| |
Interruptions & Drug Resistance
|
| |
| |
Emergence of Minor Populations of Human Immunodeficiency Virus Type 1 Carrying the M184V and L90M Mutations in Subjects Undergoing Structured Treatment Interruptions
The Journal of Infectious Diseases
2003;188:1433-1443
Karin J. Metzner,1,3 Sebastian Bonhoeffer,4 Marek Fischer,5 Rose Karanicolas,2 Kristina Allers,3 Beda Joos,5 Rainer Weber,5 Bernard Hirschel,6 Leondios G. Kostrikis,7 Huldrych F. Günthard,5 and the Swiss HIV Cohort Study, 1Aaron Diamond AIDS Research
“...drug-resistant viruses were detected as minor populations of HIV-1 in the majority of subjects undergoing STI……. such variants can emerge during periods of increased HIV-1 replication, when drug concentrations may be suboptimal…… we suggest that STI should remain restricted to controlled clinical trials, to minimize the risk of the development of drug-resistant HIV-1 variants. The influence of minor populations of drug-resistant variants on long-term treatment is currently under investigation”.
1Aaron Diamond AIDS Research Center and 2College of Physicians and Surgeons, Columbia University, New York, New York; 3University of Erlangen-Nuremberg, Institute of Clinical and Molecular Virology, Erlangen, Germany; 4ETH, Swiss Federal Institute of Technology, Ecology & Evolution and 5University Hospital Zurich, Department of Medicine, Division of Infectious Diseases and Hospital Epide m iology, Zur i ch, and 6University Hospital of Geneva, Division of Infectious Diseases, Geneva, Switzerland; 7University of Athens School of Medicine, Des type 1 (HIV-1). However, the existence of a long-lived reservoir of latently infected CD4+ T lymphocytes makes it unlikely that HIV-1 can be eradicated by currently available antiretroviral drugs. Adverse effects and long-term toxicity, associated with prolonged use of these drugs, has become a critical issue when considering the beneficial effects of HAART.
Structured treatment interruption (STI) is an alternative therapeutic strategy for HIV-1 that is currently under investigation. STI is used as a means of boosting HIV-1 specific immune responses (edit note: this utility has mostly been discounted) , while, at the same time, reducing the adverse effects, toxicity, and cost associated with HAART. Encouraging results from the use of STI in subjects with acute HIV-1 infection, the majority of whom controlled viral replication after cessation of treatment, have been reported. The clinical benefits of STI in subjects with chronic HIV-1 infection remain unclear. The largest STI trial undertaken thus far, the Swiss-Spanish Intermittent Treatment Trial (SSITT), enrolled 133 subjects with chronic HIV-1 infection who underwent STI after successful HAART. The results of that trial did not show a relationship between control of HIV-1 viremia and induction of virus-specific cytotoxic T lymphocytes. However, SSITT demonstrated that STI is safe in subjects with chronic HIV-1 infection. The risk of selecting drug-resistant HIV-1 variants during STI is unknown and remains a concern. A recent report demonstrates selection of drug-resistant viruses in the context of STI: 2 of 12 subjects with chronic HIV-1 infection developed the M184V mutation, which is associated with reduced susceptibility to lamivudine.
Genotypic mutations associated with drug resistance are generally detected by direct sequencing or hybridization to allele-specific oligonucleotides. One drawback to these techniques is their inability to detect and quantify minor (<20%) populations of either the wild-type or mutant variants. This is mainly due to the failure of polymerase chain reaction (PCR) primers to discriminate, and subsequently to amplify, variants with a low molar ratio. In a previous study, we reported a new method for differential amplification of simian immunodeficiency virus species that is based on real-time PCR using molecular beacons and selective oligonucleotides. Subsequently, other groups have reported similar assays for the selective quantification of resistant HIV-1 viral sequences. In the present study, this technique was modified to detect 2 key HIV-1 drug-resistance mutations, L90M (protease) and M184V (reverse transcriptase [RT]). This assay allows simultaneous detection of both the wild-type and drug-resistant variants in the same sample, enabling direct quantification and analysis of changes in each virus population.
To evaluate the development of minor drug-resistant HIV-1 variants, we used this assay to assess subjects with chronic HIV-1 infection during STI. Our results indicate the presence of minor populations of HIV-1 with L90M and M184V mutations in subjects undergoing STI.
