Search

Objectives: With the goal of facilitating the use of HIV-TRePS to optimize therapy in settings with limited healthcare resources, we aimed to develop computational models to predict treatment responses accurately in the absence of commonly used baseline data., Methods: Twelve sets of random forest models were trained using very large, global datasets to predict either the probability of virological response (classifier models) or the absolute change in viral load in response to a new regimen (absolute models) following virological failure. Two 'standard' models were developed with all baseline variables present and 10 others developed without HIV genotype, time on therapy, CD4 count or any combination of the above., Results: The standard classifier models achieved an AUC of 0.89 in cross-validation and independent testing. Models with missing variables achieved AUC values of 0.78-0.90. The standard absolute models made predictions that correlated significantly with observed changes in viral load with a mean absolute error of 0.65 log10 copies HIV RNA/mL in cross-validation and 0.69 log10 copies HIV RNA/mL in independent testing. Models with missing variables achieved values of 0.65-0.75 log10 copies HIV RNA/mL. All models identified alternative regimens that were predicted to be effective for the vast majority of cases where the new regimen prescribed in the clinic failed. All models were significantly better predictors of treatment response than genotyping with rules-based interpretation., Conclusions: These latest models that predict treatment responses accurately, even when a number of baseline variables are not available, are a major advance with greatly enhanced potential benefit, particularly in resource-limited settings. The only obstacle to realizing this potential is the willingness of healthcare professions to use the system., (© The Author(s) 2021. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For permissions, please email: journals.permissions@oup.com.)

Objective: Definitions of virological response vary from <50 up to 1000 copies of HIV-RNA/mL. Our previous models estimate the probability of HIV drug combinations reducing the viral load to <50 copies/mL, with no indication of whether higher thresholds of response may be achieved. Here, we describe the development of models that predict absolute viral load over time., Methods: Two sets of random forest models were developed using 50,270 treatment change episodes from more than 20 countries. The models estimated viral load at different time points following the introduction of a new regimen from variables including baseline viral load, CD4 count, and treatment history. One set also used genotypes in their predictions. Independent data sets were used for evaluation., Results: Both models achieved highly significant correlations between predicted and actual viral load changes (r = 0.67-0.68, mean absolute error of 0.73-0.74 log10 copies/mL). The models produced curves of virological response over time. Using failure definitions of <100, 400, or 1000 copies/mL, but not 50 copies/mL, both models were able to identify alternative regimens they predicted to be effective for the majority of cases where the new regimen prescribed in the clinic failed., Conclusions: These models could be useful for selecting the optimum combination therapy for patients requiring a change in therapy in settings using any definition of virological response. They also give an idea of the likely response curve over time. Given that genotypes are not required, these models could be a useful addition to the HIV-TRePS system for those in resource-limited settings.

Objectives: Optimizing antiretroviral drug combination on an individual basis can be challenging, particularly in settings with limited access to drugs and genotypic resistance testing. Here we describe our latest computational models to predict treatment responses, with or without a genotype, and compare their predictive accuracy with that of genotyping., Methods: Random forest models were trained to predict the probability of virological response to a new therapy introduced following virological failure using up to 50 000 treatment change episodes (TCEs) without a genotype and 18 000 TCEs including genotypes. Independent data sets were used to evaluate the models. This study tested the effects on model accuracy of relaxing the baseline data timing windows, the use of a new filter to exclude probable non-adherent cases and the addition of maraviroc, tipranavir and elvitegravir to the system., Results: The no-genotype models achieved area under the receiver operator characteristic curve (AUC) values of 0.82 and 0.81 using the standard and relaxed baseline data windows, respectively. The genotype models achieved AUC values of 0.86 with the new non-adherence filter and 0.84 without. Both sets of models were significantly more accurate than genotyping with rules-based interpretation, which achieved AUC values of only 0.55-0.63, and were marginally more accurate than previous models. The models were able to identify alternative regimens that were predicted to be effective for the vast majority of cases in which the new regimen prescribed in the clinic failed., Conclusions: These latest global models predict treatment responses accurately even without a genotype and have the potential to help optimize therapy, particularly in resource-limited settings.

Objectives: Optimizing antiretroviral drug combination on an individual basis in resource-limited settings is challenging because of the limited availability of drugs and genotypic resistance testing. Here, we describe our latest computational models to predict treatment responses, with or without a genotype, and compare the potential utility of global and local models as a treatment tool for South Africa., Methods: Global random forest models were trained to predict the probability of virological response to therapy following virological failure using 29 574 treatment change episodes (TCEs) without a genotype, 3179 of which were from South Africa and were used to develop local models. In addition, 15 130 TCEs including genotypes were used to develop another set of models. The 'no-genotype' models were tested with an independent global test set (n = 1700) plus a subset from South Africa (n = 222). The genotype models were tested with 750 independent cases., Results: The global no-genotype models achieved area under the receiver-operating characteristic curve (AUC) values of 0.82 and 0.79 with the global and South African tests sets, respectively, and the South African models achieved AUCs of 0.70 and 0.79. The genotype models achieved an AUC of 0.84. The global no-genotype models identified more alternative, locally available regimens that were predicted to be effective for cases that failed their new regimen in the South African clinics than the local models. Both sets of models were significantly more accurate predictors of outcomes than genotyping with rules-based interpretation., Conclusions: These latest global models predict treatment responses accurately even without a genotype, out-performed the local South African models and have the potential to help optimize therapy, particularly in resource-limited settings., (© The Author 2016. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com.)

Objectives: The optimal individualized selection of antiretroviral drugs in resource-limited settings is challenging because of the limited availability of drugs and genotyping. Here we describe the development of the latest computational models to predict the response to combination antiretroviral therapy without a genotype, for potential use in such settings., Methods: Random forest models were trained to predict the probability of a virological response to therapy (<50 copies HIV RNA/mL) following virological failure using the following data from 22,567 treatment-change episodes including 1090 from southern Africa: baseline viral load and CD4 cell count, treatment history, drugs in the new regimen, time to follow-up and follow-up viral load. The models were assessed during cross-validation and with an independent global test set of 1000 cases including 100 from southern Africa. The models' accuracy [area under the receiver-operating characteristic curve (AUC)] was evaluated and compared with genotyping using rules-based interpretation systems for those cases with genotypes available., Results: The models achieved AUCs of 0.79-0.84 (mean 0.82) during cross-validation, 0.80 with the global test set and 0.78 with the southern African subset. The AUCs were significantly lower (0.56-0.57) for genotyping., Conclusions: The models predicted virological response to HIV therapy without a genotype as accurately as previous models that included a genotype. They were accurate for cases from southern Africa and significantly more accurate than genotyping. These models will be accessible via the online treatment support tool HIV-TRePS and have the potential to help optimize antiretroviral therapy in resource-limited settings where genotyping is not generally available.

