The authors have declared that no competing interests exist.
Conceived and designed the experiments: SC HS JL WZ. Performed the experiments: SC DZ WG LL. Analyzed the data: SC JL. Contributed reagents/materials/analysis tools: SL HL YL ZB JH. Wrote the paper: SC JL HS.
This study was designed to identify common HIV-1 mutation complexes affecting the slope of inhibition curve, and to propose a new parameter incorporating both the IC50 and the slope to evaluate phenotypic resistance.
Utilizing site-directed mutagenesis, we constructed 22 HIV-1 common mutation complexes. IC50 and slope of 10 representative approved drugs and a novel agent against these mutations were measured to determine the resistance phenotypes. The values of new parameter incorporating both the IC50 and the slope of the inhibition curve were calculated, and the correlations between parameters were assessed.
Depending on the class of drug, there were intrinsic differences in how the resistance mutations affected the drug parameters. All of the mutations resulted in large increases in the IC50s of nucleoside reverse transcriptase inhibitors. The effects of the mutations on the slope were the most apparent when examining their effects on the inhibition of non-nucleoside reverse transcriptase inhibitors and protease inhibitors. For example, some mutations, such as V82A, had no effect on IC50, but reduced the slope. We proposed a new concept, termed IIPatoxic, on the basis of IC50, slope and the maximum limiting concentrations of the drug. The IIPatoxic values of 10 approved drugs and 1 novel agent were calculated, and were closely related to the IIPmax values (r > 0.95,
This study confirms that resistance mutations cannot be accurately assessed by IC50 alone, because it tends to underestimate the degree of resistance. The slope parameter is of very importance in the measurement of drug resistance and the effect can be applied to more complex patterns of resistance. This is the most apparent when testing the effects of the mutations on protease inhibitors activity. We also propose a new index, IIPatoxic, which incorporates both the IC50 and the slope. This new index could complement current IIP indices, thereby enabling predict the efficacy of pre-clinical drugs for which human pharmacokinetic is not available.
Over the last two decades, advances in antiretroviral therapy have revolutionized the management of human immunodeficiency virus (HIV) and the control of HIV epidemics [
Several pharmacodynamic properties are used to determine the activity of a drug. The currently standard measure is the IC50, the concentration of drug required for 50% inhibition
Furthermore, it is implicit in current studies that resistance mutations shift dose-response curves to the right alone without affecting their slope [
Instantaneous inhibitory potential (IIP), which is dependent on IC50, slope, and
This study was therefore designed to evaluate the effects of common HIV-1 resistance mutation complexes on the IC50 and slope, and to propose a parameter that incorporates both IC50 and slope to determine the efficacy of pre-clinical antiretroviral drug candidates.
The antiretroviral drugs used in this study included zidovudine (ZDV, AZT), lamivudine (3TC), indinavir (IDV), nelfinavir (NFV), saquinavir (SQV), and ritonavir (RTV), all from Sigma-Aldrich Co. (St. Louis, MO, USA); and didanosine (ddI), stavudine (d4T), nevirapine (NVP), and efavirenz (EFV), all from Shanghai Desano Chemical Pharmaceutical Development Co., Ltd. (Shanghai, China). All of these drugs are used commonly in China for antiretroviral treatment. DG35 was a new PI, provided by Hesi Scientific and Technology Ltd (Xi’an, Shaanxi, China).
All of these drugs were dissolved in dimethyl sulfoxide (DMSO) and stored at -20°C. Drugs were serially diluted, such that the final concentration of DMSO in cell culture medium was 0.5%.
