BRIEF REPORT ��� JID 2004:190 (1 October) ��� 1355 B R I E F R E P O R T Incidence of HIV-1 Dual Infection and Its Association with Increased Viral Load Set Point in a Cohort of HIV-1 Subtype C���Infected Female Sex Workers Jandre Grobler,1 Clive M. Gray,3 Cecilia Rademeyer,1 Cathal Seoighe,2 Gita Ramjee,4 Salim Abdool Karim,5 Lynn Morris,3 and Carolyn Williamson1 1Institute for Infectious Diseases and Molecular Medicine, Division of Medical Virology, University of Cape Town, Observatory, and 2South African National Bioinformatics Institute, Bellville, Capetown, and 3 National Institute for Communicable Diseases, Sandringham, Johannesburg, and 4 Medical Research Council, and 5 Centre for the AIDS Programme of Research in South Africa, University of Natal, Durban, South Africa This longitudinal study aimed to determine the incidence and pathogenic implications of dual human immunodefi- ciency virus type 1 (HIV-1) infection in a cohort of female sex workers. Blood samples from 31 recently infected women were screened by use of a heteroduplex mobility assay and sequencing. The median viral load set point was 5404 copies/ mL ( which was measured by use of the bDNA assay. n p 22), Within 3 months of infection, 19% (6/31) of the women were dually infected with 2 distinct HIV-1 subtype C viruses. No evidence of superinfection was detected over the course of 24 months of follow-up, indicating that the risk of dual in- fection is highest around the time of the initial infection. There was a significant association between dual infection and elevated viral load set point. Understanding HIV diversity and its relationship with the nat- ural history of disease provides valuable insights into ascer- taining the role diversity plays in disease progression, as well as understanding correlates of immunity from infection and superinfection. Shortly after infection, intraperson diversity is usually low, although 1 study of women has shown that there may be sex differences, with recently infected women harboring highly heterogeneous viral populations [1]. The virus evolves Received 8 January 2004 accepted 7 April 2004 electronically published 25 August 2004. Reprints or correspondence: Dr. Jandre Grobler (reprints) or Dr. Carolyn Williamson (corre- spondence), Institute for Infectious Disease and Molecular Medicine, Div. of Medical Virology, Univer- sity of Cape Town, Observatory, South Africa (jgrobler@curie.uct.ac.za or cwilliam@curie.uct.ac.za). The Journal of Infectious Diseases 2004 190:1355���9 2004 by the Infectious Diseases Society of America. All rights reserved. 0022-1899/2004/19007-0021$15.00 over time in response to selective pressure some studies have shown a direct relationship between diversity and disease pro- gression, and others have not [2, 3]. Recent studies have shown that dual infection with 2 HIV strains [4] or infection with multiple diverse variants [5] may be associated with faster dis- ease progression. Dual infection can be a consequence of either cotransmission (infection with 2 strains at or close to the time of seroconversion) or superinfection (subsequent HIV-1 infec- tion of an already-infected individual). Superinfection has been reported in only a few cases and is of interest since it provides information on protective immune responses [6]. Viral load and CD4+ T cell counts are prognostic markers that are pivotal in predicting disease outcome [7]. There are limited data from African non-subtype-B���infected cohorts on viral loads after seroconversion, as well as on virological factors affecting the natural history of disease. The HIV epidemic in southern Africa is predominantly associated with heterosexu- ally transmitted subtype C infection [8]. A recent study in India showed that HIV-1 subtype C viral load set point was not sig- nificantly different from that of subtype B however, differences in viral load trajectory suggested a faster disease progression in subtype C infections [9]. To assess the effect of virological fac- tors, such as dual infection, on clinical outcome, it is essential to understand the natural history of disease. Natural-history studies are relevant for future intervention, such as vaccine tri- als, in which viral load may be used as a secondary end point. Many vaccine trials are likely to take place in less-developed regions, where non���subtype B viruses dominate. The aim of the present study was to determine the incidence of dual infection at different time points after infection and to investigate the pathogenic implications of dual infection. Subjects, materials, and methods. Blood samples were ob- tained from 31 HIV-1���seropositive female sex workers (FSWs) who were followed monthly as part of a microbicide phase 3 trial [10]. Informed consent was obtained from all participants, and human-experimentation guidelines of the institutional re- view board of the Nelson R. Mandela School of Medicine of Natal University and University of Cape Town were followed. Antibodies to HIV were detected by use of recombinant HIV- Presented in part: Keystone Symposia, HIV Vaccine Development: Immunological and Biological Challenges, Banff, Canada, 29 March���4 April 2003 (abstract 221). Financial support: Technology and Human Resources for Industry Program, Foundation for Re- search and Development, South Africa International AIDS Vaccine Initiative National Institute of Allergy and Infectious Diseases (grant 19AI51794 to C.M.G., S.A.K., L.M., and C.W.) South African AIDS Vaccine Initiative, Medical Research Council (support to C.W.) UNAIDS C1O protocol (support to the Durban cohort) Wellcome Trust (International Senior Research Fellowship in Biomedical Sciences to L.M.). at Cape Town University on November 21, 2011 http://jid.oxfordjournals.org/ Downloaded from