Objectives: Genotypic HIV drug-resistance testing is typically 60%-65% predictive of response to combination antiretroviral therapy (ART) and is valuable for guiding treatment changes. Genotyping is unavailable in many resource-limited settings (RLSs). We aimed to develop models that can predict response to ART without a genotype and evaluated their potential as a treatment support tool in RLSs., Methods: Random forest models were trained to predict the probability of response to ART (≤400 copies HIV RNA/mL) using the following data from 14 891 treatment change episodes (TCEs) after virological failure, from well-resourced countries: viral load and CD4 count prior to treatment change, treatment history, drugs in the new regimen, time to follow-up and follow-up viral load. Models were assessed by cross-validation during development, with an independent set of 800 cases from well-resourced countries, plus 231 cases from Southern Africa, 206 from India and 375 from Romania. The area under the receiver operating characteristic curve (AUC) was the main outcome measure., Results: The models achieved an AUC of 0.74-0.81 during cross-validation and 0.76-0.77 with the 800 test TCEs. They achieved AUCs of 0.58-0.65 (Southern Africa), 0.63 (India) and 0.70 (Romania). Models were more accurate for data from the well-resourced countries than for cases from Southern Africa and India (P < 0.001), but not Romania. The models identified alternative, available drug regimens predicted to result in virological response for 94% of virological failures in Southern Africa, 99% of those in India and 93% of those in Romania., Conclusions: We developed computational models that predict virological response to ART without a genotype with comparable accuracy to genotyping with rule-based interpretation. These models have the potential to help optimize antiretroviral therapy for patients in RLSs where genotyping is not generally available.

Objective: Antiretroviral drug selection in resource-limited settings is often dictated by strict protocols as part of a public health strategy. The objective of this retrospective study was to examine if the HIV-TRePS online treatment prediction tool could help reduce treatment failure and drug costs in such settings., Methods: The HIV-TRePS computational models were used to predict the probability of response to therapy for 206 cases of treatment change following failure in India. The models were used to identify alternative locally available 3-drug regimens, which were predicted to be effective. The costs of these regimens were compared to those actually used in the clinic., Results: The models predicted the responses to treatment of the cases with an accuracy of 0.64. The models identified alternative drug regimens that were predicted to result in improved virological response and lower costs than those used in the clinic in 85% of the cases. The average annual cost saving was $364 USD per year (41%)., Conclusions: Computational models that do not require a genotype can predict and potentially avoid treatment failure and may reduce therapy costs. The use of such a system to guide therapeutic decision-making could confer health economic benefits in resource-limited settings.

Introduction: A major challenge in Romania is the optimisation of antiretroviral therapy for the many HIV-infected adults with, on average, a decade of treatment experience. The RDI has developed computational models that predict virological response to therapy but these require a genotype, which is not routinely available in Romania. Moreover the models, which were trained without any Romanian data, have proved most accurate for patients from the healthcare settings that contributed the training data. Here we develop and test a novel model that does not require a genotype, with test data from Romania., Methods: A random forest (RF) model was developed to predict the probability of the HIV viral load (VL) being reduced to <50 copies/ml following therapy change. The input variables were baseline VL, CD4 count, treatment history and time to follow-up. The model was developed with 3188 treatment changes episodes (TCEs) from North America, Western Europe and Australia. The model's predictions for 100 independent TCEs from the RDI database were compared to those of a model trained with the same data plus genotypes and then tested using 39 TCEs from Romania in terms of the area under the ROC curve (AUC)., Results: When tested with the 100 independent RDI TCEs, the AUC values for the models with and without genotypes were 0.88 and 0.86 respectively. For the 39 Romanian TCEs the AUC was 0.60. However, when 14 cases with viral loads that may have been between 50 and 400 copies were removed, the AUC increased to 0.83., Discussion: Despite having been trained without data from Romania, the model predicted treatment responses in treatment-experienced Romanian patients with clade F virus accurately without the need for a genotype. The results suggest that this approach might be generalisable and useful in helping design optimal salvage regimens for treatment-experienced patients in countries with limited resources where genotyping is not always available.

Objective: The optimum selection and sequencing of combination antiretroviral therapy to maintain viral suppression can be challenging. The HIV Resistance Response Database Initiative has pioneered the development of computational models that predict the virological response to drug combinations. Here we describe the development and testing of random forest models to power an online treatment selection tool., Methods: Five thousand, seven hundred and fifty-two treatment change episodes were selected to train a committee of 10 models to predict the probability of virological response to a new regimen. The input variables were antiretroviral treatment history, baseline CD4 cell count, viral load and genotype, drugs in the new regimen, time from treatment change to follow-up and follow-up viral load values. The models were assessed during cross-validation and with an independent set of 50 treatment change episodes by plotting receiver-operator characteristic curves and their performance compared with genotypic sensitivity scores from rules-based genotype interpretation systems., Results: The models achieved an area under the curve during cross-validation of 0.77-0.87 (mean = 0.82), accuracy of 72-81% (mean = 77%), sensitivity of 62-80% (mean = 67%) and specificity of 75-89% (mean = 81%). When tested with the 50 test cases, the area under the curve was 0.70-0.88, accuracy 64-82%, sensitivity 62-80% and specificity 68-95%. The genotypic sensitivity scores achieved an area under the curve of 0.51-0.52, overall accuracy of 54-56%, sensitivity of 43-64% and specificity of 41-73%., Conclusion: The models achieved a consistent, high level of accuracy in predicting treatment responses, which was markedly superior to that of genotypic sensitivity scores. The models are being used to power an experimental system now available via the Internet.

The HIV Resistance Response Database Initiative (RDI), which comprises a small research team in the United Kingdom and collaborating clinical centers in more than 15 countries, has used antiretroviral treatment and response data from thousands of patients around the world to develop computational models that are highly predictive of virologic response. The potential utility of such models as a tool for assisting treatment selection was assessed in two clinical pilot studies: a prospective study in Canada and Italy, which was terminated early because of the availability of new drugs not covered by the system, and a retrospective study in the United States. For these studies, a Web-based user interface was constructed to provide access to the models. Participating physicians entered baseline data for cases of treatment failure and then registered their treatment intention. They then received a report listing the five alternative regimens that the models predicted would be most effective plus their own selection, ranked in order of predicted virologic response. The physicians then entered their final treatment decision. Twenty-three physicians entered 114 cases (75 unique cases with 39 entered twice by different physicians). Overall, 33% of treatment decisions were changed following review of the report. The final treatment decisions and the best of the RDI alternatives were predicted to produce greater virologic responses and involve fewer drugs than the original selections. Most physicians found the system easy to use and understand. All but one indicated they would use the system if it were available, particularly for highly treatment-experienced cases with challenging resistance profiles. Despite limitations, the first clinical evaluation of this approach by physicians with substantial HIV-experience suggests that it has the potential to deliver clinical and economic benefits.

