icon-folder.gif   Conference Reports for NATAP  
  10th Conference on Retroviruses and Opportunistic Infections
Boston, Mass, Feb 10-14, 2003
Back grey_arrow_rt.gif
Review of Resistance at Retrovirus Conference
Andrew R. Zolopa, MD Stanford University
--Nucleoside Resistance: K65R (tenofovir) versus TAM's (nukes: AZT, d4T, etc)
--Protease Inhibitor-Hypersusceptibility: Prevalence and Clinical Implications
-Atazanavir and the I50L mutation: Implications for Treatment Strategies
-lopinavir/r; Fos-amprenavir/r (no PI resistance observed)
--Clinical cut-off for indinavir/r (800/200)
--Tipranavir efficacy and resistance
--T-20 resistance and T-1249
--Genotype vs Virtual Phenotype (CREST Study)
Antiretroviral drug resistance continues to be a major challenge in the treatment of HIV infection. At this year's CROI, there were a number of interesting abstracts related to drug resistance, including information on resistance to the new class of antiretrovirals - the entry inhibitors. This report will summarize the more clinically relevant of resistance abstracts. As organized at the conference we will cover resistance to reverse transcriptase inhibitors, protease inhibitors and entry inhibitors separately. In addition, we will cover resistance testing studies and viral fitness in this report.
Nucleoside Resistance: K65R versus TAM's
With the availability of tenofovir (TDF) there has been increased attention and concern raised about the K65R mutation- a signature mutation for TDF and also selected by abacavir (ABC), didanosine (DDI) and DDC. Therefore, K65R has the potential of causing cross-resistance within the nucleoside class and the concern raised has been related to the increasing use of TDF and ABC causing increased prevalence of 65R and thereby limiting treatment options.
To date K65R has been a relatively rare mutation, seen in <3% of treatment -experienced patients. In contrast thymidine-analog mutations (aka TAM's) which are selected by zidovudine (AZT) and stavudine (D4T) are seen more commonly - one to two thirds of treatment experienced patients who have had resistance testing performed possess these mutations. Moreover, as TAM's accumulate they result in broad cross-resistance within the nRTI class. With this as background, there were a number of abstracts at this year's meeting that focused on the relative impact of TAM's versus 65R on potential treatment options, viral fitness and the changing prevalence of these mutations.
Lanier and colleagues at GSK using data from the LabCorp genotype database demonstrated slight increases in the prevalence of K65R and slight decreases in TAM's from 1999 to 2002. [abstract 635] This database is comprised of nearly 38,000 clinical specimens sent for genotype testing during these years and therefore is not necessarily representative of all patients with resistance but is probably fairly representative of patients that have had resistance testing preformed during that period. Another limitation of this database is the treatment histories for these patients are not known. Nonetheless, the authors showed that K65R increased from 0.64% in 1999 to 1.69% in 2002 while TAM's decreased, for example genotypes with 41L, 210W and 215Y decreased from 17% to 11%. The authors also showed an increase in Y115F, a mutation selected for by abacavir. These trends seem to be associated temporally with increase use of abacavir and tenofovir based on prescription rates. At this time it appears that the prevalence of 65R is still quite low but may change with secular trends in the use of nucleoside/tide RT inhibitors.
In another poster presentation, Lanier evaluated the phenotypic susceptibility patterns associated with various nRTI-associated resistance mutations. [586] The investigators identified 22 clinical isolates from the ViroLogic database with K65R without other major mutations and demonstrated reduced susceptibility to DDI, DDC, 3TC as well as TDF while these isolates demonstrated hypersusceptibility to AZT. Furthermore, in 7 samples that contained both 65R and 184V there was reduced susceptibility to abacavir as well. Unfortunately, we can't draw any strong conclusions from this analysis regarding the clinical impact of K65R on the virologic response to these antiretrovirals given the lack of virologic response data.
Miller from Gilead, analyzed the replicative capacity or "fitness" of 65R-containing clinical isolates from ViroLogic's specimen bank. [616] The investigators demonstrated decreased fitness of 65R-containing specimens that was similar in magnitude to that seen with 184V - that is 56% of control. Furthermore, the fitness impact of 184V and 65R were additive such that isolates with both mutations had a fitness that was reduced to 29% of control. In contrast, the investigators showed that TAM mutations had only a minimal impact on viral fitness. These differences in viral fitness may explain, in part, the observed differences in prevalence of 65R versus TAM's. In the ViroLogic database, for example, which was used in this study, included over 1800 clinical specimens, K65R was present in less than 1% of specimens while 49% of specimens contained TAM's.
