K103N, V106M and Y188L Significantly Reduce HIV-1 Subtype C Phenotypic Susceptibility to Doravirine

Abstract

Doravirine (DOR) is a non-nucleoside reverse transcriptase inhibitor (NNRTI) with efficacy against some NNRTI-resistant mutants. Although DOR resistance mutations are established for HIV-1 subtype B, it is less clear for non-B subtypes. This study investigated prevalent NNRTI resistance mutations on DOR susceptibility in HIV-1 subtype C. Prevalent drug resistance mutations were identified from a South African genotypic drug resistance testing database. Mutations, single or in combination, were introduced into replication-defective pseudoviruses and assessed for DOR susceptibility in vitro. The single V106M and Y188L mutations caused high-level resistance while others did not significantly impact DOR susceptibility. We observed an agreement between our in vitro and the Stanford HIVdb predicted susceptibilities. However, the F227L mutation was predicted to cause high-level DOR resistance but was susceptible in vitro. Combinations of mutations containing K103N, V106M or Y188L caused high-level resistance, in agreement with the predictions. These mutations are frequently observed in patients failing efavirenz- or nevirapine-based first-line regimens. However, they are also observed in those failing a protease inhibitor-based second-line regimen, as we have observed in our database. Genotypic drug resistance testing is therefore vital prior to the initiation of DOR-based treatment for those previously exposed to efavirenz or nevirapine.

Keywords: HIV; NNRTI; doravirine; phenotypic; resistance; subtype C.

Figure 1

Phenotypic susceptibility of HIV-1 wild-type subtype B (NSX) and C (MJ4) to DOR. The HIV-1 subtype B (p8.9NSX+) and C (p8.9MJ4) backbone plasmids were used to produce the respective PSVs. In vitro phenotypic assays were performed to assess the susceptibility of each wild-type ubtype. (a) The average of the resulting IC₅₀ values of the subtype B and C PSVs were used to calculate the FC for each mutant PSV. (b) DOR susceptibility was expressed as the IC₅₀ value of the PSV compared to the average IC₅₀ value of the corresponding subtype wild-type reference. The TCO was the average wild-type IC₅₀ of the PSV compared to the 99th percentile of the wild-type IC₅₀. The average IC₅₀ values were as follows: subtype B—0.0033 µM; subtype C—0.0056 µM. The 99th percentile of the IC₅₀ values were as follows: subtype B—0.0051 µM; subtype C—0.0073 µM. The TCOs were as follows: subtype B—1.55; subtype C—1.31. The TCOs were used to classify the susceptibility/resistance of the NNRTI-mutant PSVs downstream. An unpaired Student t-test with Welch’s correction was used to compare the DOR susceptibility of NSX and MJ4: IC₅₀ = p-value < 0.0001 (****) and FC = p-value > 0.9999. ns—no significance (p-value > 0.05), IC₅₀—50% inhibitory concentration.

Figure 2

Phenotypic susceptibility of single NNRTI mutations to DOR. In vitro phenotypic assays were performed to assess the susceptibility of each mutant. The lower TCO for DOR susceptibility in the assay was 1.31. DOR susceptibility/resistance was classified as follows: FC ≤ TCO = susceptible (■); TCO > FC ≤ 2 × TCO = potential low-level resistance (■); 2 × TCO > FC ≤ 3 × TCO = low-level resistance (■); 3 × TCO > FC ≤ 4 × TCO = intermediate resistance (■); and FC > 4 × TCO = high-level resistance (■). The coloured dots above each bar graph represent Standford’s HIV Drug Resistance Mutation (DRM) score. It indicates the level of resistance predicted by their algorithm for a particular mutation (low-level resistance (■), intermediate resistance (■), high-level resistance (■)). The bars without dots (e.g., K103N, V179D) were predicted to be susceptible to DOR; as their score would be 0, the dots are not depicted. V106M and Y188L displayed high-level resistance to DOR. F227L showed low-level DOR resistance despite having a predicted high-level resistance.

Figure 3

Phenotypic susceptibility of V106M in laboratory-adapted strains to DOR. In vitro phenotypic assays were performed to assess the susceptibility of V106M in the subtype B (a) and subtype C (b) laboratory-adapted strains and were expressed as FC. The lower TCO for DOR susceptibility in the assay was 1.31 for subtype C and 1.55 for subtype B. DOR susceptibility/resistance was classified as follows: FC ≤ TCO = susceptible (■); TCO > FC ≤ 2 × TCO = potential low-level resistance (■); 2 × TCO > FC ≤ 3 × TCO = low-level resistance (■); 3 × TCO > FC ≤ 4 × TCO = intermediate resistance (■); and FC > 4 × TCO = high-level resistance (■). The Stanford HIV Drug Resistance Database predicted intermediate resistance. The subtype B and C laboratory-adapted strains displayed high-level resistance to DOR. *—p-value ≤ 0.05, **—p-value ≤ 0.01, ***—p-value ≤ 0.001.

Figure 4

Comparing V106M susceptibility in subtype B and C laboratory-adapted strains. The phenotypic responses between the laboratory-adapted strains containing the V106M mutation were compared to each other. The mean IC₅₀ of each laboratory-adapted strain was compared with a one-way ANOVA and the p-values < 0.05 were considered significant (■).

Figure 5

Phenotypic susceptibility of F227L in laboratory-adapted strains to DOR. In vitro phenotypic assays were performed to assess the susceptibility of F227L in the subtype B (a) and subtype C (b) laboratory-adapted strains and were expressed as FC. The lower TCO for DOR susceptibility in the assay was 1.31 for subtype C and 1.55 for subtype B. DOR susceptibility/resistance was classified as follows: FC ≤ TCO = susceptible (■); TCO > FC ≤ 2 × TCO = potential low-level resistance (■); 2 × TCO > FC ≤ 3 × TCO = low-level resistance (■); 3 × TCO > FC ≤ 4 × TCO = intermediate resistance (■); and FC > 4 × TCO = high-level resistance (■). Intermediate resistance was predicted by the Stanford HIV Drug Resistance Database. The subtype B laboratory-adapted strains displayed potential and low-level resistance to DOR. The subtype C laboratory-adapted strains displayed susceptibility to DOR. ns—no significance (p-value > 0.05), *—p-value ≤ 0.05, **—p-value ≤ 0.01, ***—p-value ≤ 0.001, ****—p-value ≤ 0.0001.

Figure 6

Comparing F227L susceptibility in subtype B and C laboratory-adapted strains. The phenotypic responses between the laboratory-adapted strains containing the F227L mutation were compared to each other. The mean IC₅₀ of each laboratory-adapted strain was compared with a one-way ANOVA and the p-values < 0.05 were considered significant (■).

Figure 7

Phenotypic susceptibility of combination NNRTI mutations to DOR. In vitro phenotypic assays were performed to assess the susceptibility of each mutant and susceptibility was expressed as FC. The lower TCO for DOR susceptibility in the assay was 1.31. DOR susceptibility/resistance was classified as follows: FC ≤ TCO = susceptible (■); TCO > FC ≤ 2 × TCO = potential low-level resistance (■); 2 × TCO > FC ≤ 3 × TCO = low-level resistance (■); 3 × TCO > FC ≤ 4 × TCO = intermediate resistance (■); and FC > 4 × TCO = high-level resistance (■). The coloured dots above each bar graph represent Standford’s HIV Drug Resistance score for the Drug Resistance Mutation (DRM) profile. It indicates the level of resistance predicted by their algorithm for a particular mutation (low-level resistance (■), intermediate resistance (■), high-level resistance (■)). The majority of the mutants (17 of 20) displayed high-level resistance to DOR.