Short resistance reports from Seville

11 September 2002. Related: Conference reports, Drug resistance, Intl Drug Resistance Workshop 11 Sitges 2002.

Simon Collins, HIV i-Base

CX/R5 phenotype and replicative capacity of resistant virus

As research progresses on entry inhibitors, biological phenotype (CCR5/CXCR4) may become increasingly important within treatment choices. This study of 28 HIV-infected patients failing ARV treatment found that although viral load levels were similar, people with R5 phenotype had a higher number of resistant mutations. The six isolates with >14 mutations in this group were all R5, although the R5 group retained higher CD4 counts and significantly lower replicative capacity of their HIV isolates (p=0.008).

All isolates had reduced replicative capacity compared to wild type, with lowest replicative capacity when >5 mutations were present. Highest replicative capacity was retained when NNRTI-associated K103N and Y181C mutations were present.

As expected, higher HIV isolation was reported in patients who maintained high replicative capacity.

Ref: Nicastri et al – Biological phenotype, replicative capacity and drug resistance of HIV strains isolated from patients failing antiretroviral therapy. Seville 2002. Abstract 52. Antiviral Therapy 2002; 7:S45

Superinfection shown in vitro and in model

Chakraborty and colleagues reported two studies that reinforced the plausibility of superinfection.

In vitro experiments involved infection and reinfection of peripheral blood mononuclear cells (PBMCs) with wild type (WT) and drug resistant (R) virus in the presence and absence of drug. In the absence of drug, WT virus out-competed the R virus and did not allow reinfection with the less fit strain. In the presence of 3TC, the R virus was more fit than WT virus, was able to super infect WT virus and prevented super infection from WT virus.

Analysis of five replicates of three different HIV-1 superinfections in SCID-hu Thy/Liv mice, which had been infected and then superinfected seven days later, showed reinfection on one viral combination set. This provided evidence that an animal model can be used to study superinfection.

The group have also been following 15 seropositive couples, independently infected and not practising safe sex from two-four years with longitudinal samples taken at a minimum six monthly intervals. No evidence of superinfection has been found in follow up so far, though details of ARV treatment were not included in the abstract.

Ref: Chakraborty et al – Evaluating HIV-1 superinfection in cell culture, the SCID-hu Thy/Liv model and HIV-infected individuals with high risk of exposure to the virus. Seville Abstract 55. Antiviral Therapy 2002; 7:S47.

Theoretical rationale for sequential monotherapy in salvage therapy

Andrew Phillips from the Royal Free Centre for HIV Medicine London presented a theoretical basis for treating multiple drug resistance based on a new paradigm. Continued use of 3TC despite high-level resistance due to the key M184V mutation is recognised to provide benefit (up to –0.4 log) due to reduced replicative competence and several studies at the resistance meeting looked at this for other drugs.

Given that resistant virus is less fit compared to wild type they developed a model for rotating drugs (mono, dual, triple therapy) on a daily or weekly basis, The model relying on the fact that it takes a few days for the virus which can replicate best on any one day to grow in number. Although combinations of drugs were not included, apparent antagonisms between particular drug mutations might be particularly useful to exploit

See overview of this study on page 25-26 of this HTB.

Ref: Phillips A et al – Theoretical rationale for the use of single drug sequential antiretroviral therapy regimens. Seville Abstract 73. Antiviral Therapy 2002; 7:S61.

Phenotypic and genotypic predictors of response to tenofovir

Two studies at the workshop and an oral presentation at Barcelona evaluated predictors of response to the recently approved nucleotide analogue tenofovir DF (TDF) from results of previous studies in treatment experienced patients.

Masquelier and colleagues reported on genotypic determinant of response to tenofovir in a sub-group of 191 patients treatment experienced patients enrolled in the French ATU (expanded access programme). [1] Baseline genotype on stable therapy and change in viral load after adding tenofovir for three months were correlated to evaluate the role of each RT mutation and to construct a mutation score validated by bootstrap sampling.

