| |
New NNRTI GW678248: preclinical resistance profile
|
| |
| |
Anti-HIV-1 Activity of the NNRTI GW678248 in Combination with Other Antiretrovirals against Clinical Isolate Viruses and In Vitro Selection for Resistance
Antimicrobial Agents and Chemotherapy, November 2005
...... The retention of the in vitro activity of GW678248 at low nanomolar concentrations against a panel of viruses exhibiting high levels of resistance from NNRTI-experienced patients suggests a clinical utility for GW678248 as a follow-on therapy.....
..... The data demonstrate that single mutations at V106I, E138K, or P236L caused very little loss of sensitivity to GW678248. The double mutation V106I and P236L or the triple mutation V106I, E138K, and P236L was required to cause high levels of resistance to GW678248......
...... Passages starting with WT virus or virus with the V106I or P236L mutation resulted in resistance to GW678248 and cross-resistance with NVP but not EFV. The activity of EFV on viruses with mutations at the locus at position 106 or 236 suggests that EFV may be a suitable follow-on agent following GW678248 therapy where viruses with these mutations have been selected.....
.....Currently, a prodrug of GW678248, designated GW695634, is in phase II clinical trials for the evaluation of its safety and efficacy and is intended for use in combination therapeutic regimens with other antiretroviral drugs.....
authors: Richard J. Hazen,1* Robert J. Harvey,1 Marty H. St. Clair,2 Robert G. Ferris,1 George A. Freeman,3 Jeffrey H. Tidwell,3 Lee T. Schaller,3 Jill R. Cowan,3 Steven A. Short,4 Karen R. Romines,5 Joseph H. Chan,3, and Lawrence R. Boone1
Departments of Virology,1 Clinical Virology,2 Medicinal Chemistry,3 Viral Diseases, Metabolic and Viral Diseases Center of Excellence for Drug Discovery,5Department of Gene Expression and Protein Bichemistry, GlaxoSmithKline, 5 Moore Drive, Research Triangle Park, North Carolina 277094
Phenotypic susceptibilities of 55 viruses obtained from NNRTI-experienced patients.
The ability of GW678248 to inhibit HIV-1 strains derived from patients who have experienced HIV-1 RNA rebound while taking combination regimens that included an NNRTI was assessed. The patient viruses used in this study had a mean of 13 protease mutations (range, 9 to 23 protease mutations) and 21 RT mutations (range, 5 to 33 RT mutations).
The viral genotypes for selected loci in the RT of the 55 clinical isolate viruses are listed in Table 3, and the phenotypes of the viruses are presented in Fig. 3. Eighty-five percent (47 of 55) of the viruses had a mean change in the EFV IC50 >10-fold, and 98% of the viruses (54 of 55) had a mean change in the NVP IC50 of >10-fold. Eighty-three percent of the viruses (46 of 55) were susceptible to GW678248, with a change in the IC50 10-fold. One of the 55 patient viruses had an IC50 >0.5 mM for GW678248 (range for the WT reference virus, 0.00035 to 0.0022 mM); the remaining 54 viruses had a mean IC50 of 0.0062 mM for GW678248 (range, 0.0006 to 0.109 mM). Fifty-six percent of the viruses (31 of 55) had <2.5-fold changes in IC50 values to GW678248.
The six viruses with K103N as the only NNRTI resistance mutation had a mean fold change in EFV IC50 of 43, yet they were sensitive to GW678248, with a mean fold change in the GW678248 IC50 of 1.9. Similarly, the three viruses with Y181C as the only NNRTI resistance mutation had a mean fold change in the NVP IC50 of 328 and a mean fold change in the GW678248 IC50 of 1.89. Of the 47 patient viruses with >10-fold changes in the EFV IC50, 38 (81%) had 10-fold changes in the GW678248 IC50. Of the 54 patient viruses with >10-fold changes in the NVP IC50, 45 viruses (83%) had 10-fold changes in the GW678248 IC50. Seven of the eight patient viruses most resistant to GW678248 had mutations at codon 106 or 188. The four viruses with the Y188L mutation had a mean fold change in the GW678248 IC50 of 27, the two viruses with the V106A mutation had a mean fold change in the GW678248 IC50 of 353, and the single virus with the V106I mutation had a 20.7-fold change in the GW678248 IC50. Thus, 81 to 83% of the EFV- and/or NVP-resistant viruses from this data set were susceptible to GW678248.
