XIV International HIV Drug Resistance Workshop: Subtypes, resistance pathways and implications for therapy

Several presentations relating to HIV subtypes highlighted somewhat different pathways to the accumulation of drug resistance in non-B HIV-1 subtypes compared to subtype B. Additionally, there is now an increasing body of data that shows that these differences in profiles may potentially affect clinical outcomes.

Confirmation that selection of resistance following first-line therapy can differ among subtypes came from a large, international, multi-centre cohort study involving several surveillance programmes from Africa, Thailand, Brazil, Argentina, Japan, United States, Canada and several countries in Europe including Belgium, Spain, Portugal and the United Kingdom. The study was intended to examine the response to the recommended WHO first-line regimens which suggest AZT (zidovudine, Retrovir), d4T (satvudine, Zerit), 3TC (lamivudine, Epivir), nevarapinex (NVP, Viramune) and efavirenz (EFV, Sustiva). Whilst transmitted resistance to these drugs is not expected to be high given the dearth of current treatment availability in developing countries, nevertheless the ability to select subsequent regimens where choices are finite becomes critically important.

Drug exposure, mutational frequencies and co-selection of mutations defined as the ‘mutation sum’ and genotypic sensitivity scores (GSS) were assessed in the study. The mutations included in the analysis were non-nucleoside reverse transcriptase inhibitor (NNRTI): 103, 106, 181, 188, 190 and nucleoside analogue reverse transcriptase inhibitors (NRTI) mutations: TAMs 41, 67, 70, 210, 215, 219 and 184. Patients across the study were exposed to between 2.2 and 3 NRTIs and 1.0-1.2 NNRTI with interestingly, no differences in drug-class exposure between the different subtypes. Between 67%-100% harboured resistance to NRTI with at least a single mutation and 85%-100% had resistance to NNRTI with significant differences reported in the frequency of specific mutations in both NRTI (41, 67, 70, 75, 210, 215) and NNRTI (103, 181, 190); they do not ascribe NNRTI mutations to differences in drug exposures between patient groups. Subtypes B, F and CRF01-AE reported significantly higher mutational scores compared to subtypes A. C, D and G, suggesting greater resistance to first-line regimens and less sensitivity to second-line NRTI therapy. Whilst it is difficult to compare against such heterogeneous patient cohorts, these data indicate that the rate and evolutionary pathways to drug resistance may be affected by subtype with implications for first-line and subsequent treatment strategies.

Variations in resistance pathways within one HIV sub-type

In an intriguing study presented by Florence Doualla-Bell from McGill University in Canada in association with the Botswana-Harvard Partnership for HIV Research and Education, Gaborone, Dr Doualla-Bell showed that genetic variations can prevail within subtypes, in this case between subtype C viruses from Ethiopia and Botswana. Botswana has one of the highest prevalence of HIV in the developing world with an infection rate of around 40%.

Botswana also has a national treatment programme that provides HIV therapy to comparatively more of its population than some of its neighbouring countries; 34,000 patients since 2002. The national guidelines suggest a recommended first-line regimen constructed from AZT, 3TC and NVP or EFV largely stratified by gender. Following failure, the second-line option comprises nelfinavir (NFV, Viracept) plus continuing treatment with some of the drugs from the prior regimen (most commonly zidovudine).

Forty-eight patients from the Botswana National Treatment Programme including 33 naïve and 10 NVP-treated individuals were assessed for resistance profiles against sequences from the Stanford database (subtype C from Ethiopian patients). Virological failures in these patients were associated with single major mutations in the protease gene including most prominently D30N with one patient also reporting the L90M mutation. Of interest, differences in polymorphisms and secondary mutations were observed between treated and untreated patients. This differs somewhat from previously reported patterns of resistance in subtype C by Grossmann and Schapiro where D30N was observed as a polymorphism rather than a primary mutation. This difference of 25% versus 62%, the authors suggest, may reflect variations in host genetics and response to treatment between Ethiopian and Batswana patients classified as C1 and C4 subtypes. Polymorphic variations were also apparent between these patients, most notably at positions:

These data suggest the need for greater understand of pathways leading to resistance inter-subtype as well as scrutiny on intra-subtype populations. What are currently lacking from some of the studies to date are correlations with drug exposure which may have an impact as well an urgent need to increase the numbers of patients in the sample size to allow for a more comprehensive analysis.

Viral fitness may determine transmission rate of drug-resistant virus

The modelling of predicted resistance in Botswana by Sally Blower and her colleague at University of California Los Angeles (UCLA) seems particularly opportune. Dr Blower is recognised in the field for her refinement of mathematical models to predict transmission of resistance at a population level. Using data from the Botswana national treatment programme that predicts treatment of 85,000 patients by 2009 they ascertain that the overall resistance rate by 2009 can be predicted to remain below the WHO defined threshold of 5%. This is predicated on the ability of the drug-resistant strains to be less transmissible due to reduced viral fitness caused by low level resistance. However, if the fitness of the virus due to drug resistance increases and becomes more transmissible, the rate of transmitted resistance can be expected to rise to nearly 10% by 2009. (13)

Sub-types and protease inhibitor resistance

Two interesting studies highlighting experience from Portugal compared responses in subtype B to treatment in HIV-1 subtype G. Ricardo Camacho from the Hospital Egas Moniz in Lisbon with other colleagues from Portugal and Belgium reported differences in response to tipranavir (TPV) in particular substitutions at position 82 of subtype G where the WT codon is 82I compared to 82V (isoleucines versus valine substitutions). From 240 sequences taken from patients experiencing virological failure with changes at codon 82, they found that 82A was more commonly presented in B due to a lower genetic barrier but that 82T and 82S predominate in subtype G. They also found a novel mutation at position 82 M showing a methionine substitution which has been reported to confer a high level of fitness and resistance to indinavir (IDV, Crixivan). These differences are of importance in that they relate to impact to PIs in particular TPV where 82T but not 82A is implicated in resistance. The authors suggest that I82M should be scrutinised more closely in algorithms for resistance interpretation.

Anne-Mieke Vandamme from Leuven, Belgium further analysed 2491 pol sequences to evaluate resistance to TPV. Of the 778 patients with PI experience, most of whom had been treated with NFV, IDV and lopinavir (LPV), differences in TPV mutational scores were investigated. These were based on the Boehringer-Ingelheim (BI) profile of 4 or more mutations that confer reduced sensitivity to TPV namely: 10V, 13V, 20M/R/V, 33F, 35G, 36I, 43T, 46L, 54A/M/V, 58E, 69K, 74P, 82L/T, 83D and 84V. BI studies were largely conducted with subtype B participants. The researchers from the study here found that all the subtypes not classified as B reported a significantly higher TPV mutational score. With greater PI experience, the number of mutations increased with mutations at 13V, 20R, 36I and 69K contributing most to the increased TPV mutational score. (15)