In the absence of widespread access to individualized laboratory monitoring, which forms an integral part of HIV patient management in resource-rich settings, the roll-out of highly active antiretroviral therapy (HAART) in resource-limited settings has adopted a public health approach based on standard HAART protocols and clinical/immunological definitions of therapy failure. The cost-effectiveness of HIV-1 viral load monitoring at the individual level in such settings has been debated, and questions remain over the long-term and population-level impact of managing HAART without it. Computational models that accurately predict virological response to HAART using baseline data including CD4 count, viral load and genotypic resistance profile, as developed by the Resistance Database Initiative, have significant potential as an aid to treatment selection and optimization. Recently developed models have shown good predictive performance without the need for genotypic data, with viral load emerging as by far the most important variable. This finding provides further, indirect support for the use of viral load monitoring for the long-term optimization of HAART in resource-limited settings.

The 2008 International HIV Drug Resistance Workshop explored six topics on viral resistance: new antiretrovirals; clinical implications; epidemiology; new technologies and interpretations; HIV pathogenesis, fitness, and resistance; and mechanisms of resistance. The last of these topics provided a forum for new work on resistance of hepatitis B and C viruses, which were also explored in two poster sessions. Much work focused on resistance to the two most recent antiretroviral classes (integrase inhibitors and CCR5 antagonists), a new set of entry inhibitor candidates and one new class represented by the maturation inhibitor bevirimat. Other research explored two novel non-nucleoside reverse transcriptase inhibitors, etravirine and IDX899. Epidemiological work analysed rates of transmitted resistant virus, multiclass resistance in antiretroviral-experienced patients and a heightened resistance risk in injecting drug users regardless of adherence. New research on resistance technologies involved an enhanced assay for HIV-1 coreceptor determination and improved gene-based tools for predicting coreceptor use. In the pathogenesis arena, a small study of intensification shed light on the likely source of residual viraemia in patients on successful antiretroviral therapy. A large study in Mozambique correlated the timing of infant infection with selection, transmission and persistence of nevirapine resistance mutations. Mechanistic research explored resistance to the integrase inhibitor raltegravir, K65R-mediated resistance to tenofovir and the role of connection domain mutations in resistance to zidovudine.

The fingers subdomain of human immunodeficiency virus type 1 (HIV-1) reverse transcriptase (RT) is a hotspot for nucleoside analogue resistance mutations. Some multi-nucleoside analogue-resistant variants contain a T69S substitution along with dipeptide insertions between residues 69 and 70. This set of mutations usually co-exists with classic zidovudine-resistance mutations (e.g. M41L and T215Y) or an A62V mutation and confers resistance to multiple nucleoside analogue inhibitors. As insertions lie in the vicinity of the dNTP-binding pocket, their influence on RT fidelity was investigated. Commonly occurring insertion mutations were selected, i.e. T69S-AG, T69S-SG and T69S-SS alone, in combination with 3'-azido-2',3'-deoxythymidine-resistance mutations M41L, L210W, R211K, L214F, T215Y (LAG(AZ) and LSG(AZ)) or with an alternate set where A62V substitution replaces M41L (VAG(AZ), VSG(AZ) and VSS(AZ)). Using a lacZalpha gapped duplex substrate, the forward mutation frequencies of recombinant wild-type and mutant RTs bearing each of the above sets of mutations were measured. All of the mutants displayed significant decreases in mutation frequencies. Whereas the dipeptide insertions alone showed the least decrease (4.0- to 7.5-fold), the VAG series showed an intermediate reduction (5.0- to 11.4-fold) and the LAG set showed the largest reduction in mutation frequencies (15.3- and 16.3-fold for LAG(AZ) and LSG(AZ), respectively). Single dNTP exclusion assays for mutants LSG(AZ) and LAG(AZ) confirmed their large reduction in misincorporation efficiencies. The increased in vitro fidelity was not due to excision of the incorrect nucleotide via ATP-dependent removal. There was also no direct correlation between increased fidelity and template-primer affinity, suggesting a change in the active site that is conducive to better discrimination during dNTP insertion.

Finger insertion mutations of human immunodeficiency virus type 1 (HIV-1) reverse transcriptase (RT) (T69S mutations followed by various dipeptide insertions) have a multinucleoside resistance phenotype that can be explained by decreased sensitivity to deoxynucleoside triphosphate (dNTP) inhibition of the nucleotide-dependent unblocking activity of RT. We show that RTs with SG or AG (but not SS) insertions have three- to fourfold-increased unblocking activity and that all three finger insertion mutations have threefold-decreased sensitivity to dNTP inhibition. The additional presence of M41L and T215Y mutations increased unblocking activity for all three insertions, greatly reduced the sensitivity to dNTP inhibition, and resulted in defects in in vitro DNA chain elongation. The DNA chain elongation defects were partially repaired by additional mutations at positions 210, 211, and 214. These results suggest that structural communication between the regions of RT defined by these mutations plays a role in the multinucleoside resistance phenotype.

Six structures of human immunodeficiency virus type 1 (HIV-1) reverse transcriptase (RT) containing combinations of resistance mutations for zidovudine (AZT) (M41L and T215Y) or lamivudine (M184V) have been determined as inhibitor complexes. Minimal conformational changes in the polymerase or nonnucleoside RT inhibitor sites compared to the mutant RTMC (D67N, K70R, T215F, and K219N) are observed, indicating that such changes may occur only with certain combinations of mutations. Model building M41L and T215Y into HIV-1 RT-DNA and docking in ATP that is utilized in the pyrophosphorolysis reaction for AZT resistance indicates that some conformational rearrangement appears necessary in RT for ATP to interact simultaneously with the M41L and T215Y mutations.

Tenofovir is a nucleotide analogue human immunodeficiency virus type 1 (HIV-1) reverse transcriptase (RT) inhibitor, and its oral prodrug, tenofovir disoproxil fumarate, has recently been approved for the treatment of HIV-1 infection in the United States. The objective of this study was to characterize the in vitro susceptibility profiles of a large panel of clinically derived HIV-1 isolates for tenofovir. The distribution of tenofovir susceptibilities in over 1,000 antiretroviral-naive, HIV-1-infected individuals worldwide was determined using the Virco Antivirogram assay. In addition, phenotypic susceptibilities to tenofovir and other RT inhibitors were determined in a panel of nearly 5,000 recombinant HIV-1 clinical isolates from predominantly treatment-experienced patients analyzed as a part of routine drug resistance testing. Greater than 97.5% of isolates from treatment-naive patients had tenofovir susceptibilities <3-fold above those of the wild-type controls by the Antivirogram. The clinically derived panel of 5,000 samples exhibited a broad range of antiretroviral drug susceptibilities, including 69, 43, and 16% having >10-fold-decreased susceptibilities to at least one, two, and three antiretroviral drug classes, respectively. Greater than 88% of these 5,000 clinical isolates were within the threefold susceptibility range for tenofovir, and >99% exhibited <10-fold-reduced susceptibilities to tenofovir. Decreased susceptibility to tenofovir was not directly associated with resistance to other RT inhibitors; r(2) values of log-log linear regression plots of susceptibility to tenofovir versus susceptibility to other RT inhibitors were <0.4. The results suggest that the majority of treatment-naive and treatment-experienced individuals harbor HIV that remains within the normal range of tenofovir susceptibilities and may be susceptible to tenofovir disoproxil fumarate therapy.

