Antagonism between Human Immunodeficiency ...

265

Antagonism between Human Immunodeficiency Virus Type 1 Protease Inhibitors Indinavir and Saquinavir In Vitro Debra P. Merrill, Douglas J. Manion, Ting-Chao Chou, and Martin S. Hirsch

Infectious Disease Unit, Massachusetts General Hospital, Harvard Medical School, Boston; Memorial Sloan-Kettering Cancer Center, New York, New York

Human immunodeficiency virus type 1 (HIV-1) protease inhibitors are a promising class of antiretroviral agents that compromise enzymatic function through substrate mimicry. The in vitro susceptibility of a panel of HIV-1 clinical isolates demonstrating various drug resistance phenotypes to combinations of the HIV-1 protease inhibitors saquinavir and indinavir was determined. Antiviral effect was assessed by an HIV-1 p24 antigen reduction assay in phytohemagglutinin-stimulated peripheral blood mononuclear cells after harvesting of cell-free supernatant fluids at peak antigen production (days 4 – 7). Drug interactions were determined by median-dose-effect analysis, with the combination index (CI) calculated at several inhibitory concentrations (IC50, IC75, IC90, IC95, IC99). The interactive effects ranged from synergy at low efficacy doses to antagonism at higher doses against a pan-susceptible clinical isolate of HIV-1. Against a zidovudine-resistant isolate as well as a multidrug-resistant isolate, the combination of saquinavir and indinavir demonstrated antagonism at all doses.

Received 6 November 1996; revised 30 January 1997. Presented in part: 36th Interscience Conference on Antimicrobial Agents and Chemotherapy, New Orleans, September 1996 (abstract I2); 4th Conference on Retroviruses and Opportunistic Infections, Washington, DC, January 1997 (abstract 158). Grant support: NIH (CA-12464); D.J.M. is a Research Fellow of the Medical Research Council of Canada. Reprints or correspondence: Dr. Martin S. Hirsch, Infectious Disease Unit, Massachusetts General Hospital, Gray J-504, Fruit St., Boston MA 02114. The Journal of Infectious Diseases 1997;176:265–8 q 1997 by The University of Chicago. All rights reserved. 0022–1899/97/7601–0037$02.00

protease inhibitors approved for clinical use in the United States, saquinavir and indinavir, with the hope of determining their interactive effects against an array of HIV-1 clinical isolates with varying antiretroviral susceptibilities. Materials and Methods Cells. Peripheral blood mononuclear cells (PBMC) from HIV1–seronegative donors were obtained by ficoll-hypaque density gradient centrifugation of heparinized venous blood. PBMC were treated with phytohemagglutinin (PHA-P, 2 mg/mL; Difco, Detroit), propagated in R-20 medium (RPMI 1640 supplemented with 20% heat-inactivated fetal calf serum [Sigma, St. Louis], 50 U of penicillin/mL, 50 mg of streptomycin/mL, 2 mM L-glutamine, and 10 mM HEPES buffer), supplemented with 10% interleukin-2 (Collaborative Biomedical Products, Bedford, MA), and incubated at 377C in a humidified atmosphere in 5% CO2 . Virus. Two isolates derived from an HIV-1–seropositive person before (14a-pre) and after (14a-post) 26 months of uninterrupted zidovudine monotherapy were propagated in PBMC and titrated to determine the TCID50 per milliliter of virus stock, as previously described [1]. Isolate 14a-pre is wild type at reverse transcriptase codons 41, 67, 70, 215, and 219, whereas 14a-post contains substitutions (D67N, K70R, T215F, and K219Q) that confer high-level (ú1000-fold) zidovudine resistance [1]. The multidrug-resistant HIV-1 isolate, which was obtained from a patient after 2 years of zidovudine and didanosine combination treatment (provided by T. Merigan), contains substitutions A62V, S68G/R, V75I, F77L, F116Y, Q151M, and K219E, conferring resistance to zidovudine, didanosine, zalcitabine, stavudine, and lamivudine [2]. Compounds. Saquinavir, previously known as Ro-31-8959, was obtained from Roche Products (Welwyn Garden City, UK). Indinavir, previously known as L-735,524 and MK-639, was obtained from Merck Research Laboratories (West Point, PA). Viral replication assay. Cell-free culture supernatant fluids were assayed by an HIV-1 p24 antigen ELISA (DuPont, Boston).

