Resistance testing methodologies and mechanisms of resistance.
RESISTANCE TESTING TECHNOLOGIES
Antiretroviral drug resistance testing is a valuable tool in the management of HIV-infected persons. However, there are limitations to current routine resistance assays. In particular it has become increasingly evident that standard tests cannot reliably detect resistant variants present within the viral quasispecies at prevalence below 20-25%. Limited data from persons receiving NNRTI-based regiments indicate that low-frequency resistant mutants can affect responses to antiretroviral treatment. It can be proposed, however, that the impact of low-frequency mutants depends on multiple factors, including the frequency within the quasispecies, the effects of the mutation on drug susceptibility and virus fitness, the specific drugs and drug-classes affected within a given regimen, and the patient's level of adherence.
Several new technologies allow the detection of low-frequency mutants with high sensitivity (Table 1).
It is important to keep in mind that the high degree of variability of the HIV genome dictates that point mutations conferring drug resistance emerge spontaneously during virus replication, leading to a natural background of drug resistant mutants. Thus, ultrasensitive resistance assays must be able to reliably and reproducibly detect resistance that matters, while avoiding the detection of the natural background. The current consensus is that the optimal sensitivity is around 0.1-1%, a threshold above which resistant mutants may negatively affect treatment responses, at least in the context of NNRTI-based therapy.
As an example of the limitations of routine genotypic testing, Table 2 illustrates the results of clonal sequencing of plasma viral RNA from a patient who underwent the unstructured discontinuation of efavirenz-based therapy after achieving sustained virological suppression <50 copies/ml. At the time of virological rebound, while population sequences obtained by routine genotypic resistance testing showed a wild-type virus, analysis of 200 clones derived from five separate PCR reactions showed NNRTI- as well as NRTI-resistant mutants at frequencies between 0.5% and 2%.
It is currently debated how easily the more sensitive assays can be introduced into routine diagnostic practice. Available methodologies are either extremely labour intensive (e.g. clonal analysis/single genome sequencing), complex to set up for the variety of possible resistance mutations and subtypes (e.g. allele-specific PCR), or very expensive (e.g. ultra-deep sequencing).
Among some exciting new methodologies being developed, pyrosequencing combined with DNA barcoding is an elegant approach, likely to have broad applications in the future . This method combines DNA barcoding and massively parallel pyrosequencing to achieve full sequencing of the genomic regions of interest and a sensitivity of detection of 5%. Multiple samples (seven or more) are tested grouped together in a single reaction and later identified through unique sequence tags. This approach allows multiplex analysis of many barcoded samples in a single sequencing experiment, thereby offering the potential to drive down the cost of each genotype determination.
RESISTANCE TO SMCCR5 INHIBITORS
Recent studies have identified two different mechanisms of resistance to small molecule CCR5 (smCCR5) inhibitors such as maraviroc.
In the first mechanism, changes in the virus envelope allow the virus to bind to the co-receptor despite the alteration in shape caused by the presence of the inhibitor . The phenotypic effect is shown by a plateau in the maximal percentage inhibition (MPI) with increasing drug concentrations. As a result of mutations in the stems of the V3 loop, resistant virus binds with higher affinity to the drug-bound co-receptor than wild-type virus. Mutations at codons 13 and 26 appear to be especially common in resistant strains, but changes outside the V3 loop may also be important. To date the mutation pattern remains to be clearly defined, making the genotypic identification of resistance problematic.
A second mechanism of resistance is characterised by a switch in viral tropism from R5 to X4. In most cases the switch is thought to occur within the viral population rather than at the individual virus level. One study examined 10 patients who experienced virological failure in clinical trials of maraviroc . The patients had R5 virus at study entry and X4 or dual tropic/mixed (DM) virus at failure, as determined by Monogram Bioscience's Trofile assay. The assay has a reported sensitivity of 10% for the detection of X4 virus. Using clonal analyses of the envelope gene, the investigators demonstrated that in seven patients the X4/DM virus detected at failure was phylogenetically related to X4/DM virus present at low frequency (up to 7%) in the baseline sample (Figure 1). These data provide support to the concept that tropism shift occurs at the virus population level, when selective suppression of R5 virus allows outgrowth of the X4/dual tropic population.