STUDY DESIGN & PATIENTS
Of the 29 subjects enrolled from the Zurich cohort of the SSITT, 28 participated in the present study. Of those, 15 participated in an extended protocol with frequent blood sampling. All subjects also participated in the Swiss HIV Cohort Study.
To be eligible for the SSITT, subjects had to be receiving their first combination antiretroviral therapy (ART) with 2 or 3 drugs (excluding nonnucleoside RT inhibitors) for >6 months and never have experienced therapy failure. Subjects who had received previous nonsuppressive treatment with 1 or 2 nucleoside RT inhibitors were excluded. Plasma viremia had to be <50 HIV-1 RNA copies/mL for at least 6 months, and CD4+ T cell counts had to be >300 cells/mL. The trial consisted of 4 cycles, each consisting of a 2-week treatment interruption followed by resumed treatment for 810 weeks. Subjects whose plasma viremia did not decrease to values of <50 HIV-1 RNA copies/mL of plasma during periods of resumed treatment were excluded from the study. After 4 STI cycles (week 40), treatment was stopped for at least 3 months, unless symptoms of acute HIV-1 infection occurred, CD4+ T cell counts decreased to <300 cells/mL, or virus load exceeded the following predetermined values: 3 consecutive measurements of >50,000, 2 of >100,000, or 1 of >500,000 HIV-1 RNA copies/mL of plasma. At week 52, treatment was resumed as described elsewhere. The SSITT was approved by the Ethics Committee of the University Hospital Zurich, and written, informed consent was obtained from each subject.
RESULTS
Quantification of minor populations of drug-resistant HIV-1 in subjects undergoing STI
A total of 28 subjects undergoing STI were enrolled in this study. All subjects had chronic HIV-1 infection and were successfully treated with HAART (virus load of <50 HIV-1 RNA copies/mL of plasma). At least 1 protease inhibitor was included in the treatment regimens of 27 subjects. All subjects were evaluated for HIV-1 carrying the L90M and M184V mutations. Subjects 102, 103, 109, and 116 had to be retreated, according to the protocol, between weeks 40 and 52. Subject 104 dropped out of the study because his virus load measurements were >50 HIV-1 RNA copies/mL of plasma after 10 weeks of retreatment. Analyses of viral and immunological responses in these subjects have been published elsewhere.
The assay for detection of the L90M mutation was performed on samples from 27 subjects. It was not possible to successfully amplify wild-type or mutant sequences from 2 of these subjects (129 and 130). These subjects were infected with HIV-1 subtypes A and C, respectively. The L90M mutation was detected as a minor variant in 3 of 25 subjects (subjects 102, 116, and 118). In subject 102, 5.6% of the total HIV-1 population carried the L90M mutation during the first cycle of STI. Interestingly, the L90M mutation was not detected in this subject's virus population during subsequent STI cycles. In subject 116, 0.3% of the virus population carried the L90M mutation during the third cycle of STI. Similarly, in subject 118, 0.06% of the virus population carried the L90M mutation at 1 time point during the fifth cycle of STI. No evidence was found for the presence of the L90M mutation in viruses from any other study subjects.
All 28 subjects were tested for the presence of the M184V mutation. We were unable to amplify either wild-type or M184V mutant sequences in samples from 3 subjects (103, 129, and 130). These subjects were infected with HIV-1 subtypes different from subtype B. Overall, drug-resistant HIV-1 variants carrying the M184V mutation were detected in 14 of 25 subjects. In 9 subjects, minority virus populations carrying the M184V mutation were detected at only 1 or 2 time points. In 7 subjects, this mutation was detected once, as a minority, during the fifth cycle of STI, with percentages ranging from 0.3% to 9.8% (average, 3.1%). The M184V mutation was not detected in subsequent samples from these subjects. Among viral sequences identified for subject 127, 5.4% carried the M184V mutation during the second STI cycle, but not during subsequent STI cycles. In subject 106, the M184V mutation represented 56.9% of the sequences detected during the beginning of the fifth STI cycle. This population decreased to 1% 2 weeks later and decreased to <0.3% after an additional 2 weeks. Lower virus loads in subsequent samples did not allow us to discriminate between wild-type and mutant variants below the range of 2.6%6.6%. Because of low virus loads during the first 4 STI cycles, only time points from the fifth STI cycle were measured. However, calculating the positive values of the M184V mutation relative to the virus load revealed that 56.9% corresponded to 782 HIV-1 RNA copies/mL of plasma. Because of a rapid increase in virus load in this subject (106), 1% corresponds to 451 HIV-1 RNA copies/mL of plasma. This indicates that the wild-type population was increasing during this period, whereas the M184V mutant population was slowly decreasing.
