Mechanisms of HIV-1 drug resistance

Introduction HIV-1 drug resistance is a result of mutations occurring within the wild-type viral genome that produce variants capable of efficiently replicating in the presence of antiretroviral agents. The first report of HIV-1 drug resistance was to zidovudine (ZDV) in 1989 [1]. Subsequently, drug resistance to all therapeutic antiretroviral agents has been observed (for a review, see [2]). Although highly active antiretroviral therapy for the treatment of HIV-1 infection has produced substantial decreases in morbidity and mortality in recent years [3,4] (for a review, see [5]), as many as 50% of patients fail therapy within 1 year of initiation [6]. As drug-resistant HIV-1 variants are often selected during the course of antiretroviral therapy [2], drug resistance is considered a major contributor to treatment failure. Furthermore, cross-resistance has a considerable negative impact on future treatment options. To address this, current consensus guidelines recommend the use of HIV-1 drug resistance testing as a clinical management tool to help guide the choice of new regimens, particularly after treatment failure and for guiding therapy for pregnant women [7,8]. The high rate of treatment failure due to the evolution of drug resistance mutations makes the development of new drugs that are active against resistant variants of paramount importance. Another resistance issue that has recently received considerable attention is the transmission of drug-resistant HIV-1 in primary infection. Wide ranging prevalence estimates from 1 to 26% [9-13] have led to considerable debate as to the seriousness of this problem. Unfortunately, comparison between studies is difficult due to a lack of standard procedures for defining resistance. A re-evaluation of previously published data using uniform criteria found that the prevalence of HIV-1 resistance among treatment-naïve subjects ranged from 1 to 11% [14]. In an effort to further clarify this issue, we examined over 1000 HIV-1 isolates in plasma samples collected from treatment-naïve individuals in the United States, Germany, Canada, and South Africa. Based on these samples, in the large majority of cases (approximately 97.5%) we found that the ranges of phenotypic drug susceptibility were < 2.5-fold to 4.0-fold, < 3.0-fold to 4.5-fold, and < 5-fold to 10-fold decreases in susceptibility to five protease inhibitors (PI), six nucleoside reverse transcriptase (RT) inhibitors, and three non-nucleoside reverse transcriptase inhibitors (NNRTI), respectively [15]. These data suggest that transmission of drug-resistant HIV-1 is still relatively rare. Nevertheless, at least in developed countries where antiretroviral drugs are widely available, the transmission of drug-resistant HIV-1 will probably increase. Given the fact that individuals are being treated for longer and longer, there is likely to be continuing evolution of virus and increases in transmission of resistant virus. In addition, as the use of antiretrovirals increases in the less developed countries, drug resistance issues are likely to become ever more acute. Defining phenotypic drug resistance: cut-off values Various criteria have been used to define drug phenotypic resistance. Most typically, 2.5-fold, 4-fold, or 10-fold decreases in drug susceptibility have been considered indicative of drug resistance [16,17]. However, these values have largely been empirically derived and based on inherent variation in the susceptibility assay (the 'technical' cut-off). There has been recent debate regarding the relevance of cut-off values currently in use, particularly with respect to the dideoxy nucleoside drugs. These cut-offs are usually the same value for each drug tested and are not established by clinical criteria. We now have good evidence that arbitrary cut-offs do not accurately reflect the range of drug susceptibility found in virus from treatment-naïve individuals for the currently available antiretrovirals. Rather, the 'normal' ranges of drug susceptibility vary from 2.5-fold for saquinavir and amprenavir to 10-fold for delaviridine [15]. These 'normal' ranges can thus be used to define an individual 'biological cut-off' for each drug. The derivation of biologically relevant cut-offs is a significant advance and should help in the interpretation of susceptibility data. However, it does not answer the ultimate question of whether a patient is likely to respond to a particular drug. The development of clinical cut-offs, which define the resistance level above which a patient is no longer likely to respond to treatment, is considered the next step forward in the interpretation of resistance information. Not surprisingly, this is currently the subject of numerous studies relating changes in susceptibility, as measured by phenotypic resistance tests, to virological response. Although the availability of biological and clinical cutoffs will undoubtedly aid the physician in making informed treatment decisions, it is important to appreciate that a single cut-off for a drug derived from one particular clinical study is unlikely to be broadly applicable to all treatment situations and all patients. Viral fitness The term viral 'fitness' usually refers to the relative replicative capacity of a particular HIV-1 variant in a particular environment. In untreated patients, the predominant, fittest virus