Secondary Logo

Journal Logo

A phylotype-based analysis highlights the role of drug-naive HIV-positive individuals in the transmission of antiretroviral resistance in the UK

Mourad, Raphaëla; Chevennet, Françoisa,b; Dunn, David T.c; Fearnhill, Estherc; Delpech, Valeried; Asboe, Davide; Gascuel, Oliviera; Hue, Stéphanef,g on behalf of the UK HIV Drug Resistance Database & the Collaborative HIV, Anti-HIV Drug Resistance Network

doi: 10.1097/QAD.0000000000000768
BASIC SCIENCE
Free
SDC

Objective: Antiretroviral-naive HIV-positive individuals contribute to the transmission of drug-resistant viruses, compromising first-line therapy. Using phylogenetic inference, we quantified the proportion of transmitted drug-resistance originating from a treatment-naive source.

Methods: Using a novel phylotype-based approach, 24 550 HIV-1 subtype B partial pol gene sequences from the UK HIV Drug Resistance database were analysed. Ongoing transmission of drug resistance amongst HIV-positive individuals was identified as phylotypes of at least three sequences with at least one shared drug resistance mutation, a maximum intra-clade genetic distance of 4.0% and a basal branch support at least 90%. The time of persistence of the transmission chains was estimated using a fast least-squares molecular clock inference approach.

Results: Around 70% of transmitted drug-resistance had a treatment-naive source. The most commonly transmitted mutations were L90M in the protease gene and K103N, T215D and T215S in reverse transcriptase. Reversion to wild type occurred at a low frequency and drug-independent reservoirs of resistance have persisted for up to 13 years.

Conclusion: These results illustrate the impact of viral fitness on the establishment of resistance reservoirs and support the notion that earlier diagnoses and treatment of HIV infections are warranted for counteracting the spread of antiretroviral resistance. Phylotype-based phylogenetic inference is an attractive approach for the routine surveillance of transmitted drug resistance in HIV as well as in other pathogens for which genotypic resistance data are available.

aInstitut de Biologie Computationnelle, LIRMM, UMR5506 CNRS

bMIVEGEC, CNRS 5290, IRD 224, Université de Montpellier, Montpellier, France

cMRC Clinical Trial Unit at UCL

dPublic Health England

eChelsea & Westminster Hospital

fUniversity College London

gLondon School of Hygiene and Tropical Medicine, London, UK.

Correspondence to Olivier Gascuel, PhD, Institut de Biologie Computationnelle, LIRMM, UMR5506 CNRS Université de Montpellier, CC 05016, 860 rue de St Priest, 34095 Montpellier cédex 5, France. Tel: +33 4 67 41 85 47; e-mail: gascuel@lirmm.fr and Stéphane Hué, PhD, Department of Infectious Disease Epidemiology, London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, UK. Tel: +44 20 7927 2469; e-mail: stephane.hue@lshtm.ac.uk

Received 23 January, 2015

Revised 2 April, 2015

Accepted 5 June, 2015

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Website (http://www.AIDSonline.com).

Back to Top | Article Outline

Introduction

Combination antiretroviral therapy has proven highly effective in controlling HIV-1 infections and has significantly improved patients’ survival and quality of life [1]. However, resistance to antiretroviral drugs is known to develop in treated individuals, leading to viral load rebound and treatment failure. Drug-resistant viruses emerge through the selective pressure induced by antiretrovirals in treated individuals but are also transmitted from patients failing therapy to susceptible, treatment-naive recipients [2]. In the UK, transmitted drug resistance (TDR) peaked at 14% of new HIV diagnoses in 2002 but has since declined to 8–9%, a change attributed to more stringent testing guidelines [3]. The continued high prevalence of TDR since 2002 suggests that antiretroviral resistance may increasingly originate from individuals who are not enrolled in HIV care. In line with this hypothesis, it was demonstrated that sustainable reservoirs of resistance persist in the HIV-infected population through continuous transmission of resistant viruses amongst treatment-naive individuals [4]. Despite the replicative impairment associated with most drug resistance mutations (DRMs), resistant viruses endure longer when transmitted to drug-naive individuals than when emerging in treated ones, due to the lack of competition with more replicative competent, wild-type subpopulations [5]. This persistence in the HIV-infected population increases the chances of transmission and it is imperative to characterize the source of transmitted resistant viruses for an optimal management of the epidemic.