Two large, independent human immunodeficiency virus type 1 resistance databases containing >7700 reverse-transcriptase (RT) sequences were used to analyze the epidemiology of amino acid substitutions at codons 44 and 118, which confer moderate lamivudine resistance in the presence of zidovudine resistance. As expected, E44A/D and V118I mutations were strongly associated with M41L, D67N, L210W, and T215Y but also with other mutations, including K43E/N/Q, T69D, V75M, H208Y, R211K, and K219R. Both E44D and V118I were more frequently associated with stavudine and didanosine than with zidovudine and lamivudine treatment. However, selection of E44A/D and V118I was also detected in association with a switch to other nucleoside RT inhibitors, including zalcitabine and abacavir. Site-directed mutagenesis confirmed that 44D and 118I can decrease phenotypic susceptibility not only to lamivudine but also to most other nucleoside analogues, particularly stavudine and abacavir. Thus, substitutions at RT codons 44 and 118 have broad implications in nucleoside RT inhibitor resistance in the setting of several nucleoside-associated mutations.

The phenomenon of cross-resistance to antiretroviral agents used to treat human immunodeficiency virus type 1 infection is well known but so far has been only qualitatively described. Here, we quantitate the degree of cross-resistance among all commonly prescribed antiretroviral agents in almost 5,000 clinically derived recombinant isolates collected in the United States since January 2000.

Human immunodeficiency virus type 1 (HIV-1) isolates from 50 plasma specimens were analyzed for phenotypic susceptibility to licensed reverse transcriptase inhibitors and protease inhibitors by the Antivirogram and PhenoSense HIV assays. Twenty of these specimens were from recently seroconverted drug-naïve persons, and 30 were from patients who were the sources of occupational exposures to HIV-1; 16 of the specimens in the latter group were from drug-experienced patients. The phenotypic results of the Antivirogram and PhenoSense HIV assays were categorized as sensitive or reduced susceptibility on the basis of the cutoff values established by the manufacturers of each assay. Data for 12 to 15 drugs were available by both assays for 38 specimens and represented a total of 529 pairs of results. The two data sets had a 91.5% concordance by phenotypic category. The discordant results (n = 45) were distributed randomly among 26 specimens and included 28 results (62.2%) which were within a twofold difference of the assay cutoff values. None of the discordant results were associated with primary resistance mutations that predicted high-level (>20-fold) resistance. Discordant results were distributed equally among specimens from drug-experienced and drug-naïve individuals and were slightly higher for protease inhibitors than for nonnucleoside reverse transcriptase inhibitors or nucleoside reverse transcriptase inhibitors. The findings of the present study demonstrate that the results of the Antivirogram and PhenoSense HIV assays correlate well, despite the use of different testing strategies.

Objectives: To examine the natural phenotypic variability in drug susceptibility among recombinant HIV-1 isolates from a large number of untreated HIV-positive individuals from wide-ranging geographic locations, and to use this information to establish biologically relevant cut-off values for phenotypic antiretroviral susceptibility testing., Methods: Phenotypic susceptibility to 14 antiretroviral agents was determined for HIV-1 samples from > 1000 treatment-naive individuals in seven clinical trials. Samples were from the USA (n = 351), Germany (n = 306), Canada (n = 265), and South Africa (n = 358). Geometric mean fold-resistance and confidence intervals were determined relative to a standard laboratory wild-type virus., Results: Baseline fold-resistance was approximately log-normally distributed for all antiretroviral agents examined. There was no evidence of large geographical differences in average antiviral susceptibility. Geometric mean fold-resistance for each of 14 antiviral agents was similar (+/- 0.5-fold) for samples derived from the USA, Canada, Germany, or South Africa. The non-nucleoside reverse transcriptase inhibitors (NNRTI) exhibited the broadest distribution of susceptibility; approximately 97.5% of all isolates had < 2.5-4.0, < 3.0-4.5, and < 5-10 fold-decrease in susceptibility to five protease inhibitors, six nucleoside analogues, and three NNRTI, respectively. No consistent geographic pattern or clade effect (B versus C) in either the mean or the distribution of baseline antiretroviral susceptibility was observed., Conclusions: Phenotypic drug susceptibility of HIV-1 in untreated individuals varies markedly from drug to drug, with broadly similar patterns world-wide. These results have important implications in defining the 'normal range' of phenotypic susceptibility to antiretroviral agents and establish biologically relevant cut-off values for this phenotypic drug susceptibility test.

Using a large panel of human immunodeficiency virus type 1 site-directed mutants, we have observed a higher correlation than has previously been demonstrated between zidovudine (AZT)-triphosphate resistance data at the reverse transcriptase (RT) level and corresponding viral AZT resistance. This enhanced-resistance effect at the RT level was seen with ATP and to a lesser extent with PP(i) when ATP was added at physiological concentrations. The ATP-dependent mechanism (analogous to pyrophosphorolysis) appears to be dominant in the mutants bearing the D67N and K70R or 69 insertion mutations, whereas the Q151M mutation seems independent of ATP for decreased binding to AZT-triphosphate.

We have found a close correlation between viral stavudine (d4T) resistance and resistance to d4T-triphosphate at the human immunodeficiency virus type 1 reverse transcriptase (RT) level. RT from site-directed mutants with 69S-XX codon insertions and/or conventional zidovudine resistance mutations seems to be involved in an ATP-dependent resistance mechanism analogous to pyrophosphorolysis, whereas the mechanism for RT with the Q151M or V75T mutation appears to be independent of added ATP for reducing binding to d4T-triphosphate.