Baseline viral load was 4.9 log 10 copies/ml (3.6-6.3) and mean decrease at month 3 was 0.9 log copies/ml (range –3.2 – +1.8). RT mutations individually associated with a poorer response were M41L, E44D, D67N, T69D/N/S, V118I, L210W and T215Y/F. The number of these mutations (excluding V118I) formed the basis for a TDF mutation score corresponding to the most significant p value. In their analysis, a score of <3 predicted absence of resistance, 3-4 possible resistance, and >4 resistance. Corresponding reductions in viral load for each group was –1.3 ±1.1 log, -0.8±1.1 log and –0.4 ±0.9 log copies/ml respectively.

The score was found significant in all bootstrap samples, although multivariate analysis was not presented to validate these results. Also, it is important to remember that unlike the Gilead 907 Study, patients in this EAP programme may have changed or added other drugs at the same time as adding TDF. Viral load responses were not solely related to TDF, and the analysis for the contributory role of TDF is not currently available.

Gilead’s analysis of the role of genotypic mutations from virology sub studies of their 902 and 907 studies (n=332 genotypes) reported two mutational pathways. Pathway 41-210-215 was associated with higher TDF resistance particularly of additional TAMS were present. Although K67R is associated with resistance and limited response addition mutations in the presence of the 67-70-219 pathway did not produce further resistance. Viral load reductions for patients without thymidine analogue mutations (TAMS), with 1-2 TAMS and with ³ 3 TAMS without M41L or L210W of –0.67 (n=61) of –0.8 (n=97) and -0.66 (n=88) and respectively. Patients with ³ 3TAMS including M41L or L210W had responses of –0.21 log (n=86). [2]

Phenotypic resistance tests are increasingly used when genotype results are unclear and as an additional tool in heavily treated patients. These tests rely on determining individual sensitivity cut-offs for each drug, and these cut-offs frequently require revision and often vary depending on the assay being used.

The Virco Antivirogram previously established that a cut-off of fourfold for tenofovir was equivalent to 24-week reduction in RNA viral load of only –0.24 log copies/ml.

In this study Miller and colleagues from Gilead Sciences, looked at phenotypic cut-offs, using ViroLogic PhenoSense assay from 112 randomly selected patient from the Gilead 907 study which added TDF for 24 weeks to a failing background therapy (viral load >500 copies/ml). [3]

Mean baseline susceptibility was 1.55-fold from wild-type control (range 0.3-13) and mean week 24 response was 0-63 log copies/ml (range –2.03 – +1.16). Categorical response variables were created corresponding to ³3, ³4 and ³5 log average reductions in viral load and well as for nadir ³5log reduction from baseline. Bootstrap analysis (n=3000) revealed two splits for susceptibility at values of 4.0 and 1.1.

The biological cut-off for the PhenoSense assay is 1.4-fold (ie 99% of wild-type virus falls within 1.4-fold sensitivity of the drug sensitive reference virus). Using cut-offs of  <1.4-fold, 1.4-4-fold and >4-fold showed median viral load responses of –0.77 (n=78), -0.37 (n=26) and –0.06 (n=8) respectively.

COMMENT

Given the range of results even with mutation score >4, some patients may still benefit from TDF. However, phenotypic resistance >4-fold does appear to predict lack of clinical response.

References:

1. Masquelier B et al – Indentification of genotypic determinants of the virological response to tenofovir-including regimens in nucleoside reverse transcriptase inhibitor-experienced patients. Seville. Abstract 126. Antiviral Therapy 2002; 7:S105.
2. M.D. Miller et al – Expanded response analyses of tenofovir DF Therapy by Baseline Resistance Genotype and Phenotype – Barcelona [ThOrB1390]
3. Miller M et al – Determination of clinical cut-offs for reduced response to tenofovir DF therapy in antiretroviral experienced patients. Seville. Abstract 125. Antiviral Therapy 2002; 7:S104.