[FIGURE 1 OMITTED]
One issue of significant current interest is related to the ideal sensitivity of tropism assays for the detection of X4 virus. Reflecting these concerns, the Trofile assay has been recently modified to achieve increased sensitivity. However, in analogy with the issues discussed in the context of low-frequency drug-resistance species, the threshold prevalence of X4 or dual tropic virus that impacts on responses to smCCR5 inhibitors remains to be clearly established. Virus strains using the CXCR4 co-receptor are likely to be present in most patients with established HIV infection even if at low or very low frequency and an assay with too great a sensitivity may, paradoxically, lead to unnecessary avoidance of smCCR5 inhibitors. The threshold above which responses are affected is likely to depend on multiple determinants, including the activity of other drugs in the regimen. Although data are available to demonstrate the pre-existence of low-frequency X4/DM virus in patients subsequently failing maraviroc-based therapy, there are no published data on the prevalence of such low-frequency variants among patients who achieved virological suppression. While in many circumstances it may be sufficient to use a tropism test with good rather than 'perfect' sensitivity to determine whether the dominant quasi-species is R5, it will always remain an important principle of antiretroviral therapy to ensure that the smCCR5 inhibitor is well supported by other active drugs in the regimen, and that prolonged virological failure is avoided.
Reassuringly, limited data indicate that after smCCR5 discontinuation, resistant virus reverts to minority species and in cases of tropism switch, R5 virus again becomes the dominant quasispecies. These observations suggest that the viral population selected during therapy appears to have no inherent advantage over the pre-treatment population in the absence of drug pressure. The long-term impact of the possible enrichment of resistant variants or X4 and dual tropic virus remains unknown at present.
PATHWAYS OF TENOFOVIR RESISTANCE
The K65R mutation in reverse transcriptase confers resistance to tenofovir and intermediate levels of cross-resistance to abacavir, didanosine, and lamivudine, and also has resistance effects for stavudine (not evident in phenotypic testing), while conferring hypersusceptibility to zidovudine. There has been some recent evidence showing a greater propensity for HIV-1 subtype C to develop the K65R mutation under in vitro selection with tenofovir compared with subtype B or other non-B subtypes . Whether this is a true subtype effect remains to be clearly shown. A single nucleotide change is required for a K [right arrow] R substitution in both subtype B (AAA to AGA) and subtype C (AAG to AGG).
A mechanistic explanation has been recently proposed, and is related to the way the sequence of the virus affects the points at which the reverse transcriptase stops or pauses before continuing its activity [5,6]. When HIV replicates, the viral RNA template is used by the reverse transcriptase to produce a DNA copy. This process is characterised by disruptions or dissociations of the enzyme and there are indications that the places at which these occur are the points at which mutations can also occur. It has been proposed that the sequence of the template influences the likelihood of the reverse transcriptase stopping, pausing or dissociating from the template. In particular, it has been suggested that subtype C sequences favour pausing at or near position 65, which will then favour the evolution of K65R. This is interesting because it reveals how the sequence variability associated with subtypes may influence the development of resistance. The clinical significance of these findings, however, remains to be determined. Currently available, albeit limited, clinical data from tenofovir-treated patients do not suggest a higher rate of K65R development among patients with subtype C . Prospective studies of patients infected with different HIV-1 subtypes are required to determine whether they show a different propensity for the development of K65R.
A less common pathway of tenofovir resistance is characterised by the emergence of the K70E mutation, which shows resistance effects overall similar to those observed with K65R. It had been previously proposed that the K70E mutation may appear transiently prior to the emergence of K65R. Recent data, however, indicate that under continuous virological pressure with tenofovir the K70E mutation evolves into the rarer mutation K70G . The change from K70E to K70G is a two-step process in which the wild-type lysine at position 70 changes to glutamic acid and then to glycine (K[AAA] [right arrow E[GAA] [right arrow] G[GGA]). This unusual resistance pattern emerged in an HIV-infected individual receiving dual antiretroviral therapy with tenofovir/emtricitabine due to poor compliance with triple therapy. Ongoing virus replication led to the initial emergence of K70E with the emtricitabine resistance mutation M184V, followed by the loss of K70E and emergence of K70G. Clonal analysis showed that K70G and M184V co-existed on the same viral genome. The phenotypic effects of the K70G mutation in isolation or with M184V had not been previously described. Using the Antivirogram assay (Virco, Belgium) phenotypic resistance to lamivudine, emtricitabine, abacavir, didanosine and tenofovir was observed, with preserved susceptibility to zidovudine and stavudine (Table 3).
RESISTANCE TO ETRAVIRINE (TMC125)
The DUET 1 and 2 studies randomised treatment-experienced patients (n = 600 per trial) with viral load >5000 copies/ml while on stable antiretroviral therapy for at least 8 weeks to receive TMC125 or placebo, each combined with ritonavir-boosted darunavir (DRV/r) and an optimised NRTI backbone, with optional use of enfuvirtide (T20). Patients had [greater than or equal to] 1 NNRTI resistance mutation (either at screening or in documented historical genotype) and [greater than or equal to] 3 primary protease mutations. Pooled data from the two studies were analysed to identify NNRTI mutations significantly associated with reduced responses to the TMC125 arm .