Phylogenetics is a powerful and well established method to identify transmission of resistant viruses. The prevailing approach relies on the identification of clusters in a phylogeny of viruses sampled from the infected population, the resistant nature of which is established through subsequent computational steps (e.g. [4–7]). The timing and duration of transmission chains is often estimated through further inference (e.g. [8]), adding computational burden to the procedure. However, traditional approaches do not inform the rate at which resistance mutations are lost in the population through reversion to wild type.

We present a novel phylogenetic method designed to quantify transmitted drug-resistance originating from a drug-naive source in a HIV diagnosed population. Our method is based on the recent concept of phylotype [9] and can model transmission lineages of drug-resistances on the basis of genetic and trait similarities between contemporary HIV sequences [10]. We applied this approach to a large set of HIV-1 subtype B sequences sampled in the UK and found that 70% of the resistant variants detected at diagnosis have a treatment-naive source. This study emphasizes the importance of early HIV diagnosis and viral genotyping.

Back to Top | Article Outline

Materials and methods

Study cohort and sequence data

A total of 24 550 HIV-1 subtype B partial pol gene sequences from the UK HIV Drug Resistance database were analysed. Detail of the database is given elsewhere [11]. In brief, sequences were sampled from HIV-positive individuals across the UK through routine genotypic resistance testing between 1997 and 2011. Sequences span the entire protease (297 nucleotides) and first 1248 nucleotides of the reverse transcriptase genes. For patients with multiple sequences, only the first available sequence was included in the analysis. Sequences were linked to pseudo-anonymised clinical and demographic information on the infected patient gathered by the UK Collaborative HIV Cohort Study (UK CHIC) [11] or the HIV and AIDS Reporting System at Public Health England (HARS) (http://www.hpa.org.uk/). This information included sex, risk group, ethnicity, age group and treatment status. The sequences curated by the UK HIV Drug Resistance database represent around 50% of cumulative HIV infections in the UK since 1996 and are highly representative of the UK epidemic in terms of risk groups, sex, ethnicity and age at diagnosis (see for instance [12]). At the time of the analysis, 75% (n = 18 354) and 25% (n = 6196) of the studied patients were antiretroviral treatment naive and experienced, respectively. According to the WHO's HIV resistance surveillance list [13], 11% (1936/18 354) and 56% (3469/6196) drug-naive and drug-experienced sequences, respectively, harboured at least one DRM.

Back to Top | Article Outline

Phylogenetic reconstruction

Sequences were aligned with ClustalX [14] and manually edited with Se-Al version 2.0a11 (http://tree.bio.ed.ac.uk/software/seal/). In order to assess the impact of phylogenetic uncertainty on our estimates, 10 alternative phylogenies were reconstructed from the sequence alignment. First, 10 000 equivalently most parsimonious trees were generated using the software TNT version 1.1 [15]. A set of 100 trees were uniformly drawn from the tree distribution and used as starting topologies for maximum likelihood inference with FastTree v2.1.5 [16] under the General Time Reversible model of nucleotide substitutions and varying evolutionary rates across sites (GTR + CAT). The 10 trees with the highest likelihoods were selected and analysed independently. Mean estimates over the 10 phylogenies are reported.

Back to Top | Article Outline

Phylotype identification

A phylotype is a set of monophyletic sequences sharing a common trait (e.g. a particular DRM) in a phylogeny [9]. We postulate that a phylotype of treatment-naive HIV sequences sharing a specific DRM represents the ongoing transmission of drug resistance in the untreated HIV positive population. Note that a treatment-naive resistant phylotype can either comprise treatment-naive individuals only (thereafter called naive phylotypes; Fig. 1a) or a combination of both treatment-naive and experienced individuals (thereafter mixed phylotypes).

Fig. 1

Fig. 1

First, an ancestral character state reconstruction of all positions associated with drug resistance (n = 93) was performed along each of the 10 maximum likelihood trees, using Fitch's maximum parsimony with the DELTRAN option [17], as implemented in Phylotype version 7.0 (http://phylotype.org/). From these, drug-resistant phylotypes were identified on the basis of the following criteria: at least three sequences harbouring the same DRM (phylotype option: size ≥3); a maximum intra-clade genetic distance of 4.0% (diversity ≤0.02); and a basal branch support at least 90% (support ≥90%; branch support calculated by SH-like test, as implemented in FastTree). Phylotypes for which the number of sequences without the shared DRM exceeded the number of sequences with the shared DRM were excluded (size/different ≥1). Treatment-naive resistant phylotypes are schematized in Fig. 1a.