The human T-cell line MT-2 [
Variant viruses bearing single or multiple amino-acid substitutions, which are commonly found in Chinese HIV drug resistance surveillance programs, were obtained by site-directed mutagenesis on a pNL4-3 wild-type background. In brief, because of the large size of the pNL4-3 plasmid (~15 Kb), the entire protease (PR) (codons 1–99) and reverse transcriptase (RT) (codons 1–312) regions of this plasmid were amplified and ligated into the pMD18T vector (TaKaRa Biotechnology Co., Ltd., Dalian, Liaoning, China), followed by site-directed mutagenesis using the Phusion™ Site-directed Mutagenesis Kit (New England Biolabs, Ltd., Beijing, China). M46I, I54V, V82A, M46I\N88S, G48V\I54V, M46I\V82T\I84V, and G48V\I54V\V82A mutations were introduced into the PR coding region, and K103N, Y181C, K103N\Y181C, K101Q, K101Q\Y181C, K101Q\H221Y, K101Q\H221Y\Y181C, V179E\T215Y, V179E\Y181C\T215Y, V179E\H221Y\T215Y, V179E\Y181C\H221Y\T215Y, K103N\Y181C\T215Y, K103N\H221Y\T215Y, K103N\H221Y\Y181C\T215Y, and M41L\L210W\T215Y\K103N\K238T mutations were introduced into the RT coding region. The ligations harboring the desired mutations were digested with the restriction enzymes
For experiments testing RT inhibitors, triple serial dilutions spanning empirically determined ranges for each drug were added to wells of 384-well plates. TZM-bl cells (10,000 cells/well) were infected with recombinant virus at an MOI of 0.02 in plates containing pre-plated antiretroviral drugs. After 48 h, the expression of the luciferase reporter gene was measured using a Bright-Glo Luciferase Assay (Promega Co., Madison, WI, USA).
As the inhibitory effect of PIs cannot be detected in TZM-bl cells after 48 h, the protocol for PI susceptibility assays was based on a modification of the reporter gene assay for determining antiretroviral activity. During the first round of infection, 200 TCID50s of each viral stock were used to infect 13,500 MT-2 cells (MOI, 0.01) in diaphanous 384-well plates containing 3-fold serial dilutions of each tested PI. After 72 h in culture, 20 μL of supernatant containing
The potential cellular toxicities of drugs in MT-2 cells were determined by measuring cellular ATP levels in the presence of various concentrations of these compounds. MT-2 cells (13,500 cells/well) were cultured with each compound at 37°C and 5% CO2. After 72 h, CellTiter-Glo reagent (Promega Co., Madison, WI, USA) was added to each well, and chemiluminescence was measured. The maximum nontoxic concentration was defined as the concentration of inhibitor that had no effect on cellular ATP levels.
Percent inhibition was calculated as [1 − (virus production in the presence of drug / virus production in the absence of drug)] × 100. The IC50 and slope of the inhibition curve of each inhibitor were determined by fitting the inhibition curves to the data using nonlinear regression analysis to generate a four parameter sigmoid dose-response equation (GraphPad Prism, version 6.02). This step was performed in triplicate for duplicate plates of each concentration of antiretroviral drug. The mean IC50 and slope were calculated using all of the replicates for each virus and were expressed as mean ± standard deviation. The slope parameter is analogous to the Hill coefficient, which determines the degree of cooperativeness of the ligand binding to the enzyme or receptor [
IIP, which incorporates IC50, slope, and drug concentration, was used to better evaluate the antiviral activity of drugs [
The mean and standard deviations for IC50, slope, and IIP were calculated using Microsoft Office Excel 2013 software. Pearson’s correlation coefficient was used to determine correlations between parameters. All statistical analyses were performed using SPSS version 16.0 software, and
Drug susceptibility was assayed based on the MT-2/TZM-bl cells assay system.
Utilizing site-directed mutagenesis, we constructed viruses bearing common mutations that confer drug resistance.
Fold changes in IC50 for mutants were relative to those for the wild-type virus, and fractional changes in slope were computed using equation (2). The drugs tested are grouped by class: NRTIs, NNRTI, and PIs. Within each class, different shapes indicate the various mutants.
x axes indicate log of drug concentration (nM), y axes indicate the inhibitor rate of virus (%). (A) is the curve of the dose-response of 3 viruses (V179E\H221Y\T215Y, V179E\Y181C\T215Y and WT NL4.3) in d4T. (B) is the curve of the dose-response of 2 viruses (K103N\H221Y\Y181C\T215Y and WT NL4.3) in EFV. (C) is the curve of the dose-response of 3 viruses (G48V\I54V\V82A, M46I\N88S and WT NL4.3) in IDV. (D) is the curve of the dose-response of 3 viruses (G48V\I54V, V82A and WT NL4.3) in SQV.
