1143

Human Immunodeficiency Virus Type I (HIV-I) Inhibitory Interactions between Protease Inhibitor Ro 31-8959 and Zidovudine, 2',3'-Dideoxycytidine, or Recombinant Interferon-aA against Zidovudine-Sensitive or -Resistant HIV-I In Vitro Victoria A. Johnson, Debra P. Merrill, Ting-Chao Chou, and Martin S. Hirsch

Infectious Disease Unit. Massachusetts General Hospital and Harvard Medical School. Boston; Laboratory of Biochemical Pharmacology. Memorial Sloan-Kettering Cancer Center. New York. New York

There is a critical need for the development of agents that inhibit replication of human immunodeficiency virus (HIV) type I by novel, virus-specific mechanisms of action. Ro 318959 is a potent inhibitor of HI V-I and -2 proteases in vitro [1-3]. It affects viral replication in both acute and chronically infected cells [1-3]. For mature infectious virions to be formed, the viral precursor polyproteins Pr55gag and Pr 16(~~ag-pol must undergo a posttranslational cleavage reaction encoded by HIV-I protease to form the virion core structural proteins ofgag(pI7, p24, p15, p9, p6) and the essential enzymes of pol (reverse transcriptase, ribonuclease H, integrase). This occurs only after the HIV-I protease has autocatalytically cleaved itself from Prl6~ag-po' [4]. The HIV-l aspartyl protease is enzymatically active when it self-assembles monomers into a dimeric form, which may occur very late in the HIV-I replicative cycle, even after virions have been released from cells.

Received 5 March 1992; revised 25 June 1992. Presented in part: Fourth Conference of the National Institutes of Health National Cooperative Drug Discovery Groups for the Treatment of HIV Infection. San Diego. November 1991 (abstract 38). Written informed consent was obtained from all subjects according to protocols approved by the Subcommittee on Human Studies. Division of Research Affairs. Massachusetts General Hospital and Harvard Medical School. Financial support: National Institutes of Health (CA-12464. CA-35020. CA-5474 I. AI-26056) and Elsa U. Pardee Foundation. Reprints or correspondence (present address): Dr. Victoria A. Johnson. Division ofinfectious Diseases, University of Alabama at Birmingham. 229 Tinsley Harrison Tower, 1900 University Blvd., UAB Station, Birmingham, AL 35294-0006. The Journal of Infectious Diseases 1992;166:1143-6 © 1992 by The University of Chicago. All rights reserved. 0022-1899/92/6605-0026$01.00

On the basis of results of clinical studies with available drugs to date, single-agent regimens will probably not be sufficient to provide long-term virus suppression in HIV-linfected individuals [5, 6]. Therefore, combination therapy will likely be required [7]. Anti-HIV-1 strategies that demonstrate favorable drug interactions (e.g., synergy or additive effects) may allow the use of individual agents below their toxic concentrations, provide more complete viral suppression, and limit the emergence of drug-resistant HIV-l mutants. We have shown previously that favorable interactions may occur in vitro when anti-HIV-I agents that act by similar or different mechanisms of action are combined. We have also demonstrated that two-drug combined regimens may inhibit replication of zidovudine-resistant nrv-1 synergistically in vitro, even in regimens that include zidovudine [8, 9]. To define the potential role ofa novel protease inhibitor in combination anti-HIV-l therapy, we evaluated single-agent and two-drug regimens ofRo 31-8959 plus the reverse transcriptase inhibitors zidovudine and 2',3'-dideoxycytidine (ddC) and recombinant interferon-aA (rIFN-aA), which may act at different sites. Either zidovudine-sensitive or -resistant HIV-l isolates were used to acutely infect peripheral blood mononuclear cells (PBMC) in vitro.