With the availability of tenofovir, clinicians now have new options in structuring the "nuc backbone" of ARV regimens. The studies summarized here, begin to put into place the variables that will help to determine the optimal strategies for using this class of antivirals in clinical practice. However, these studies are not sufficient in and of themselves to make any definitive statements about how the nucleoside/tide RTI's should best be used. In the end, relative prevalence of mutations, the rate at which they develop in failing regimens, as well as relative degree of cross-resistance they confer and their impact on viral fitness will all play into decisions about how best to utililize this class of antiretrovirals. However, further study is required -in particular additional studies of "TAM-sparing" approaches to ARV therapy.
Protease inhibitor Resistance: Hypersusceptibility and Impact of PI-Boosting on resistance
Protease Inhibitor-Hypersusceptibility: Prevalence and Clinical Implications

Hypersusceptibility is an in vitro phenotypic measurement that was first observed with the NNRTI's and reported by investigators at ViroLogic. Hypersusceptibility is defined by a patient isolate that requires lower drug concentrations to inhibit growth compared to the control virus, that is the fold change is less than one (strictly defined as FC <0.4). Several studies have now demonstrated that virologic response to an NNRTI-containing regimen is improved in patients who have NNRTI hypersusceptable virus compared to those with susceptible virus [Shulman AIDS 2001, Haubrich AIDS 2003, Mellors 9th CROI 2002].
Hypersusceptibility to protease inhibitors has been previously reported although it's prevalence and clinical implications were unknown. Andrew Leigh Brown and colleagues demonstrated that in a cohort of newly infected individuals, as many as 20% have hypersusceptibility to amprenavir and/or ritonavir.[594] These isolates also appear to have reduced replicative capacity by the PhenoSense assay. However, the question remains whether PI-hypersusceptibility will result in improved virologic response to PI-based ARV as appears to be the case for NNRTI-related hypersusceptibility.
Chip Schooley on behalf of the GSK sponsored study team-ESS40006, presented an oral abstract that preliminarily seems to show hyper-susceptibility to amprenavir (APV) resulting in an improved virologic response to APV/r-boosted ARV. [143] The study enrolled patients who experienced virologic failure on a PI-containing regimen for at least 12 weeks. The analysis was restricted to those patients who were NNRTI-experienced and received a regimen of APV/r (either 600mg/100mg or 900mg/100mg BID) along with ABC, TDF and one other nRTI based on baseline phenotypic susceptibility. In multivariate analysis APV hypersusceptibility (defined as FC<0.66 - which was median value for APV FC in the baseline isolates) and baseline viral load were the two independent predictors of virologic response at 24 weeks. The odds ratio for a viral load of <200 copies at 24 weeks was 7.24 for APV hypersusceptibility (p=0.015). If this result is confirmed with further analysis then this would mean that like NNRTI- hypersusceptibility, PI- hypersusceptibility could potentially be strategically utilized to improve response PI-based regimens.
Atazanavir and the I50L mutation: Implications for Treatment Strategies
Atazanavir (ATV) is the new once a day protease inhibitor available through BMS's expanded access program. Colonno and colleagues from BMS demonstrated that a unique mutation -I50L is associated with virologic failure in patients receiving ATV as their first PI-containing regimen. [597] This mutation resulted in a median fold change increase of 8.8 for reduced susceptibility to ATV but the striking finding in this study was that all the specimens that contained the 50L mutation either maintained full susceptibility or in many cases hypersusceptibility to the other 6 approved protease inhibitors. It appears that the 50L is the signature mutation for ATV and that it results in selective resistance to ATV while increasing susceptibility to the other PI's. The clinical implications of these in vitro observations have yet to be determined but these findings support ATV as a first line PI - that in addition to once a day dosing, a favorable lipid profile there may also be advantages from a resistance point of view.
Resistance Profiles in Boosted PI ARV Regimens
There were two poster presentations that demonstrated less drug resistance in patients experiencing viral rebound on boosted PI regimens compared to un-boosted PI regimens. In the first of these studies, Kempf and colleagues from Abbott, updated the resistance analysis from their study M98-863, the head to head comparison of Lopinavir/ritonavir (LPV/r) versus nelfinavir- both use in combination with D4T and 3TC in treatment naive patients. [600] Through 96 weeks the researchers found a cumulative probability of resistance to 3TC in the NFV arm of 29% among all treated patients compared to 7% in the LPV/r arm. Major resistance mutations to NFV were seen in 20% compared to no major protease resistance mutations in the LPV/r, although some secondary PI-related resistance mutations were seen. TAM mutations related to D4T resistance developed in 5% of the NFV-treated patients compared to 0 in the LPV/r group.