In order to test the statistical significance of the phylogeny/DRM association found in the selected phylotypes, DRM annotations were reassigned at the phylogenies’ tips of the trees through 1000 random permutations, simulating the null hypothesis that the distribution of DRMs in the tree is independent to the sequences relatedness. Phylotypes for which the phylogeny/trait association did not reach statistical significance at the P = 0.01 (10/1000) level were excluded from the analysis.

As phylotypes are topology-dependent, the size and composition of a given phylotype may vary across the 10 alternative maximum likelihood trees. The stability of each phylotype was further tested by calculating pairwise similarity values between two phylotypes A (from tree %plane1D;4AF;1) and A’ (from tree %plane1D;4AF;2) as:

with sequences (X) being the set of sequences found in phylotype X. The resulting similarity values range from 0 to 1, with 1 indicating absolute similarity between the two phylotypes. A matrix of similarity values was computed for each possible pair of trees. For a given pair, the percentage of matched phylotypes was calculated as:

where a pair of phylotypes A and A’ is considered as matching if

was also averaged over all DRM positions. Using this approach, we were able to assess to which extent the same phylotypes were found in the 10 maximum likelihood trees, as a stable phylotype is expected to be found in all trees.

Back to Top | Article Outline

Quantification of naive-to-naive transmission of drug resistance

Naive-to-naive transmission of resistance was defined as the percentage of treatment-naive individuals who inherited a drug-resistant virus from another treatment-naive individual. Not all transmitted DRMs in a treatment-naive resistant phylotype have a treatment-naive source. As DRMs exclusively emerge through treatment-associated selective pressure in a treated individual, at least one treatment-naive individual in the transmission chain has a treated source. We therefore assumed that the number of naive-to-naive DRM transmissions in a phylotype comprising treatment-naive individuals only (i.e. naive phylotypes) corresponds to the number of treatment-naive sequences in the resistant phylotype minus one. For resistant phylotypes comprising both treatment-naive and experienced individuals (i.e. mixed phylotypes), the proportion of naive-to-naive DRM transmission was quantified using a parsimony ancestral reconstruction framework specifically designed for the study (see supplementary methods, http://links.lww.com/QAD/A728).

For each DRM, the overall rate of naive-to-naive transmission was calculated as:

where the numerator is the number of sequences that possess a mutation resulting from naive-to-naive transmission, and the denominator is the number of sequences from naive patients in phylotypes (naive or mixed).

Back to Top | Article Outline

Rate of drug-resistance loss in the untreated population

In a phylotype, the most recent common ancestors of subclades without the shared trait (i.e. a particular DRM) are called exceptions (see Fig. 1a). Exceptions are connected to a phylotype but do not belong to it [9]. In treatment-naive resistant phylotypes, an exception represents the loss of a DRM in the population through reversion to wild type (Fig. 1a). For each phylotype, the population loss rate (PLR) of a mutation was calculated as:

For a given DRM, the PLR was averaged over all treatment-naive phylotypes found for that mutation:

with weight wi the number of phylotype sequences along with the number of exceptions for a phylotype i.

Back to Top | Article Outline

Time of persistence in the drug-naive population

The persistence of DRMs in treatment-naive phylotypes (both 100% naive and mixed) in the HIV-infected population was estimated using molecular clock inference. Fast least-squares inference was used to date the ancestral nodes of the each phylotype, as implemented in the Least-Squares Dating program (LSD) version 1.0 (http://www.atgc-montpellier.fr/LSD/). LSD time estimates were compared with those obtained by Bayesian Markov chain Monte Carlo inference using BEAST v1.8.0 [18], under the SRD06 model of nucleotide substitution [19], an uncorrelated lognormal (UCLN) model of molecular evolutionary rate heterogeneity and a Bayesian skyline tree coalescent prior. As BEAST inference is not scalable to large phylogenies, phylotypes were pooled in groups of 150–200 sequences and divergence times were estimated as previously described [12].

