Insights to HIV-1 coreceptor usage by estimating HLA adaptation with Bayesian generalized linear mixed models

The mechanisms triggering the human immunodeficiency virus type I (HIV-1) to switch the coreceptor usage from CCR5 to CXCR4 during the course of infection are not entirely understood. While low CD4+ T cell counts are associated with CXCR4 usage, a predominance of CXCR4 usage with still high CD4+ T cell counts remains puzzling. Here, we explore the hypothesis that viral adaptation to the human leukocyte antigen (HLA) complex, especially to the HLA class II alleles, contributes to the coreceptor switch. To this end, we sequence the viral gag and env protein with corresponding HLA class I and II alleles of a new cohort of 312 treatment-naive, subtype C, chronically-infected HIV-1 patients from South Africa. To estimate HLA adaptation, we develop a novel computational approach using Bayesian generalized linear mixed models (GLMMs). Our model allows to consider the entire HLA repertoire without restricting the model to pre-learned HLA-polymorphisms. In addition, we correct for phylogenetic relatedness of the viruses within the model itself to account for founder effects. Using our model, we observe that CXCR4-using variants are more adapted than CCR5-using variants (p-value = 1.34e-2). Additionally, adapted CCR5-using variants have a significantly lower predicted false positive rate (FPR) by the geno2pheno[coreceptor] tool compared to the non-adapted CCR5-using variants (p-value = 2.21e-2), where a low FPR is associated with CXCR4 usage. Consequently, estimating HLA adaptation can be an asset in predicting not only coreceptor usage, but also an approaching coreceptor switch in CCR5-using variants. We propose the usage of Bayesian GLMMs for modeling virus-host adaptation in general.

A large population sample of African HIV genomes from the 1980s reveals a reduction in subtype D over time associated with propensity for CXCR4 tropism

We present 109 near full-length HIV genomes amplified from blood serum samples obtained during early 1986 from across Uganda, which to our knowledge is the earliest and largest population sample from the initial phase of the HIV epidemic in Africa. Consensus sequences were made from paired-end Illumina reads with a target-capture approach to amplify HIV material following poor success with standard approaches. In comparisons with a smaller 'intermediate' genome dataset from 1998 to 1999 and a 'modern' genome dataset from 2007 to 2016, the proportion of subtype D was significantly higher initially, dropping from 67% (73/109), to 57% (26/46) to 17% (82/465) respectively (p < 0.0001). Subtype D has previously been shown to have a faster rate of disease progression than other subtypes in East African population studies, and to have a higher propensity to use the CXCR4 co-receptor ("X4 tropism"); associated with a decrease in time to AIDS. Here we find significant differences in predicted tropism between A1 and D subtypes in all three sample periods considered, which is particularly striking the 1986 sample: 66% (53/80) of subtype D env sequences were predicted to be X4 tropic compared with none of the 24 subtype A1. We also analysed the frequency of subtype in the envelope region of inter-subtype recombinants, and found that subtype A1 is over-represented in env, suggesting recombination and selection have acted to remove subtype D env from circulation. The reduction of subtype D frequency over three decades therefore appears to be a result of selective pressure against X4 tropism and its higher virulence. Lastly, we find a subtype D specific codon deletion at position 24 of the V3 loop, which may explain the higher propensity for subtype D to utilise X4 tropism.

CD4 is expressed on a heterogeneous subset of hematopoietic progenitors, which persistently harbor CXCR4 and CCR5-tropic HIV proviral genomes in vivo

