December 14, 2006
In our armamentarium of highly active antiretroviral therapy (HAART) agents, non-nucleoside reverse transcriptase inhibitors (NNRTIs) are a potent and important class of drugs. Three licensed NNRTIs are currently available for use as HIV treatment: efavirenz (EFV, Sustiva, Stocrin), nevirapine (NVP, Viramune) and delavirdine (DLV, Rescriptor).
In addition, there is one agent (etravirine [TMC125]) that is in advanced clinical trials and expanded access, one agent (TMC278) that is in dose finding clinical trials and another agent (BILR 355 BS) that is in early clinical development. Some agents in the NNRTI class see little use (e.g., delavirdine) or are not considered "preferred" by current U.S. Department of Health and Human Services (DHHS) HIV treatment guidelines (e.g., nevirapine).1 However, efavirenz-based HAART is one of the most commonly used therapies in antiretroviral-naive individuals. Due to extensive data supporting the efficacy, tolerability and low rates of toxicity of efavirenz-based HAART, this regimen remains "recommended" by the International AIDS Society (IAS)-USA guidelines2 and "preferred" by the U.S. HIV treatment guidelines for the treatment of antiretroviral-naive individuals.
The data available on the newest agents in the NNRTI class, the experimental drugs etravirine and TMC278, indicate that they share some of the attributes of efavirenz; namely, they are effective and generally well-tolerated antiretroviral therapies.3,4 However, these second-generation NNRTIs appear to have potential resistance advantages over efavirenz and the other first-generation NNRTIs in that they remain active despite the presence of NNRTI mutations that lead to high-level virologic resistance to those agents. Etravirine and TMC278 are both being developed by the same company, and in most respects these agents are similar in their overall profiles. However, since etravirine is in the later phases of development, this article will focus on etravirine as well as issues regarding the use of first-generation NNRTIs, the evolution of NNRTI resistance and the integral role that second-generation NNRTIs may play in the future of HIV therapy.
Mechanisms of NNRTI Activity and HIV NNRTI Resistance
Adapted from François Clavel, M.D., et al.5N Engl J Med. 2004;350(10):1023-1035.
The NNRTIs primarily bind to codons 98-108 and 179-190 within the HIV reverse transcriptase hydrophobic pocket.6 Although efavirenz, nevirapine and delavirdine all bind in the same general area of the hydrophobic pocket, some differences exist in their specific codon interactions. As a result, on virologic failure some NNRTI mutations are more likely to occur with certain NNRTIs. K103N, for example, is typically associated with initial efavirenz resistance, while Y181C is associated with initial nevirapine resistance. These mutations, although considered to be drug specific, can still lead to significant cross-resistance among the three approved NNRTIs, discussed in more detail below.
Structural studies of etravirine, a diarylpyrimidine (DAPY) analog,6 indicate that it possesses several characteristics that enable it to bind in the hydrophobic pocket of the reverse transcriptase enzyme using multiple conformations and remain active against most HIV quasispecies with mutations that lead to high-level cross resistance among the first-generation NNRTIs.
These characteristics include:
Newly diagnosed, treatment-naive patients have been found to harbor antiretroviral resistance due to primary infection with drug-resistant virus. A number of recent studies determined that the prevalence of primary resistance in newly infected patients ranges from 7% to 16%.8-15 The prevalence of primary resistance shows variability over time, between different populations and between different geographic regions.
Some studies indicate that the prevalence of resistance in antiretroviral-naive individuals has increased in recent years. One large study identified 377 individuals who were newly infected with HIV between 1995 and 2000.16 A comparison of the prevalence of resistant virus in individuals infected from 1995 to 1998 with the prevalence for those infected from 1999 to 2000 revealed that the frequency of resistance-associated mutations increased from 8.0% to 22.7% and that the frequency of multi-class resistance increased from 1.1% to 6.2%.
Another study evaluated primary resistance among 949 antiretroviral-naive patients who were recently (i.e., within the past four to six months) or chronically (i.e., six or more months) infected with HIV in 10 U.S. cities.17 Resistance was fairly common in both groups, but significantly higher rates of NNRTI resistance and multi-class resistance were identified in newly infected patients.
The prevalence and persistence of drug-resistant virus in individuals with established disease have led to the DHHS recommendation in October 20061 that all HIV-infected, antiretroviral-naive patients be tested for resistance prior to starting antiretroviral therapy. Studies indicate that this approach is both cost-effective and improves treatment outcomes.18,19
Although there are no data yet describing the prevalence of primary second-generation NNRTI resistance in antiretroviral-naive individuals, given the low rate of use of etravirine and other second-generation NNRTIs, transmission of such primary resistance seems unlikely.