Five subjects showed diverse patterns, with regard to the appearance of the M184V mutation. Analysis of samples from subject 102 revealed a continuous decrease in viruses carrying the M184V mutation, during the first 3 cycles of STI (20.1%, 5.2%, and 4.9%, respectively). The M184V mutation was not detected within the range of detectable minor variants during subsequent STI cycles. This subject resumed ART (zidovudine, lamivudine, and indinavir) on day 329. The majority of viruses identified for subject 104 carried the M184V mutation at baseline, with 93.9% detected by day 8 of the first STI cycle. After reintroduction of HAART, the virus load decreased; however, the M184V mutation was detectable at levels of 60%-80%, up to day 70. This subject subsequently dropped out of the STI study because his virus load did not decrease to <50 HIV-1 RNA copies/mL of plasma after the first STI cycle. Viremia continued to decrease but remained detectable at low levels (<500 HIV-1 RNA copies/mL of plasma). Five months later, virus load rebounded to >1000 HIV-1 RNA copies/mL of plasma. Population sequencing revealed the M184V mutation and the protease inhibitorassociated mutations 46L, 54V, 63P, and 82A. A salvage regimen was started with abacavir, saquinavir, ritonavir (changed to lopinavir/ritonavir after 20 months), and stavudine. Efavirenz was added to the regimen after 1 month. Virus load decreased and remained undetectable in plasma for the duration of follow-up.
Viral variants carrying the M184V mutation were detectable as a minority population in samples obtained from subject 105 during the fifth STI cycle. The percentage of M184V mutants varied between 0.5% and 11% during this time but did not disappear, despite the fact that the subject was not receiving ART. Time points before the fifth STI cycle were not analyzed, because of low virus load. The first sample from subject 112 that was measured was obtained during the third STI cycle; samples of earlier STI cycles have not been used for differential amplification, because of low virus loads. Approximately 50% of the virus population from subject 112 carried the M184V mutation during the third STI cycle. During the fourth STI cycle, the M184V mutation was not detectable (<2.1%), but it appeared again during the fifth STI cycle, showing an increase over time. In subject 116, the M184V mutation was detectable as a minor variant during the first STI cycle (0.9%), the second STI cycle (0.3%), and the third STI cycle, at days 148 (5.9%) and 154 (0.8%). During the fourth STI cycle, the level of viruses carrying the M184V mutation reached 27.3%. HAART was reintroduced in this subject 33 days after the fifth STI.
All subjects had received their first ART before they were enrolled in SSITT. To determine whether minor populations carrying the L90M and M184V mutations were already present, before initiation of HAART, in our subjects' virus populations, we also tested plasma samples obtained from 24 of these subjects when they were still drug naivethat is, before any ART was started. No successful amplification of wild-type or mutant sequences was possible in samples from 3 (L90/M) and 2 (M184/V) subjects, respectively. In the remaining samples, no minor variants of viruses carrying the L90M or M184V mutations were detectable. With regard to the L90M mutation, frequencies were <0.03%<1% (median, <0.2%) in samples from 21 subjects. The M184V mutation was not detectable in 22 subjects, at frequencies of <0.2%9.1% (median, <0.5%).
DISCUSSION BY AUTHORS
The present study was undertaken to analyze the potential emergence of drug-resistant HIV-1 in subjects undergoing STI. Genotypic analysis of minor populations was performed by use of a novel, quantitative real-time PCR assay for differential amplification with selective oligonucleotides that detect 2 key resistance mutations: L90M (protease) and M184V (RT). This methodology is based on the ARMS. Gene-specific oligonucleotides contained a deoxyinosine at the -2 position of the 3 end that, because of destabilizing effects on the formation of duplexes, increased the discriminatory ability 510-fold (data not shown). In addition, since deoxyinosine is less tolerant of mismatches, variability in the second position, among different viral sequences, is less critical. This methodology is able to detect a low percentage of minor populations of L90 wild-type and L90M drug-resistant variants (0.01%), as well as M184 wild-type (0.1%) and M184V drug-resistant variants (0.2%).