Latent HIV infection of long-lived cells is a barrier to viral clearance. Hematopoietic stem and progenitor cells are a heterogeneous population of cells, some of which are long-lived. CXCR4-tropic HIVs infect a broad range of HSPC subtypes, including hematopoietic stem cells, which are multi-potent and long-lived. However, CCR5-tropic HIV infection is limited to more differentiated progenitor cells with life spans that are less well understood. Consistent with emerging data that restricted progenitor cells can be long-lived, we detected persistent HIV in restricted HSPC populations from optimally treated people. Further, genotypic and phenotypic analysis of amplified env alleles from donor samples indicated that both CXCR4- and CCR5-tropic viruses persisted in HSPCs. RNA profiling confirmed expression of HIV receptor RNA in a pattern that was consistent with in vitro and in vivo results. In addition, we characterized a CD4^high HSPC sub-population that was preferentially targeted by a variety of CXCR4- and CCR5-tropic HIVs in vitro. Finally, we present strong evidence that HIV proviral genomes of both tropisms can be transmitted to CD4-negative daughter cells of multiple lineages in vivo. In some cases, the transmitted proviral genomes contained signature deletions that inactivated the virus, eliminating the possibility that coincidental infection explains the results. These data support a model in which both stem and non-stem cell progenitors serve as persistent reservoirs for CXCR4- and CCR5-tropic HIV proviral genomes that can be passed to daughter cells.

People who are effectively treated with antiretroviral medication harbor persistent forms of HIV that are integrated into the cellular genome. While HIV is cytopathic to most cells, transcriptionally silent, latent forms do not express toxic HIV gene products and can survive in the host for years. When conditions change, the latent virus can be activated to reinitiate infection. Because of the capacity for virus to spread, cure of HIV will require that we identify and eradicate all cells harboring functional HIV provirus. CD4⁺ T cells are abundant and easily identified as harboring proviral genomes. However, rare cell types that express HIV receptors, such as bone marrow hematopoietic progenitor and stem cells can also be infected by the virus potentially serving as barriers to cure strategies. We found that HIV can infect and persist in progenitor sub-types that were previously thought to be short lived, which expands the types of cells that can support reservoir formation. In addition, we found that HIV can spread by proliferation and cellular differentiation without the need for viral gene expression and virion production that could reveal the infection to the immune system. A deeper understanding of viral reservoirs is critically important for developing strategies that will succeed in viral eradication.

Monophylogenetic HIV-1C epidemic in Ethiopia is dominated by CCR5-tropic viruses-an analysis of a prospective country-wide cohort

Background

CCR5 coreceptor using HIV-1 subtype C (HIV-1C) has been reported to dominate the Ethiopian epidemic. However, almost all data have been obtained from two large cities in the central and north-west regions and recent data is lacking.

Methods

Plasma were obtained from 420 treatment-naïve patients recruited 2009–2011 to a large country-wide Ethiopian cohort. The V3 region was sequenced and the co-receptor tropism was predicted by the clinical and clonal models of the geno2pheno tool at different false positive rates (fpr) and for subtype. In an intention to treat analysis the impact of baseline tropism on outcome of antiretroviral therapy was evaluated.

Results

V3 loop sequencing was successful in 352 (84%) patients. HIV-1C was found in 350 (99.4%) and HIV-1A in two (0.6%) patients. When comparing the geno2pheno fpr10% clonal and clinical models, 24.4% predictions were discordant. X4-virus was predicted in 17.0 and 19.0%, respectively, but the predictions were concordant in only 6%. At fpr5%, concordant X4-virus predictions were obtained in 3.1%. The proportion of X4-tropic virus (clonal fpr10%) increased from 5.6 to 17.3% (p < 0.001) when 387 Ethiopian V3 loop sequences dated from 1984 to 2003 were compared with ours. In an intention to treat analysis, 67.9% reached treatment success at month 6 and only 50% at month 12. Only age and not tropism predicted therapy outcome and no difference was found in CD4+ cell gain between R5-tropic and X4-tropic infected patients. At viral failure, R5 to X4 switch was rare while X4 to R5 switch occurred more frequently (month 6: p = 0.006; month 12: p = 0.078).

Conclusion

The HIV-1C epidemic is monophylogenetic in all regions of Ethiopia and R5-tropic virus dominates, even in patients with advanced immunodeficiency, although the proportion of X4-tropic virus seems to have increased over the last two decades. Geno2pheno clinical and clonal prediction models show a large discrepancy at fpr10%, but not at fpr5%. Hence further studies are needed to assess the utility of genotypic tropism testing in HIV-1C. In ITT analysis only age and not tropism influenced the outcome.

