1. Home>
2. Specialties>
3. HIV/AIDS Clinical Care>
4. Feature

Clinical Implications of Resistance to Antiretroviral Drugs

The treatment of HIV infection is one of the most rapidly evolving fields in medicine. Advances in basic research, the development of new technologies for monitoring therapy, the continuous introduction of new drugs into clinical practice, and the dissemination of the results of recent trials all raise expectations and suggest new strategies for optimal management of HIV disease. It is clear that, as more antiretrovirals become available, clinicians will face more and more complex decisions on when to start antiretroviral treatment and with what combinations and, in the long term, how to define therapeutic failure.

Although analysis of the relation between drug resistance and treatment failure has in the past been complicated by concomitant interaction with other virological factors, the emergence of drug resistance appeared, until a few years ago, to be the inevitable consequence of all antiretroviral treatments and the major cause of therapeutic failure. Because of expanded knowledge of the dynamics of HIV replication that has become available within the last two to three years, we now understand that the rapid emergence of resistance is a direct consequence of the incomplete viral suppression obtainable with single or double nucleoside therapy.

Indeed, in just a very short time, new virological concepts have emerged and results from trials with potent combinations have demonstrated that drug resistance can be at least delayed, if not completely overcome, by appropriate treatment strategies.

Definition and Measures of Resistance

Drug resistance can be more appropriately termed "altered drug susceptibility." It is a phenomenon that can occur in vivo and in vitro, in response to the exposure of the HIV virus to a drug or to a combination of drugs. The phenotypic resistance of HIV virus can be assessed in vitro. It is roughly represented by the need to use increased concentrations of drugs (inhibitory concentration, or IC) to inhibit the growth of the resistant virus, as compared with the concentrations generally needed for the inhibition of a "reference" virus (or the "wild" virus isolated from the patient before the start of a particular drug regimen).

Phenotypic susceptibility is usually quantified in terms of IC₅₀ or IC₉₀ (a measure of the concentration of drug needed to inhibit 50% or 90%, respectively, of viral growth). If the IC₅₀ (or IC₉₀) characteristic of the so-called wild-type virus is known, the IC₅₀ (or IC₉₀) for a resistant virus will be X-fold greater. The increase in IC₅₀ needed to define a virus as resistant to a particular drug is often established empirically. For example, a virus highly resistant to AZT is assumed to have an IC₅₀ of at least greater than 1.00 uM, whereas wild-virus generally has an IC₅₀ of 0.01uM to 0.05 uM.

The molecular mechanism underlying phenotypic susceptibility can be identified in changes in the sequence of the gene coding for the enzyme target of our antiviral intervention (so far, the reverse transcriptase or the protease). The characteristic genetic changes (at the codon level) that predict the subsequent resistant phenotype are referred to as genotypic resistance.

Determination of phenotypic resistance (assessment of the drug susceptibility of the virus harbored by an HIV-infected individual) can be accomplished by growing out the virus (preferably isolated from plasma, which has a better representation of the actively growing virus population) in the presence of various concentrations of the drug under study. However, because phenotypic resistance is the consequence of specific mutations in target genes (e.g., the reverse transcriptase gene or the protease gene), PCR and gene-sequencing methods to directly detect these mutations have been developed to analyze genotypic resistance. Moreover, although phenotypic and genotypic resistance are often found on analysis to be directly related, in some cases the relation is altered by the fact that the emergence of resistance is a dynamic process, and multiple strains of virus, with various sensitivities, often coexist in an organism.

A rather controversial issue is whether determination of the resistance pattern in patient isolates would be useful in current clinical practice, now that a number of relatively easy and reproducible techniques (at least for viral genotyping) have become available. In a case of documented antiviral therapy failure in a patient taking a three-drug combination, identifying the drug to which the virus has become resistant could aid the design of salvage therapy and prevent the need to change all three drugs. This may be considered a very reasonable option, if we consider the scarcity of available therapeutic options. However, some experts feel strongly that all components of a failing combination should be changed; this strategy would not benefit from the availability of quick genotypic resistance testing.

Current data suggest that the evaluation of drug resistance in individual patients might best be used in the near future as a data point collected before starting treatment, to be used as a baseline against which to measure efficacy of therapy. In other words, treatment could be planned on the basis of baseline susceptibility or cross-resistance profiles. In fact, mutations that confer resistance to nucleoside analog reverse transcriptase inhibitors, nonnucleoside reverse transcriptase inhibitors, and protease inhibitors, have all been identified in persons who have never been treated with antiretroviral drugs, perhaps as a consequence of transmission of already resistant strains, or of natural polymorphism. With the wider and wider use of antiretrovirals, pretreatment resistance will probably occur with increasing frequency.

