Prevalence, genetic diversity and antiretroviral drugs resistance-associated mutations among untreated HIV-1-infected pregnant women in Gabon, central Africa

Gabon occupies 26 7667 km² in the Gulf of Guinea on the Equator, with tropical forest covering three quarters of the territory. The population is around 1.5 million, 73% of whom live in urban areas. Gabon is divided into nine provinces, in which the main cities are: Libreville (Estuaire), the capital of Gabon, in the northwest; Port-Gentil (Ogooué-Maritime), the main harbor and economic capital, in the west; Lambaréné (Moyen-Ogooué) in the centre west; Oyem (Woleu-Ntem) in the north; Makokou (Ogooué-Ivindo) in the northeast; Franceville (Haut-Ogooué) in the southeast; Koulamoutou (Ogooué-Lolo) in the centre east; Mouila (Ngounié) in the centre and Tchibanga (Nyanga) in the south of the country (Figure 1).

In 2005 and 2008, the retrovirology laboratory at the International Centre for Medical Research of Franceville (CIRMF) and the National Program to Fight Against AIDS in Libreville conducted epidemiologic surveys of pregnant women at their first antenatal examination or first postnatal medical consultation, according to the guidelines of the World Health Organization. The study reported here was conducted in the main cities of the nine provinces, thus representing the general population of pregnant women in Gabon. The survey was conducted by a multidisciplinary team, including physicians from the Gabonese Ministry of Health. Anonymous venous blood samples were collected by locally trained health staff. Before testing, fully informed consent was obtained from each woman and when women were younger than 18 years, informed consent was obtained from their parents. The samples were anonymous, but age and geographic origin were retained. The study obtained ethical clearance from the Gabonese public health authorities and from the Gabonese scientific and ethical committees.

A total of 1098 samples were collected in 2005 and 2916 samples in 2008. The mean age of the women was 24.5 ± 6.3 years (range, 14-44 years) in 2005 and 24.0 ± 6.4 years (range, 14-47 years) in 2008. To determine the circulating subtypes and HIV-resistance to ARVs, samples were taken from 107 untreated HIV-1-infected pregnant women aged 26.6 ± 6.4 years (range, 15-40 years) and included for molecular and phylogenetic analysis (data not shown). The women was selected on the basis of a high viral load (≥ 4.0 log₁₀ copies/mL; mean, 5.7 ± 6.3 log₁₀ copies/mL; range, 4.0-7.3 log₁₀ copies/mL) in order to obviate failure of pol gene amplification.

Plasma samples were separated from venous blood and tested for the presence of antibodies to HIV-1 or HIV-2 by ELISA (Genscreen^® HIV-1/2 version 2, Biorad, Marne la Coquette, France) and by a rapid test (Determine™ HIV-1/2, Abbott, Chicago, Illinois, USA), according to the manufacturers' instructions. Antibody pattern was confirmed by western blot (New LAV Blot I & II, BioRad, Marne la Coquette, France).

RNA was extracted from plasma samples with a QIAamp^® Viral RNA Mini Kit (Qiagen, Courtaboeuf, France), and viral plasma RNA from HIV-positive samples was used to quantify the viral load (Generic HIV Charge Viral, Biocentric, Bandol, France). In order to determine the HIV subtypes and resistance-associated mutations in protease (P) and reverse transcriptase (RT) coding regions, pol gene targets of the main ARVs, were amplified by the French National Agency for AIDS Research consensus polymerase chain reaction (PCR) technique (AC-11 Resistance Group, http://​www.​hivfrenchresista​nce.​org).

HIV genotype was determined by local alignment of partial pol sequences with the HIV Drug Resistance Database of Stanford University http://​dbpartners.​stanford.​edu/​RegaSubtyping/​ and the HIV Database of the National Institutes of Health http://​www.​hiv.​lanl.​gov/​content/​sequence/​BASIC_​BLAST/​basic_​blast.​html. To confirm the assignment of HIV subtypes, reference sequences obtained from the Los Alamos Sequence Database and each Gabonese sequence were aligned with the ClustalX program [13]. Manual corrections were performed with the editor program of the MEGA package, and phylogenetic relations were reconstructed by the neighbor-joining method [14, 15]. Bootstrapping was performed with 1000 replicates.