The time of persistence of a DRM was estimated as the difference between the tMRCA date and the sampling date of the most recently sampled drug-naive sequence. Persistence estimates were averaged over all phylotypes found for a given DRM. The date of emergence of the phylotypes was defined for each DRM as the most ancient tMRCA over all phylotypes found for that mutation. A detailed description of the overall methodology is available as supplementary information, http://links.lww.com/QAD/A728.

Back to Top | Article Outline

Evolution of drug-naive, resistant viral reservoirs over time

In order to investigate the potential growth of treatment-independent resistant reservoirs, we compared the phylotypes observed in the present study with the treatment-independent, drug-resistant transmission clusters identified from an earlier version of the UK Drug Resistance Database in 2009 [4]. The 38 viral sequences representing naive-to-naive transmission of resistance in the original study were aligned to the phylotypes’ sequences from the present analysis and a maximum likelihood phylogeny was inferred with FastTree v2.1.5 [16] under the General Time Reversible model of nucleotide substitutions and varying evolution rates across sites (GTR + CAT).

Back to Top | Article Outline

Results

Identification of treatment-naive resistant phylotypes

On average, 98 resistant phylotypes were identified. Of these, 65 comprised drug-naive individuals only (naive phylotypes), and 19 comprised both drug-naive and drug-experienced individuals (mixed phylotypes) (Table 1). On average, 17% (333/1936) of the drug-naive resistant sequences included in the study were involved in one resistant phylotype. These phylotypes were mostly stable across the 10 alternative maximum likelihood topologies, with an average percentage of matched phylotypes of 80%. The median size of naive and mixed phylotypes were four individuals, with 5% (three out of 65) and 0% (zero out of 19) of the naive and mixed phylotypes, respectively, involving more than 10 individuals. The largest phylotype contained 24 patients infected with a strain harbouring the K219Q mutation in reverse transcriptase.

Table 1

Table 1

Of the 93 DRMs tested, 19 (20%) were found in at least one treatment-naive resistant phylotype (Table 1). Mutations conferring resistance to the three main classes of antiretroviral drugs prescribed in the UK (i.e. protease inhibitors; nucleoside analogue reverse transcriptase inhibitors, nRTIs; and nonnucleoside analogue reverse transcriptase inhibitors, nnRTIs) were found in treatment-naive resistant phylotypes. The most commonly found DRMs were L90M (33 individuals forming six phylotypes) in protease; K103N (59 individuals forming 7 phylotypes), T215D (60 individuals forming 14 phylotypes) and T215S (59 individuals forming 14 phylotypes) in reverse transcriptase. We observed a significant linear positive correlation (R = 0.83, P = 1×10–5) between the frequency of DRMs in drug-naive phylotypes and in the naive cohort (Fig. 1b). This observation suggests that although there are differences in transmission rates between DRMs in the drug-naive cohort, naive-to-naive transmission of resistance is frequency dependent.

Individuals involved in treatment-naive resistant phylotypes were representative of the original cohort. Those for which demographic information was available were more likely to be male (71%), MSM (65%) and of white ethnicity (33%).

Back to Top | Article Outline

Transmission of drug-resistance from a treatment-naive source

On average, 70% (232/333) of the (resistant) viruses found in resistant phylotypes (either 100% naive or mixed) had a drug-naive source. The highest rates of naive-to-naive transmission were observed for the mutations V82L (73% of the drug-naive sequences in V82L phylotypes) in protease, K103 (86%) and Y188L (81%) in reverse transcriptase (Fig. 1c). The lowest rates of naive-to-naive transmission were found for the mutations M46L and V82A (67%) in protease, L210W (57%) and G190A (50%) in reverse transcriptase.

A total of 15% (50/333) of the viral sequences in treatment-naive resistant phylotype harboured more than one DRM (Fig. 1d), which is significantly less than the number of sequences with more than one DRM in the drug-naive cohort (472/1609 sequences; 29%; P = 2 x 10–9). Of these, 14% (46/333) involved two DRMs, 1% (2/333) involved three DRMs and none had more than three DRMs. The mutations most commonly found in multiresistant drug-naive phylotypes were M41L (23 sequences) and T215D (11 sequences).