General Mechanisms of Resistance

HIV variability represents the most important factor in the emergence of resistant strains. The high rate of HIV replication throughout the course of the infection and the occurrence of many mutations during each replication cycle because of the inaccuracy of reverse transcriptase (a phenomenon common to all single-stranded RNA viruses) are the basis for the emergence of drug-resistant variants under the selective pressure of antiretrovirals. In fact, a Darwinian model could be applied to HIV dynamics, with the continuous production of variants and the continuous selection of the "fittest" virus.

With the daily production of perhaps 109 to 1010 virions, a mutation rate of 3 x 10³ per nucleotide per replication cycle, and the HIV genome being approximately 10,000 nucleotides in length, any single mutation could exist before any drug is introduced. The relative levels of mutants is probably determined by three concurrent factors: the forward mutation frequency (i.e., the number of copying errors on a particular codon), the cost of the mutation (i.e., the replicative capability of the mutated virus), and the age of the quasispecies (i.e., how long ago, in the individual patient, the viral population with a particular resistance mutation was generated).

It is clear today that the appearance of genetic variants is a function of the number of cycles that take place during infection, and that combination therapy that suppresses HIV replication to undetectable levels can delay or prevent the emergence of resistant strains. In fact, variants that are resistant to three drugs (i.e., that harbor many contemporary mutations) are unlikely to pre-exist, the development of new mutations depends on "new viral cycles," and, often, high-level resistance requires the presence of multiple mutations. Therefore, we may conclude that the more effectively HIV is suppressed the fewer opportunities there will be for new mutations to emerge. That is why it is recognized that the goal of antiretroviral therapy should be to suppress HIV replication at least to levels undetectable by the more sensitive HIV RNA assays.

However, because this goal might not be attainable in all patients, the tendency of HIV to generate drug-resistant variants may remain the major factor limiting the ability of antiretroviral therapy to fully control HIV replication and completely reverse the natural history of the disease.

Resistance Patterns

Nucleoside Analogues

Decreased susceptibility to AZT in clinical isolates of HIV variants were first reported in 1989. This was followed in subsequent years by reports of resistance to other compounds of the same class. In general, advanced-stage disease, low baseline CD4-cell count, and high HIV RNA plasma level strongly predicted the development of resistance. For AZT, re- sistance appears to be the consequence of a stepwise accumulation of mutations, at codons 215, 70, 41, 67, and 219. (The same phenomenon has been seen with some of the protease inhibitors.) For other drugs, such as didanosine and zalcitabine, the mechanisms and the mo-lecular correlates of resistance are less clear, although a number of mutations responsible for a reduced susceptibility have been identified.

The clinical significance of resistance to some dideoxynucleosides is still not completely defined. High-level resistance to didanosine, stavudine, or zalcitabine is very difficult to document. However, cross-resistance has been reported between AZT and other azido-nucleosides, ddC and ddI/3TC (with the 65 and 184 mutations involved), ddI and ddC (codon 74), d4T and ddI/ddC (codon 75), and between 3TC and ddI/ddC (codon 184). Multiple resistance to nucleoside analogues has also been observed after long-term combination therapy using these agents. The mutations mainly responsible for multiple resistance to AZT, ddI, ddC, and D4T include codons 75, 77, 116, 62, and in particular, 151.

During treatment with lamivudine (3TC), resistance occurs rapidly in vivo, and is associated with a single substitution at codon 184. Although this mutation leads to high-level resistance to 3TC and to some cross-resistance to didanosine and zalcitabine, this codon change may antagonize AZT resistance mutations, leading to a restored phenotypic sensitivity to AZT. This mechanism, however, does not seem to be effective in all cases; dual AZT/3TC resistance has also been observed. Concerns have also been raised about the use of 3TC-containing regimens as first-line therapy because of the limitations it may place on subsequent nucleoside options, particularly ddI.

