Time-scale of minor HIV-1 complex circulating recombinant forms from Central and West Africa

Abstract

Background: Several HIV-1 circulating recombinant forms with a complex mosaic structure (CRFs_cpx) circulate in central and western African regions. Here we reconstruct the evolutionary history of some of these complex CRFs (09_cpx, 11_cpx, 13_cpx and 45_cpx) and further investigate the dissemination dynamic of the CRF11_cpx clade by using a Bayesian coalescent-based method.

Results: The analysis of two HIV-1 datasets comprising 181 pol (36 CRF09_cpx, 116 CRF11_cpx, 20 CRF13_cpx and 9 CRF45_cpx) and 125 env (12 CRF09_cpx, 67 CRF11_cpx, 17 CRF13_cpx and 29 CRF45_cpx) sequences pointed to quite consistent onset dates for CRF09_cpx (~1966: 1958-1979), CRF11_cpx (~1957: 1950-1966) and CRF13_cpx (~1965: 1958-1973) clades; while some divergence was found for the estimated date of origin of CRF45_cpx clade [pol = 1970 (1964-1976); env = 1960 (1952-1969)]. Phylogeographic reconstructions indicate that the HIV-1 CRF11_cpx clade most probably emerged in Cameroon and from there it was first disseminated to the Central Africa Republic and Chad in the early 1970s and to other central and western African countries from the early 1980s onwards. Demographic reconstructions suggest that the CRF11_cpx epidemic grew between 1960 and 1990 with a median exponential growth rate of 0.27 year^-1, and stabilized after.

Conclusions: These results reveal that HIV-1 CRFs_cpx clades have been circulating in Central Africa for a period comparable to other much more prevalent HIV-1 group M lineages. Cameroon was probably the epicenter of dissemination of the CRF11_cpx clade that seems to have experienced a long epidemic growth phase before stabilization. The epidemic growth of the CRF11_cpx clade was roughly comparable to other HIV-1 group M lineages circulating in Central Africa.

Keywords: Africa; Complex CRFs; HIV-1; Phylodynamics; Phylogeography.

Fig. 1

Time-scaled Bayesian MCC tree of the HIV-1 CRFs09/11/13/45_cpx pol gene fragment. Branch color indicates the subtype classification obtained in this study, according to the legend in top left. The external circular segments highlight the position of each specific clade as indicated at the line. Asterisks point to key nodes with a high (> 0.90) PP support. Branch lengths are drawn to scale with the concentric circles indicating years. The tree was automatically rooted under the assumption of a relaxed molecular clock

Fig. 2

Time-scaled Bayesian MCC tree of the HIV-1 CRFs09/11/13/45_cpx env gene fragment. Branch color indicates the subtype classification obtained in this study, according to the legend in top left. The external circular segments highlight the position of each specific clade as indicated at the line. Asterisks point to key nodes with a high (> 0.90) PP support. Branch lengths are drawn to scale with the concentric circles indicating years. The tree was automatically rooted under the assumption of a relaxed molecular clock

Fig. 3

Time-scaled Bayesian MCC phylogeographic trees of HIV-1 CRF11_cpx pol (a) and env (b) datasets. The color of each branch represents the most probable location origin according to the map given in the figure. Nodes with a relative-high (PP > 0.80 and < 0.90) and high support (PP > 0.90) are marked with black dots and asterisks, respectively. The red dots represent Cameroon as the ancestral root state with location posterior probabilities of 0.99 and 0.98 for pol and env datasets, respectively. Branch lengths are drawn to scale with the concentric circles indicating years. Localities represented are: DRC (CD), Central African Republic (CF), Cameroon (CM), Gabon (GA), Equatorial Guinea (GQ), Chad (TD) and West African countries (WA)

Fig. 4

Spatiotemporal dispersion of the HIV-1 CRF11_cpx in Central and West Africa. Viral migration events were estimated for pol (a) and env (b) fragments. Arrows between locations represent branches in the Bayesian MCC tree along which location transitions occurred. Each panel represents a time interval of locations transitions as reported. Locations are colored according to the legend. Localities codes: CD (DRC), CF (Central African Republic), CM (Cameroon), GA (Gabon), GQ (Equatorial Guinea), TD (Chad) and West African countries (WA)

Fig. 5

Demographic history of the HIV-1 CRF11_cpx epidemic. Non-parametric estimates of effective number of infections through time of the HIV-1 CRF11_cpx epidemic in pol (a) and env (b) datasets are represented by Bayesian skyline plots. Demographic history of HIV-1 CRF11_cpx epidemic based on pol (c) and env (d) datasets were reconstructed using a logistic growth coalescent model. Median estimate of the effective number of infections (solid line) and 95% confidence limits of the estimate (dashes lines) are shown. Vertical axes indicate the estimated effective number of infections and were represented on logarithmic scale. Time scale is in calendar years