FIG. 3. Phenotypic susceptibilities to efavirenz, nevirapine, and GW678248 of 55 clinical isolate viruses obtained from NNRTI-experienced patients
In vitro serial passage of HIV-1 in the presence of dose-escalating concentrations of GW678248.
In the first series of dose-escalating serial passage experiments, six parallel lineages of HIV-1 strain HXB2 were carried for eight passages to select for resistance to GW678248. The results of genotypic analysis of the six viral lineages at passage 8 (64 nM GW678248) indicate that all lineages selected for V106I, E138K, and P236L. One of six lineages yielded a 50-50 mixture at residue 41 of Met and Leu. The significance of detection of an M-L mixture at residue 41 in one sample is unknown, and reproducibility will be monitored in future studies. This mutation was not seen in the second series of passage experiments (see below). Isolates from passages earlier than passage 8 were unavailable for genotype analysis, and therefore, temporal sequence data could not be obtained.
In a second series of HIV-1 serial passage experiments, either the WT or strains of HIV-1 containing NNRTI-resistance related mutations were used to initiate the passages (Table 4). The selection of the starting genotypes was based on key mutations that confer resistance to current NNRTIs (e.g., K103N, Y181C, and V106A) or specific mutations identified as being among those selected in the presence of GW678248 in the series one passage (e.g., V106I, P236L, and V106I-P236L).
WT virus passaged in the presence of up to 128 nM GW678248 developed three mutations, K102E, V106A, and P236L, after 10 passages (Table 4). V106A emerged rapidly at passage 2 (2 nM) and was retained throughout all subsequent passages. P236L emerged at passage 5 (16 nM), and K102E appeared at passage 8 (32 nM) as a K-E mixture at passages 8 and 9, followed by only as an E at passage 10. Notably, in this passage series, V106A emerged rather than V106I, and there was no selection for a mutation at locus 138 compared with the selection for a mutation at locus 138 in the virus generated in the first serial passage series.
In the series that was initiated with virus containing K103N, the only additional mutation selected was V106A, which emerged at passage 5 (8 nM). K103N was retained for the duration of the passage series.
In the series that was initiated with virus containing Y181C, the mutations V90I and K101E emerged in passage 4 (4 nM), and M230L emerged in passage 5 (8 nM) as a M-L mixture but only as an L from passage 6 onwards. Finally, F227C accumulated at passage 8 (64 nM). Y181C was retained for the duration of the passage series.
In the passage series that was initiated with virus containing V106A, the drug pressure resulted in the emergence of D237N at passage 2 (2 nM) as a mixture of D-N but only as an N from passage 3 onwards. The appearance of P225H was detected at passage 4 (8 nM) and was maintained for the duration of the series. By passage 7 (16 nM; note that 32 nM was present at passage 6) the E138K mutation emerged. The initial V106A mutation was maintained for the duration of the passage series.
In the passage series that was initiated with virus containing V106I, which was chosen because of selection in the series one passage study, the first additional mutation to be selected was E138K at passage 3 (4 nM). This mutation was also selected in the series one passage study. Interestingly, in the presence of these mutations, continued passage led to the selection of Y181C by passage 6 (32 nM), as a mixture of Y-C in passages 6 and 7, but only as a C by passage 9. Continued passage led to the emergence of I132L at passage 8 (64 nM). V106I remained throughout the passage series.
In the series that was initiated with the virus containing P236L, a G196R mutation was selected at passage 3 (4 nM) and E138K plus I142T was selected at passage 5 (16 nM). With the exception of I142T not being detected in the passage 7 sample, these mutations were maintained for the duration of the passage series. V106A was selected in passage 6 (32 nM) as a V-A mixture but as an A from passage 7 onwards. The K173Q mutation was observed only in passage 7 (64 nM). The significance of the transient appearance of K173Q is unknown, and reproducibility will be monitored in future studies. The P236L mutation was maintained throughout this passage series.
In the series that was initiated with the mutant with the V106I and P236L double mutation, only one additional mutation (Y188L) was selected at passage 7 (64 nM).
Drug sensitivity assays of viruses selected in the series two passages were performed on the last passage in each series (Table 3). These viruses were passaged once in the absence of compound before assay. The virus selected from the passage initiated with WT HIV was resistant to GW678248 and cross-resistant to NVP but not to EFV. The selected virus was sensitive to AZT and APV, as expected.