Back to Top | Article Outline

Rate of drug-resistance loss in the untreated population

The rate at which DRMs are lost in the population was estimated. Nine DRMs showed evidence of reversions through transmissions within the untreated population: L90M in protease; and M41L, D67N, K103N, T215C, T215D, T215S, T215V and K219Q in reverse transcriptase (Table 1). Among these mutations, T215D presented the highest PLR (six out of 66 sequences; 9%). Inversely, K103N had the lowest PLR (three out of 54 sequences; 5%). The PLR of mutations conferring resistance to nRTIs (18/246 sequences; 7%) was higher than that of mutations conferring resistance to the two other classes, that is protease inhibitor (one out of 33 sequences; 3%) and nnRTI (three out of 73 sequences; 3%).

Back to Top | Article Outline

Time of persistence in the drug-naive population

Using fast least-squares inference, we estimated the date of emergence and time of persistence of the treatment-naive resistant phylotypes (Table 1 and Fig. 2). The longest time of persistence was found for V82L in protease (13.1 years) and for K103N (13.5 years), T215S (10.2 years) in reverse transcriptase. Inversely, mutations V82A (5.4 years) in protease and G190A (<1 year) in reverse transcriptase showed the shortest persistence times in the untreated population. No correlation was found between a DRM time of persistence and the median time to resistance loss of that mutation within patients, as estimated by Castro et al. (data not shown; correlation coefficient R2 = 0.13; P = 0.7) [20].

Fig. 2

Fig. 2

Times estimates obtained with fast least-squares inference were similar with those from Bayesian Markov Chain Monte Carlo (MCMC) inference, as implemented in BEAST (Supplementary Figure 1, http://links.lww.com/QAD/A728), while LSD significantly reduced computational time. The main difference between the two dating methods lied in a slightly longer time of persistence and emergence (∼2 years) with LSD.

Back to Top | Article Outline

Comparison with the 2009 study

We compared the phylotypes observed in this study with the treatment-independent resistant transmission clusters reported by Hué et al. in 2009 [4]. Three of the five clusters found in 2009 were seen in the present analysis: cluster A (K219Q, 22 sequences in 2009), cluster B (L90M, 13 sequences in 2009) and cluster C (V82A; L90M; D67N; K219Q, four drug-naive sequences in 2009) (Fig. 3). All three clusters showed evidence of expansion between 2006 and 2011. Cluster A expanded from 22 to 40 linked sequences, experiencing an 82% increase, with 11 of the 18 new infections forming a strongly supported subcluster. Cluster B also expanded from 13 to 22 sequences (69% increase), with six of the 13 new infections forming a strongly supported subcluster including a single individual from the previous study, a possible source of the outbreak. Cluster C increased by two sequences (50% increase). The two other clusters identified in 2009 (i.e. clusters D and E) were not observed amongst the treatment-naive resistant phylotypes, probably due to the use of different inclusion criteria in the two analyses.

Fig. 3

Fig. 3

Back to Top | Article Outline

Discussion

We present a novel phylogenetic framework designed to quantify transmission of antiretroviral resistance, and estimated that 70% of TDRs have a drug-naive source in the UK. We further show that reservoirs of resistance have persisted in the drug-naive population for over a decade, representing a hindrance on the management of HIV infections. The improved effectiveness of new antiretroviral regimens has led to a decline in the emergence of DRMs in many countries [21–23], further reducing the risk of transmission of resistant variants. Our results are in line with the postulate that, as a result of this decline, an increasingly larger proportion of transmitted antiretroviral resistance will originate from untreated individuals infected with resistant viruses [4]. This trend further suggests that the waning of TDR is limited unless a larger proportion of the HIV infected population is diagnosed and enrolled in HIV care. This study therefore supports the notion that earlier diagnoses and treatment of HIV infections are warranted for preventing the spread of antiretroviral resistance.