Downloaded from jid.oxfordjournals.org by guest on July 13, 2011

With an increased understanding of the pathogenesis of human immunodeficiency virus type 1 (HIV-1) – induced immunodeficiency, new strategies are emerging to slow and perhaps even eradicate viral replication in infected persons. Central to this shift in strategies is new information concerning interactions between viral replication kinetics, immune organ destruction, and disease progression. These advances come at a time when major improvements in terms of virus quantitation techniques allow for ongoing monitoring of viral replication status. The advent of a novel class of potent antiviral agents, the aspartyl protease inhibitors, ushers in a new era of antiretroviral chemotherapy. Many members of this class are peptidomimetic agents competing with the HIV-1 – encoded gag-pol polypeptide precursor at the enzyme substrate binding site. Unlike reverse transcriptase inhibitors, these agents work after virus integration. Laboratory and early clinical results indicate significantly greater suppressive effects on viral replication by protease inhibitors over most nucleoside and nonnucleoside reverse transcriptase inhibitors. Optimism regarding improved methods to minimize viral replication has extended to the possible use of different protease inhibitors in combination. We thus undertook a study of two

266

Concise Communications

Results All virus isolates demonstrated susceptibility to both protease inhibitors, with IC50s – IC99s being similar for the wild type isolate and the 2 isolates resistant to reverse transcriptase inhibitors (table 1). There was no significant toxicity in uninfected PBMC with concentrations of saquinavir or indinavir up to 0.16 mM. All combination assays were performed with equimolar concentrations of both drugs over a range of 0.01 – 0.16 mM. The CIs for all virus isolates tested at IC50 , IC75 , IC90 , IC95 , and IC99 are shown in figure 1, assuming both mutually nonexclusive (figure 1A) and exclusive effects (figure 1B). As expected, CIs are slightly lower assuming the mutually exclusive effect hypothesis over the more conservative mutually nonexclusive effect hypothesis. Overall, the combination of saquinavir and indinavir displayed increasing antagonism with increasing drug concentration regardless of the isolate tested.

Against the 14a-pre isolate, the combination of protease inhibitors demonstrated slight synergy at lower drug concentrations (IC50 , IC75), additive effect at IC90 , and antagonism at higher doses (IC95 , IC99), assuming mutually exclusive effects, with CIs ranging from 0.68 to 1.54. Assuming mutually nonexclusive effects, the range shifted to 0.79 – 2.11, demonstrating antagonism at all concentrations except the IC50 (additive effect). The highest CI in this study occurred with the pan-susceptible 14a-pre isolate at the IC99 . The combination of indinavir and saquinavir demonstrated an antagonistic effect against the zidovudine-resistant isolate, 14a-post, at all active concentrations regardless of assumptions regarding exclusivity of effect, with CIs ranging from 1.24 to 1.88. Saquinavir and indinavir demonstrated additive effects at lower drug concentrations (IC50 , IC75 , IC90) against the multidrug-resistant isolate and low-level antagonism at higher concentrations (IC95 , IC99), assuming mutually exclusive interactions, with a CI range of 0.84 – 1.21. However, assuming mutually nonexclusive interactions, the combination demonstrated antagonism at all concentrations except the IC50 , with a CI range of 1.14 – 1.38. Discussion The potential benefits of combining agents active against HIV-1 include increased efficacy in suppressing viral replication, decreased emergence of resistant isolates, and broadening of effect in terms both of isolates susceptible to each agent and of cellular and tissue reservoirs accessible to the antiretrovirals. These potential benefits must be weighed against the potential adverse effects of combination therapy, including additive toxicity, increased cost, and emergence of replication-competent multidrug-resistant isolates. In the present study, we have demonstrated mostly antagonistic interactions of combining the HIV-1 protease inhibitors saquinavir and indinavir against a panel of clinical isolates displaying varying susceptibilities to nucleoside and nonnucleoside reverse transcriptase inhibitors. The isolates chosen for this study reflect a range of susceptibility phenotypes likely to be encountered by clinicians treating HIV-infected patients. Each represents a pedigree of differing past antiretroviral experience: drug-naive (14a-pre), zidovudine monotherapy – experienced (14a-post), and reverse transcriptase inhibitor combination therapy – experienced (multidrug-resistant isolate). The range of drug concentrations used in this study also reflects levels achievable in patients’ intravascular compartments [6, 7]. In this study, we defined a combination as synergistic when the mean of the CI was õ1.0, taking the variance (SD) into account, and antagonistic when that value exceeded 1.0. CIs straddling the zero effect line of 1.0 were considered additive. The optimal inhibitory concentration at which the CI should be assessed is unclear, since concentrations in various body