We assessed the reproducibility of human immunodeficiency virus type 1 (HIV-1) reverse transcriptase (RT) and protease sequencing using cryopreserved plasma aliquots obtained from 46 heavily treated HIV-1-infected individuals in two laboratories using dideoxynucleotide sequencing. The rates of complete sequence concordance between the two laboratories were 99.1% for the protease sequence and 99.0% for the RT sequence. Approximately 90% of the discordances were partial, defined as one laboratory detecting a mixture and the second laboratory detecting only one of the mixture's components. Only 0.1% of the nucleotides were completely discordant between the two laboratories, and these were significantly more likely to occur in plasma samples with lower plasma HIV-1 RNA levels. Nucleotide mixtures were detected at approximately 1% of the nucleotide positions, and in every case in which one laboratory detected a mixture, the second laboratory either detected the same mixture or detected one of the mixture's components. The high rate of concordance in detecting mixtures and the fact that most discordances between the two laboratories were partial suggest that most discordances were caused by variation in sampling of the HIV-1 quasispecies by PCR rather than by technical errors in the sequencing process itself.

A variety of key mutations in HIV reverse transcriptase (RT) have been associated with nucleoside reverse transcriptase inhibitor (NRTI) exposure, which give rise to a diverse range of effects in terms of altered drug susceptibilities, viral replicative capacity and RT biochemistry. There are three basic mechanisms of resistance conferred by specific mutations in the coding region of RT. The first is drug discrimination, whereby a particular drug or drugs are either selectively excluded from uptake or from the RT-primer-template catalytic complex. Drug discrimination is, for the most part, relatively specific for individual drugs. Repositioning of the template-primer to prevent a catalytically competent complex in the presence of a bound drug molecule has also been observed in some instances, and forms a second mechanism. The third, and potentially most significant for long-term efficacy of the NRTIs, is pyrophosphorolysis, the primary mode of resistance to zidovudine. Mutations selected by this drug or stavudine serve to elevate the natural rate of the reverse reaction for RT. Pyrophosphorolysis uncouples the last nucleoside monophosphate added to the proviral transcript, and attaches it to either a free pyrophosphate (regenerating a deoxynucleoside triphosphate) or to a nucleoside di- or triphosphate (usually ATP). Uncoupling a chain-terminating NRTI residue therefore rescues reverse transcription and reduces drug susceptibility across the class, since the process is not specific for the selecting drug. Of all the nucleoside-associated mutations, the best known and most studied are the six associated with thymidine analogue exposure. These six mutations (M41L, D67N, K70R, L210W, T215Y/F, K219Q) enhance RT pyrophosphorolysis to confer high-level viral resistance to zidovudine, and clinically significant loss of response to stavudine and didanosine. They have also been found to confer reduced susceptibility to lamivudine and abacavir, particularly when present alongside other NRTI-induced changes. Other key mutations generally confer more limited resistance to specific agents, although the primary lamivudine- and abacavir-associated M184V substitution generates a broad spectrum of drug-dependent phenotypes, and uncommon mutational complexes conferring resistance across the entire class are well known. In addition to 'classical' multi-nucleoside-resistant genotypes, database-driven 'virtual phenotyping' for accumulations of NRTI-associated mutations around a core of thymidine analogue-induced changes predicts drug susceptibilities below wild-type across the entire NRTI class, even in the absence of key mutations associated with individual agents. When the natural range of drug susceptibilities for treatment-naive isolates is used as the basis for defining resistance, retrospective analysis of clinical isolates in the Virco database shows a significantly increased incidence of reduced susceptibility for the dideoxy NRTIs (didanosine, stavudine and zalcitabine) that was undetected in previous assays. These data imply a cumulative degradation of response to

Objective: Although the use of HIV-1 protease inhibitors (PI) has substantially benefited HIV-1-infected individuals, new PI are urgently needed, as broad PI resistance and therapy failure is common., Methods: The antiviral activity of tipranavir (TPV), a non-peptidic PI, was assessed in in vitro culture for 134 clinical isolates with a wide range of resistance to currently available peptidomimetic PI. The susceptibility of all 134 variants was then re-tested with the four PI simultaneously with TPV, using the Antivirogram assay., Results: Of 105 viruses with more than tenfold resistance to three or four PI and an average of 6.1 PI mutations per sample, 95 (90%) were susceptible to TPV; eight (8%) had four- to tenfold resistance to TPV and only two (2%) had more than tenfold resistance., Conclusions: The substantial lack of PI cross-resistance to TPV shown by highly PI-resistant clinical isolates makes TPV an attractive new-generation HIV inhibitor.

Objective: To evaluate in HIV-1 the extent of phenotypic and genotypic antiretroviral drug resistance and cross-resistance towards the protease inhibitors (PIs) saquinavir, ritonavir, indinavir and nelfinavir among a set of patient samples originating from European and US routine clinical practice and submitted for phenotypic drug resistance testing and/or genotypic analysis. The mutational pattern(s) underlying both resistance and cross-resistance to PIs was investigated., Method: Over 6000 patient isolates with plasma viral load greater than 1000 copies/ml plasma were analysed. Phenotypic resistance was evaluated by a recombinant virus assay. Phenotypic resistance is expressed as the fold-increase of the 50% inhibitory concentration (IC50) value of a compound for a patient-derived recombinant virus isolate compared with that for a wild-type laboratory virus. Genotypic analysis is reported as amino acid changes at positions in the HIV-1 protease compared to a wild-type reference., Results: Phenotypic resistance to any single PI was observed in 17 to 25% of the clinical isolates investigated. Phenotypic cross-resistance among PIs (> 10-fold increase in IC50 value) was detected in 59 to 80% of the samples resistant (> 10-fold increase in IC50 value) to at least one PI. The prevalent mutations in PI-resistant isolates involved substitutions at codons 10, 36, 46, 54, 71, 77, 82 and 90. The most frequent mutational pattern in samples with PI cross-resistance involved combined substitutions at positions 10 and 90, extended with substitutions at positions 54, 71, 77, 82 or 84., Conclusions: Extensive use of first-generation PIs leads to the emergence of HIV-1 isolates possessing cross-resistance to all members of this class. Identification of particular mutational profiles among these isolates may assist in the design of new generation inhibitors with specific activity against protease-mutant HIV strains.

We describe a new human immunodeficiency virus type 1 (HIV-1) mutational pattern associated with phenotypic resistance to lamivudine (3TC) in the absence of the characteristic replacement of methionine by valine at position 184 (M184V) of reverse transcriptase. Combined genotypic and phenotypic analyses of clinical isolates revealed the presence of moderate levels of phenotypic resistance (between 4- and 50-fold) to 3TC in a subset of isolates that did not harbor the M184V mutation. Mutational cluster analysis and comparison with the phenotypic data revealed a significant correlation between moderate phenotypic 3TC resistance and an increased incidence of replacement of glutamic acid by aspartic acid or alanine and of valine by isoleucine at residues 44 and 118 of reverse transcriptase, respectively. This occurred predominantly in those isolates harboring zidovudine resistance-associated mutations (41L, 215Y). The requirement of the combination of mutations 41L and 215Y with mutations 44D and 44A and/or 118I for phenotypic 3TC resistance was confirmed by site-directed mutagenesis experiments. These data support the assumption that HIV-1 may have access to several different genetic pathways to escape drug pressure or that the increase in the frequency of particular mutations may affect susceptibility to drugs that have never been part of a particular regimen.