The level of naive-to-naive transmission of DRM found in the UK cohort is consistent with estimates reported by a similar study conducted in Switzerland [7]. Drescher et al.[7] analysed the sources of TDR in over 50 patients carrying resistant viruses and estimated that 84.8% of these infections had a drug-naive source. Both studies highlight the significant role of untreated or HIV undiagnosed individuals in the spread of DRMs. The DRMs most frequently found in naive-to-naive transmission chains were amongst the most prevalent in the cohort. These mutations (e.g. L90M in PRT; K103N, T215D and T215S in reverse transcriptase) are associated with a weak replicative impairment [24–26], which probably explains their sustainability in a drug-free environment, high frequency in the cohort and high probability of naive-to-naive transmission. As comparative studies on DRMs fitness costs are scarce, we were unable to establish a correlation between viral fitness and transmissibility for all DRMs found in naive-to-naive transmission chains. However, we note that resistant mutations associated with a severe fitness cost (e.g. tenofovir-associated mutations K65R and M184V) were not found in treatment-naive resistant phylotypes. Furthermore, naive-to-naive transmission of multiresistant strains was rare and inversely correlated to the number of DRM harboured by the transmitted virus. We also established that the rate at which these DRMs are lost in the untreated population is low, and their time of persistence up to a decade long. Remarkably, we did not to find a strong correlation between DRM persistence within a host [20] and at the population level. This can probably be explained by the timing of HIV transmissions. If, as is generally admitted (e.g. [27]), most infections occur early on in the acute phase of the transmitter, most transmissions of resistance will occur before differences in intra-host persistence play a role (e.g. before the least stable mutations revert to wild type in the host). All together, these observations illustrate the role of viral fitness on the establishment of resistance reservoirs in the untreated population.

Incomplete data, a problem inherent to all convenience datasets, are unlikely to compromise the validity of our conclusions. It could be possible, for instance, that unsampled treated individuals were involved in the observed transmission chain. This would result in a lower prevalence of TDR with a drug-naive source that what estimated here. However, as a vast majority of infections are believed to occur while the transmitter is undiagnosed or within the acute phase of infection [27–29], and therefore untreated, such scenario remains unlikely. A statistical framework designed to estimate the probability that a phylotype is incomplete would permit to rigorously test and adjust our findings with that regard. Of note, our approach concentrates on characterized transmission chains, thus excluding for the analysis sequences with no linkage in the cohort. Further statistical development would also involve the extrapolation of inferred trends to the entire dataset.

Our phylotype-based method offers several advantages over traditional phylogenetic studies of TDR. It allows the simultaneous identification of drug-resistant lineages and reversion events; it does not rely on the inference of infection dates from external sources; it takes into account phylogenic uncertainty; finally, it is time and computation-inexpensive, and scalable to large HIV sequence datasets. These assets make a phylotype-based approach an attractive tool for routine surveillance of TDR in HIV as well as in other pathogens for which genotypic resistance data are available.

Back to Top | Article Outline

Acknowledgements

LabEx NUMEV supported the postdoctoral grant for R.M., and the Computational Biology Institute (Institut de Biologie Computationnelle, IBC) provided research environment for R.M..

UK HIV Drug Resistance Database Steering Committee included Celia Aitken (Gartnavel General Hospital, Glasgow); David Asboe, Anton Pozniak (Chelsea & Westminster Hospital, London); Patricia Cane (Public Health England, Porton Down); David Chadwick (South Tees Hospitals NHS Trust, Middlesbrough); Duncan Churchill (Brighton and Sussex University Hospitals NHS Trust); Duncan Clark (St Bartholomew's and The London NHS Trust); Simon Collins (HIV i-Base, London); Valerie Delpech (Centre for Infections, Public Health England); Samuel Douthwaite (Guy's and St. Thomas’ NHS Foundation Trust, London); David Dunn, Esther Fearnhill, Kholoud Porter, Anna Tostevin, Ellen White (MRC Clinical Trials Unit at UCL, London)*; Christophe Fraser (Imperial College London); Anna Maria Geretti (Institute of Infection and Global Health, University of Liverpool); Antony Hale (Leeds Teaching Hospitals NHS Trust); Stéphane Hué (London School of Hygiene & Tropical Medicine); Steve Kaye (Imperial College, London); Paul Kellam (Wellcome Trust Sanger Institute & University College London Medical School); Linda Lazarus (Expert Advisory Group on AIDS Secretariat, Public Health England); Andrew Leigh-Brown (University of Edinburgh); Tamyo Mbisa (Virus Reference Department, Public Health England); Nicola Mackie (Imperial NHS Trust, London); Samuel Moses (King's College Hospital, London); Chloe Orkin (St. Bartholomew's Hospital, London); Eleni Nastouli, Deenan Pillay, Andrew Phillips, Caroline Sabin (University College London Medical School. London); Erasmus Smit (Public Health England, Birmingham Heartlands Hospital); Kate Templeton (Royal Infirmary of Edinburgh); Peter Tilston (Manchester Royal Infirmary); Daniel Webster (Royal Free NHS Trust, London); Ian Williams (Mortimer Market Centre, London); Hongyi Zhang (Addenbrooke's Hospital, Cambridge).