A new nucleoside analogue (1592U89), with antiretroviral potency that is apparently superior to that of other available compounds, is currently under clinical development. Although data are still scarce, the main mutations associated with decreased susceptibility to this compound seem to be associated with codons 184, 65, 74, and 115. No cross-resistance to AZT or d4T is induced. However, AZT/3TC-resistant viruses are also highly resistant to 1592U89. In general, for antiretroviral-experienced patients, the virological response to this new agent varies widely with the type of antiretroviral combinations previously used by the patient, suggesting that this promising new compound might perform better as first-line therapy than as salvage therapy.

NNRTIs

The very rapid development of resistance to non-nucleoside reverse transcriptase inhibitors, whether used in monotherapy or in double combination, suggested, until a few months ago, only a limited clinical utility for this class of compounds. However, the results of two trials (INCAS and ISS-047 studies) in which the NNRTI nevirapine was used in triple-combination regimens with nucleoside analogues, have shown that resistance to NNRTIs can be significantly delayed if viral load suppression is obtained and sustained. This is important for two reasons. One, it gives rise to the possibility that NNRTIs may be incorporated into clinical practice, adding a new class of agents to the clinical armamentarium. Two, these trial results confirm, with a class of compounds other than protease inhibitors, the concept that resistance occurs as a direct consequence of viral replication.

Despite this, it is worth noting that NNRTIs are often associated by a common pattern of resistance, which may in some cases limit their sequential use. In general, although cross-resistance is a more common phenomenon among NNRTIs, with in vitro studies indicating several mutations shared by different compounds, some new NNRTIs (DNP-266 and MKC-442) show distinct resistance profiles that may make them suitable candidates for effective subsequent therapeutic regimens. Moreover, mutually counteracting mutations have also been detected among NNRTIs; the clinical correlates of these observations are being investigated.

Protease Inhibitors

Reduced sensitivity has been reported for all tested protease inhibitors. The patterns of mutations, however, appear to be more complex than for reverse transcriptase inhibitors, with a high natural polymorphism, a larger number of sites involved, and greater variability in the temporal patterns and in the combinations of mutations leading to "phenotypic" resistance.

Mechanisms conferring resistance to protease inhibitors are an exemplary model of the Darwinian dynamics of HIV resistance. Resistance patterns may evolve from mutations that reduce inhibitor-enzyme binding to mutations with "compensatory" activity, i.e., mutations that improve the "fitness" of the virus by compensating for the disadvantageous changes in the functionality of the protease enzyme. The compensatory mutations may include new changes in the protease enzyme, mutations that drive the increased production of the "less fit" enzyme, or even mutations that modify the protein cleavage sites.

As far as the individual protease inhibitors are concerned, the codon 82 mutation is the leading one in reducing sensitivity to indinavir, although high-level resistance to indinavir develops only as a consequence of multiple codon changes. Resistance to ritonavir also seems to occur as a consequence of the accumulation of different mutations, the most relevant being 82, 46, and 84, which are also common to indinavir, confirming the cross-resistance between these two compounds. Cross-resistance has also been reported between indinavir and saquinavir. Saquinavir may have a partially different resistance profile (the main mutations being at codons 48 and 90). However, results of a recent controlled trial seem to suggest that the virological response to indinavir is rather weak in patients switched to indinavir after treatment with saquinavir.

Because we have learned we should probably avoid changing from ritonavir to indinavir or vice versa, and avoid changing from long-term saquinavir to indinavir, it appears that the emergence of broad cross-resistance between protease inhibitors, which had been feared, is indeed becoming a problem, and is complicating the design of correct sequencing of protease inhibitors in case of therapeutic failure.

Nelfinavir seems to be characterized by a distinct genetic pattern of mutations, with codons 30, 35, 36, 46, 71, 77, and 88 most frequently involved. Because 60% of viral isolates from indinavir- or ritonavir-treated patients seem also to be resistant to nelfinavir, but, conversely, nelfinavir-resistant strains seem to retain sensitivity to other protease inhibitors, nelfinavir may be a candidate for first-line use in antiretroviral-naive patients. Changing from ritonavir or indinavir to nelfinavir should therefore be avoided, whereas change from nelfinavir to ritonavir might be acceptable.

Another compound under clinical development, 141W94, also seems to be characterized by a partially different resistance profile (codons 10, 46, 47, 50, 84), although data are mainly from in vitro experiments.

A final issue -- hypothetical and still unproven -- is the possibility of increasing efficacy by using protease/ protease combination regimens that would induce mutually counteracting, drug-induced mutations. This might convert the unavoidable selection of mutant viruses into an at least partially favorable phenomenon. This possibility is currently addressed by the development of a new compound (ABT-378) designed to act on ritonavir-resistant viruses.