Thirteen protease inhibitor-naive patients with HIV-1 infection, and 12 patients with a median of 58 months prior treatment with saquinavir (SQV) monotherapy, were treated with SQV (400 mg twice daily) and ritonavir (RIT, 500 mg twice daily) in a study designed to assess the effect of prior treatment with SQV monotherapy on the antiretroviral activity of RIT-SQV combination therapy. Median baseline viral load and CD4+ cell counts were 155,000 and 262,000 copies/ml and 333 and 225 cells/mm3 in the naive and experienced groups, respectively. Mean viral load changes at 24 weeks were -1.63 and -0.27 log copies/ml in the naive and SQV-experienced groups, respectively (intent-to-treat analysis). Baseline genotype by point mutation assay and sequencing in the SQV-experienced group was highly predictive of virological response. Eight of 11 SQV-experienced patients had evidence of phenotypic resistance to RIT at baseline, despite previous treatment with SQV only. There was strong correlation between phenotypic resistance to RIT and the presence of the L90M mutation. We conclude that prolonged prior treatment with saquinavir monotherapy may produce cross-resistance to ritonavir and reduce the subsequent response to ritonavir-saquinavir in combination. In this study, both phenotypic resistance to ritonavir and presence of the L90M mutation predicted the viral load response to ritonavir-saquinavir.

To investigate the occurrence of multinucleoside analog resistance during therapy failure, we surveyed the drug susceptibilities and genotypes of nearly 900 human immunodeficiency virus type 1 (HIV-1) samples. For 302 of these, the 50% inhibitory concentrations of at least four of the approved nucleoside analogs had fourfold-or-greater increases. Genotypic analysis of the reverse transcriptase (RT)-coding regions from these samples revealed complex mutational patterns, including the previously recognized codon 151 multidrug resistance cluster. Surprisingly, high-level multinucleoside resistance was associated with a diverse family of amino acid insertions in addition to "conventional" point mutations. These insertions were found between RT codons 67 and 70 and were commonly 69Ser-(Ser-Ser) or 69Ser-(Ser-Gly). Treatment history information showed that a common factor for the development of these variants was AZT (3'-azido-3'-deoxythymidine, zidovudine) therapy in combination with 2',3'-dideoxyinosine or 2',3'-dideoxycytidine, although treatment patterns varied considerably. Site-directed mutagenesis studies confirmed that 69Ser-(Ser-Ser) in an AZT resistance mutational background conferred simultaneous resistance to multiple nucleoside analogs. The insertions are located in the "fingers" domain of RT. Modelling the 69Ser-(Ser-Ser) insertion into the RT structure demonstrated the profound direct effect that this change is likely to have in the nucleoside triphosphate binding site of the enzyme. Our data highlight the increasing problem of HIV-1 multidrug resistance and underline the importance of continued resistance surveillance with appropriate, sufficiently versatile genotyping technology and phenotypic drug susceptibility analysis.

HIV reverse transcriptase (RT) is one of the main targets for the action of anti-AIDS drugs. Many of these drugs [e.g., 3'-azido-3'-deoxythymidine (AZT) and 2',3'-dideoxyinosine (ddI)] are analogues of the nucleoside substrates used by the HIV RT. One of the main problems in anti-HIV therapy is the selection of a mutant virus with reduced drug sensitivity. Drug resistance in HIV is generated for nucleoside analogue inhibitors by mutations in HIV RT. However, most of these mutations are situated some distance from the polymerase active site, giving rise to questions concerning the mechanism of resistance. To understand the possible structural bases for this, the crystal structures of AZT- and ddI-resistant RTs have been determined. For the ddI-resistant RT with a mutation at residue 74, no significant conformational changes were observed for the p66 subunit. In contrast, for the AZT-resistant RT (RTMC) bearing four mutations, two of these (at 215 and 219) give rise to a conformational change that propagates to the active site aspartate residues. Thus, these drug resistance mutations produce an effect at the RT polymerase site mediated simply by the protein. It is likely that such long-range effects could represent a common mechanism for generating drug resistance in other systems.

Recent clinical trials examining 3'-azido-3'-deoxythymidine (AZT, zidovudine, or Retrovir) combined with L-2', 3'-dideoxy-3'-thiacytidine (3TC or lamivudine) have shown that combination therapy with these nucleoside analogs affords significant virological and clinical benefits. The addition of 3TC to AZT delays AZT resistance in therapy-naive patients and can restore viral AZT susceptibility in patients who previously received AZT alone. In some AZT-experienced patients, the virological response to AZT-3TC therapy is not sustained and virus resistant to both drugs can be identified. To gain insight into the possible mechanism of dual resistance, we studied a recently described variant resistant to both AZT and 3TC and obtained by simultaneous passage of an AZT-resistant clinical isolate in cell culture with AZT and 3TC. Genetic mapping and site-directed mutagenesis experiments demonstrated that a polymorphism at codon 333 (Gly to Glu) of human immunodeficiency virus type 1 reverse transcriptase (RT) was critical in facilitating dual resistance in a complex background of AZT and 3TC resistance mutations. To assess the potential clinical relevance of RT codon 333 changes, we studied dually resistant viruses from patients taking AZT and 3TC. Genetic mapping of RT molecular clones derived from patients' plasma samples demonstrated that in some cases polymorphism at codon 333 was responsible for facilitating dual resistance.

Human immunodeficiency virus type 1 (HIV-1) strains dually resistant to zidovudine and lamivudine (3TC) may arise during zidovudine-3TC combination therapy. The objective of this cross-sectional study (n = 43 patients) was to test the association between therapy response (clinical and immunologic) to zidovudine-3TC and the level of phenotypic zidovudine resistance and zidovudine resistance-associated genotype of 3TC-resistant isolates. Other variables included were baseline CD4+ cell count, baseline Centers for Disease Control and Prevention (CDC) classification, virus load, and time receiving zidovudine. Phenotypic resistance was assessed using a recombinant virus assay. Genotypic analysis was based on population sequencing of plasma HIV-1. In a univariate analysis using a logistic regression model, it was found that therapy response was significantly associated with phenotypic and genotypic zidovudine resistance, baseline CD4+ cell count, and virus load. After adjustment for all variables, phenotypic resistance to zidovudine remained the only significantly associated factor, independent of baseline CD4+ cell count, baseline CDC classification, and virus load.