In the second study, investigators from GSK compared development of resistance across two separate studies of the amprenavir prodrug - Fos-APV or 908. [598] In the NEAT study 908 was dosed at 1400mg BID without ritonavir boosting and in the SOLO study, 908 was boosted with ritonavir (1400mg/200mg qD). Both studies used ABC and 3TC in addition to fos-APV and both studies enrolled ART-naive subjects and were compared to nelfinavir (NFV), ABC and 3TC. There was no difference in the incidence of resistance to 3TC or PI-related resistance between the 908 arm and NFV arm in those patients experiencing virologic failure in the NEAT study. By contrast, in the SOLO trial there was significantly less resistance to 3TC in the 908/r arm compared to the NFV arm, 13% vs. 56% respectively and less resistance to the PI as well (0% vs. 50%). No selection of resistance by 908/r was observed in virus from 31 (0/31) subjects analyzed.
These two posters combined suggest that even in the setting of virologic failure, boosted PI regimens appear to result in less resistance to all components of the failing regimen. These results have to be weighed against the ability to obtain a virologic response after failure of a NFV-based regimen or perhaps a ATV-based regimen in which the PI-resistance profile is likely to be responsive to a boosted PI regimen.
Clinical Cutoff for IDV/r established
Investigators from Merck in collaboration with ViroLogic and others defined the so called clinical cutoff for Indinavir boosted regimens. Using data from several clinical cohorts and a clinical trial, patients receiving IDV/r 800/200 BID were evaluated for virologic response. [603] Szumiloski showed that a FC of >10 for IDV was associated with reduced virologic response. Based on this analysis, it is likely that the cutoff for IDV/r will be revised in the PhenoSense assay.
Tipranavir response in PI-resistant patients
Tipranavir is a novel non-peptidic protease inhibitor that holds promise as a PI with activity against PI-resistant HIV. In Study BI 1182.52, multiple PI-experienced patients were treated with three different doses of TPV boosted with ritonavir [ 500mg/100mg, 500/200 and 750/200]. [179] In this study, patients on failing regimens continued their nRTI's and switched the PI in the failing regimen with one of the 3 doses TPV/r for 14 days then the nRTIs and/or NNRTI's were changed based on the resistance profile taken at study entry. Despite the presence of multiple PI-related mutations at baseline TPV phenotype FC was only slightly increased compared to wildtype, only 12% had a FC >4. The switch to TPV/r resulted in about a 1-log decline in viral load at day 14 in all three arms with slightly better viral load response in the highest dose group. The response seemed to hold up even in the presence of multiple PI-related mutations.
In an analysis of baseline phenotypic susceptibility, virologic response to TPV/r seemed to fall off in the groups with baseline TPV FC of greater than 2. [596] However, nearly 70% of subjects in this trial had FC of < 2. Therefore, it appears that a large number of PI-related mutations are required to push TPV phenotype FC above 2 which appears to be the clinical cutoff for TPV.
Entry Inhibitor Resistance
Entry inhibitors are the newest class of antiretrovirals to have widespread use in the clinic. Enfuvirtide (ENF) formerly T-20 is expected to be approved by the FDA in the near term. Greenberg from Trimeris presented results from the TORO trials evaluating the impact of baseline phenotypic susceptibility to ENF and virologic response. [141] The investigators using a phenotype assay recently developed by ViroLogic, demonstrated a wide variation in baseline susceptibility to ENF - IC50's ranged from 0.007 to 7.526 ug/ml. Despite this wide variation in baseline susceptibility it did not appear to influence virologic response to ENF, nor did cell trophism or viral clade seem to influence response in these trials.
Resistance to ENF is known to occur and mutations seem to cluster in the amino acids 36-45 of gp41. Mutations in this region are associated with virologic failure in patients receiving ENF-containing regimens and in this analysis were associated with a 21 fold decreased susceptibility to ENF. In the TORO trials, 99% of isolates with a FC >4 had one or more mutations in this region of gp41. Of note, Miralles of Trimeris presented an interim analysis of T1249 in patients who experienced virologic failure on a T20-containing regimen and showed median viral load declines of 1 log by day 11. Response was better in those who had been failing T20 for less than 48 weeks compared to those failing T20 for >48 weeks. [14lb] This study suggests that these two fusion inhibitors may be able to sequenced if resistance develops to ENF.
Clinical Trials of Resistance testing
Virtual Phenotype provides no additional benefit beyond a Genotype with Expert Interpretation
CREST is an Australian study in which patients on failing ARV were randomly assigned to have a new ARV regimen guided by a genotype with a standard rules-based interpretation versus genotype with interpretation plus the Virtual Phenotype. [592] There was no difference in viral load response over 48 weeks between the two arms. It remains unclear what role the Virtual phenotype test will play in the management of HIV infection. Previous studies have suggested that the virtual phenotype is as good as a real phenotype (VIRCO), yet this study seems to indicate that there is no added value to the virtual phenotype when a genotype with standard interpretation is used.