Virological response and resistance profile in HIV-1-infected patients starting darunavir-containing regimens

OBJECTIVES

We evaluated the virological response in patients starting a regimen based on darunavir/ritonavir (DRV/r), which is currently the most widely used ritonavir-boosted protease inhibitor.

METHODS

Data from 206 drug-naïve and 327 PI-experienced patients starting DRV/r 600/100 mg twice daily (DRV600) or 800/100 mg once daily (DRV800) were examined. The probabilities of virological success (VS) and virological rebound (VR) were evaluated in survival analyses. Baseline DRV/r resistance and its evolution at failure were also examined.

RESULTS

DRV600 was preferentially administered in patients with complex requirements (older age, higher viraemia, lower CD4 cell count and DRV/PI resistance) compared with DRV800. By 12 months, the probability of achieving VS was 93.2% and 84.3% in drug-naïve and PI-experienced patients, respectively. The higher the baseline viraemia, the longer was the time required to achieve VS, both in drug-naïve patients [>500 000 HIV-1 RNA copies/mL: median [interquartile range (IQR)] 6.1 (5.1-10.3) months; 100 000-500 000 copies/mL: median (IQR) 4.9 (3.8-6.1) months; <100 000 copies/mL: median (IQR) 3.9 (3.5-4.8) months; P < 0.001] and in PI-experienced patients [≥100 000 copies/mL: median (IQR) 7.2 (5.7-11.6) months; <100 000 copies/mL: median (IQR) 2.8 (2.4-3.3) months; P < 0.001]. In PI-experienced patients, the probability of VR was higher for higher viraemia levels (22.3% for ≥100 000 copies/ml vs. 9.7% for <100 000 copies/mL; P = 0.007). Baseline resistance did not affect the virological response. At failure, a high percentage of patients maintained virus susceptible to all PIs (drug-naïve: 95%; PI-experienced: 80%). Despite being used more often in patients with more complex requirements, DRV600 performed as well as DRV800.

CONCLUSIONS

In clinical practice, use of DRV/r (with its flexible dosage) results in high rates of virological response. These data support the use of PI/r in patients whose characteristics require potent drugs with a high genetic barrier.

Position-specific automated processing of V3 env ultra-deep pyrosequencing data for predicting HIV-1 tropism

HIV-1 coreceptor usage must be accurately determined before starting CCR5 antagonist-based treatment as the presence of undetected minor CXCR4-using variants can cause subsequent virological failure. Ultra-deep pyrosequencing of HIV-1 V3 env allows to detect low levels of CXCR4-using variants that current genotypic approaches miss. However, the computation of the mass of sequence data and the need to identify true minor variants while excluding artifactual sequences generated during amplification and ultra-deep pyrosequencing is rate-limiting. Arbitrary fixed cut-offs below which minor variants are discarded are currently used but the errors generated during ultra-deep pyrosequencing are sequence-dependant rather than random. We have developed an automated processing of HIV-1 V3 env ultra-deep pyrosequencing data that uses biological filters to discard artifactual or non-functional V3 sequences followed by statistical filters to determine position-specific sensitivity thresholds, rather than arbitrary fixed cut-offs. It allows to retain authentic sequences with point mutations at V3 positions of interest and discard artifactual ones with accurate sensitivity thresholds.

Characteristics and spread to the native population of HIV-1 non-B subtypes in two European countries with high migration rate

Background

Non-B subtypes account for at least 50 % of HIV-1 infections diagnosed in Belgium and Luxembourg. They are considered to be acquired through heterosexual contacts and infect primarily individuals of foreign origin. Information on the extent to which non-B subtypes spread to the local population is incomplete.

Methods

Pol and env gene sequences were collected from 410 non-subtype B infections. Profound subtyping was performed using 5 subtyping tools and sequences of both pol and env. Demographic information, disease markers (viral load, CD4 count) and viral characteristics (co-receptor tropism) were compared between subtypes. Maximum likelihood phylogenetic trees were constructed and examined for clustering.