is the 'wild type'. However, the combination of a rapid replication rate and the lack of a proofreading system for correcting replication errors mean that a large subpopulation of genetic variants exists at any one time as quasi-species. In the presence of antiretroviral agents, the predominant variant evolves through selection of the fittest species. The rate at which drug-resistant variants arise is related to the virus replication rate, and evolution of resistance mutations is minimized by combination therapies that are potent inhibitors of viral replication [2]. In vivo, viral load does not typically return to pre-therapy levels, suggesting that some residual antiviral activity remains and/or that resistant virus is not as 'fit' as susceptible virus. Indeed, studies of viral fitness have shown that the viruses initially selected are less fit than wild-type virus [18,19]. Thus, the study of viral fitness is enhancing our understanding of the evolution of drug-resistant HIV-1 [20,21]. A number of in vitro fitness assays have been described [19-22]; however, this approach only provides information about HIV-1 evolution in a fixed environment that may not accurately reflect the environment in vivo. Nevertheless, if in vitro HIV-1 fitness can be demonstrated to predict disease progression (or lack of progression), it may add value to viral load, CD4 cell count and resistance testing as a prognostic tool [23]. Genetic basis of drug resistance A comprehensive list of documented mutations that have been associated with HIV-1 drug resistance is available and periodically updated [24]. It is important to appreciate that not all of these mutations have been verified, via site-directed mutagenesis, to confer resistance to a specific drug or drugs. Such information is critical to build an accurate picture of those mutations that are linked to resistance, rather than genetic polymorphisms that do not play a role. It is generally accepted that while resistance to some drugs is conferred by a single point mutation [e.g., lamivudine (3TC) and NNRTI] [25,26], this resistance can be exacerbated by additional mutations and the causation of resistance to other inhibitors is highly complex. Given the ever-expanding list of resistance mutations, predicting phenotypic resistance from mutational patterns is far from straightforward. Resistance to nucleoside analogues The accepted view of resistance to nucleoside analogues has been that discrete mutations, or groups of mutations, confer specific resistance to the different nucleosides. For example, the group of ZDV resistance mutations (M41L, D67N, K70R, L210W, T215Y/F, 219Q/E) is distinct from the dominant 3TC resistance mutation (M184V). Indeed, these mutations may even cause interactions that result in the suppression of resistance to one drug. The most notable example of this is the effect of the M184V mutation on ZDV resistance in the context of ZDV resistance mutations [27]. However, recent observations have made this picture much less clear. These include the description of multi-nucleoside resistance (MNR) mutations and the broad effect of ZDV resistance mutations on resistance to other nucleosides. The first MNR mutation complex (A62V, V75I, F77L, F116Y, and Q151M) was described some time ago [28-30]. More recently, a second pathway to MNR has been described. This is the curious result of amino acid insertions (typically two) in the codon 68-70 region in the HIV-1 RT [31-33]. The insertions are quite heterogeneous in nature but are usually between codons 69 and 70, and commonly occur in a background of ZDV resistance mutations. This observation provided a clue that the effect of ZDV resistance mutations may not just be confined to ZDV. A second, unrelated observation added weight to this notion. Although much of the observed 3TC resistance can be attributed to the effect of M184V, a significant degree of low-level 3TC resistance had been seen in the absence of the 184 mutation. Interrogation of a large genotype-phenotype database with subsequent verification by site-directed mutagenesis confirmed that mutations E44D and/or V118I together with ZDV resistance mutations conferred this resistance [34]. The genetic basis of resistance to the dideoxy nucleoside analogue stavudine (d4T) has been the subject of considerable recent debate. It has been difficult to associate resistance to d4T with any specific mutations [35,36]. However, more recently, a picture has emerged whereby d4T resistance seems associated with ZDV resistance mutations (M41L, D67N, K70R, L210W, T215Y/F, 219Q/E) [30,36,37] in addition to both of the MNR complexes. Confusion regarding the contribution of particular mutations to drug resistance makes predicting d4T resistance from genotypic data particularly challenging. By developing a systematic method using artificial intelligence neural network systems, we have been able to identify a panel of 15 mutations associated with d4T resistance [38]. These include most of the ZDV resistance mutations, E44D,V118I and a number of other mutations that have previously been associated with resistance to various nucleoside analogues. More recently, it has been shown that a number of the nucleoside analogues, including d4T, become resistant as a result of selection of the 44/118 mutations in the context of ZDV resistance mutations (L. Romano, G. Venturi, S. Bloor, et al., manuscript