July 24, 2007
Doug Richman provided a comprehensive recap on HIV resistance from the annual HIV resistance conference, held this year in June in Barbados.1 His focus was on four major areas:
With every new medication it's critical to understand which mutations are predictive of a less than optimal response. A group led by Vingerhoets looked at the impact of baseline NNRTI mutations on the virologic response to etravirine in the phase 3 clinical trials DUET-1 and DUET-2, the phase 3 studies evaluating the efficacy of etravirine.2
The researchers evaluated 44 mutations in the reverse transcriptase (RT) gene that are thought to impact NNRTI activity (14 common mutations, 27 additional mutations and three that appear to be specific for etravirine). Of the 44 mutations, 26 were present in more than five people, and 13 were found to specifically affect the activity of etravirine (V90I, A98G, L100I, K101E/P, V106I, V179D/F, Y181C/I/V, G190A/S). Results of the study showed that virologic response was significantly decreased when either the Y181C or G190A RT mutations were present with additional RT mutations.
For example, if these mutations were present with an additional four RT mutations thought to affect etravirine activity, less than 20% of those patients had a virologic response.
Dr. Richman reviewed data from Hazuda and colleagues. They analyzed resistance to the HIV integrase inhibitor raltegravir from the phase 2 P005 study, a 24-week, four-arm, three-dose escalation (+ placebo arm) study in treatment-experienced people. This study demonstrated significant virologic potency (about 60% of patients experienced a viral load drop to below 50 copies/mL) of this integrase inhibitor when used with an optimized background HIV medication regimen.3 However, virologic failure was observed in 38/133 (28%) patients. Of these, 35 demonstrated mutations in the HIV-1 integrase gene.
Two different pathways for resistance were found, the N155H (observed in 14 patients) and Q148R (20 patients). One person developed Y143R. The most common pattern was Q148R-G140S (13 patients). A similar number developed resistance across the three-dosing arms. Resistance was more likely to occur in those with very high viral loads, no use of enfuvirtide (T-20, Fuzeon) in the background regimen and no other HIV medications that had activity, as demonstrated by phenotypic resistance testing.
Dr. Richman then showed a slide that looked at the impact of secondary or additional mutations in addition to the N155H or Q148R mutation, demonstrating that some additional mutations are more powerful in creating high level resistance to raltegravir than others. Another slide showed that these mutations also affect the activity of elvitegravir and result in a loss of activity.
The last topic Dr. Richman discussed involved the CCR5 inhibitor class. A lot of discussion revolved around the fact that in many patients, HIV is a mixture of viruses that either use R5 or X4 or both as the co-receptor along with CD4 to gain entry into cells. The question posed was, in patients who are treated with R5 inhibitors and fail the regimen, is it the result of mutations in the R5 receptor or just the emergence of previously unrecognized X4 virus?
Data from Lewis and colleagues was presented on one of the maraviroc (MVC; brand name expected to be Selzentry) studies, where 20 trial participants (16 who received active drug) were studied who had early or late failure on a maraviroc-containing regimen.4 Dr. Richman then spent some time describing the issues around the sensitivity of assays that determine whether R5 or X4 virus is present in a blood sample. In order to be able to detect a 1% X4 population with 99% accuracy, one would have to analyze hundreds of sequences from a single sample, which of course is not practical.
Dr. Richman gave an example of a patient who by the Monogram Trofile tropism assay was determined to have R5 virus, but by clonal sequencing analysis of over a hundred sequences had 7% X4 virus already present at baseline before receiving maraviroc. After receiving maraviroc, no clones sequenced were found to be an R5 virus, and the tropism assay result showed a dual-mixed pattern (meaning that R5 and/or X4 using virus were present according to this assay). The sequencing analysis indicated that these were not mutated R5 viruses, but rather pre-existing X4 viruses.
In another study, Mori and colleagues determined that maraviroc-resistant virus recognizes the R5 receptor differently.5 Even though maraviroc binds to R5, the binding site for maraviroc-resistant virus is not disrupted (as it would be for a non-resistant virus) and so the virus could still infect the cell. Dr. Richman briefly reviewed data looking at a new R5 inhibitor from Roche (RO 1752) that still appears to have activity against maraviroc-resistant virus.
Dr. Richman also mentioned a review by Harrigan that looked at the epidemiology of R5 and X4 tropism.6 In the seven studies that have been published related to this, about 80% of antiretroviral-naive patients have R5 virus, 12% to 20% have dual-mixed virus, and less than 1% have X4 virus.
In treatment-experienced patients, about 50% to 70% had R5, 22% to 48% had dual-mixed, and 2% to 5% had X4 only. Prevalence of dual-mixed or X4 virus was also correlated with lower CD4+ cell counts. A study by Gill and colleagues looked at 920 virus samples that were R5 or dual-mixed and compared the Trofile phenotype result to different sequencing genotype assays interpretive rules. All had good correlation with phenotype. But all had poor sensitivity for presence of X4 virus. When a clonal analysis was done on the samples, they all showed that X4 virus was indeed present.
It is further currently recognized that the Monogram Trofile assay is 100% sensitive in picking up 10% minor variants in a mixed population, but only 83% sensitive at the 5% level. Monogram is working on a newer assay that will be about 10 times more sensitive to pick up small amounts of X4.
Dr. Richman ended with some final comments on using genotype sequencing assays for determining tropism of viruses from patient samples. The sensitivity of bulk sequencing or population sequencing will be poor for picking up small amounts of X4 virus. At this point, interpretation rules have poor sensitivity for picking up subtype B-X4 virus and no data exists for looking at this with non-subtype B virus.