Results

The majority of non-B infections were diagnosed in patients originating from Africa (55.8 %), individuals born in Western Europe represented 30.5 %. Heterosexual transmission was the most frequently reported transmission route (79.9 %), MSM transmission accounted for 12.2 % and was significantly more frequently reported for Western Europeans (25.7 % versus 4.3 % for individuals originating from other regions; p < 0.001). Subtypes A and C and the circulating recombinant forms CRF01_AE and CRF02_AG were the most represented and were included in the comparative analysis. Native Western Europeans were underrepresented for subtype A (14.5 %) and overrepresented for CRF01_AE (38.6 %). The frequency of MSM transmission was the highest for CRF01_AE (18.2 %) and the lowest for subtype A (0 %). No differences in age, gender, viral load or CD4 count were observed. Prevalence of CXCR4-use differed between subtypes but largely depended on the tropism prediction algorithm applied. Indications for novel intersubtype recombinants were found in 20 patients (6.3 %). Phylogenetic analysis revealed only few and small clusters of local transmission but could document one cluster of CRF02_AG transmission among Belgian MSM.

Conclusions

The extent to which non-B subtypes spread in the native Belgian-Luxembourg population is higher than expected, with 30.5 % of the non-B infections diagnosed in native Western Europeans. These infections resulted from hetero- as well as homosexual transmission. Introduction of non-B variants in the local high at risk population of MSM may lead to new sub-epidemics and/or increased genetic variability and is an evolution that needs to be closely monitored.

Chemokine co-receptor usage in HIV-1-infected treatment-naïve voluntary counselling and testing clients in Southern Taiwan

OBJECTIVE

The goal of this present study was to determine the proportion of CCR5-tropic and CXCR4-tropic viruses and impact of tropism test on clinical presentation, CD4 cell counts, viral load and genotypic drug resistance from drug-naïve, voluntary counselling and testing (VCT) clients in southern Taiwan.

DESIGN

This was a cross-sectional study. Plasma samples were collected from HIV-1-infected patients from January 2013 to December 2013; subjects were recruited from free VCT centres in southern Taiwan.

SETTING

Taiwan.

PARTICIPANTS

Plasma samples from 108 HIV-1-infected, treatment-naïve, VCT clients were analysed. HIV-1 strains were sequenced, genotype resistance was determined by a commercial kit (Viro-seq) and co-receptor tropism (CRT) was predicted by an internet tool geno2pheno[coreceptor], with a 10% false-positive rate as the cut-off. Differences in progression markers, patient characteristics, VCT questionnaires and HIV subtype distribution were evaluated statistically.

RESULTS

All the 108 VCT clients were male with 90% between the ages of 20 and 40 years. Eighty-eight per cent of the patients were men who have sex with men (MSM). The median (IQR) CD4 cell count was 342 cells/µL (221-454) and the viral load was 4.6 log (4.0-5.0). HIV-transmitted drug resistance was found in 9.3% (10/108) of the patients. CRT predictions indicated that 74% of the patients had only R5-tropic strains. CRT was not associated with CD4 cell counts, patient characteristics, VCT questionnaire and transmitted drug resistance. There was a significant difference with regard to viral load at the time of presentation, showing that patients with R5 more often had a higher viral load as compared with those with X4/DM strains (4.6±0.6 log vs 4.33±0.7 log, p=0.036).

CONCLUSIONS

We found that 74% of the VCT clients were infected with R5-tropic virus strains. HIV-transmitted drug resistance was not associated with CRT predictions. Higher viral load at presentation was predictive of R5 co-receptor usage.