submitted). Site-directed mutagenesis confirmed that 44D and 118I could decrease phenotypic susceptibility not only to 3TC, but also to most nucleoside analogues, particularly d4T and abacavir (L. Romano, G. Venturi, S. Bloor, et al., manuscript submitted). Thus, substitutions at RT codons 44 and 118 in the context of ZDV resistance mutations have broad implications for nucleoside RT inhibitor resistance. In fact, it appears that rather than ZDV resistance mutations (so-called 'thymidine analogue mutations') being restricted to conferring ZDV and d4T resistance, it is the co-selection of other mutations such as 44/118 that results in much broader nucleoside resistance. This set of ZDV resistance mutations plus associated mutations should perhaps more accurately be referred to as 'nucleoside associated mutations'. Resistance to NNRTI It is generally accepted that there is broad cross-resistance among the NNRTI class due to the selection of common mutations in the RT (for a review, see [39]). Furthermore, NNRTI resistance mutations do not appear to have significant effects on viral fitness, and a single point mutation can produce a highly resistant variant of similar fitness to wild-type virus (for a review, see [40]). However, there are exceptions. Abnormalities in RNase cleavage are commonly found in HIV-1 variants harbouring NNRTI resistance mutations [41,42] that, in some cases (e.g., V106A or P236L mutations), cause significant reductions in fitness [41]. In contrast, a particular NNRTI resistance mutation (M230L) has been associated with dose-dependent stimulation of HIV replication [43,44]. Interestingly, another new study has shown that mutations in the nucleoside binding pocket may interfere with MNR, producing a partial reversal of the MNR phenotype [45]. Until recently, all NNRTI resistance mutations have been found within specific regions of the HIV-1 RT (notably, codons 98-108 and 179-190). In addition, a number of delaviridine resistance mutations are located in the RT codon 230 region. However, a newly described NNRTI resistance mutation (Y318F) is a notable exception [46]. Like the 44/118 nucleoside mutations, this mutation was also discovered through a process of database interrogation and the analysis of site-directed mutant variants. Y318F alone has a significant effect on delaviridine susceptibility but alters susceptibility to nevirapine or efavirenz only in combination with other NNRTI resistance mutations. Specifically, in the presence of K103N,Y318F appears to significantly enhance the degree of resistance to efavirenz [46]. Resistance to PI Multiple mutations in the HIV-1 protease have been associated with PI resistance [24]. Furthermore, crossresistance within the PI class of antiretroviral agents has been described for some time (for reviews, see [47,48]). PI cross-resistance frequently involves combined mutations at codons 10 and 90 with at least five additional protease mutations emerging as drug-resistant HIV-1 evolves [49]. The degree of cross-resistance varies with the number and type of mutations. For example, recent studies have shown that HIV-1 isolates from patients initially receiving nelfinavir were less likely to be cross-resistant to other protease inhibitors than HIV-1 isolates from patients treated with indinavir [50]. Recent interest has focused on the patterns of mutations in the protease that confer resistance to the newer PI such as lopinavir. By comparing lopinavir phenotypic susceptibility to patterns of mutations, it has become clear that lopinavir shares many mutations with previously approved PI such as indinavir and ritonavir [51]. In addition, we have also used the neural network approach to identify a panel of 28 protease mutations that are associated with lopinavir resistance [52]. This study confirmed the important role of many familiar PI resistance mutations for lopinavir resistance. Furthermore, additional genetic changes in the protease, such as 55R, 85V and 95L, were also identified as substitutions that, in concert with established PI mutations, are likely to enhance the degree of resistance to lopinavir. Although many preexisting HIV-1 clinical variants have been identified that are cross-resistant to lopinavir, the precise pattern of resistance mutations selected during initial clinical use of Kaletra (lopinavir combined with low dose ritonavir) still remains to be determined. The association between HIV-1 protease resistance mutations and genetic changes in Gag cleavage sites is still of considerable interest. It is assumed that these cleavage site changes occur in response to subtle changes in the substrate specificity of proteases containing drug resistance mutations [53]. Such protease enzymes have been associated with impaired replicative capacity due to multiple defects in the processing of Gag and Gag-Pol, that subsequently lead to Gag precursor maturation defects [54]. The, subsequent amino acid substitution at a number of Gag cleavage sites is presumably primarily driven by antiretroviral drug pressure [55]. In some cases [56], but not all [56,57], these mutations at least partially compensate for the impaired replicative capacity of PI-resistant viruses. However, it is clear that Gag adaptation does not have an impact on the susceptibility of HIV-1 to protease inhibitors or clinical progression of HIV-1 disease, at least in the short term [58]. Interestingly, in