A total of 28 subjects were enrolled in the present study; 181 plasma samples from 25 subjects were analyzed for the presence of the L90M mutation. This mutation was detected as a minor population in 3 samples from 3 of 25 subjects. A total of 216 plasma samples from 25 subjects were analyzed for the presence of the M184V mutation: 40 samples from 14 subjects were positive for the M184V mutation. The predominance of the M184V mutation, compared with the L90M mutation, was not unexpected. The selection of mutations associated with drug resistance to protease inhibitors is a slow, stepwise process, compared with the emergence of M184V, which is a single-point mutation selected by lamivudine. Thus, the more frequent detection of the M184V mutation, compared with the L90M mutation, might represent a very early event in the emergence of drug resistance in our subjects undergoing several periods of significant HIV-1 replication.
Sporadic occurrence of the L90M mutation was detected in 3 subjects during STI. This mutation appeared only once in each subject, during different STI cycles. The percentage of viruses carrying the L90M mutation was <0.3% in subjects 116 and 118. We cannot rule out the possibility that a minor population of L90M variants within this range may be present at other time points; however, for most of the other samples, low virus loads did not allow us to discriminate <1%. In 9 subjects, the M184V mutation was detected during the fifth STI cycle in only 1 or 2 samples. It is possible that M184V was also selected during prior STI cycles, but, because of lower virus loads during these periods, the M184V mutation was not detectable. On the other hand, more samples were suitable for testing during the last (fifth) STI cycle, increasing the probability of detecting mutations, because of the occurrence of a full rebound in virus load, compared with the shorter, prior STI cycles. Interestingly, 2 of these subjects were treatment naive with respect to lamivudine. Instead, their regimen contained didanosine. That virus strains harboring the M184V mutation show reduced sensitivity against didanosine in vitro may potentially lead to selection of minor populations carrying the M184V mutation in subjects receiving didanosine treatment.
Baseline samples (from the first STI cycle, if available) from 3 subjects were positive for the M184V mutation. This mutation was not detected in subsequent samples obtained from subject 102 during successive STI cycles. In subject 116, the frequency of variants carrying M184V increased in subsequent STI cycles. Drug-specific pharmacokinetic factors were excluded as an explanation for the divergent outcomes, because subjects 102 and 116 each received the same drugs. However, we cannot rule out the possibility that other pharmacokinetic factors, such as differences in drug absorption, led to concentrations that favored the outgrowth of the M184V mutation in this subject. Of the 25 subjects studied, only subject 104 developed drug-resistance mutations associated with treatment failure. On day 8 of the first STI cycle, the majority (>93%) of viruses from this subject already harbored the M184V mutation. This finding was confirmed by direct sequencing (L. Perrin, personal communication). The M184V mutation was found, in addition to multiple mutations associated with resistance to protease inhibitors. The subject dropped out of the study with a virus load of >50 HIV-1 RNA copies/mL of plasma after the first STI cycle. A salvage regimen was subsequently introduced with successful suppression of plasma viremia to <50 HIV-1 RNA copies/mL for >27 months.
There is one potential caveat to our study: the very high sensitivity of our assay to detect the 2 mutations at very low frequencies, in theory, may have lead to detection of randomly occurring minor populations present in subjects with chronic infection, independent of treatment and, in particular, STI. However, that the rapidly emerging single-point mutation M184V was detected in 14 of 25 subjects, compared with only 3 of 25 subjects harboring the slow-emerging L90M mutation (P = .02, Fisher's exact test), suggests that our results are causally linked to treatment and STI. The finding of a higher frequency of the M184V mutation, compared with the L90M mutation, is even more prominent if one considers the 20-times higher sensitivity of the L90M assay. Furthermore, both mutations have not been detected in plasma samples obtained from these subjects at time points before they received their first ART.
In conclusion, drug-resistant viruses were detected as minor populations of HIV-1 in the majority of subjects undergoing STI. It is not yet known whether a similar distribution occurs in subjects who stop successful therapy for longer periods without consistent reintroduction of HAART. However, our data suggest that such variants can emerge during periods of increased HIV-1 replication, when drug concentrations may be suboptimal. We observed tremendous variation in the appearance and disappearance, as well as in the timing of the emergence, of these mutations. On the basis of these observations, we suggest that STI should remain restricted to controlled clinical trials, to minimize the risk of the development of drug-resistant HIV-1 variants. The influence of minor populations of drug-resistant variants on long-term treatment is currently under investigation.
|
|
| |
| |
|
|
|
