Ho SY, Phillips MJ, Rambaut A (2006) Relaxed phylogenetics and dating with confidence. PLoS Biol 4: e88.
  38. 38. Hue S, Pillay D, Clewley JP, Pybus OG (2005) Genetic analysis reveals the complex structure of HIV-1 transmission within defined risk groups. Proc Natl Acad Sci U S A 102: 4425–4429.
  39. 39. Zehender G, Ebranati E, Lai A, Santoro MM, Alteri C, et al. (2010) Population dynamics of HIV-1 subtype B in a cohort of men-having-sex-with-men in Rome, Italy. J Acquir Immune Defic Syndr 55: 156–160.
  40. 40. Chen JH, Wong KH, Chan KC, To SW, Chen Z, et al. (2011) Phylodynamics of HIV-1 subtype B among the men-having-sex-with-men (MSM) population in Hong Kong. PLoS ONE 6: e25286.
  41. 41. Lemey P, Rambaut A, Drummond AJ, Suchard MA (2009) Bayesian phylogeography finds its roots. PLoS Comput Biol 5: e1000520.
  42. 42. Ferreira MAR, Suchard MA (2008) Bayesian analysis of elapsed times in continuous-time Markov chains. Canadian Journal of Statistics 26: 355–368.
  43. 43. Minin VN, Suchard MA (2008) Counting labeled transitions in continuous-time Markov models of evolution. J Math Biol 56: 391–412.
  44. 44. Talbi C, Lemey P, Suchard MA, Abdelatif E, Elharrak M, et al. (2010) Phylodynamics and human-mediated dispersal of a zoonotic virus. PLoS Pathog 6: e1001166.
  45. 45. Nunes MR, Faria NR, Vasconcelos HB, Medeiros DB, Silva de Lima CP, et al. (2012) Phylogeography of dengue virus serotype 4, Brazil, 2010–2011. Emerg Infect Dis 18: 1858–1864.
  46. 46. Rambaut A, Drummond A (2007) Tracer v1.6. Available: http://treebioedacuk/software/tracer/. Accessed 2014 June 20.
  47. 47. Bielejec F, Rambaut A, Suchard MA, Lemey P (2011) SPREAD: spatial phylogenetic reconstruction of evolutionary dynamics. Bioinformatics 27: 2910–2912.
  48. 48. Drummond AJ, Rambaut A, Shapiro B, Pybus OG (2005) Bayesian coalescent inference of past population dynamics from molecular sequences. Mol Biol Evol 22: 1185–1192.
  49. 49. Baele G, Lemey P, Bedford T, Rambaut A, Suchard MA, et al. (2012) Improving the accuracy of demographic and molecular clock model comparison while accommodating phylogenetic uncertainty. Mol Biol Evol 29: 2157–2167.
  50. 50. de Arazoza H, Joanes J, Lounes R, Legeai C, Clemencon S, et al. (2007) The HIV/AIDS epidemic in Cuba: description and tentative explanation of its low HIV prevalence. BMC Infect Dis 7: 130.
  51. 51. Sepulveda-Torres Ldel C, De La Rosa A, Cumba L, Boukli N, Rios-Olivares E, et al. (2012) Prevalence of Drug Resistance and Associated Mutations in a Population of HIV-1(+) Puerto Ricans: 2006–2010. AIDS Res Treat 2012: 934041.
  52. 52. Deren S, Kang SY, Colon HM, Andia JF, Robles RR, et al. (2003) Migration and HIV risk behaviors: Puerto Rican drug injectors in New York City and Puerto Rico. Am J Public Health 93: 812–816.
  53. 53. Deren S, Kang SY, Colon HM, Robles RR (2007) The Puerto Rico-New York airbridge for drug users: description and relationship to HIV risk behaviors. J Urban Health 84: 243–254.
  54. 54. Figueras MA (2001) International Tourism and the Formation of Productive Clusters in the Cuban Economy. Latin American Studies Association (LASA) XXII Congress, Washington, DC Available: http://www3.dogus.edu.tr/memek/DersNotlari/TOUR%20229/TOUR%20229-2004%20fall/cuba_contribution.pdf. Accessed 2014 June 20.
  55. 55. Rojas P, Malow R, Ruffin B, Rothe EM, Rosenberg R (2011) The HIV/AIDS Epidemic in the Dominican Republic: Key Contributing Factors. J Int Assoc Physicians AIDS Care (Chic) 10: 306–315.
  56. 56. Cohen J (2006) HIV/AIDS: Latin America & Caribbean. Dominican Republic: a sour taste on the sugar plantations. Science 313: 473–475.