*Coordinating Centre

The UK HIV Drug Resistance Database is supported by the Medical Research Council This project was also funded by the University College London Hospitals Biomedical Research Centre and the Collaborative HIV and Anti-HIV Drug Resistance Network (CHAIN), by the Wellcome Trust (grant number 092807) and a studentship from the Biotechnology and Biological Science Research Council (MR-C). The LabEx NUMEV supported the postdoctoral grant of R.M.

Centres contributing data included Clinical Microbiology and Public Health Laboratory, Addenbrooke's Hospital, Cambridge (Jane Greatorex); Guy's and St. Thomas’ NHS Foundation Trust, London (Siobhan O'Shea, Jane Mullen); PHE – Public Health Laboratory, Birmingham Heartlands Hospital, Birmingham (Erasmus Smit); PHE – Virus Reference Department, London (Tamyo Mbisa); Imperial College Health NHS Trust, London (Alison Cox); King's College Hospital, London (Richard Tandy); Medical Microbiology Laboratory, Leeds Teaching Hospitals NHS Trust (Tracy Fawcett); Specialist Virology Centre, Liverpool (Mark Hopkins, Lynn Ashton); Department of Clinical Virology, Manchester Royal Infirmary, Manchester (Peter Tilston); Department of Virology, Royal Free Hospital, London (Claire Booth, Ana Garcia-Diaz); Edinburgh Specialist Virology Centre, Royal Infirmary of Edinburgh (Jill Shepherd); Department of Infection & Tropical Medicine, Royal Victoria Infirmary, Newcastle (Matthias L Schmid, Brendan Payne); South Tees Hospitals NHS Trust, Middlesbrough (David Chadwick); Department of Virology, St Bartholomew's and The London NHS Trust (Spiro Pereira, Jonathan Hubb); Molecular Diagnostic Unit, Imperial College, London (Steve Kaye); University College London Hospitals (Stuart Kirk); West of Scotland Specialist Virology Laboratory, Gartnavel, Glasgow (Rory Gunson, Amanda Bradley-Stewart, Celia Aitken).

Back to Top | Article Outline

Conflicts of interest

There are no conflicts of interest.