The viruses recovered from passages initiated with the various NNRTI-resistant mutants were resistant to GW678248, cross-resistant to NVP, and variably cross-resistant to EFV, depending on the specific mutations selected. The viruses selected from the passages initiated with the NNRTI-resistant mutants were sensitive to the nucleoside inhibitor AZT or the protease inhibitor APV, with the one exception of a fourfold elevated APV IC50 against the virus selected from the P236L series. The significance of this finding is uncertain and will be further explored and monitored for reproducibility in future studies.
Passage of HIV-1 strains that contain RT mutations conferring resistance to NVP or EFV at the initiation did not result in viruses that had regained sensitivity as a consequence of accumulating resistance to GW678248.
The sensitivity of isogenic, site-directed mutant viruses in HeLa MAGI assay with many of these selected mutations, alone and in combination, have been examined and are reported separately (8a). The viruses containing the triple mutation V106I, E138K, and P236L showed a greatly diminished sensitivity to GW678248 (165-fold less sensitive), while they were sensitive to NVP and less than 3-fold more resistant to EFV.
DISCUSSION
NNRTI-based HAART, especially regimens containing EFV, are becoming the standard of treatment. Several studies have demonstrated the equivalence or superiority and low toxicity of NNRTI-based HAART compared with those of protease inhibitor-based HAART (26; R. Levy, D. Labriola, and N. Ruiz, Eighth Conference on Retroviruses and Opportunistic Infections, abstr. 325, 2001; K. Tashima, S. Staszewski, M. Nelson, A. Rachlis, D. Skiest, R. Stryker, L. Bessen, V. Wirtz, S. Overfield, and D. Sahner, Fifteenth International AIDS Conference, abstr. TuPeB4547, 2004), as well as a prolonged time to virologic failure (26). The emergence of resistance to anti-HIV drugs, even when they are given as HAART, has necessitated the development of new, highly potent antiretroviral agents active against drug-resistant viral strains. In a separate study we have shown that GW678248 is a potent inhibitor of both WT and NNRTI-resistant HIV-1 in HeLa-CD4 MAGI cells and merits further development (8a). These results were also presented, in part, at the 43rd Interscience Conference on Antimicrobial Agents and Chemotherapy (R. J. Hazen, R. Harvey, R. Ferris, K. Creech, M. St. Clair, K. Romines, G. Freeman, L. Schaller, J. Chan, R. Dornsife, and L. Boone, Abstr. 43rd Intersci. Conf. Antimicrob. Agents Chemother., abstr. H-445, 2003). We have shown in the present study that GW678248 has anti-HIV-1 potency similar to that of EFV and superior to that of NVP in both PBMC and MT-4 cell assays. Like the other NNRTIs tested, GW678248 was not active against HIV-2. This absence of activity against HIV-2 is a common characteristic of the NNRTI class, although there are reports that NNRTI has activity against certain strains of HIV-2 (22, 27).
In combination with the marketed anti-HIV-1 agents, the activity of GW678248 was found to be additive or synergistic with the activities of the other agents, with the exception of slight antagonism with DLV. The slight antagonism seen with DLV is not considered detrimental to the development of GW678248, as the treatment guidelines do not recommend the use of two NNRTIs in combination. It is not clear why we have observed both additive and synergistic activities with GW678248 in combination with different members within a single drug class. The effects of another NNRTI in development, TMC125, has been found to be additive with the effects of APV, IDV, LPV, NFV, RTV, or SQV in an MT-4 cell assay (2). The effect of EFV has been found to be additive to synergistic with the effect of NFV in a yield reduction assay with HIV-1 RF and MT-2 cells (16). In this test, the degree of synergy between EFV and NFV increased when the combination index was determined at either the IC75 or the IC90 compared with additive effects when the combination index was determined at the IC50. Thus, in one assay system, it was possible to demonstrate various degrees of interaction between EFV and NFV that ranged from additivity to synergy at 50, 75, and 95% inhibition levels. While it may be tempting to speculate that in vitro synergy would predict that a certain combination might be more beneficial than one which is only additive, there are other more important factors which determine the suitability of using antiretroviral drugs in combination, including pharmacokinetics, interactions with drug-metabolizing enzymes, and toxicities. Many of these factors can be evaluated only in animal or human studies, and the efficacy of a combination regimen can be evaluated only in clinical trials. The greatest value of the in vitro combination studies is to identify the antagonism of the antiviral activity that would indicate that certain drugs should not be used clinically in combination. In this regard, the in vitro combination antiviral data support clinical testing of GW678248 with other approved antiretroviral drugs otherwise suitable for use in combination with NNRTIs.