Objective: To investigate the prevalence and magnitude of M184V-mediated changes in susceptibility to zalcitabine, didanosine, stavudine and abacavir (1592U89 succinate) in a cohort of lamivudine-treated patients., Design and Methods: A total of 255 samples from patients treated with lamivudine and zidovudine with or without other nucleoside reverse transcriptase inhibitors (NRTI) were analysed for susceptibility to zidovudine, lamivudine, zalcitabine, didanosine and stavudine using a recombinant virus assay. Seventy-three samples originated from patients exposed to zidovudine and lamivudine only. A subset of 27 samples was investigated for cross-resistance to abacavir. Resistance was defined as a change in median inhibitory concentration more than fivefold compared with wild-type (high-level resistance, > 10-fold). A genotypic analysis of plasma-derived reverse transcriptase coding regions was carried out in samples with cross-resistance., Results: The majority of samples displayed wild-type or greater than wild-type sensitivity to zalcitabine, didanosine and stavudine: resistance was seen in 17.2, 9 and 6.3% of the total sample population, respectively. Of these, 1.2, 2.7 and 2.4%, respectively, showed high-level resistance. The prevalence of resistance to a particular NRTI was lower in samples from patients not pretreated with that NRTI and in samples from patients exposed to zidovudine-lamivudine only. Cross-resistance was more prevalent in samples with high ZDV resistance. There was no obvious correlation between cross-resistance and genotype; all but two samples were mutant at codon 184. There were no consistent changes at positions associated with zidovudine resistance. The majority of samples from a subset (n=27) were four- to eightfold less sensitive to abacavir. There were no other genotypic changes in addition to M184V known to be associated with abacavir resistance., Conclusions: Cross-resistance was not commonly observed in this lamivudine-treated cohort. M184V per se is not expected to compromise subsequent treatment with NRTI such as didanosine-stavudine or combinations containing abacavir.

Replication of mixtures of two or more human immunodeficiency virus type 1 (HIV-1) variants would be expected to result in the eventual selection of the fittest virus due to Darwinian competition among the variants. The relative proportions of known HIV-1 variants (which may differ only by a single nucleotide from a standard "wild-type" virus, HIV-1HXB2) in mixed viral cultures were quantified by analysis of automated sequence signals of reverse transcriptase PCR products. With this method, the relative levels of replicative fitness of several zidovudine (3'-azidothymidine)-resistant HIV-1HXB2 variants were estimated under controlled in vitro conditions by measuring the rate of change in the proportions of viral variants as they replicated in cell cultures both in the presence and in the absence of drug selection pressure. These variants were engineered to contain commonly observed zidovudine resistance mutations in the HIV-1 reverse transcriptase (M41L, K70R, T215Y, and M41L+T215Y). In the absence of zidovudine, all variants tested displayed reduced replicative fitness compared to wild-type HIV-1HXB2. The order of relative fitness was wild type > K70R >> T215Y = M41L+T215Y > M41L. Mixed cultures in the presence of zidovudine showed a dose-dependent selection pressure against the wild-type virus which varied according to the resistance profile of each virus. The information gathered from this approach provides insight into competition among multiple HIV-1 variants, which likely occurs in vivo with drug selection pressure, and may be applicable in more complex mathematical models for predicting the emergence of HIV-1 variants after the initiation of antiretroviral therapy.

Amino acid variation at reverse transcriptase (RT) codon 210 (generally Leu-210 to Trp [L210W], TTG-->TGG) is occasionally detected after the initiation of azidothymidine (AZT) therapy. The impact of this variation on AZT resistance and viral replication was addressed by four different approaches. The frequency and genetic background of the L210W mutation in vivo were assessed by analyzing sera of AZT-naive and AZT-experienced patients by RT-PCR and DNA sequencing. The degree of AZT resistance (50% infective concentration [IC50]) of recombinant viruses constructed by using the RT of 21 clinical isolates was stratified by the presence or absence of the 210 mutation. The AZT IC50S of a panel of mutant viruses (with or without W-210) constructed by site-directed mutagenesis in an HXB2 background were assayed by using a HeLa CD4 plaque reduction assay. Finally, the effect of the 210 mutation on viral replication was assessed by replication competition of an AZT-resistant virus, RTMN (L-41/Y-215), and RTMN with the W-210 mutation in the presence and in the absence of AZT. In AZT-naive patients, tryptophan at RT residue 210 was rare. After AZT exposure, W-210 appeared in a minority of those patients, most commonly in association with L-41 and Y-215. The presence of W-210 increased the AZTIC50 by two- to fourfold, as determined by both the recombinant virus assay and site-directed mutagenesis. A significant replication advantage in favor of the wild-type L-210 over W-210 was observed, although the selection against the 210 mutant was two- to threefold lower when the viruses were grown in the presence of 5 microM AZT. In summary, the L210W mutation appears to be of marginal significance, conferring approximately two- to fourfold-reduced sensitivity to AZT compared with similar AZT-resistant genomes with L-210. The selection pressure against W-210 may account for the modest proportion of patients in which W-210 appears in vivo.

Human immunodeficiency virus type 1 (HIV-1) isolates obtained prior to and during a combination therapy trial comparing zidovudine (AZT; 3'-azidothymidine) monotherapy with AZT plus 2',3'-dideoxyinosine (ddI) or AZT plus 2',3'-dideoxycytidine (ddC) were assessed for the development of drug resistance. Drug susceptibility was measured by using two different phenotypic assays, one that requires infection of peripheral blood mononuclear cells with HIV-1 isolated from cocultures and a second based on infection of HeLa CD4+ cells with recombinant virus containing the reverse transcriptase (RT) of the clinical isolate. In addition, genotypic assessment of resistance was obtained by DNA sequencing of the RT coding region. No difference in the development of AZT resistance was noted in isolates from individuals receiving AZT monotherapy or combination therapy. However, a low frequency of ddI or ddC resistance was seen in isolates from the combination arms, which may at least partially explain the enhanced efficacy observed with these drug combinations compared with monotherapy. It was noted from DNA sequencing that a relatively high frequency of the nonnucleoside RT inhibitor resistance mutation, codon 181 changed from encoding Tyr to encoding Cys, was present in some isolates both before and during nucleoside analog combination therapy. Since these patients were unlikely to have access to nonnucleoside RT inhibitors, it is probable that this mutation preexisted at a reasonable level in the wild-type virus population. Comparisons of the AZT susceptibility assays indicated a good correlation between the phenotypic and genotypic determinations. However, direct numerical comparisons between the phenotypic assays were not reliable, suggesting that valid comparisons of different resistance data sets will require the use of the same assay procedure.

