Recent increased identification and transmission of HIV-1 unique recombinant forms in Sweden

Abstract

A temporal increase in non-B subtypes has earlier been described in Sweden by us and we hypothesized that this increased viral heterogeneity may become a hotspot for the development of more complex and unique recombinant forms (URFs) if the epidemics converge. In the present study, we performed subtyping using four automated tools and phylogenetic analysis by RAxML of pol gene sequences (n = 5246) and HIV-1 near full-length genome (HIV-NFLG) sequences (n = 104). A CD4⁺ T-cell decline trajectory algorithm was used to estimate time of HIV infection. Transmission clusters were identified using the family-joining method. The analysis of HIV-NFLG and pol gene described 10.6% (11/104) and 2.6% (137/5246) of the strains as URFs, respectively. An increasing trend of URFs was observed in recent years by both approaches (p = 0·0082; p < 0·0001). Transmission cluster analysis using the pol gene of all URFs identified 14 clusters with two to eight sequences. Larger transmission clusters of URFs (BF1 and 01B) were observed among MSM who mostly were sero-diagnosed in recent time. Understanding the increased appearance and transmission of URFs in recent years could have importance for public health interventions and the use of HIV-NFLG would provide better statistical support for such assessments.

Figure 1

(A) Maximum-likelihood phylogenetic analysis. The HIV-NFLG sequenced in our study (n = 104) and a unique set of 175 reference sequences downloaded from Los Alamos database were used. (B) Distribution of HIV-1 pure subtypes (n = 80), circulating recombinant forms (CRFs) (n = 13) and unique recombinant forms (URFs) (n = 11) based on three subtyping tools (REGA v3, RIP 3.0 and COMET-HIV) and ML-phylogenetic analysis. (C) Proportion of pure subtypes/CRFs and URFs among patients infected in or outside Sweden, according the reports from the treating physician. A higher proportion of URFs was identified among patients infected in Sweden (n = 8; 18%) than outside (n = 3; 5%) (Fisher’s exact test p = 0·0523). (D) Distribution of pure subtypes/CRFs and URFs in relation to year of transmission predicted by the CD4⁺ T-cell decline trajectory algorithm (n = 79) or by a serologically verified primary HIV infection (n = 9). The trends indicate a significant increase of URFs in newly diagnosed patients in Sweden over time (Chi-square test: 6·986; p = 0·0081).

Figure 2

(A) Mosaic pattern of three HIV-1 A1D recombinants. Precise inter-subtype recombination analysis was performed using bootscanning analysis and similarity plot analysis implemented in SimPlot ver. 3.5.1 with 500 bp window size and 20 bp step size, Recombination Detection Program (RDP) ver.4 and jumping profile Hidden Markov Model (jpHMM). The recombination breakpoints are indicated as an HXB2 position. ML-phylogenetic analysis was used to confirm the recombination events. Patients were self-reported to be men who have sex with men and infected in Sweden. (B) Mosaic pattern of eight other HIV-1 URFs. These URFs did not belong to any identified transmission cluster. SE: patient infected in Sweden; EE: patient infected in East-Europe; EA: patient infected in East Africa.

Figure 3

The evolutionary relationships based on HIV-NFLG using family-joining. Transmission clusters (n = 7) were constructed for trees at a threshold of 0·08 substitutions per site and are indicated within the figure with red colour. SE: patient reported to have been infected in Sweden; EA: patient reported to have been infected in East Africa. MSM: men who have sex with men; HET: patient self-reported to have been infected heterosexually; PWID: patient with intravenous drug use.