Effect of lysine to arginine mutagenesis in the V3 loop of HIV-1 gp120 on viral entry efficiency and neutralization

HIV-1 infection is characterized by an ongoing replication leading to T-lymphocyte decline which is paralleled by the switch from CCR5 to CXCR4 coreceptor usage. To predict coreceptor usage, several computer algorithms using gp120 V3 loop sequence data have been developed. In these algorithms an occupation of the V3 positions 11 and 25, by one of the amino acids lysine (K) or arginine (R), is an indicator for CXCR4 usage. Amino acids R and K dominate at these two positions, but can also be identified at positions 9 and 10. Generally, CXCR4-viruses possess V3 sequences, with an overall positive charge higher than the V3 sequences of R5-viruses. The net charge is calculated by subtracting the number of negatively charged amino acids (D, aspartic acid and E, glutamic acid) from the number of positively charged ones (K and R). In contrast to D and E, which are very similar in their polar and acidic properties, the characteristics of the R guanidinium group differ significantly from the K ammonium group. However, in coreceptor predictive computer algorithms R and K are both equally rated. The study was conducted to analyze differences in infectivity and coreceptor usage because of R-to-K mutations at the V3 positions 9, 10 and 11. V3 loop mutants with all possible RRR-to-KKK triplets were constructed and analyzed for coreceptor usage, infectivity and neutralization by SDF-1α and RANTES. Virus mutants R₉R₁₀R₁₁ showed the highest infectivity rates, and were inhibited more efficiently in contrast to the K₉K₁₀K₁₁ viruses. They also showed higher efficiency in a virus-gp120 paired infection assay. Especially V3 loop position 9 was relevant for a switch to higher infectivity when occupied by R. Thus, K-to-R exchanges play a role for enhanced viral entry efficiency and should therefore be considered when the viral phenotype is predicted based on V3 sequence data.

A diagnostic HIV-1 tropism system based on sequence relatedness

Key clinical studies for HIV coreceptor antagonists have used the phenotyping-based Trofile test. Meanwhile various simpler-to-do genotypic tests have become available that are compatible with standard laboratory equipment and Web-based interpretation tools. However, these systems typically analyze only the most prominent virus sequence in a specimen. We present a new diagnostic HIV tropism test not needing DNA sequencing. The system, XTrack, uses physical properties of DNA duplexes after hybridization of single-stranded HIV-1 env V3 loop probes to the clinical specimen. Resulting "heteroduplexes" possess unique properties driven by sequence relatedness to the reference and resulting in a discrete electrophoretic mobility. A detailed optimization process identified diagnostic probe candidates relating best to a large number of HIV-1 sequences with known tropism. From over 500 V3 sequences representing all main HIV-1 subtypes (Los Alamos database), we obtained a small set of probes to determine the tropism in clinical samples. We found a high concordance with the commercial TrofileES test (84.9%) and the Web-based tool Geno2Pheno (83.0%). Moreover, the new system reveals mixed virus populations, and it was successful on specimens with low virus loads or on provirus from leukocytes. A replicative phenotyping system was used for validation. Our data show that the XTrack test is favorably suitable for routine diagnostics. It detects and dissects mixed virus populations and viral minorities; samples with viral loads (VL) of <200 copies/ml are successfully analyzed. We further expect that the principles of the platform can be adapted also to other sequence-divergent pathogens, such as hepatitis B and C viruses.

Performance of commonly used genotypic assays and comparison with phenotypic assays of HIV-1 coreceptor tropism in acutely HIV-1-infected patients

OBJECTIVES

Although founder viruses in primary HIV-1 infections (PHIs) typically use the CCR5 coreceptor (R5-tropic), 3%-19% of subjects also harbour CXCR4-using viruses (X4-tropic), making tropism determination before CCR5 antagonist usage mandatory. Genotypic methods can be used to accurately determine HIV-1 tropism in chronically infected patients.

METHODS

We compared the results of genotypic methods [geno2pheno, PSSMx4r5 including a novel nucleotide-input version (ntPSSM) and distant segments (ds)Kernel] to predict coreceptor usage in a cohort of 67 PHIs. Specimens with discrepant results were phenotypically tested after cloning the V3 gene region into proviral backbones. Recombinant viruses were used to infect U87 indicator cell lines bearing CD4 and either CCR5 or CXCR4.

