printer friendly

CME Expired 4/18/08

return to FreeHIVinfo.com 

return to ResistanceWATCH home

HIV Resistance in Context: A Historical Perspective Integrating Highlights from 14^th CROI to Support Current HIV Practice

HIV RESISTANCE: A HISTORICAL PERSPECTIVE

When zidovudine (AZT)^[1] was first approved in 1987, early laboratory experiments could not show that HIV selected resistance to it. Unfortunately, within a couple of years, Larder reported isolating resistant strains from people who had been on AZT for a year or more[2]. And regrettably, there was no test to measure HIV resistance to AZT. It just seemed to stop working after about a year or so.

AZT and the other early antiviral drugs inhibit the activity of reverse transcriptase, which HIV uses to copy its genetic material into a form that can be easily inserted into a chromosome of the human host cell, thus enabling HIV to use the host cell's machinery to replicate. AZT interferes with reverse transcriptase's gene-copying capacity. Resistance develops when HIV evades AZT's interference by reshaping its reverse transcriptase protein through mutations. Each mutation changes a key amino acid of the reverse transcriptase protein, which enhances excision or removal of AZT, reducing its effectiveness. This mutation process is enhanced because HIV does not have any “proofreading” function as reverse transcriptase makes copies of the HIV genetic code, so mistakes are common. Some of these mistakes may confer resistance to antiretroviral drugs.

This was a challenging phase in our understanding of resistance. It was clear, for example, that AZT stopped working as well when resistant mutations appeared, 6 to 12 months after starting AZT. However, was it the disease progression that drove the process, or the accumulation of mutations? There was no way of knowing. By 1992, and prior to any formal studies of dual antiviral therapy, there were two studies that suggested that HIV was selecting mutations in response to AZT and that switching from AZT to didanosine[3] might be beneficial[4],[5].

The next major phase of exploration of drug resistance was the cataloguing of mutations in HIV that were determined to be the most relevant to the development of resistance. As each new antiviral drug was developed, data were generated from laboratory and clinical studies of patients whose therapy was failing, and manufacturers produced studies to determine which mutations were the most likely to be associated with viral resistance to that drug. Ultimately, Stanford University created a mammoth database of resistance-associated mutations (see http://hivdb.stanford.edu). This database provides tables that categorize the impact of viral resistance to each antiretroviral drug. There are many interpretation systems available today. Commonly, resistance is categorized as high, medium or low for reverse transcriptase inhibitors (NRTIs), non-nucleoside reverse transcriptase inhibitors (NNRTIs), and protease inhibitors (PIs). Often, the manufacturer claimed a “unique resistance profile” for its drug, implying that it might be best used in a particular sequence in HIV therapy. Often, the “unique resistance profile” identified did not correlate with clinical results as some degree of cross-resistance frequently existed. Hence, clinicians grew to be much more cautious of claims by the manufacturer[6],[7],[8],[9].

The first HIV viral load test was approved by the FDA in 1996. The amount of HIV RNA in the blood sample gives a rough idea of how actively HIV is replicating. Similar molecular tools allowing amplification of viral genetic material and its analysis lead the way to HIV drug resistance testing. Within the next few years, there were randomized controlled trials using drug resistance testing to guide antiviral treatment decisions, with positive results. Among the most prominent were the GART[10], VIRADAPT[11] and Havana[12] trials, which explored the clinical utility of HIV-1 genotyping and expert advice. As these clinical trials progressed, they addressed the issue of the role of expert interpretation of resistance test results.

It soon became clear that specific primary resistance mutations such as M184V and K103N clearly conferred high level resistance the picture was much less clear, especially for protease inhibitors; HIV drug resistance was going to require ongoing refinement. As our knowledge of resistance patterns evolved, especially with the newer PIs, expanding lists of resistance-associated mutations are being developed, with a sliding scale of drug effectiveness depending on how many of these mutations, and which ones, are present. In 2003, the approval of the first entry inhibitor, enfuvertide (T-20)[13], gave new life to those who had run out of options and time to wait for new drugs to become available.

EVOLVING CONSIDERATION OF SUBTYPE INFLUENCE

The general feeling among experts today is that although there are differences among HIV subtypes, the effect of subtype is not usually major and it is NOT usually the key factor in deciding on drug therapy. Furthermore, subtype considerations should never prevent access to drug despite the unknowns: Recently, questions have been raised about non-subtype-B virus and data on the effect of different subtypes on HIV drug resistance has been increasing. It is important to note that modern assays produce relatively reliable results for most subtypes with the possible exception of subtype D and are predictive of clinical disease similar to subtype B.  At the recent Conference for Retroviruses and Opportunistic Infections (14^th CROI 2007), two presentations addressed this topic: Data from a cohort of Kenyan women followed from the time of HIV-1 acquisition, infection with HIV-1 subtype D was associated with a faster rate of CD4 decline and a >2-fold higher risk of death than with subtype A infection, in spite of similar HIV-1 plasma viral loads[14].  In another presentation, Jonathan Schapiro discussed how individual subtypes may differ in their susceptibility to specific drugs, frequency of resistance pathways usage, selection of unique (non-B-related) mutations, as well as influence of specific resistance patterns. The clinical effect of these findings will remain unclear unless there are large collaborative efforts to pool data to facilitate further progress[15].

TRANSMITTED DRUG RESISTANT HIV (TDR)

 Soon after the awareness of viral resistance, the question was raised as to whether viral resistance could be transmitted along with HIV infection. The answer soon proved to be “yes.” This answer was complicated by the fact that in many cases, transmitted resistance can fade over time. However, this is a very different situation from the “decay” of resistance selected by antiviral drugs that are then discontinued.