The effect of resistance mutations on IC50 and slope was more evident for PIs than for NRTIs and NNRTIs
Mutations | IDV | NFV | RTV | SQV | ||||
---|---|---|---|---|---|---|---|---|
Fold change in IC50 |
Fractional change in slope |
Fold change in IC50 |
Fractional change in slope |
Fold change in IC50 |
Fractional change in slope |
Fold change in IC50 |
Fractional change in slope |
|
M46I | 1.12±0.06 | -0.37±0.08 | 2.15±0.37 | -0.72±0.06 | 1.05±0.32 | 0.38±0.15 | 0.16±0.03 | -0.13±0.04 |
I54V | 0.95±0.01 | -0.11±0.13 | 0.97±0.01 | 0.18±0.08 | 0.33±0.12 | 0.71±0.11 | 0.10±0.03 | 0.44±0.08 |
V82A | 1.09±0.17 | -0.27±0.10 | 1.59±0.33 | 0.06±0.12 | 2.58±1.30 | 0.04±0.01 | 0.48±0.23 | 0.59±0.05 |
M46I\N88S | 5.75±0.73 | -0.22±0.05 | 16.54±9.03 | 0.29±0.02 | 2.11±0.26 | 0.11±0.02 | 0.37±0.04 | 0.34±0.18 |
G48V\I54V | 2.20±0.71 | 0.49±0.02 | 51.78±6.00 | -0.36±0.10 | 1.66±0.05 | 0.59±0.11 | 10.87±0.67 | 0.37±0.01 |
M46I\V82T\I84V | 0.70±0.06 | 0.63±0.15 | 5.42±2.11 | 0.13±0.06 | 7.56±2.44 | 0.53±0.15 | 0.25±0.01 | 0.56±0.05 |
G48V\I54V\V82A | 4.12±1.90 | 0.62±0.07 | 68.14±1.45 | 0.28±0.14 | 12.72±1.86 | 0.56±0.01 | 8.07±2.54 | 0.40±0.07 |
WT |
1.00 | 0.00 | 1.00 | 0.00 | 1.00 | 0.00 | 1.00 | 0.00 |
* Values are expressed as means ± standard deviations of three independent experiments.
# WT indicates the wild-type HIV NL4-3, as a control.
IIP was calculated using equation (1), which included the IC50, slope and concentration of antiviral drugs. To simulate the real activity
IIP was calculated using equation (1), with the concentrations used being peak plasma concentrations for IIPmax and the maximum nontoxic concentration for IIPatoxic. Fractional changes in IIP were calculated using equation (3). The drugs tested are grouped by class: NRTIs, NNRTI, and PIs. Within each class, different shapes indicate the various mutants.
A lack of correlation was observed between fold changes in IC50 and fractional changes in IIPmax for EFV and PIs (
A problem with using IC50 as an indicator of antiviral activity is that it obscures differences in antiviral activity at higher drug concentrations. The slope of the dose-response curve, however, provides a better indication of antiviral activity at high drug concentrations. Slopes can distinguish antiretroviral drugs from different classes with the same IC50 and are intrinsically drug class-specific [
Many studies have reported that longer cumulative exposure to HAART correlates with higher rates of HIV resistance in China [
Taken together, our results revealed that resistance cannot be accurately assessed by IC50 alone, and that the effect of slope can be applied to more complex patterns of resistance. Furthermore, depending on drug class, there were intrinsic differences in the method by which resistance mutations affect drug parameters. The effects of mutants on slope were the most apparent when assessing the effects of mutants on PIs. The differences in the effects of the mutations on the parameters may be a consequence of the different mechanisms of drug inhibition. Mutations may lower enzyme efficiency, which affects the number of PR molecules needed to complete maturation. The exact mechanisms underlying the intrinsic differences in drug resistance are not yet fully understood and need further investigation.
For licensed drugs, the activity mainly depends on the intrinsic pharmacodynamic properties (IC50 and slope) and the pharmacokinetic properties of each drug
This study confirmed that resistance mutations cannot be accurately assessed by IC50 alone, because it tends to underestimate the degree of resistance. The slope parameter is of very importance in the measurement of drug resistance and the effect can be applied to more complex patterns of resistance. This was the most apparent when testing the effects of mutants on PI activity. We also added a new parameter, IIPatoxic, to IIP indices for novel, pre-clinical drugs. The new parameter incorporates both the IC50 and the slope, thereby enabling predict the efficacy of pre-clinical drugs for which human pharmacokinetic is not available.
(XLSX)
(XLSX)
We thank Dr. Lu for providing the cells and virus, and we thank Xi’an Hesi Scientific and Technology Ltd for providing the DG35.