Tests with 55 viruses obtained from NNRTI treatment-experienced patients showed a high degree of in vitro resistance to either EFV or NVP. These same viruses were found to be 81 to 83% susceptible to in vitro inhibition by GW678248. While most patients receiving NNRTI-containing HAART regimens have sustained antiviral responses, rebounds in plasma viral loads have been observed in some patients concomitant with the emergence of resistant strains of HIV-1. Reduced susceptibility to NVP is often associated with Y181C and Y188L substitutions (23), while reduced susceptibility to EFV is often associated with K103N, G190S, and Y188L substitutions. Viruses with K103N or Y188L substitutions exhibit reduced susceptibilities to all of the presently available NNRTIs (3). The retention of the in vitro activity of GW678248 at low nanomolar concentrations against a panel of viruses exhibiting high levels of resistance from NNRTI-experienced patients suggests a clinical utility for GW678248 as a follow-on therapy.
The results from in vitro serial passage experiments with either WT virus or virus with the K103N mutation as the starting virus indicate that GW678248 selects for V106A, among other mutations. In a companion paper (8a), using a panel of viruses with site-directed mutations identified in the series one serial passage experiment, we were able to demonstrate the contribution of each mutation to the overall resistance pattern of the virus. The data demonstrate that single mutations at V106I, E138K, or P236L caused very little loss of sensitivity to GW678248. The double mutation V106I and P236L or the triple mutation V106I, E138K, and P236L was required to cause high levels of resistance to GW678248. The presence of the K103N mutation at the initiation of passage does not seem to alter the pattern of GW678248-induced mutation and may prevent the selection of the P236L mutation seen with WT virus passage. Evidence in the literature supports the fact that the P236L mutation may be of minor concern in the clinic. The P236L mutation is rapidly selected in in vitro serial passage in the presence of DLV, yet it is infrequently observed in the viruses from patients receiving DLV therapy (11), where the K103N mutation is commonly observed. It is thought that the replicative fitness of viruses containing the K103N mutation is higher than that of viruses containing the P236L mutation, accounting for the dominance of the K103N mutation in the patient under therapy. The start of the passage with Y181C did not lead to a substitution at V106 in the resistant virus but, rather, led to other mutations of V90I, K101E, F227C, and M230L, in addition to the maintenance of Y181C. This virus sample demonstrated cross-resistance to all three NNRTIs tested. The contribution of each mutation in this virus to resistance has not been determined. The M230L mutation is generally associated with virologic failure of treatment with all NNRTIs (13), but it is seen in only approximately 1% of the NNRTI-treated patient population. In our panel of 55 viruses from NNRTI-experienced patients, no instance of the M230L mutation was seen. Passages starting with WT virus or virus with the V106I or P236L mutation resulted in resistance to GW678248 and cross-resistance with NVP but not EFV. The activity of EFV on viruses with mutations at the locus at position 106 or 236 suggests that EFV may be a suitable follow-on agent following GW678248 therapy where viruses with these mutations have been selected. None of the 55 treatment-experienced patient viruses in our study had the P236L mutation.
Currently, a prodrug of GW678248, designated GW695634, is in phase II clinical trials for the evaluation of its safety and efficacy and is intended for use in combination therapeutic regimens with other antiretroviral drugs.
ABSTRACT
GW678248, a novel nonnucleoside reverse transcriptase inhibitor, has been evaluated for anti-human immunodeficiency virus activity in a variety of in vitro assays against laboratory strains and clinical isolates.
When GW678248 was tested in combination with approved drugs in the nucleoside and nucleotide reverse transcriptase inhibitor classes or the protease inhibitor class, the antiviral activities were either synergistic or additive. When GW678248 was tested in combination with approved drugs in the nonnucleoside reverse transcriptase inhibitor class, the antiviral activities were either additive or slightly antagonistic.