Mathematical analysis of single-agent IC50 and combined-drug interactions. The 50% inhibitory concentration (IC50) of each single-agent therapy in experiments was determined by use of the computer software program of Chou and Chou [3]. The multipledrug-effect equation of Chou and Talalay, based on the mediandose-effect principle and the isobologram technique, was used to analyze combined-drug effects. This method involves plotting dose-response curves for each agent and for fixed-ratio combinations of the agents as previously described [4]. The CI values were based on the median-dose-effect equation for both the mutually exclusive and nonexclusive effect hypotheses. Mean CI values were determined as well as standard deviations. In this study, we define a combination as synergistic if the range of the mean and standard deviation is õ1 (i.e., greater than the expected additive effect when two agents are combined), additive if the range straddles 1.0, and antagonistic if the range is ú1 (i.e., less than the expected additive effect when two agents are combined). Experimental design. In each drug study, 3- or 4-day PHAstimulated PBMC were exposed to the HIV-1 inoculum (1000– 5000 TCID50/106 cells) without a subsequent wash. Drugs were added simultaneously. Cells were suspended in 1.0-mL final volume of R-20 medium supplemented with 10% interleukin-2 in 24well tissue culture plates and incubated in a humidified atmosphere with 5% CO2 at 377C. In all experiments, culture medium was changed twice weekly so that 0.5 mL of cell suspension was resuspended in 1.0 mL of fresh medium that contained the original drug concentration(s). Multiply diluted fixed-ratio combinations of the drugs or single drugs were added to each well, as previously described [5]. Each drug or combination was tested in duplicate, and each experiment was repeated at least twice. In addition, uninfected drug-treated toxicity controls were maintained at the highest concentration of each agent studied (either alone or in combination). Cell-free culture supernatant fluids were harvested and assayed for HIV-1 p24 antigen production on day 3 or 4 of culture and again on day 7. Cell proliferation and viability were assessed by the trypan blue exclusion method on uninfected drug-treated cultures maintained in parallel.

JID 1997;176 (July)

JID 1997;176 (July)


267

Table 1. Inhibitory concentrations (mM) { standard deviations of saquinavir and indinavir against HIV-1 clinical isolates. HIV-1 isolate, drug 14a-pre Saquinavir Indinavir 14a-post Saquinavir Indinavir MDR Saquinavir Indinavir

IC50

IC75

IC90

IC95

IC99

0.031 { 0.022 0.025 { 0.013

0.039 { 0.025 0.038 { 0.020

0.049 { 0.028 0.055 { 0.032

0.058 { 0.030 0.072 { 0.043

0.083 { 0.035 0.130 { 0.084

0.028 { 0.004 0.027 { 0.001

0.039 { 0.002 0.041 { 0.000

0.054 { 0.001 0.064 { 0.003

0.068 { 0.005 0.087 { 0.008

0.114 { 0.021 0.169 { 0.027

0.019 { 0.004 0.055 { 0.029

0.024 { 0.006 0.075 { 0.041

0.032 { 0.008 0.102 { 0.061

0.038 { 0.009 0.127 { 0.080

0.056 { 0.015 0.206 { 0.146

NOTE. 14a-pre and 14a-post represent paired clinical isolates, prior to therapy (wild type) and after zidovudine therapy (showing both phenotypic and genotypic resistance to zidovudine), respectively. Multidrug-resistant isolate (MDR) is multi – reverse transcriptase inhibitor – resistant strain isolated from patient treated with zidovudine and didanosine for 2 years.


Figure 1. Combination indices between saquinavir and indinavir assuming (A) mutually nonexclusive and (B) mutually exclusive effect hypotheses. 14a-pre, pan-susceptible isolate; 14a-post, zidovudine-resistant isolate; MDR, multidrug-resistant isolate. Data are combination index { SD.

268


nificantly greater suppression of viral replication and may even lead to antagonism. This suggests another reason for caution in combining protease inhibitors, in addition to separate concerns about selection of multiply cross-resistant viruses. However, it is only through the careful design and implementation of therapeutic trials that the clinical validity of these in vitro studies can be determined. Acknowledgments

We thank Emilio A. Emini, Noel Roberts, Don M. Coen, Ron Swanstrom, Joan Kaplan, and Richard T. D’Aquila for critical review of this manuscript and Daryld A. Strick for technical support.