The genotypic score included both mutations with well-known effects on NNRTI susceptibility such as Y181C and others with unknown effects such as V90I. Overall, 13 mutations showed a highly statistically significant correlation with responses in the multivariate analysis, after adjusting for baseline parameters including phenotypic susceptibility to DRV (Table 4).
In general, as the number of mutations increased, the response to TMC125 was reduced, with the greatest loss of responses seen in the presence of three or more mutations. Given the prominent role of mutations at codon 181 in conferring resistance to TMC125, it can be concluded that a history of failure on nevirapine, particularly if protracted, may be considered predictive of a reduced response to TMC125.
These data are extremely interesting as they indicate that NNRTI sequencing may now be proposed as a valid treatment strategy. To test the hypothesis fully and evaluate the impact of low-frequency NNRTI resistant mutants, clinical data are eagerly awaited on the use of second generation NNRTIs (TMC125 or TMC278) in combination with NRTIs (without ritonavir-boosted PIs or other drug classes) in patients who failed standard first-line efavirenz or nevirapine therapy.
PATHWAYS OF LOPINAVIR RESISTANCE
Studies have reported that the L76V mutation in protease is selected by lopinavir (LPV) use in first-line therapy. The MONARK study randomised treatment-naive patients to receive lopinavir/ritonavir (LPV/r) either alone or with a backbone of zidovudine/lamivudine. Low-level viraemia (>50 but < 400 copies/ml) occurred more commonly in the monotherapy arm than the triple therapy arm over 48 weeks. During the study period, 32 (38.5%) and 7 (13.2%) patients, respectively, underwent genotypic resistance testing, which was initiated at the clinician's discretion when viral load was >500 copies/ml . Overall five patients in the monotherapy arm and none in the triple therapy arm developed major PI resistance mutations. The resistance analysis in the five patients is summarised in Table 5. Three of the five patients showed the emergence of L76V, either alone or with M46I.
The phenotypic effects of the L76V + M46I mutation pattern on PI susceptibility have also been investigated by site-directed mutagenesis . Overall intermediate levels of resistance to LPV and amprenavir were observed (Table 6).
While its phenotypic effect appears to be modest, the L76V mutation is part of the darunavir score and may therefore reduce the genetic barrier to the drug. Although the mutation has been observed in patients infected with subtype CRFAG, it remains to be demonstrated whether there exists a subtype-related effect on its emergence.
[1.] Hoffmann C, Minkah N, Leipzig J et al. DNA bar coding and pyro-sequencing to identify rare HIV drug resistance mutations. Nucleic Acids Res, 2007, 35, 91.
[2.] Mori J, Mosley M, Lewis M et al. Characterization of maraviroc resistance in patients failing treatment with CCR5-tropic virus in MOTIVATE 1 and MOTIVATE 2. XVI International HIV Drug Resistance Workshop, June 12-16, 2007, Barbados. Abstract 10.
[3.] Lewis M, Simpson P, Fransen S et al. CXCR4-using virus detected in patients receiving maraviroc in the phase III studies MOTIVATE 1 and 2 originates from a pre-existing minority of CXCR4-using virus. XVI International HIV Drug Resistance Workshop, 12-16 June, 2007, Barbados. Abstract 56.
[4.] Brenner BG, Oliveira M, Doualla-Bell F et al. HIV-1 subtype C viruses rapidly develop K65R resistance to tenofovir in cell culture. AIDS, 2006, 20, F9-F13.
[5.] Harrigan PR, Chihwei Sheen T, Wynhoven B et al. Silent mutations at RT codons 65 and 66 are very strongly associated with treatment and drug resistance and are implicated in reverse transcription efficiency. XVI International HIV Drug Resistance Workshop, 12-16 June, 2007, Barbados. Abstract 113.
[6.] Coutsinos D, Invernizzi CF, Xu H et al. In vitro molecular characterization of the development of the K65R and M184V mutations in subtype C HIV-1. XVI International HIV Drug Resistance Workshop, 12-16 June, 2007, Barbados. Abstract 114.
[7.] Miller MD, Margot N, McColl D, Cheng AK. K65R development among subtype C HIV-1-infected patients in tenofovir DF clinical trials. AIDS, 2007, 21, 265-266.
[8.] Bradshaw D, Malik S, Booth C et al. Novel drug resistance pattern associated with the mutations K70G and M184V in Human Immunodeficiency Virus type 1 reverse transcriptase. Antimicrob Agents Chemother, 2007, 51, 4489-4491.