addition to protease processing defects, PI-resistant viruses have been reported to also exhibit abnormalities in protease processing of RT [59]. Biochemical mechanisms of nucleoside analogue resistance There is increasing biochemical evidence for two distinct mechanisms of resistance to nucleoside analogues. First, there are mutations that interfere with the incorporation of nucleosides by HIV-1 RT (for a review, see [60]). Second, at least with ZDV-resistant variants, there is increased pyrophosphorylysis (the back-reaction of nucleotide incorporation) that increases the capacity of mutant virus to remove the chain-terminating nucleotides from newly synthesized viral DNA, thus rescuing DNA synthesis [61,62]. Recent studies have demonstrated that several ZDV resistance mutations enhance adenosine triphosphate (ATP) binding so that excision of chain-terminating nucleotides is relatively efficient. The incorporated ZDV monophosphate is removed from the growing DNA chain by combining with ATP, and thus ATP appears to function as a 'recipient' of the reverse polymerase reaction. This mechanism also appears to occur with the MNR insertion mutants. In particular, insertion of two serine residues between codons 69 and 70 appears to be critical for enhanced ZDV resistance due to rescue of ZDV-terminated DNA synthesis [63]. It has been proposed that the specificity of this mechanism for ZDV excision is due to the specific interaction of the region around the HIV-1 RT active site with the azido group of ZDV-triphosphate [64]. However, recent data indicates this mechanism may also contribute to d4T resistance [65]. Interestingly, it was found that RT variants harbouring ZDV resistance mutations showed a degree of d4T resistance at the enzyme level mediated via pyrophosphorylysis [65]. Of note, the re-sensitization of some ZDV-resistant variants to ZDV, by the addition of a single mutation (M184V) that confers resistance to 3TC [27], may be explained by severely compromised rescue of ZDV-terminated DNA synthesis by these variants [62]. Resistance to drugs in clinical development There are numerous antiretroviral drugs currently in various stages of clinical development. These include new-generation inhibitors of RT and protease, in addition to inhibitors of new targets such as integrase and gp41 (fusion). Gaining insight into the resistance profiles of these new inhibitors is an essential aspect of their development. For example, it is important to determine the likely degree of cross-resistance of a new PI to strains that are resistant to existing PI in clinical use [66]. With inhibitors of new targets such as integrase, it is important to assess the ease of development of resistance and the patterns of mutations responsible for conferring resistance. To this end, a combination of in vitro cell culture selection studies, together with the assessment of susceptibility of panels of pre-existing resistant variants, are both important activities during the development of new antiretrovirals. The following summarizes the status of a number of promising new inhibitors, but is by no means a comprehensive survey. Nucleoside analogues 1-β-D-2,6-Diaminopurine (DAPD) is a dioxolane guanosine nucleoside analogue that is deaminated to 1-β-D-dioxolane guanosine (DXG). In the triphosphate form, DXG is an inhibitor of HIV-1 RT [67]. This inhibitor has stimulated considerable interest as it appears to inhibit many strains of nucleoside analogue resistant HIV-1 [68]. These include viruses with standard ZDV resistance mutations and viruses with MNR codon 69 insertion mutations. However, virus harbouring the MNR codon 151 cluster is resistant to DAPD. Cell-culture drug passage studies have resulted in the selection of virus containing either the L74V or K65R mutations in RT [68,69]. Similar studies have been performed with the nucleotide analogue, tenofovir [70]. In vitro passage studies also resulted in the selection of the K65R mutation in RT, which caused a fivefold decrease in susceptibility. Profiling studies using preexisting nucleoside analogue resistant HIV-1 clinical strains revealed very little cross-resistance. In fact, substantial levels of resistance were only seen with mutant strains carrying the MNR codon 69 insertion mutations and, interestingly, not with the MNR codon 151 complex [71]. Viruses containing multiple ZDV resistance mutations showed a modest degree of resistance to tenofovir, although the additional presence of the M184V mutation appeared to abrogate this effect. Non-nucleoside reverse transcriptase inhibitors The development of new-generation NNRTI is focused on inhibitors that are able to suppress HIV-1 variants containing the common NNRTI mutations such as K103N. To this end, there are now a number of such inhibitors currently undergoing clinical evaluation that fit these criteria. These include capravirine (AG1549), TMC-120 and TMC-125. The development of resistance to capravirine during serial passage only occurred after the acquisition of multiple NNRTI resistance mutations [72]. Recombinant HIV-1 strains containing K103N, V106A, or L100I showed no significant reduction in susceptibility to capravirine. Furthermore, strains with multiple NNRTI mutations, including L100I/K103N or K103N/P225H, were also inhibited by capravirine. The activity of two more interesting NNRTI, TMC120 and TMC125, have recently been