Efavirenz resistance. Due, in large part, to the widespread use of efavirenz-based HAART in the developed world, efavirenz-related NNRTI mutations are some of the most commonly encountered. Mutations leading to early virologic failure on efavirenz include K103N, Y188L and G190S/A; K103N is by far the most common and confers high-level efavirenz resistance.
Rates of efavirenz resistance can be seen in a few recent trials. The GS903 trial was designed to compare the safety and efficacy of two efavirenz-based regimens in treatment-naive individuals: efavirenz + tenofovir (TDF, Viread) + lamivudine (3TC, Epivir) and efavirenz + stavudine (d4T, Zerit) + lamivudine.20 The overall rate of treatment failure was low, involving about 15% of the 600 patients enrolled in the trial. However, among the participants who did not achieve complete viral suppression, the likelihood of selecting for efavirenz resistance (defined as the presence of K103N, V106M, Y188C/L or G190A/S/E/Q) was quite high at 55% and 49% for the tenofovir-containing and stavudine-containing arms, respectively. It was further noted that 76% of the individuals with efavirenz resistance had the K103N mutation present.
|Table 1. Development of Resistance Mutations Through Week 144 in GS903|
|d4T + 3TC + EFV|
(n = 301)
|TDF + 3TC + EFV|
(n = 299)
|n||% of Total||% of Failures||n||% of Total||% of Failures|
|Wild-Type or as Baseline||23||7.6||47||18||6.0||38|
Consistent with GS903 are the results from another study, known as GS934,21 that were presented by Damian J. McColl et al at the 8th International Congress on Drug Therapy in HIV Infection (Glasgow 2006), which was held in Glasgow, United Kingdom, from Nov. 12 to 16, 2006. GS934 compared two other nucleoside/nucleotide reverse transcriptase inhibitor (NRTI) pairs -- tenofovir + emtricitabine (FTC, Emtriva) versus zidovudine/lamivudine (AZT/3TC, Combivir) -- both of which were combined with efavirenz, in order to evaluate their efficacy. The 96-week results revealed that 71% and 74% of the participants who experienced virologic failure in the two respective arms developed efavirenz resistance, defined as the presence of K103N, A98G, K101E, K103E, V108I, V179D, Y188H, G190A/S/E, P225H or M230L. Moreover, 82% of these individuals harbored the K103N mutation.
|Table 2. Resistance Rates Among Virologic Failures in GS934|
|Patient Outcomes||TDF + FTC + EFV|
(n = 244)
|AZT/3TC + EFV|
(n = 243)
|n||% Total||% Gen*||n||% Total||% Gen*|
|Patients in Resistance Population||14||6||29||12||.017|
|Patients With Genotypic Data||14||6||27||11||.035|
|Wild-Type or as Baseline||4||2||29||7||3||26||.38|
|* % genotyped|
Nevirapine resistance. Studies have shown that patients who have incomplete viral suppression on nevirapine-based HAART usually develop the K103N mutation and/or the Y188C/L/H mutation. However, the V106A/M or Y181C/I mutation emerges as the initial NNRTI mutation in about 20% of patients.22
The Y181C mutation is the most common to develop in patients who receive nevirapine monotherapy (nevirapine monotherapy is typically used to prevent mother-to-child transmission of HIV) or in those receiving nevirapine in combination with stavudine rather than in combination with zidovudine (AZT, Retrovir), which is relatively uncommon today in resource-rich settings.23,24
Since efavirenz phenotypically remains active in the presence of Y181C, it was hoped that patients who began on an NNRTI-based regimen could begin with nevirapine in combination with stavudine, which would then allow them to be switched over to an efavirenz-based regimen following the development of the Y181C mutation and virologic failure. Unfortunately, clinical data failed to bear this strategy out. Several studies clearly demonstrated that switching patients to an efavirenz-based regimen following the development of Y181C while on nevirapine was ineffective, as resistance to efavirenz quickly developed, thereby leading to rapid virologic failure. Thus patients who develop the Y181C mutation while on nevirapine must be switched to another active agent, such as a boosted PI or etravirine.25-27
Delavirdine resistance. Although rarely used for treatment, the most commonly observed NNRTI mutations that develop during delavirdine failure are K103N and Y181C.28
Other first-generation NNRTI mutations. In addition to the mutations discussed above, all three approved NNRTIs may select for Y188L, which confers high-level resistance to all three agents when present as a single mutation. V106M also causes broad cross-resistance. This mutation is seen more frequently in HIV subtype C, which is predominant in Africa, India and Nepal, than subtype B, which is predominant in the Americas, Europe, Japan and Australia. A number of other NNRTI-associated mutations -- namely, L100I, V106A, Y181I, G190S/A and M230L -- reduce the efficacy of the approved NNRTIs when two or more are present in the same virus.5,6