RESULTS

Geno2pheno10%, PSSMx4r5 and (ds)Kernel gave identical predictions in 85% of cases. Geno2pheno10% predicted the presence of CXCR4 viruses in 18% of patients. Two patients were predicted to carry X4-tropic viruses by all algorithms and X4-tropic viruses were detected in at least one of the recombinant AD8 or NL4-3 backbone-based assays. Ten samples resulted in discordant predictions with at least one algorithm. Full concordance between tropism prediction by using population sequencing and phenotypic assays was observed only with ntPSSM. Geno2pheno prediction and the phenotypic assay gave the same results in a minority of 'discordant' patients.

CONCLUSIONS

Compared with both PSSMx4r5 versions, (ds)Kernel and our phenotypic assay, geno2pheno10% overestimated the frequency of X4-tropic viruses (18% versus 3%). ntPSSM was able to detect one additional X4 virus compared with (ds)Kernel that was confirmed with the phenotypic assay.

Defining differential genetic signatures in CXCR4- and the CCR5-utilizing HIV-1 co-linear sequences

The adaptation of human immunodeficiency virus type-1 (HIV-1) to an array of physiologic niches is advantaged by the plasticity of the viral genome, encoded proteins, and promoter. CXCR4-utilizing (X4) viruses preferentially, but not universally, infect CD4⁺ T cells, generating high levels of virus within activated HIV-1-infected T cells that can be detected in regional lymph nodes and peripheral blood. By comparison, the CCR5-utilizing (R5) viruses have a greater preference for cells of the monocyte-macrophage lineage; however, while R5 viruses also display a propensity to enter and replicate in T cells, they infect a smaller percentage of CD4⁺ T cells in comparison to X4 viruses. Additionally, R5 viruses have been associated with viral transmission and CNS disease and are also more prevalent during HIV-1 disease. Specific adaptive changes associated with X4 and R5 viruses were identified in co-linear viral sequences beyond the Env-V3. The in silico position-specific scoring matrix (PSSM) algorithm was used to define distinct groups of X4 and R5 sequences based solely on sequences in Env-V3. Bioinformatic tools were used to identify genetic signatures involving specific protein domains or long terminal repeat (LTR) transcription factor sites within co-linear viral protein R (Vpr), trans-activator of transcription (Tat), or LTR sequences that were preferentially associated with X4 or R5 Env-V3 sequences. A number of differential amino acid and nucleotide changes were identified across the co-linear Vpr, Tat, and LTR sequences, suggesting the presence of specific genetic signatures that preferentially associate with X4 or R5 viruses. Investigation of the genetic relatedness between X4 and R5 viruses utilizing phylogenetic analyses of complete sequences could not be used to definitively and uniquely identify groups of R5 or X4 sequences; in contrast, differences in the genetic diversities between X4 and R5 were readily identified within these co-linear sequences in HIV-1-infected patients.

A very low geno2pheno false positive rate is associated with poor viro-immunological response in drug-naïve patients starting a first-line HAART

BACKGROUND

We previously found that a very low geno2pheno false positive rate (FPR ≤ 2%) defines a viral population associated with low CD4 cell count and the highest amount of X4-quasispecies. In this study, we aimed at evaluating whether FPR ≤ 2% might impact on the viro-immunological response in HIV-1 infected patients starting a first-line HAART.

METHODS

The analysis was performed on 305 HIV-1 B subtype infected drug-naïve patients who started their first-line HAART. Baseline FPR (%) values were stratified according to the following ranges: ≤ 2; 2-5; 5-10; 10-20; 20-60; >60. The impact of genotypically-inferred tropism on the time to achieve immunological reconstitution (a CD4 cell count gain from HAART initiation ≥ 150 cells/mm(3)) and on the time to achieve virological success (the first HIV-RNA measurement <50 copies/mL from HAART initiation) was evaluated by survival analyses.