References 1. Johnson VA, Merrill DP, Videler JA, et al. Two-drug combinations of zidovudine, didanosine, and recombinant interferon-aA inhibit replication of zidovudine-resistant human immunodeficiency virus type 1 synergistically in vitro. J Infect Dis 1991; 164:646 – 55. 2. Shafer RW, Iversen AK, Winters MA, Aguiniga E, Katzenstein DA, Merigan TC. Drug resistance and heterogeneous long-term virologic responses of human immunodeficiency virus type 1 – infected subjects to zidovudine and didanosine combination therapy. J Infect Dis 1995; 172:70 – 8. 3. Chou J, Chou TC. Dose-effect analysis with microcomputers: quantitation of ED50 , LD50 synergism, antagonism, low-dose risk, receptor-ligand binding and enzyme kinetics: a computer software for IBM-PC and manual. Cambridge, UK: Biosoft, 1989. 4. Chou TC. The median-effect principle and the combination index for quantitation of synergism and antagonism in chemotherapy. In: Chou TC, Rideout DC, ed. Synergism and antagonism in chemotherapy. New York: Academic Press, 1991:61 – 102. 5. Johnson VA. Evaluation of candidate anti-HIV agents in vitro. In: Walker BD, Aldovini A, ed. Techniques in HIV research. New York: Stockton Press, 1990:225 – 37. 6. Vacca JP, Dorsey BD, Schleif WA, et al. L-735,524: an orally bioavailable human immunodeficiency virus type 1 protease inhibitor. Proc Natl Acad Sci USA 1994; 91:4096 – 100. 7. Kitchen VS, Skinner C, Ariyoshi K, et al. Safety and activity of saquinavir in HIV infection. Lancet 1995; 345:952 – 5. 8. Rose RE, Gong YF, Greytok JA, et al. Human immunodeficiency virus type 1 viral background plays a major role in development of resistance to protease inhibitors. Proc Natl Acad Sci USA 1996; 93:1648 – 53. 9. Deminie CA, Bechtold CM, Stock D, et al. Evaluation of reverse transcriptase and protease inhibitors in two-drug combinations against human immunodeficiency virus replication. Antimicrob Agents Chemother 1996; 40:1346 – 51.


compartments will vary widely depending on multiple factors, including dose regimens, absorption, and distribution of individual components. However, since the achievement of higher inhibitory concentrations (e.g., the maximum tolerated dose) is increasingly the aim of antiretroviral therapy, it is of concern that antagonism is greatest at higher concentrations. This analysis raises questions concerning whether increased antiviral effect can be achieved by the addition of saquinavir to indinavir. Moreover, combining these two agents in full concentrations may increase toxicity and lead to cross-resistance. Although use of lower concentrations of each drug led to additive or synergistic interactions in some situations, suboptimal dosing may also increase the likelihood of emergence of multidrug-resistant variants [8]. The antagonism observed between indinavir and saquinavir should not be extrapolated to other members of the protease inhibitor class without further testing. When other combinations of protease inhibitors were studied in vitro using pansusceptible laboratory and clinical HIV-1 isolates in either a T-lymphoblastoid cell line or donor PBMC, predominantly synergistic interactions were observed [9]. In that study, however, the combination of indinavir with other protease inhibitors displayed antagonism in both cell types, which increased as the ratio of the compounds approached 1:1, as used here. The combination of saquinavir and ritonavir is currently being studied clinically on the basis of favorable pharmacokinetic interactions and hypotheses concerning apparent lack of cross-resistance. Although initial results suggest that this combination can lower plasma HIV-1 RNA levels, there have been no comparisons with monotherapy regimens; moreover, interactions between saquinavir and ritonavir have not yet been reported in vitro. The mechanisms underlying the antagonistic interaction between saquinavir and indinavir were not assessed in this study. These compounds may compete with each other for binding at the active site of the HIV-1 protease enzyme. Other mechanisms, including competitive cellular uptake or inactivation, may also be involved. Our findings do not indicate that saquinavir and indinavir cannot be used together in patients. In vitro antagonism does, however, suggest that combining these agents provides antiviral activity less than one would expect were they additive. The addition of a second protease inhibitor may not result in sig-

JID 1997;176 (July)