A randomized controlled study was done to determine whether initial combination therapy with zidovudine and zalcitabine or zidovudine and didanosine would delay the emergence of zidovudine-resistant virus. Human immunodeficiency virus (HIV)-1-infected patients with <300 CD4 cells/mm3 and <4 weeks of prior zidovudine therapy were randomized to zidovudine, zidovudine plus zalcitabine, or zidovudine plus didanosine. Combination therapy did not delay the emergence of zidovudine-resistant virus isolates. However, combination therapy resulted in a significant increase in CD4 cells through 72 weeks compared with zidovudine monotherapy and a greater and more sustained decline in serum HIV-1 RNA. Although this trial was not designed as a clinical end-point study, patients assigned to zidovudine plus didanosine combination therapy experienced a significant delay in time to first AIDS-defining event or death compared with those assigned to zidovudine monotherapy.

Combinations of antiretroviral drugs that prevent or delay the appearance of drug-resistant human immunodeficiency virus-type 1 (HIV-1) mutants are urgently required. Mutants resistant to 3'-azidothymidine (AZT, zidovudine) became phenotypically sensitive in vitro by mutation of residue 184 of viral reverse transcriptase to valine, which also induced resistance to (-)2'-deoxy-3'-thiacytidine (3TC). Furthermore, AZT-3TC coresistance was not observed during extensive in vitro selection with both drugs. In vivo AZT-3TC combination therapy resulted in a markedly greater decreased in serum HIV-1 RNA concentrations than treatment with AZT alone, even though valine-184 mutants rapidly emerged. Most samples assessed from the combination group remained AZT sensitive at 24 weeks of therapy, consistent with in vitro mutation studies.

Genetic recombination between viral genomes has been shown to contribute to the generation of genetic diversity during retrovirus infections. The role of recombination in the development of human immunodeficiency virus type 1 (HIV-1) zidovudine resistance was investigated as a possible cause of the formation of the linked Leu-41/Tyr-215 resistance genotype. Zidovudine resistance is conferred by the presence of subsets of four or five amino acid substitutions in the HIV-1 reverse transcriptase. Zidovudine therapy of asymptomatic HIV-1-infected individuals results in the selection of drug-resistant variants that posses defined combinations of the five zidovudine resistance mutations. The linked Leu-41/Tyr-215 resistance genotype appears central to the continued development of high-level zidovudine resistance. By using genetically tagged mutant viruses, it was possible readily to select recombinant viruses from mixed infections of Leu-41 and Tyr-215 single mutants in the presence of zidovudine drup pressure. After three passages of a mixed infection in the presence of drug, 38% of clones screened were recombinant double mutants. In the absence of zidovudine selection, little change in the mixed virus populations was noted. No evidence of de novo generation of mutations at codons 41 and 215 was seen during any in vitro passage. This provides the first example of the role of retroviral recombination in the development of HIV-1 variants with increased drug resistance.

Mounting scientific evidence suggests that viral replication and evolution of diversification, with the resulting emergence of variants including drug-resistant strains, is responsible for the gradual destruction of the immune system and is the mechanism of disease progression in HIV-infected patients. Monotherapy does not give long-term suppression of viral replication and evolution, and combination therapy is viewed as a potentially more effective long-term approach based on increased and more durable suppression of HIV replication. Because of the large number of drugs that could be used in combination therapy regimens, it is important to investigate, in clinical studies, only those combinations likely to be most effective. In vitro studies are therefore critical in the selection of such combinations. Resistance mutations to many antiretroviral agents have been documented and their patterns of emergence elucidated. A number of studies have since demonstrated the phenomenon of cross-resistance, in which resistant virus emerging under the selective pressure of one therapy is also cross-resistant to a second antiviral agent. This information can be used to avoid combining agents to which cross-resistance can occur. Suppression or "phenotypic reversal" of zidovudine resistance by the lamivudine-resistance mutation at codon 184 has been demonstrated in in vitro studies. It is encouraging that this finding has led to the clinical evaluation of these two agents with promising results. Moreover, in vitro scientific data are critical in facilitating the identification of potentially effective combination therapies.

We have selected a human immunodeficiency virus type 1 (HIV-1) mutant strain with a moderate (sevenfold) level of resistance to the nucleoside analog 2',3'-didehydro-2',3'-dideoxythymidine (D4T or stavudine). After serial passage of the HXB2 strain of HIV-1 in MT4 cells, a novel mutation involving two nucleotide substitutions in codon 75 of the viral reverse transcriptase, altering valine to threonine, was seen. When introduced into a wild-type HIV-1 background by site-directed mutagenesis, the T-75 mutation conferred cross-resistance to the dideoxynucleosides dideoxyinosine and dideoxycytosine as well as to 2',3'-didehydro-2',3'-dideoxycytosine.

Mutation in the human immunodeficiency virus type 1 reverse transcriptase (RT) at codon 215 has been shown to play a significant role in resistance to zidovudine (AZT). Substitution of threonine with tyrosine or phenylalanine alone confers decreased susceptibility to the inhibitor. In this study we constructed a panel of 10 viruses with different amino acids at this codon, including 7 novel mutants, and assessed their susceptibilities to AZT. The majority of the new mutants were AZT sensitive, whereas the Thr-215-->Trp mutant was partially resistant (threefold less susceptible). A combination of the Thr-215-->Trp with the other AZT resistance mutations Lys-70-->Arg and Met-41-->Leu gave additive resistance. The Thr-215-->Phe virus was less AZT resistant than the Thr-215-->Tyr mutant, both on its own and when each was combined with the Met-41-->Leu mutant. These observations confirm the general hypothesis that increased bulk of the amino acid side chains at this position confers decreased AZT sensitivity. A leucine-to-valine substitution at codon 74 has previously been found to confer dideoxynucleoside resistance. We constructed mutants with five novel amino acid substitutions (Ala, Gly, Glu, Met, and Asp) at codon 74. Of these, only one (that with the Met substitution) retained enough RT activity to yield viable virus. It thus appears that there are severe structure-function constraints on the amino acid side chains at this position in the enzyme. The activities of the Leu-74-->Ala and Leu-74-->Met RT enzymes expressed in Escherichia coli appeared to have reduced susceptibility to ddGTP compared with the wild-type enzyme. The mutants described in this work may prove useful for correlation with structural studies of the human immunodeficiency virus type 1 RT.

During zidovudine therapy, human immunodeficiency virus type 1 (HIV-1) acquires a distinctive set of mutations that diminish the sensitivity of the virus to this drug in vitro. An AIDS patient is described who, while being treated with zidovudine, transmitted HIV-1 bearing a drug resistance mutation to a young woman who had never received zidovudine treatment. DNA sequencing of HIV-1 proviruses confirmed that these 2 persons shared HIV genetic variants, including a mutation at codon 70 in the reverse transcriptase gene associated with reduced in vitro sensitivity to zidovudine. This mutation persisted in the woman > 1 year in the absence of antiretroviral therapy. HIV-1 with genetic markers of zidovudine resistance can be transmitted heterosexually, but it is uncertain whether dissemination of drug-resistant virus will substantially reduce the usefulness of this drug.