Back to Top | Article Outline

References

1. May MT, Gompels M, Delpech V, Porter K, Orkin C, Kegg S, et al. Impact on life expectancy of HIV-1 positive individuals of CD4+ cell count and viral load response to antiretroviral therapy. AIDS 2014; 28:1193–1202.
2. Tang JW, Pillay D. Transmission of HIV-1 drug resistance. J Clin Virol 2004; 30:1–10.
3. Dolling D, Sabin C, Delpech V, Smit E, Pozniak A, Asboe D, et al. Resistance UKCGoHD. Time trends in drug resistant HIV-1 infections in the United Kingdom up to 2009: multicentre observational study. BMJ 2012; 345:e5253.
4. Hue S, Gifford RJ, Dunn D, Fernhill E, Pillay D. Resistance UKCGoHD. Demonstration of sustained drug-resistant human immunodeficiency virus type 1 lineages circulating among treatment-naive individuals. J Virol 2009; 83:2645–2654.
5. Wagner BG, Garcia-Lerma JG, Blower S. Factors limiting the transmission of HIV mutations conferring drug resistance: fitness costs and genetic bottlenecks. Sci Rep 2012; 2:320.
6. Audelin AM, Gerstoft J, Obel N, Mathiesen L, Laursen A, Pedersen C, et al. Molecular phylogenetics of transmitted drug resistance in newly diagnosed HIV Type 1 individuals in Denmark: a nation-wide study. AIDS Res Hum Retroviruses 2011; 27:1283–1290.
7. Drescher SM, von Wyl V, Yang WL, Boni J, Yerly S, Shah C, et al. Treatment-naive individuals are the major source of transmitted HIV-1 drug resistance in men who have sex with men in the Swiss HIV Cohort Study. Clin Infect Dis 2014; 58:285–294.
8. Lewis F, Hughes GJ, Rambaut A, Pozniak A, Leigh Brown AJ. Episodic sexual transmission of HIV revealed by molecular phylodynamics. PLoS Med 2008; 5:e50.
9. Chevenet F, Jung M, Peeters M, de Oliveira T, Gascuel O. Searching for virus phylotypes. Bioinformatics 2013; 29:561–570.
10. Yang Z, Rannala B. Molecular phylogenetics: principles and practice. Nat Rev Genet 2012; 13:303–314.
11. UKCHCS Committee. The creation of a large UK-based multicentre cohort of HIV-infected individuals: the UK Collaborative HIV Cohort (UK CHIC) Study. HIV Med 2004; 5:115–124.
12. Hue S, Brown AE, Ragonnet-Cronin M, Lycett SJ, Dunn DT, Fearnhill E, et al. Phylogenetic analyses reveal HIV-1 infections between men misclassified as heterosexual transmissions. AIDS 2014; 28:1967–1975.
13. Bennett DE, Camacho RJ, Otelea D, Kuritzkes DR, Fleury H, Kiuchi M, et al. Drug resistance mutations for surveillance of transmitted HIV-1 drug-resistance: 2009 update. PLoS One 2009; 4:e4724.
14. Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res 1997; 25:4876–4882.
15. Goloboff PA, Farris J, Nixon K. T.N.T.: tree analysis using new technology. http://www.zmuc.dk/public/phylogeny/TNT 2003 [Accessed 17 June 2015].
16. Price MN, Dehal PS, Arkin AP. FastTree xs2 - approximately maximum-likelihood trees for large alignments. PLoS One 2010; 5:e9490.
17. Fitch WM. Toward defining the course of evolution: minimum change for a specific tree topology. Syst Zool 1971; 20:406–416.
18. Drummond AJ, Rambaut A. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol Biol 2007; 7:214.
19. Shapiro B, Rambaut A, Drummond AJ. Choosing appropriate substitution models for the phylogenetic analysis of protein-coding sequences. Mol Biol Evol 2006; 23:7–9.
20. Castro H, Pillay D, Cane P, Asboe D, Cambiano V, Phillips A, et al. Persistence of HIV-1 transmitted drug resistance mutations. J Infect Dis 2013; 208:1459–1463.
21. Schmidt D, Kollan C, Fatkenheuer G, Schulter E, Stellbrink HJ, Noah C, et al. Estimating trends in the proportion of transmitted and acquired HIV drug resistance in a long term observational cohort in Germany. PLoS One 2014; 9:e104474.
22. Charest H, Doualla-Bell F, Cantin R, Murphy DG, Lemieux L, Brenner B, et al. A significant reduction in the frequency of HIV-1 drug resistance in Quebec from 2001 to 2011 is associated with a decrease in the monitored viral load. PLoS One 2014; 9:e109420.
23. Siliciano JD, Siliciano RF. Recent trends in HIV-1 drug resistance. Curr Opin Virol 2013; 3:487–494.
24. de Ronde A, van Dooren M, van Der Hoek L, Bouwhuis D, de Rooij E, van Gemen B, et al. Establishment of new transmissible and drug-sensitive human immunodeficiency virus type 1 wild types due to transmission of nucleoside analogue-resistant virus. J Virol 2001; 75:595–602.
25. Cong ME, Heneine W, Garcia-Lerma JG. The fitness cost of mutations associated with human immunodeficiency virus type 1 drug resistance is modulated by mutational interactions. J Virol 2007; 81:3037–3041.
26. Martinez-Picado J, Savara AV, Sutton L, D’Aquila RT. Replicative fitness of protease inhibitor-resistant mutants of human immunodeficiency virus type 1. J Virol 1999; 73:3744–3752.
27. Phillips AN, Cambiano V, Nakagawa F, Brown AE, Lampe F, Rodger A, et al. Increased HIV incidence in men who have sex with men despite high levels of ART-induced viral suppression: analysis of an extensively documented epidemic. PLoS One 2013; 8:e55312.
28. Hall HI, Holtgrave DR, Maulsby C. HIV transmission rates from persons living with HIV who are aware and unaware of their infection. AIDS 2012; 26:893–896.
29. Powers KA, Ghani AC, Miller WC, Hoffman IF, Pettifor AE, Kamanga G, et al. The role of acute and early HIV infection in the spread of HIV and implications for transmission prevention strategies in Lilongwe, Malawi: a modelling study. Lancet 2011; 378:256–268.
Keywords:

drug resistance; HIV; phylotype; treatment-naive

Supplemental Digital Content

Back to Top | Article Outline
Copyright © 2015 Wolters Kluwer Health, Inc.