  57. 57. Collins-Fairclough AM, Charurat M, Nadai Y, Pando M, Avila MM, et al. (2011) Significantly longer envelope V2 loops are characteristic of heterosexually transmitted subtype B HIV-1 in Trinidad. PLoS ONE 6: e19995.
  58. 58. Bartholomew C, Saxinger WC, Clark JW, Gail M, Dudgeon A, et al. (1987) Transmission of HTLV-I and HIV among homosexual men in Trinidad. JAMA 257: 2604–2608.
  59. 59. Murphy EL, Gibbs WN, Figueroa JP, Bain B, LaGrenade L, et al. (1988) Human immunodeficiency virus and human T-lymphotropic virus type I infection among homosexual men in Kingston, Jamaica. J Acquir Immune Defic Syndr 1: 143–149.
  60. 60. Murillo W, Veras N, Prosperi M, de Rivera IL, Paz-Bailey G, et al. (2013) A single early introduction of HIV-1 subtype B into Central America accounts for most current cases. J Virol 87: 7463–7470.
  61. 61. Abecasis AB, Vandamme AM, Lemey P (2009) Quantifying differences in the tempo of human immunodeficiency virus type 1 subtype evolution. J Virol 83: 12917–12924.
  62. 62. Paraskevis D, Magiorkinis E, Magiorkinis G, Sypsa V, Paparizos V, et al. (2007) Increasing prevalence of HIV-1 subtype A in Greece: estimating epidemic history and origin. J Infect Dis 196: 1167–1176.
  63. 63. Salemi M, de Oliveira T, Ciccozzi M, Rezza G, Goodenow MM (2008) High-Resolution Molecular Epidemiology and Evolutionary History of HIV-1 Subtypes in Albania. PLoS ONE 3: e1390.
  64. 64. Dalai SC, de Oliveira T, Harkins GW, Kassaye SG, Lint J, et al. (2009) Evolution and molecular epidemiology of subtype C HIV-1 in Zimbabwe. Aids 23: 2523–2532.
  65. 65. Faria NR, Suchard MA, Abecasis A, Sousa JD, Ndembi N, et al. (2012) Phylodynamics of the HIV-1 CRF02_AG clade in Cameroon. Infect Genet Evol 12: 453–460.
  66. 66. Veras NM, Santoro MM, Gray RR, Tatem AJ, Lo Presti A, et al. (2011) Molecular epidemiology of HIV type 1 CRF02_AG in Cameroon and African patients living in Italy. AIDS Res Hum Retroviruses 27: 1173–1182.
  67. 67. Bello G, Passaes CP, Guimaraes ML, Lorete RS, Matos Almeida SE, et al. (2008) Origin and evolutionary history of HIV-1 subtype C in Brazil. AIDS 22: 1993–2000.
  68. 68. Passaes CP, Bello G, Lorete RS, Matos Almeida SE, Junqueira DM, et al. (2009) Genetic characterization of HIV-1 BC recombinants and evolutionary history of the CRF31_BC in Southern Brazil. Infect Genet Evol 9: 474–482.
  69. 69. Delatorre E, Bello G (2013) Spatiotemporal dynamics of the HIV-1 CRF06_cpx epidemic in western Africa. AIDS 27: 1313–1320.
  70. 70. Delatorre EO, Bello G (2012) Phylodynamics of HIV-1 Subtype C Epidemic in East Africa. PLoS ONE 7: e41904.
  71. 71. Delatorre E, Couto-Fernandez JC, Guimaraes ML, Vaz Cardoso LP, de Alcantara KC, et al. (2013) Tracing the origin and northward dissemination dynamics of HIV-1 subtype C in Brazil. PLoS ONE 8: e74072.
  72. 72. de Oliveira T, Pillay D, Gifford RJ (2010) The HIV-1 subtype C epidemic in South America is linked to the United Kingdom. PLoS ONE 5: e9311.
  73. 73. Bello G, Aulicino PC, Ruchansky D, Guimaraes ML, Lopez-Galindez C, et al. (2010) Phylodynamics of HIV-1 circulating recombinant forms 12_BF and 38_BF in Argentina and Uruguay. Retrovirology 7: 22.
  74. 74. Bello G, Eyer-Silva WA, Couto-Fernandez JC, Guimaraes ML, Chequer-Fernandez SL, et al. (2007) Demographic history of HIV-1 subtypes B and F in Brazil. Infect Genet Evol 7: 263–270.
  75. 75. Robbins KE, Lemey P, Pybus OG, Jaffe HW, Youngpairoj AS, et al. (2003) U.S. Human immunodeficiency virus type 1 epidemic: date of origin, population history, and characterization of early strains. J Virol 77: 6359–6366.