Initially, the transmission of resistant virus was feared to be a rapidly growing threat. Indeed strains resistant to multiple drugs and even drug classes have been encountered. Data suggest transmitted drug resistance poses an especially relevant threat to patients starting NRTI + NNRTI regimens. Impact on patients starting NRTI + boosted PI regimens is less clear. From a more global perspective, although the rate of TDR does not appear to be increasing in many areas, it may indeed increase in resource-limited settings. In the past, the highest rates of TDR were seen to occur in the NRTI class until the rates in the NNRTI class increased a few years ago, an alarming—but expected—event. Thus, in many cases the absolute rates of NRTI and NNRTI resistance do not differ that much. Transmitted drug resistance as studied by Susan Little was in the range of 11-15%, which is above the threshold of 4-5% at which primary resistance testing is considered cost-effective. Genotypic testing is suggested due to faster turnaround time[16].  Little reported that resistant virus persisted about 2 years before there was any evidence of the emergence of wild type virus. Interestingly, transmission of drug resistant virus was not associated with decreased replication capacity. Patients with drug resistant virus overall did not have viral loads significantly different from patients with wild type virus. Patients with resistant virus took longer to achieve complete suppression; however, most patients in this study began treatment with more than 3 active agents, making it difficult to assess these results. Taken together with other research studies, it does not currently appear that the transmission of resistant virus is on a one-way upward slope. Part of the initial high rate may be due to the practice of sequential monotherapy where physicians quickly changed one medication at a time when viral loads increased. This is much less common today, although there are some areas and circumstances where rates remain substantially high: (a) settings where patients are not well monitored, (b) areas where rates remain substantially high, e.g., San Francisco.

NEWER CONSIDERATIONS

 Newer drugs appear able to exert some control of HIV despite the presence of a few—or even several—resistance mutations. Pharmacology and toxicity profiles often determine how many mutations it will take for HIV to become resistant. For example: tipranavir[17] is an elegant antiviral molecule active against virus with multiple PI mutations. obtained by boosting with ritonavir[17],[18], Duranavir[18] is another recently approved example. Some data suggests cross-resistance between these two compounds is not absolute, and after tipranavir failure duranavir retains activity[19]; the higher the PI level, the greater its ability to overcome mutations. Here pharmacology and toxicity come in to play: If you could achieve a drug level that is twice as high without toxicity, you would probably need far more mutations to select for enough resistance mutations to affect its efficacy. Some of the newest drugs in development are said to be “resistance resistant” and feature flexible molecular scaffolding that the manufacturers claim help them adjust to resistance mutations and continue to bind the antiviral medications. For example, Joe D. Bauman presented a cogent structural rationale for the superiority of the NNRTI TMC278 with respect to drug resistance[20]; TMC278 retains activity in vitro against NNRTI-resistant HIV-1 strains, including L100I/K103N and K103N/Y181C double mutants, which are resistant to all approved NNRTIs. And so the concept of structural flexibility in overcoming the effects of drug-resistance mutations evolved from systematic structural studies throughout the drug-development process.  In order to achieve these results, laboratory studies created HIV RT mutations based on other HIV-1 strains, and then built and rebuilt RT inhibitor molecules to select drugs with enhanced contacts to the viral enzyme. By creating pure crystals of drug and RT enzyme in the lab, the scientists could take pictures of the drug binding the enzyme at a resolution of 2.0-Angstroms (two 10-millionths of a millimeter). This process, called "crystal engineering" by the scientists, allowed the design of an RT inhibitor with a finger-like structure to fit precisely into a tunnel-like groove in the RT enzymes active site, an interaction that is difficult for an enzyme mutation to break up. Although TMC278 seems, preliminarily, at least the equal of efavirenz[21] with perhaps fewer side effects, the "resistance-resistant" character of this new candidate will require more clinical validation[22].

Another research team led by John Erickson has claimed the identification of a specific constellation of interactions that is a necessary condition for endowing PIs with the property of being “resistant-repellant”. They have incorporated these findings into a predictive algorithm for the efficient design of resistant-repellant PIs and believe that their algorithm is applicable to other drug targets[23]. More validation will also be required to move this claim into reality.

Another consideration is “clinical cutoffs.” Where at first it was conceived that viral resistance was virtually all-or-nothing, it is now clear that there are degrees of resistance. Based on clinical outcomes, levels of “fold change” have been defined that correlate to lower clinical cutoff (sometimes defined as 20% reduction in antiviral activity) or higher clinical cutoff (sometimes defined as 80% reduction in antiviral activity).  

The odds of resistance are significantly higher in patients with a history of antiretroviral drug use, advanced HIV disease, higher plasma HIV viral load and lowest CD4 cell count by self-report, and has significant implications for HIV treatment and transmission[24]. See Figure 1.

LESSONS LEARNED

A clear lesson from this history is the folly of adding just one active agent at a time. The probabilities for viral control are much greater when two or more antiviral agents are added at the same time. This knowledge has given rise to some exciting developments: first, some of the newest clinical trials and/or expanded access programs permit the use of more than one experimental agent at a time, rather than restricting participants to a single new agent. Second, resistance testing has been shown to be a key factor in limiting the impact of HIV drug resistance; testing of HIV strains for drug sensitivity allows the clinician to choose the drug or drug combination that will have the greatest chance for success, thus providing the most durable reductions in viral load. When all else fails, some clinicians are actively using the strategy of  a “holding regimen” where HIV viral load is at least partially controlled, and waiting until new agents are approved so that two new agents may be used together. This has most often been practiced by discontinuing protease inhibitors—whose continued use may continue the accumulation of resistance mutations—but continuing NRTI therapy where viral resistance has already been extensively selected and further mutations are less likely. And finally, although we still have a lot to learn about HIV viral resistance and how to interpret its test results, patient outcomes are clearly much better than they were just a couple of years ago.