Clinical Implications

The dynamics of HIV-1 replication in vivo strongly suggest that early, aggressive antiretroviral therapy is needed to minimize the negative consequences of HIV replication. Recent findings have shown that combinations of potent agents can reduce circulating free virus to undetectable levels in many individuals for prolonged periods of time. A theoretical possibility has also been raised that, in individuals treated at very early stages (i.e., before seroconversion), HIV infection might even be eradicated, because, after some years of theoretical "zero-replication," HIV infection might "burn itself out," as chronically infected cells die off. Long-term data are yet to come, however, and we know that HIV might survive, even in the presence of powerful treatment, within latently infected cells in tissue reservoirs such as lymphoid tissues, bone marrow, and other macrophage-rich tissues and organs, or in "sanctuary sites" inaccessible to treatment.

Indeed, the best opportunity to accomplish maximal suppression of virus replication and to minimize the risk of drug resistance is to use potent combinations in individuals with no prior history of antiretroviral use. However, very early intervention is, unfortunately, not possible for the majority of our patients. Careful therapeutic intervention should therefore be designed with the aim of keeping HIV viral load at undetectable or minimal levels indefinitely, to prevent disease progression and transform HIV disease into a chronic disease with minimal negative impact on the duration and quality of life of the infected persons.

With that as an aim, then, rational criteria must be adopted for 1) selecting drug combinations with the best chance of long-term efficacy, but also 2) preserving subsequent therapeutic options if the initial choice fails to achieve the desired results.

Because any choice has an impact on later options, therapeutic plan design should include, whenever possible, both initial therapy and predefined alternative antiretroviral regimens. Unfortunately, new resistance mutations continue to be discovered and many of the promising new drugs seem to be better for first-line use. Therefore, the issue of the many patients who are failing aggressive antiretroviral regimens remains unsolved.

As far as resistance is concerned, a number of rules should be followed in planning therapeutic strategies:

Clinical Guidelines

1. Sequential use of combination drugs that share clear cross-resistance phenomena should be avoided.

2. Potent antiretroviral drugs to which HIV readily develops high-level resistance should not be used in regimens that are expected to yield incomplete suppression of viral replication.

3. Decisions to alter antiretroviral therapy need to be made carefully, because the number of effective drugs available is still very limited. In fact, an increase in HIV RNA levels in persons receiving fully suppressive antiretroviral therapy can be due to a number of factors, one of which is lack of full adherence to a particular drug combination. In fact, the complexity of the regimens may jeopardize the expected benefits of potent combinations started early. The problem of drug resistance, far from being overcome, might dramatically strike back.

Dr. Vella is an associate editor of ACC.

— Stefano Vella, MD

Published in AIDS Clinical Care June 1, 1997

Citation(s):

Carpenter CJ et al. Antiretroviral therapy for HIV infection in 1996. Recommendation of an international panel, International AIDS Society-USA. JAMA 1996 276 146-154.

• Medline abstract (Free)

Carpenter CJ et al. Antiretroviral therapy for HIV infection in 1997: Updated recommendations of the International AIDS Society-USA panel. JAMA 1997 277 1962-1969.

• Medline abstract (Free)

Ho DD et al. Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1 infection. Nature 1995 373 123-126.

• Medline abstract (Free)

Mellors JW et al. Prognosis in HIV-1 infection predicted by the quantity of virus in plasma. Science 1996 272 1167-1170.

• Medline abstract (Free)

Perelson AS et al. HIV-1 dynamics in vivo: virion clearance rate, infected cell life-span, and viral generation time. Science 1996 271 1582-1586.

• Medline abstract (Free)

Richman DD. HIV therapeutics. Science 1996 272 1886-1888.

• Original article (Subscription may be required)
• Medline abstract (Free)

Shirasaka T et al. Emergence of human immunodeficiency virus type 1 variants with resistance to multiple dideoxynucleosides in patients receiving therapy with dideoxynucleosides. Proc Natl Acad Sci U S A 1995 92 2398-2402.

• Medline abstract (Free)

Vella S and Floridia M. Combination antiretroviral therapy. Baillieres Clin Infect Dis 1996 3 79-101.

Wei X et al. Viral dynamics in human immunodeficiency virus type 1 infection. Nature 1995 373 117-122.

• Medline abstract (Free)