1356 ��� JID 2004:190 (1 October) ��� BRIEF REPORT Table 1. Dual infection compared with viral load set point. Patient Sample date, days from estimated time of infectiona Amplifiable copy no., copiesb Infectionc Viral load set point, copies/mLd Du21 51 18 Single 2089 Du23 43 132 Single 50 Du62 7 1200 Dual 46,984 Du114 133 18 Single 3280 Du115 14 1200 Single 83 Du123 104 1200 Single 6613 Du151 30 1200 Dual 321,040 Du156 24 140 Single 10,561 Du172 23 18 Single 1042 Du174 22 140 Single 9545 Du179 14 18 Dual 1359 Du204 90 140 Single 5443 Du258 90 1200 Single 5365 Du281 154 18 Single 28,504 Du301 48 140 Single 11,797 Du332 28 140 Single 1707 Du422 28 1200 Single 16,098 Du457 95 140 Single 15,854 Du467 97 140 Dual 40,824 a Midpoint between the last HIV-seronegative and the first HIV-seropos- itive test result, except for Du115 and Du179 the first HIV-seropositive test result was based on detection of HIV RNA by reverse-transcriptase poly- merase chain reaction, and the date after infection was estimated at 14 days. b Amplifiable copy no. at first time point (within 6 months of infection). c Number of infections at the first time point. d Viral load measurement closest to 12 months (9���25 months) after infection. 1/HIV-2 ELIZAQ (Abbott) and confirmed by use of the Vi- ronostika HIV Uniform II micro-ELISA 4 system (Omnimed). Viral loads were measured by use of bDNA (version 3.0 Bayer), and lymphocyte subsets were measured on fresh samples by use of FACscount (Becton Dickinson). HIV-1 env C2���C3 was amplified from viral RNA extracted from plasma. Primers for nested polymerase chain reaction (PCR) were envAR (5 -TGCTGCTCCCAAGAACCCAA-3 , at position 7783���7802 of the HIV HXB2 genome), envAF (5 - GAAAGAGCAGAAGACAGTGGC-3 , at position 6203���6223), BR (5 -AATTCTAGGTCCCCTCCTGA-3 , at position 7317��� 7337), and BF (5 -TAACACAAGCCTGTCCAAAGG-3 , at po- sition 6826���6847). RNA was reverse transcribed to cDNA by use of ThermoScript reverse-transcriptase PCR (Invitrogen). To circumvent resampling and to ensure that the amplified variants reflected the viral populations in the blood, limiting dilution assays were performed, and amplifiable copies were determined by use of QUALITY (availableat:http://ubik.microbiol.washington .edu/computing/quality) (table 1). Intraperson viral diversity was assessed by use of heterodu- plex mobility assays (HMAs) [11]. Samples containing diverse populations or low amplifiable copy numbers were cloned. A total of 20���134 clones were screened, depending on the am- plifiable copy number ( 8���1000 copies). Total populations or cloned PCR products were sequenced (ABI PRISM dye termi- nator cycle-sequencing kit Applied Biosysytems) (GenBank ac- cession numbers AY585270���AY585306). Sequences were aligned by use of CLUSTALX (available at: ftp://ftp-igbmc.u-strasbg. fr/pub/ClustalX/) Maximum-likelihood trees were construct- ed with PAUP* 4.0b10 (Sinauer Associates) by use of tree bi- section/reconnection branch swapping and substitution rate��� matrix parameters estimated via maximum likelihood (param- eters used are available on request). Sequence gaps were ig- nored. DNA distance was determined under the Kimura Two Parameter model by use of MEGA (version 2.1 Molecular Evo- lution Genetic Analysis). Results. Blood samples were obtained from 31 HIV-1���se- ropositive FSWs, of whom 25 were monitored for dual infection over the course of 25 months. Only 22 women had viral load and CD4+ T cell counts determined. Thirty-one women were enrolled within 6 months of HIV-1 infection. The median time from infection to enrollment was 36.5 days time of infection was taken as the midpoint between the last HIV-seronegative and the first HIV-seropositive test result. Intraperson diversity was assessed by use of an HMA with the 510-bp C2���C3 fragment. The resolution of HMA was determined by comparing the HMA banding patterns of total viral popu- lations with the DNA distance between clones. In 8 samples in which the total-population PCR products showed a single band on HMA gels, the mean DNA distance between clones was !1%, with a range of 0.04%���0.86%. HMA of individual clones hy- bridized to an intraperson probe showed that PCR products could be resolved into 2 bands if the genetic difference was 11%. In cases in which a smear or a distinct slow-migrating band of total-population PCR products was visualized on gels, the DNA distance had a range of 2.0%���12.5% ( It was therefore n p 10). assumed that a single heteroduplex band reflected diversity !2%. All 31 FSWs were infected with HIV-1 subtype C (figure 1). A total of 81% (25/31) of the women enrolled within 6 months of infection harbored highly homogenous viral populations (!2% in C2���C3) 19% (6/31) harbored diverse viral populations, in- cluding Du36, Du467, Du95, Du151, Du179, and Du62, with C2���C3 diversity of 9.21%, 7.0%, 8.83%, 11.21%, 11.0%, and 15.72%, respectively. Dual infection with the same subtype can be defined as an infection with 2 phylogenetically distinct viruses that are no more closely related to each other than are another pair of epidemiologically unlinked viruses, with a mean pairwise DNA distance at least as distant as that to a group of unlinked se- quences [4]. Viruses from 5 of these 6 women (Du36, Du95, Du151, Du179, and Du62) were grouped separately on a phy- logenetic tree, demonstrating that these women were dually infected (figure 1). The sixth woman (Du467) harbored di- vergent strains that did not separate phylogenetically how- at Cape Town University on November 21, 2011 http://jid.oxfordjournals.org/ Downloaded from