Etravirine resistance. Since etravirine has rarely been used in a first-line treatment setting, the development of etravirine resistance in individuals treated with etravirine as their first NNRTI has not been well documented. Data from one in vitro serial passaging study found that etravirine selected for T39A, E138K, V179F, Y181C, L214F, F227L and M230I/L, with polymorphisms at E40K, K70R, Q91L, L109M, R125G, A158T, Q174P, G196R, N265T, D256G and E291K.29 Clonal analysis of HIV passaged in the presence of etravirine 360 nM identified nine of 24 clones containing the triple mutant E138K + V179F + Y181C, which showed decreased susceptibility to etravirine by more than 700-fold (50% etravirine inhibitory concentration = 514 nM). At 10 µM, etravirine selected for mutations at E138K, V179F, Y181C, L214F and M230L. The quadruple mutant V179F + Y181C + L214F + M230L showed decreased susceptibility to efavirenz and etravirine by 133-fold and more than 850-fold, respectively. Thus, etravirine appears to select for the V179F, Y181C, L214F and M230L mutations in vitro. When these mutations are present in combination, they confer high-level resistance to etravirine as well as cross-resistance to efavirenz.
Other studies of etravirine have demonstrated that the agent has an increased genetic barrier to resistance in comparison with the first-generation NNRTIs, since multiple mutations are required before decreased etravirine susceptibility is observed. For example, in another in vitro study,30 this one by Johan Vingerhoets et al, both wild-type HIV and NNRTI-resistant HIV were serially passaged in the presence of etravirine to identify possible mutations selected by the agent. The investigators found that the development of resistance to etravirine required at least two or three mutations, and these mutations frequently conferred cross-resistance to efavirenz and nevirapine. Furthermore, the investigators found that etravirine selected for known NNRTI-associated mutations (L100I, Y181C, G190E, M230L and Y318F) as well as novel NNRTI mutations (V179I and V179F). The ability of these individual mutations to produce etravirine resistance was highly dependent on the presence and identity of coexisting mutations. Thus, it appears clear that, at least relative to the first-generation NNRTIs, etravirine has a unique resistance profile and a higher genetic barrier to the development of resistance.
As discussed above, among the three licensed first-generation NNRTIs -- efavirenz, nevirapine and delavirdine -- there is considerable overlap among the major resistance mutations associated with each agent.31
As a result, aside from a few exceptions, selection of a single point mutation in the presence of one agent usually confers cross-resistance to all three approved NNRTIs.
Fortunately, the second-generation NNRTIs in development are able to maintain activity against HIV that has resistance to the first-generation NNRTIs. Koen Andries and colleagues evaluated the potency of etravirine against wild-type and resistant HIV strains carrying clinically relevant NNRTI mutations.33 They found that etravirine was highly active against a panel of 25 viruses carrying single and double reverse transcriptase amino acid substitutions associated with NNRTI resistance, including the double mutants K101E + K103N and K103N + Y181C. Etravirine also had activity against 97% of 1,081 clinically derived recombinant viruses resistant to at least one of the currently licensed NNRTIs.
Several clinical studies of etravirine that have been performed in HIV-infected individuals support the in vitro finding that etravirine retains activity in the presence of NNRTI mutations that would otherwise lead to complete cross-resistance among the first-generation NNRTIs. One of the initial studies to assess etravirine efficacy was TMC125-C207, an open-label, phase 2a study by Brian Gazzard et al.34 The study involved 16 HIV-infected individuals with documented efavirenz resistance (10-fold to 500-fold), all of whom were taking an NNRTI-containing (efavirenz or nevirapine) regimen. Patients received etravirine 900 mg twice daily in place of their current NNRTI for seven days. The median HIV RNA decline after seven days of etravirine treatment was 0.89 log10 copies/mL (Figure 7), with seven individuals achieving a viral load decrease of greater than 1 log10 copies/mL. In addition, there was no relationship between baseline genotype or phenotype and the response to etravirine. These preliminary data first established the clinical effectiveness of etravirine in patients who had resistance to first-generation NNRTIs.