[9.] Vingerhoets J, Buelens A, Peeters M et al. Impact of baseline NNRTI mutations on the virological response to TMC125 in the phase III clinical trials DUET-1 and DUET-2. XVI International HIV Drug Resistance Workshop, 12-16 June, 2007, Barbados. Abstract 32.
[10.] Delaugerre C, Flandre P, Chaix ML et al. Protease gene mutations in a trial comparing first-line lopinavir/ritonavir monotherapy to lopinavir/ritonavir + zidovudine/lamivudine (MONARK Trial). XVI International HIV Drug Resistance Workshop, 12-16 June, 2007, Barbados. Abstract 75.
[11.] Nijhuis M, Wensing AMJ, Bierman W et al. A novel genetic pathway involving L76V and M46I leading to lopinavir/ritonavir. XVI International HIV Drug Resistance Workshop; 12-16 June, 2007, Barbados. Abstract 127.
Dr Anna Maria Geretti, MD, PhD, FRCPath, Department of Virology, Royal Free Hospital and Royal Free & University College Medical School, Pond Street, London NE3 2QG, UK. Email: email@example.com
Table 1: Ultrasensitive methods for the detection of drug-resistant mutants. Allele-specific real-time PCR Parallel allele-specific sequencing (PASS) Pyrosequencing Ultra-deep sequencing Clonal analysis Single genome sequencing Table 2: Clonal analysis (n=200 clones) of plasma HIV RNA (reverse transcriptase codons 1-250) from a treatment-experienced patient lacking resistance by routine genotyping. Mutation detected No of clones (%) Drugs affected A62V 1 (0.5) Possibly NRTIs V75A 1 (0.5) Stavudine K103N 4 (2.0) Efavirenz, nevirapine V106A 1 (0.5) Efavirenz, nevirapine Y115C 2 (1.0) Unknown Q151R 1 (0.5) Unknown V179G 1 (0.5) Possibly NNRTIs M184T 1 (0.5) Lamivudine Y188D 1 (0.5) NNRTIs Table 3: Phenotypic resistance (fold-changes in [IC.sub.50]) of HIV-1 clinical isolates with the reverse transcriptase mutations K70E+M184V, K70G+M184V or K70G alone. (Adapted from reference 8.) Drug K70E+M184V K70G+M184V K70 BCO (a) G Lamivudine >29.3 >84.5 4.0 2.4 Emtricitabine >40.8 >63.4 NA 3.5 Zidovudine <0.3 <0.8 1.0 2.7 Stavudine 0.4 0.9 1.3 2.3 Didanosine 1.6 5.3 2.1 2.2 Abacavir 1.3 3.2 2.1 2.2 Tenofovir 1.7 1.5 1.9 2.1 Drug VircoType CCO (b) Lower Upper Lamivudine 1.0 3.4 Emtricitabine NA NA Zidovudine 1.2 9.6 Stavudine 0.9 2.0 Didanosine 0.9 2.6 Abacavir 0.8 1.9 Tenofovir 0.9 2.1 (a) Antivirogram biological cut-off; (b) VircoTYPE HIV-1 lower and upper clinical cut-off; NA, not available. Table 4: Mutations associated with reduced responses to etravirine in the DUET 1 and 2 studies. V90I A98G L100I K101E/P V106I V179D/F Y181C/I/V G190A/S Table 5: Resistance analysis of monotherapy arm of the MONARK trial. (Adapted from reference 10.) Viral load ([log. Study sub.10] Patient Subtype visit copies/ml) 1 B Baseline 4.8 week 40 2.9 2 CRFAG Baseline 4.8 week 44 2.8 3 CRFAG Baseline 5.0 week 62 2.6 4 B Baseline 4.4 week 76 3.1 5 CRFAG Baseline 4.6 week 90 2.5 Lopinavir fold Patient Protease genotype (a) change 1 L63P/S V771 1.2 M46I L63P V77I 1.4 2 I13V K20I M36I H69K 0.6 I13V K20I M36I H69K L76V 1.3 3 I13V K20I M36I H69K 0.6 I13V K20I M36I M46I H69K L76V 2.7 4 L10I/L L63P A71A/T 1.5 L10F V82A/V L63P 1.1 5 I13V K20I M36I H69K L89M 0.9 I13V K20I M36I H69K L76V L89M NA (a) Emerging mutations in bold; NA, not available. Table 6: Phenotypic susceptibility to protease inhibitors of M46I+L76V site-directed mutants. Fold-change Mutations Lopinavir Atazanavir Saquinavir M46I + L76V 12 0 0 Fold-change Mutations Tipranavir Amprenavir Darunavir M46I + L76V 1 5 1