Rising prevalence of non-B HIV-1 subtypes in North Carolina and evidence for local onward transmission

Abstract

HIV-1 diversity is increasing in North American and European cohorts which may have public health implications. However, little is known about non-B subtype diversity in the southern United States, despite the region being the epicenter of the nation's epidemic. We characterized HIV-1 diversity and transmission clusters to identify the extent to which non-B strains are transmitted locally. We conducted cross-sectional analyses of HIV-1 partial pol sequences collected from 1997 to 2014 from adults accessing routine clinical care in North Carolina (NC). Subtypes were evaluated using COMET and phylogenetic analysis. Putative transmission clusters were identified using maximum-likelihood trees. Clusters involving non-B strains were confirmed and their dates of origin were estimated using Bayesian phylogenetics. Data were combined with demographic information collected at the time of sample collection and country of origin for a subset of patients. Among 24,972 sequences from 15,246 persons, the non-B subtype prevalence increased from 0% to 3.46% over the study period. Of 325 persons with non-B subtypes, diversity was high with over 15 pure subtypes and recombinants; subtype C (28.9%) and CRF02_AG (24.0%) were most common. While identification of transmission clusters was lower for persons with non-B versus B subtypes, several local transmission clusters (≥3 persons) involving non-B subtypes were identified and all were presumably due to heterosexual transmission. Prevalence of non-B subtype diversity remains low in NC but a statistically significant rise was identified over time which likely reflects multiple importation. However, the combined phylogenetic clustering analysis reveals evidence for local onward transmission. Detection of these non-B clusters suggests heterosexual transmission and may guide diagnostic and prevention interventions.

Keywords: HIV-1; molecular epidemiology; southeastern United States; transmission.

Figure 1.

(A) Change in the proportion of non-B subtypes over time among 15,246 HIV-1 pol sequences collected in North Carolina from 1997 to 2014. Proportions are based on sampling date for the first sequence available per individual by calendar year. (B) Percentage of non-B subtypes by total number of non-B subtype sequences by calendar year of sample acquisition.

Figure 2.

Maximum Likelihood phylogenetic tree of HIV-1 non-subtype B pol sequences (n=461) from 325 HIV-infected men and women in North Carolina and 428 references from the Los Alamos National Laboratory (LANL) HIV database (http://www.hiv.lanl.gov). Subtype and Circulating Recombinant Forms (CRF) references are indicated by diamonds with black outline (n=159). Color coding indicates the common subtypes and CRFs among study sequences. LANL references identified through BLAST search of study sequences are shown with grey diamonds (n=269). Putative transmission clusters are highlighted in red and indicated with red dots. Number label at tips are cluster identifiers. Branch lengths represent nucleotide substitutions per site.

Figure 3.

Non-B subtype transmission clusters (n=32) involving 81 HIV-1 pol sequences from individuals sampled in North Carolina. Clusters were confirmed through Bayesian Markov Chain Monte Carlo inference in BEAST. The scale bar (x-axis) is calendar years. Black circles at nodes indicate cluster origin and estimated time of the most recent common ancestor (tMRCA) and tips correspond to sampling date. Nodes with posterior probabilities >0.90 are labeled. Region of origin (if known) is indicated by tip colors. Numbers at tips indicate year of diagnosis (if available). HET indicates reported heterosexual transmission.