In TMC125-C209, a 48-week, open-label trial, seven individuals with three-class experience received etravirine in combination with an optimized background regimen containing a protease inhibitor (PI).35 Results from the trial were presented by Adriano Lazzarin et al at Glasgow 2006. All the patients in the study were heavily NNRTI-resistant at baseline: 50%, 83% and 100% were resistant to delavirdine, efavirenz and nevirapine, respectively. One individual discontinued treatment at week 12, but all the others remained on therapy throughout the entire 48 weeks. At 48 weeks, HIV RNA levels decreased by a median of 1.4 log10 copies/mL, with 66% of patients achieving a viral load decline of at least 1 log10 copies/mL from baseline.
Finally, in the TMC125-C223 trial presented at the XVI International AIDS Conference in Toronto, Canada, by Cal Cohen (also one of the authors of this review),36 199 HIV-infected individuals with documented NNRTI resistance and at least three primary PI mutations were randomly assigned to receive:
Baseline data are shown below.
At week 48, the mean reductions in HIV RNA levels -- which was the primary endpoint of the study -- were 0.14, 0.88 and 1.01 log10 copies/mL in the active control, etravirine 400 mg twice daily and etravirine 800 mg twice daily arms, respectively. Both etravirine arms were statistically superior to the active control arm (Figure 9). In addition, 8%, 28% and 30% of patients in the active control, etravirine 400 mg twice daily and etravirine 800 mg twice daily arms, respectively, achieved a viral load below 400 copies/mL, and 0%, 23% and 22%, respectively, achieved a viral load below 50 copies/mL.
It was found that as the number of baseline NNRTI mutations increased, the effectiveness of etravirine decreased.
Thus, although etravirine is highly effective against HIV with one or even two first-generation NNRTI mutations, it is relatively less effective against HIV with three or more of these mutations. Clinicians will need to contemplate the level of NNRTI resistance when considering etravirine as a treatment option for patients who have experienced viral rebound while on a first-generation NNRTI.
Some evidence suggests that the first-generation NNRTIs have enhanced activity in the presence of thymidine analog mutations (TAMs) and some other select NRTI resistance mutations (Table 3 and Figure 11).37,38
|Table 3. RTI Susceptibility of Site-Directed Mutants|
|M41L + T215Y||2.0||2.6||11.2||1.5||0.9||0.3||0.5||1.0|
|M41L + M184V + T215Y||>100||6.9||1.7||1.3||1.3||0.1||0.3||0.4|
|5M + M184V||>100||3.4||4.8||1.4||1.4||0.4||0.3||0.3|
|M41L + A62V + T69SSA + T215Y||8.1||9.7||>1,000||8.0||2.6||0.2||0.3||0.6|
| Decreased susceptibility: >2.5-fold resistance
Hypersusceptible: <0.4-fold resistance
* 5M = M41L + D67N + K70R + T215Y + K219Q/E
Such NNRTI "hypersusceptibility" has been demonstrated to some extent to improve treatment outcomes when an NNRTI was administered after the development of NRTI resistance. In a study by Richard Haubrich et al, patients with and without efavirenz hypersusceptibility were treated with efavirenz-based HAART. As shown in Figure 12,39 at the end of two months of treatment, the patients with efavirenz hypersusceptibility had a significantly greater decline in HIV RNA levels compared to those patients without efavirenz hypersusceptibility; however, while overall suppression remained higher in the efavirenz hypersusceptible patients during months 4 to 10 of treatment, the differences were not significant.
Due to questions regarding the benefits of NNRTI hypersusceptibility and a variety of other reasons, including the consistent success, tolerability and low rates of high-grade toxicity of NNRTI-based initial therapy, observations such as those by Richard Haubrich et al have not led to a change in treatment strategy. That is, NNRTIs usually are not reserved for use in second- or third-line regimens to take advantage of NRTI resistance mutations. Such a strategy would require the accumulation of TAMs or other NRTI mutations, and early modification of therapy is now preferred to, if possible, avoid the development of TAMs.1
Recent data provide additional insight into the impact of NRTI mutations on NNRTI-based regimens. Importantly, these data indicate that NRTI mutations may have a significant detrimental impact on the efficacy of etravirine and likely other second-generation NNRTIs.