RESULTS

Overall, at therapy start, 27% of patients had FPR ≤ 10 (6%, FPR ≤ 2; 7%, FPR 2-5; 14%, FPR 5-10). By 12 months of therapy the rate of immunological reconstitution was overall 75.5%, and it was significantly lower for FPR ≤ 2 (54.1%) in comparison to other FPR ranks (78.8%, FPR 2-5; 77.5%, FPR 5-10; 71.7%, FPR 10-20; 81.8%, FPR 20-60; 75.1%, FPR >60; p = 0.008). The overall proportion of patients achieving virological success was 95.5% by 12 months of therapy. Multivariable Cox analyses showed that patients having pre-HAART FPR ≤ 2% had a significant lower relative adjusted hazard [95% C.I.] both to achieve immunological reconstitution (0.37 [0.20-0.71], p = 0.003) and to achieve virological success (0.50 [0.26-0.94], p = 0.031) than those with pre-HAART FPR >60%.

CONCLUSIONS

Beyond the genotypically-inferred tropism determination, FPR ≤ 2% predicts both a poor immunological reconstitution and a lower virological response in drug-naïve patients who started their first-line therapy. This parameter could be useful to identify patients potentially with less chance of achieving adequate immunological reconstitution and virological undetectability.

Mutations in gp41 are correlated with coreceptor tropism but do not improve prediction methods substantially

BACKGROUND

The main determinants of HIV-1 coreceptor usage are located in the V3-loop of gp120, although mutations in V2 and gp41 are also known. Incorporation of V2 is known to improve prediction algorithms; however, this has not been confirmed for gp41 mutations.

METHODS

Samples with V3 and gp41 genotypes and Trofile assay (Monogram Biosciences, South San Francisco, CA, USA) results were taken from the HOMER cohort (n=444) and from patients screened for the MOTIVATE studies (n=1,916; 859 with maraviroc outcome data). Correlations of mutations with tropism were assessed using Fisher's exact test and prediction models trained using support vector machines. Models were validated by cross-validation, by testing models from one dataset on the other, and by analysing virological outcome.

RESULTS

Several mutations within gp41 were highly significant for CXCR4 usage; most strikingly an insertion occurring in 7.7% of HOMER-R5 and 46.3% of HOMER-X4 samples (MOTIVATE 5.7% and 25.2%, respectively). Models trained on gp41 sequence alone achieved relatively high areas under the receiver-operating characteristic curve (AUCs; HOMER 0.713 and MOTIVATE 0.736) that were almost as good as V3 models (0.773 and 0.884, respectively). However, combining the two regions improved predictions only marginally (0.813 and 0.902, respectively). Similar results were found when models were trained on HOMER and validated on MOTIVATE or vice versa. The difference in median log viral load decrease at week 24 between patients with R5 and X4 virus was 1.65 (HOMER 2.45 and MOTIVATE 0.79) for V3 models, 1.59 for gp41-models (2.42 and 0.83, respectively) and 1.58 for the combined predictor (2.44 and 0.86, respectively).

CONCLUSIONS

Several mutations within gp41 showed strong correlation with tropism in two independent datasets. However, incorporating gp41 mutations into prediction models is not mandatory because they do not improve substantially on models trained on V3 sequences alone.

In vivo analysis of a new R5 tropic SHIV generated from the highly pathogenic SHIV-KS661, a derivative of SHIV-89.6

Although X4 tropic SHIVs have been studied extensively, they show distinct infection phenotypes from those of R5 tropic viruses, which play an important role in HIV-1 transmission and pathogenesis. To augment the variety of R5 tropic SHIVs, we generated a new R5 tropic SHIV from the highly pathogenic X4 tropic SHIV-KS661, a derivative of SHIV-89.6. Based on consensus amino acid alignment analyses of subtype B R5 tropic HIV-1, five amino acid substitutions in the third variable region successfully changed the secondary receptor preference from X4 to R5. Improvements in viral replication were observed in infected rhesus macaques after two passages, and reisolated virus was designated SHIV-MK38. SHIV-MK38 maintained R5 tropism through in vivo passages and showed robust replication in infected monkeys. Our study clearly demonstrates that a minimal number of amino acid substitutions in the V3 region can alter secondary receptor preference and increase the variety of R5 tropic SHIVs.