Importance of intermediate mutations at 215 in treatment naive patients

Two studies provided conflicting conclusions on the importance of less frequent substitutions at codon 215 (‘revertant’ C, D, E, N, A, V and S substitutions).

Riva and colleagues from the Italian ICoNA study looked at 405 treatment naive individuals who had a genotypic test prior to starting a thymidine analogue combination. Virological failure was defined as time to the first of two consecutive viral load tests after >24 weeks therapy. Thirteen (3%) patients were found to have 215 variants at baseline (D=6, c=1, E=2, N=1, A=1, V=1 and S=1) and two patients (0.5%) additionally had the more significant T215Y mutation. Treatment was started with 2RTIs and either a PI or NNRTI in over 90% cases, with AZT used in approximately 75% and d4T in 25% regimens.

Multivariate Cox regression analysis, including variables pre-HAART viral load, use of Invirase as only PI, use of AZT, number of prior RT and PI mutations, and time from seroconversion to date of test showed that patients carrying 215 substitutions had an increased risk of virological failure compared to patient without these mutations (adjusted RH=2.44, 95% CI 1.01-5.89, p=0.04). 9/13 patients with 215 substitutions experienced virological failure. Resistance test at time of failure in five of these patients showed reversion to T215Y in three cases.

Lanier and colleagues from GlaxoSmithKline looked retrospectively at baseline genotypes from 574 treatment naïve patients from three studies using AZT/3TC with either indinavir, abacavir or efavirenz. Genotypes at failure (confirmed viral load >400 copies/ml) were available for >90% of patients.

2.8% of patients showed baseline mutations at 215 D/C/S (n=8, 4, 4 respectively). M41L was detected simultaneously in 5 cases and L210W in one. 6/16 patients with mutations at baseline later failed virologically, and in three cases the dominant species became 215Y. The small numbers of cases didn’t allow for an analysis of a preferential regimen, and the continued detection of revertant mutations was not always associated with treatment failure.

References:

Riva et al – Transmitted virus with substitutions at position 215 and risk of virological failure in antiretroviral naive patients starting HAART. Seville. Abstract 124. Antiviral Therapy 2002; 7:S103.

Lanier E et al – Effect of baseline 215D/C/S ‘revertant’ mutations to lamivudine/zidovudine-containing regimens and emergence of 215Y upon virological failure. Seville. Abstract 146. Antiviral Therapy 2002; 7:S120.

Genotypic resistance to nelfinavir detected 4 months prior to viral rebound >500 copies

Descamps and colleagues reported on 16/21 antiretroviral naive patients failing first line nelfinavir containing regimens with successfully amplified protease genotypic results. Failure was defined as a viral load >500 copies/ml over the first year of follow up and genotypic testing (ViroSeq) with additional nested PCR performed at -4 month (S1) and -2 month (S2) prior to failure and time of rebound (baseline).

In 8/16 patients, wild type virus was present at all three time points. In the other eight patients mutations were detected prior to rebound at S1 or S2. In two cases resistance at rebound matched the earlier samples but major or minor differences were detected in the remaining six indicating minor viral species can evolve with some drugs even at levels <500 copies/ml. At time of rebound 3/8 remained wild type, and 5/8 harboured D30N, associated with N88D or M46I+V77I in two cases.

Amplification of RT gene in 15 patients revealed no RT mutations in seven patients receiving ddI/d4T/nelfinavir. M184V mutation was detected at S1 or S2 in 8/9 evaluable patients receiving 3TC-containing regimens.

Without full details on more sensitive levels of suppression – and the treatment goal for all naive patients is now recognised as suppression to <50 copies/ml – it is difficult to draw conclusions from this study. However it should reinforce the concern that suppression to only <500 carries high risk of subsequent viral rebound and evolution of drug resistance while still <500 copies/ml

Ref: Descamps et al – Detection of resistance mutations in patients receiving first line protease inhibitor containing regimen with undetectable viral load. Seville. Abstract 139. Antiviral Therapy 2002; 7:S115.