Genotypic resistance to ARVs was defined according to the 2011 Stanford Resistance Surveillance list of critical mutations http://​hivdb.​stanford.​edu, the latest versions of the algorithms of the French National Agency for AIDS Research and the International AIDS Society http://​www.​iasusa.​org. We confirmed our results by determining transmitted ARVs resistance-associated mutations in untreated women, as recommended by Bennett et al. [16].

The newly sequenced HIV-1 strains are available under GenBank accession numbers HQ541872-HQ541944 and HQ541945-HQ541956.

Antibodies to HIV were found in 69 of 1098 pregnant women tested in 2005 (6.3%; 95% CI, 4.9-7.7) and 174 of 2916 in 2008 (6.0%; 95% CI, 5.7-6.3) (Table 1). No significant difference between 2005 and 2008 was observed by age group or by province (data not shown). During both years of the study, seroprevalences increased by age group, with the highest among women aged ≥ 36 years and the lowest among those aged ≤ 15 years. ELISA and western blotting indicated some HIV-2 (n = 3, data not shown); no HIV-1 group O or N was found in our population. HIV-2-infected women were therefore excluded from the molecular studies.

To evaluate the circulating HIV-1 subtypes, a > 1000 base-pair (bp) fragment (n = 95; range, 1004-1073 bp) of the pol gene (P and RT coding regions) was sequenced and phylogenetically analyzed (Figure 2A). The long fragment of the pol gene could not be obtained for 12 of the 107 HIV-1 samples, and analyzed either the RT (n = 2; 785 bp) or the P (n = 10; range, 467-479 bp) coding region (Figure 2B and 2C).

Phylogenetic analysis of the pol gene showed wide genetic diversity (Figure 2A, B and 2C). The genotypes distribution is shown in Table 2. The main subtype identified was CRF02_AG (n = 50, 46.7%), followed by subtypes A (n = 21, 19.6%), G (n = 11, 10.3%), F (n = 5, 4.7%), H (n = 2, 1.9%) and D (n = 1, 0.9%). Complex recombinant forms were also found, consisting of CRF06_cpx and CRF11_cpx (n = 2, 1.9%). Thirteen subtypes could not be characterized (12.1%).

Resistance to ARVs was genotyped in 107 HIV-1-positive samples. Data on mutations conferring resistance to protease inhibitors (PIs), nucleoside reverse transcriptase inhibitors (NRTIs) and non-NRTIs are summarized in Table 2. In all, 44/107 (41.1%) HIV-1 strains presented minor or major genotypic resistance to ARVs, of which 3/107 (2.8%) were major mutations for resistance to PIs (n = 2/107, 1.9%) and NRTIs (n = 1/107, 0.9%). No major mutation for resistance to non-NRTIs was found.

The most frequent mutations observed in the P coding region corresponded to subtype polymorphism (data not shown), while minor mutations, such as L10V/I, V11I, E35G, A71T and T74S, were associated with low-level resistance to some PIs (Table 2). L33F and M46I indicate major PI resistance. In the RT coding region, strong polymorphism was found (data not shown), with common mutations such as V90I, V106I and E138A (Table 2). A62V and T69S are minor mutations, which alone do not affect susceptibility to NRTIs, like K103R and V179I to non-NRTIs. L210W contributes to resistance to NRTIs, and K101Q causes low-level reduction in susceptibility to non-NRTIs. However, the only mutation known to be associated with resistance to NRTIs was V118I.

Six of the 44 (13.6%) HIV-1 strains had mutations for resistance to PIs and non-NRTIs, and one (2.3%) was resistant to both NRTIs and non-NRTIs. Five strains (11.4%) cumulated mutations to PIs, and three (6.8%) cumulated mutations to non-NRTIs (Table 2).