In study TMC125-C227, an exploratory, phase 2 trial presented by Brian Woodfall at Glasgow 2006, etravirine was tested in PI-naive, HIV-infected individuals who experienced virologic failure on an NNRTI-based regimen.40 Individuals were randomized to receive either a PI (n = 57; 95% ritonavir [RTV, Norvir] boosted) or etravirine (n = 59), each in combination with two NRTIs. The enrolled individuals had a baseline viral load of 4.3 log10 copies/mL and a CD4+ cell count of 229 cells/mm3.
The study was prematurely stopped by a data safety monitoring board, because the outcome overwhelmingly favored the PI arm. At week 8 and after, there was a clear separation of antiretroviral activity between the two arms. The PI arm showed the expected rate of viral load decline, whereas the etravirine arm did not. Although the etravirine arm had an initial viral load response of about 1.5 log10 copies/mL at week 8, and thereafter, viral load quickly rebounded back toward baseline.
Viral rebound was most apparent in individuals who had more than one NRTI mutation at baseline. In contrast, the subset of patients with no NRTI mutations demonstrated a more durable viral load suppression of about 1.75 log10 copies/mL below baseline as of week 16. Those patients with four or more NRTI mutations at baseline showed virtually no response to etravirine. The number of NNRTI mutations also had an impact on etravirine efficacy, and there was transient suppression in some individuals who had just one NNRTI mutation at baseline.
Thus, although etravirine is a potent drug, the findings from TMC125-C227 demonstrate that it requires an active background regimen to avoid the rapid development of resistance. In TMC125-C227, the background regimen of NRTIs was compromised by baseline resistance, since nearly half of the patients were using one or two recycled NRTIs in their regimen. Adding these results together with the data from the TMC125-C223 study, it appears that etravirine will be most effective when combined with a background regimen containing other active drugs, such as in combination with boosted PIs. Studies are currently underway to test this hypothesis.
As discussed above, the number of NNRTI mutations at baseline has a significant impact on the efficacy of etravirine-based HAART. Thus, if second-generation NNRTIs such as etravirine are to have a role in treatment, it is important to limit NNRTI resistance evolution as much as possible. Initial NNRTI resistance mutations tend to emerge quickly in patients with detectable viremia during NNRTI-based therapy.20,21 While single NNRTI mutations have the potential to lead to significant NNRTI resistance and cross-resistance, it is not uncommon for additional mutations to accumulate in patients who remain on the same antiretroviral regimen following a rebound in HIV viremia.
For example, although early virologic failure with efavirenz usually involves a single mutation (e.g., K103N), continuation of the regimen following virologic failure may lead to the accumulation of multiple mutations, including L100I, V108I, Y181C/I and P225H.41 This suggests that these mutations confer some additional benefit to HIV by producing increased resistance or cross-resistance.42,43 Although these mutations may have little impact on the efficacy of first-generation NNRTIs, data now show that they have the potential to reduce the utility of next-generation agents in this class, including etravirine. In addition, there appears to be no benefit to the continued use of these drugs: NNRTI resistance mutations have minimal impact on viral fitness, they do not increase susceptibility to other drugs, and the NNRTIs do not maintain partial virologic activity in the presence of most NNRTI mutations.6,7 Therefore, there is no justification for continuing an NNRTI after resistance has emerged, especially given the potential loss of future drug options with the emergence of additional NNRTI resistance mutations.
Although the first generation of NNRTIs -- particularly efavirenz -- has proven to be highly effective as a part of HAART, several pitfalls to this drug class have prompted the development of second-generation NNRTI agents. These drawbacks include:
Second-generation NNRTIs, notably etravirine, remain active against most HIV variants with first-generation NNRTI resistance. However, the presence of numerous NNRTI mutations (three or more mutations) or NRTI mutations may limit the use of second-generation NNRTIs. Clinicians should therefore be sure to carefully monitor HIV-infected individuals on first-generation NNRTIs for the development of viremia. In the event of virologic failure, first-generation NNRTIs should be promptly stopped and switched with other active agents (e.g., PIs, etravirine) so as to avoid the development of further NNRTI resistance mutations that could compromise the efficacy of the second-generation NNRTIs.
Although the benefits of these newer agents are still being defined, the available results suggest that the second generation of NNRTIs can produce significant antiretroviral activity in appropriately defined circumstances.
This is part one of a two-part article. Please click here to read part two.