Studies of HIV-1 co-receptor tropism, which have been conducted primarily in populations where subtype B infections predominate, have demonstrated a relationship between HIV-1 co-receptor use and disease stage. In general, early stages of infection and disease are characterized by greater prevalence of only C-C chemokine type 5 (CCR5)-tropic (R5) HIV-1, which has been associated with slower progression to AIDS [8-12]. The emergence of C-X-C chemokine receptor type 4 (CXCR4)-using virus (X4) has been associated with greater treatment experience and higher risk of death, and coincides with more rapid CD4⁺ T-cell depletion and disease progression [6,8,9,12,13]. Some variants of HIV-1 can use either co-receptor (dual/mixed-tropic [DM] HIV-1); these can be found in all stages of infection, but are more common in infections of longer duration, with lower CD4⁺ cell counts and higher viral loads [12-14]. Despite the emergence of X4-using variants in some patients, only R5 infection typically persists in the majority of patients. Nearly 50% of patients who die of HIV-1 disease have only R5 HIV-1 detectable at the time of their death, indicating that CCR5 remains a critical co-receptor throughout the course of HIV infection [12,15].

Although HIV-1 co-receptor usage and its relationship to disease stage have been studied in the developed world, where subtype B predominates, such relationships are less well understood for subtypes A, C and D. The R5 phenotype is predominant in subtype C HIV-1 infections, whereas X4-using virus has been reported infrequently, even in advanced disease. R5-using virus is more common in subtype A than subtype D HIV-1 infections, and a high proportion of subtype D infections shows D/M tropism throughout the course of disease [16-24]. However, some of these previous studies have been limited by small sample sizes.

The introduction of the CCR5 antagonist, maraviroc, for HIV-1 therapy [25] has increased interest in the epidemiology of tropism and relationships with HIV-1 subtype. A greater understanding of the tropism of non-B subtype HIV-1 is key for the optimal use of CCR5 antagonists in the treatment of these infections in the developing world, and HIV-1 prevention strategies, such as topical microbicides and systemic pre- or post-exposure prophylaxis. In addition, this information will be important for management of clinic populations in the developed world that include individuals with non-B subtype infections who have migrated from endemic countries [26]. HIV-1 tropism can be determined by genotypic and phenotypic methods. While genotypic assays may have lower specificity and sensitivity, retrospective analyses have found that they are comparable to phenotypic tropism assays for prediction of response to treatment with CCR5 antagonists, in populations pre-screened with a phenotypic assay [27,28].

The clinical development programme for maraviroc, the first-in-class CCR5 antagonist, used the Trofile^® phenotypic assay (Monogram Biosciences, South San Francisco, California) [29-31], which determines tropism via the expression of full-length env genes of multiple viruses isolated from patient plasma and can detect 10% of X4 variants with 100% sensitivity. More recently, a Trofile^® assay with enhanced sensitivity to improve detection of low-level X4-using variants has been developed that can detect 0.3% of these variants with 100% sensitivity. Otherwise, assay validation performance characteristics are equivalent between the original and enhanced Trofile^® assays [31]. The enhanced Trofile^® assay has been validated in a number of studies by re-testing the co-receptor tropism of clinical samples that were initially determined using the original assay [31-34]: in a re-analysis of samples from the Phase 3 MERIT study, which evaluated the efficacy of maraviroc in treatment-naïve patients with CCR5-tropic virus, 15% of enrolled patients (n = 106/721) were reclassified as having D/M virus with the enhanced assay [34].

Most (83%) study subjects with reportable tropism results had only R5 HIV-1 detectable, including 90% of TN and 72% of TE patients. However, tropism distribution varied by subtype, CD4⁺ cell count and treatment experience, which can be considered a marker of advanced disease. Indian subjects with subtype C virus were almost exclusively infected with R5 regardless of their CD4⁺ count, HIV-1 disease status, or treatment experience. The prevalence of R5 virus was similarly high in TN South African patients with subtype C infection but, while most common and relatively well-preserved, was somewhat lower in the South African TE cohort of this study (71%). The difference observed between R5 prevalence in India and South Africa may be a result of differences in CD4+ cell count. Compared with the Indian cohort, there was a higher proportion of individuals in the South African cohort with a D/M phenotype (25%) despite similar time since HIV diagnosis and similar exposure to prior therapies. Exclusive use of X4 was rare (4%).

Our findings are consistent with previous reports of subtype C strains from India and elsewhere, where R5 HIV-1 is maintained even in advanced stages of disease. For example, some similarities can be observed between the findings of the present study and a previous study that assessed co-receptor tropism of HIV-1 strains in China: mainly that R5 tropism was seen exclusively in subtype C infections, and remained stable over time [24]. These data highlight the persistence of CCR5 tropism even in long-term infections, possibly related to differences within the env variable loops [37], and contrasts with other subtypes, especially subtype B, where X4 HIV-1 is more frequent during later stages of infection [6,8-10,16,18,21,37-40].

The distribution of HIV-1 subtypes in India and South Africa was consistent with published epidemiologic data [6]. However, this study found more subtype diversity within the Ugandan cohort than previously reported [41,42]. Despite the existence of multiple subtype variations in Uganda, the predominant subtypes were A/A1 (49.0%) and D (28.7%), and there were no marked differences in demographics or HIV disease characteristics. Among Ugandan subjects with subtype A/A1 virus, R5 infection predominated in TN individuals but was reduced in TE individuals, similar to the pattern seen with subtype B. While subtype D TN patients were predominantly R5, X4-using virus (D/M and X4) was detected in nearly 70% of the small number of TE patients. This agrees with earlier reports, where Ugandan individuals infected with subtype D virus showed predominantly D/M tropism [6,16-19,21,22].

Multivariate analyses of other cohorts have identified higher CD4⁺ cell count as a predictor of R5 infection in both TN and TE patients with subtype B and C HIV-1 infections [43,44]. In the present study, multivariate analysis identified CD4⁺ count as a significant predictor of CCR5 tropism for subtypes C, A/A1, and D. In Ugandan patients with subtype A/A1, R5 use was negatively associated with the presence of CDC category B disease, with a negative trend for category C.

This study provided an opportunity to assess the ability of the Trofile^® assay to provide results in different non-B subtype HIV-1 viral subtypes collected in developing countries. While initial testing using standard amplification primers with just a single 1 mL aliquot of patient plasma yielded a lower reportable rate (67-86%) than that expected from experience with subtype B virus, non-reportable samples that were re-tested with the original Trofile^® assay (South Africa and Uganda) or enhanced assay (India) using modified primers optimized for non-B subtype virus, yielded a much higher percentage of reportable results (90-97%), and the distribution of R5 and non-R5 HIV-1 did not change. These improvements in amplification primers - which also retain high performance for subtype B specimens and are currently the primers used regularly for RNA amplification in the Trofile assay - may be advantageous when Trofile testing is used in clinical populations where non-subtype B virus is common [45].

The authors thank the study participants, as well as the investigators: Dr JK Maniar, Jaslok Hospital and Research Centre, Mumbai; Dr GD Ravindran, St. John's Medical College & Hospital, Bangalore; Dr Tripathi, Chattrapati Shahuji Maharaj Medical University (Upgraded King George's Medical College), Lucknow; Dr S Pujari, Ruby Hall Clinic, Pune; Dr N Kumarasamy, YRG Center for AIDS Research & Education, Chennai; Dr AR Pazare, Seth GS Medical College & King Edward Memorial Hospital, Mumbai; Dr N Khippal, S.M.S. Medical College and Hospital, Jaipur; and staff at the participating study sites.