Clinical use of ddI in patients with M184V mutation, with and without 3TC

In vitro data have suggested that M184V may reduce the efficacy of ddI. Several studies suggested that this should not be a concern in clinical practice, whether or not 3TC is included in the subsequent regimen.

Pozniak and colleagues performed a retrospective analysis of the proportion of patients achieving viral load <400 copies and the average area under the curve (AUC) viral load response in 281 patients followed for at least 48 weeks who switched to ddI-containing regimens in a cohort at the Chelsea and Westminster Hospital in London. [1] Baseline characteristics and univariate analysis of changes in viral load were performed using Wilcoxon rank sum test. Multivariate analyses were performed using the Van Elteren test. Additional variables included baseline RT mutations, previous PI and NNRTI use and whether 3TC was used in the subsequent regimen.

Patients with the M184V mutation at treatment switch (105/281) had greater median fold change in phenotypic susceptibility to ddI than patients without 184 (2.2 vs. 1.2-fold, p<0.001). However in both univariate and multivariate analyses, median change in viral load and percentage of patients achieving <400 copies was similar irrespective of 184 at baseline. Additionally patients with 184 at baseline showed a significantly better AUC response if treated with ddI-containing regimens compared to those who didn’t use ddI (p<0.03).

A second study from Eron and colleagues looked at treatment response in patients from ACTG 307. This double-blinded eight-week study compared ddI monotherapy to using ddI with two different doses of hydroxyurea in 134 treatment naive and experienced patients. Previous treatment was discontinued 14 days prior to the study. [2]

16/24 3TC-experienced patients had baseline genotype, 13 of which included 184V. Median baseline viral load were 4.2 (IQR 3.8-5.1), 4.6 (4.2-5.0) and 4.6 (4.3-4.9) in the 3TC-naive, 3TC experienced (without 184) and 3TC-experienced with 184 arms. Decreases in HIV RNA at week 8 were 1.81 (n=52), 1.42 (n=24) and 1.42 (n=13) respectively. In analysis by treatment arm, decreases of 0.89 (19), 0.77 (7) and 0.77 (4) were recorded for patients receiving ddI alone and 1.94 (19), 1.69 (7) and 1.63 (9) for patients receiving ddI/HU.

Aside from showing a substantial benefit from hydroxyurea, and despite slightly higher decreases in viral load in 3TC-naive individuals, previous 3TC therapy appeared to have minimal short-term effect on antiviral activity of ddI, with or without hydroxyurea.

Winters and colleagues looked at the effect of M184V on the response to ddI in 63 nucleoside experienced (AZT/3TC or AZT/3TC/ddC) patients who rolled over to ddI/d4T plus nelfinavir, indinavir or nelfinavir + indinavir in the ACTG 364 Study. Forty one additional patients on AZT/3TC/ddI were randomised to d4T/ddI or d4T/3TC plus either nelfinavir, indinavir or nelfinavir + indinavir. [3]

Although reporting in multivariate analyses that M184V did not significantly contribute to ddI resistance in vivo, the overlapping use of other nucleosides, complexity of switched treatments. Low numbers of patients and a high definition of virological failure at >2000 copies/ml probably limits the practical results from this study.

References:

1. Pozniak et al – Influence of M184V mutation on virological outcome of HAART with or without didanosine. Seville. Abstract 152. Antiviral Therapy 2002; 7:S124.
2. Eron et al – Antiretroviral activity of didanosine in lamivudine-experienced subjects in comparison to activity in subjects who were lamivudine-naive. Abstract 123. Antiviral Therapy 2002; 7:S102.
3. Winters et al – Clinical impact of the M184V mutation on switching to didanosine or maintaining lamivudine treatment in nucleoside experienced patients. Seville. Abstract 122. Antiviral Therapy 2002; 7:S101.