Clinical isolates from antiretroviral drug-experienced patients were selected for evaluation of sensitivity to GW678248 in a recombinant virus assay. Efavirenz (EFV) and nevirapine (NVP) had 10-fold increases in their 50% inhibitory concentrations (IC50s) for 85% and 98% of the 55 selected isolates, respectively, whereas GW678248 had a 10-fold increase in the IC50 for only 17% of these isolates. Thus, 81 to 83% of the EFV- and/or NVP-resistant viruses from this data set were susceptible to GW678248.
Virus populations resistant to GW678248 were selected by in vitro dose-escalating serial passage. Resistant progeny viruses recovered after eight passages had amino acid substitutions V106I, E138K, and P236L in the reverse transcriptase-coding region in one passage series and amino acid substitutions K102E, V106A, and P236L in a second passage series.
RESULTS
Antiviral activity of GW678248 in MT-4 cells and human PBMCs.
A comparison of the activities of the NNRTI compound GW678248 against WT HIV-1 or HIV-2 in MT-4 cell and PBMC assays is shown in Table 1. These data indicate that GW678248 inhibits WT HIV-1 replication in MT-4 cells and PBMCs with IC50 values of 1.0 and 0.4 nM, respectively. The anti-HIV-1 activity of GW678248 is essentially equivalent to the activity of EFV, 18-fold greater than that of DLV, and more than 90-fold greater than that of NVP against WT virus. Neither GW678248 nor the other NNRTIs tested were active against HIV-2.
Antiviral activities of GW678248 in combination with other marketed anti-HIV agents. Table 2 presents the values for the deviation from additivity for combinations of GW678248 and currently approved anti-HIV-1 agents. These effects are further detailed in the graphical presentation of the isobolograms for the interaction of GW678248 with the other agents in Fig. 2.
INTRODUCTION
Highly active antiretroviral therapy (HAART) combination therapeutic regimens have dramatically decreased the morbidity and increased the life expectancy of patients infected with human immunodeficiency virus (HIV). Nonnucleoside reverse transcriptase (RT) inhibitors (NNRTIs) have become important components of combination therapies and have been shown to be effective at decreasing plasma viremia, are well tolerated, and may reduce the pill burden compared with that of other regimens (21). Unlike the nucleoside reverse transcriptase inhibitors (NRTIs), the NNRTIs do not require anabolism to the active triphosphate for activity and bind in a region of the HIV RT that is away from the catalytic site (8). Although very effective, the current NNRTIs have been described as having a "low genetic barrier" to resistance; i.e., the presence of one or two key mutations produces resistance, and the rate of cross-resistance to other NNRTIs is high (7, 10). Despite the high potential for resistance, NNRTIs are used extensively in first-line combination therapies. HAART regimens that combine the NNRTI efavirenz (EFV) with two NRTIs have been shown to be more effective than similar regimens that combine the protease inhibitor indinavir (IDV) with the same two NRTIs (26).
Recently, we have described the activities of a new class of NNRTIs, the benzophenones (4). Members of the benzophenone series showed low-nanomolar potencies against wild-type (WT) HIV type 1 (HIV-1) and a wide spectrum of NNRTI-resistant HIV strains, including strains containing Y181C and K103N mutations, which are often found in patients failing NNRTI-based antiviral therapy. From an extensive study of the benzophenone structure-activity relationship, we selected GW678248 (Fig. 1) for further development (K. R. Romines, G. A. Freeman, L. T. Schaller, J. R. Cowan, S. S. Gonzales, J. H. Tidwell, C. W. Andrews III, R. D. K. Stammers, R. J. Hazen, R. G. Ferris, S. A. Short, J. H. Chan, and L. R. Boone, submitted for publication). The initial biochemical and antiviral characterization of GW678248 indicates that this compound is a potent inhibitor of HIV RT, is active against a wide variety of NNRTI-resistant mutant strains, is only modestly affected by human serum proteins, and has a high selectivity index (8a).
This report describes the additional anti-HIV activities of GW678248, a novel benzophenone NNRTI. In this study, the activities of GW678248 in combination with approved anti-HIV agents, the sensitivities of 55 clinical isolates from NNRTI-experienced patients, and the genotypic and phenotypic patterns of resistant viruses selected during serial passage in the presence of GW678248 are reported.
|
|
| |
| |
|
|
|