described [73]. Like capravirine, these inhibitors both generally caused the selection of multiple NNRTI mutations during in vitro passage. With TMC120, the Y181C/Y188L combination was observed and the L100I/Y181C combination was seen with TMC125. TMC120, a showed a high degree of activity against viruses harbouring single NNRTI mutations such as K103N, and inhibitors PI are also being developed that activity against common strains of PI-resistant is a PI that appears to in the active site of the protease A large panel of PI-resistant clinical isolates were for susceptibility to in to determine the degree of cross-resistance with approved PI Of isolates with 10-fold resistance to three or PI and an of PI mutations, susceptibility, had to 10-fold resistance, and only two had 10-fold resistance. of PI mutations, including primary mutations, were to confer even modest of resistance. More recent site-directed mutagenesis studies have been in an to identify specific of PI resistance mutations that confer resistance However, no common of the PI resistance mutations were identified that conferred resistance. passage of a pre-existing PI-resistant but clinical resulted in the selection of resistance. of the genotypic changes that with resistance was an substitution in the mutant virus protease is an HIV-1 PI that also and is currently undergoing clinical evaluation In vitro passage of wild-type virus showed that selected for resistant variants, more than nelfinavir or analysis of variants that an protease substitution was common during the selection The to resistance was distinct for each of three strains suggesting there are multiple to resistance. HIV-1 variants selected for resistance to the approved PI showed patterns of cross-resistance to The and variants but and viruses modest of resistance. A recent study using panels of PI-resistant clinical isolates that of with cross-resistance to at least three approved PI A further two and have recently been described that also to the of HIV-1 variants that are resistant to existing PI These inhibitors the replication of mutant strains containing three to five substitutions that resistant to other A panel of PI-resistant isolates that three to PI mutations were also tested for susceptibility to these new The value for these inhibitors was in the range which was significantly than the values for the approved PI ranged from for lopinavir, to 1 for of other targets Although the HIV-1 integrase is essential for viral it has difficult to integrase inhibitors that are also active against the virus in cell a group of acid inhibitors of HIV-1 integrase was described that have antiviral activity due to their effect on The antiviral activity of these is due to the of which is one of the two of the of the of of these inhibitors from the that serial passage of wild-type virus in cell culture resulted in the selection of resistant virus harbouring specific mutations in the integrase region. The single mutations and all appeared to cause a modest reduction in susceptibility to these In some of these mutations resulted in even resistance, although there was evidence for impaired of these variants. Another essential function for HIV-1 replication is cell mediated by the gp41 of the viral This for has been using of the region of the gp41 HIV-1 variants with resistance to the first of these inhibitors, have been selected by serial passage in cell culture analysis of the resistant isolates that two codon changes and within the of the gp41 could be responsible for the observed resistance. Site-directed mutagenesis studies confirmed that these mutations were for development of resistance to the Of it was confirmed that mutations appeared in in the same during clinical with A has been for cross-resistance to patient It was found that was able to inhibit the replication of suggesting that the resistance profiles of these inhibitors may be distinct It is clear that the HIV drug resistance to at a rapid new mutations are discovered and that appear in response to antiretroviral The picture is now emerging that the virus may a these mutations are selected to the effects of ever more potent the complex between the development of resistance, viral and impaired replication capacity is an that will be in in the the clinical relevance of relatively subtle changes in viral replication in cell culture due to resistance development will be an issue for some time to The of resistance mutations has been a with the mechanisms by which the virus is drug This is the for a number of the nucleoside analogue RT in our understanding of nucleoside resistance have been made recently, by a combination of complex studies, and the of artificial intelligence such as neural This to much more in the future as to which mutations are and the HIV-1 RT resistant to these As the number of new antiretrovirals to the capacity of HIV-1 to and their effects seems to have no We have that it is to new drugs against targets that are able to inhibit virus resistant to drugs in clinical As new targets are we new mutations and new resistance The that future antiretroviral drug development include the study of drug resistance has been more we will to the understanding of drug resistance mechanisms to resistance testing in the clinical In this we can that therapy is and with the evolution of the virus.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.