Methods Cells. PBMC from a single HIV-I-seronegative donor were obtained by ficoll-hypaque density gradient centrifugation of heparinized venous blood and propagated as described previously [8]. Virus stocks. Paired HIV-I isolates were derived from an HIV-I-infected subject with AIDS before and after 26 months

Downloaded from http://jid.oxfordjournals.org/ at University of Birmingham on June 8, 2015

Protease inhibitor Ro 31-8959, a compound that interrupts human immunodeficiency virus (HIV)-specific formation of infectious virions, was evaluated in two-drug combined regimens with zidovudine, 2',3'-dideoxycytidine (ddC), or recombinant interferon-aA (rIFN-aA) against HIV-1 replication in vitro. By using peripheral blood mononuclear cells infected with HIV-1, drug interactions wereevaluated by the median-effectprinciple and the isobologram technique. A zidovudine-sensitive and -resistant HIV-1 isolate pair was studied. Additive to synergistic antiHIV-1 interactions were seen with 7.5-30 oM Ro 31-8959 and 0.005-0.02 p.M zidovudine (for the zidovudine-sensitive HIV-l isolate), 0.25-;-1.0 p.M zidovudine (for the zidovudine-resistant HIV-l isolate), 0.025-0.1 ,."M ddC, and 8-32 units/ml, rIFN-aA, without additive toxicity. Phase 1/11 clinical trials of Ro 31-8959 for therapy of HIV-1 infection are in progress. Ifresults are favorable, combined regimens including Ro 31-8959 deserve consideration for future clinical trials.

1144

JID 1992; 166 (November)

Concise Communications

Results Effects ofone- and two-drug regimens in PBMC using a zidovudine-sensitive HIV-I isolate. The initial study (experiment 1) sought to identify the subinhibitory concentrations

of Ro 31-8959 that would best demonstrate two-drug interactions in the subsequent combined therapy experiments 2-4 at the selected virus inoculum. In experiment 1, there was significant inhibition of HIV-l p24 antigen production at 10 nM Ro 31-8959 and 100 nM Ro 31-8959 on day 7 in culture in PBMC compared with the untreated HIV -L-infected control (data not shown). The ED 50 for Ro 31-8959 was 3.5 nM. Zidovudine at 0.1 /.lM also inhibited HIV-I replication significantly in this experiment; the 0.0008% DMSOtreated no-drug control did not show any anti-HlV-I effects (data not shown). In experiment 1, there was no toxicity of 100 nM Ro 31-8959, 0.0008% DMSO, or O.IIJ-M zidovudine in the drug-treated un infected control flasks. Synergistic interactions were seen between Ro 31-8959 and zidovudine, ddC, or rIFN-aA at ED90 to ED95 (data not shown). CI values were < 1 (table 1), with one exception of CI = 1.02. More complete viral suppression was achieved by each of the two-drug regimens compared with each agent used singly on day 7 in culture. There were no significant differences in cell proliferation or cell viability among these drug-treated flasks (either as single-agent or combined regimens). Table 1. Combination index (Cl) values for two-drug combined regimens including protease inhibitor Ro 31-8959 plus zidovudine (AZT), 2',3'-dideoxycytidine (ddC), or recombinant interferon (IFN)-aA against human immunodeficiency virus type I (HlV-l) replication in peripheral blood mononuclear cells. HIV-I isolate. experiment no. 14a-4/87 2

3

4

CI at % of HIV-I inhibition Two-drug regimen

90

95

Ro 31-8959 + AZT Ro 31-8959 + ddC Ro 31-8959 + IFN Ro 31-8959 + AZT Ro 31-8959 + ddC Ro 31-8959 + JFN Ro 31-8959 + AZT Ro 31-8959 + ddC Ro 31-8959 + JFN

1.02 0.98 0.71 0.78 0.56 0.80 0.57 0.69 0.83

0.89 0.88 0.67 0.76 0.56 0.88 0.54 0.59 0.84

Ro 31-8959 + AZT Ro 31-8959 + ddC Ro 31-8959 + JFN Ro 31-8959 + AZT Ro 31-8959 + ddC Ro 31-8959 + JFN Ro 31-8959 + AZT Ro 31-8959 + ddC Ro 31-8959 + JFN

0.63 0.94 0.81 0.77 1.02 0.84 0.61 1.05 0.75

0.64 0.86 0.80 0.77 0.90 0.84 0.66 1.02 0.80

14a-6/89 6

7

8

NOTE. CI values of -c l , I. and> I indicate synergism. additive effects. and antagonism. respectively. CI values were determined from median-effect plot parameters 111 (slope) and Dm (EDso) of eachdrugand eachcombination based on classic isobologram equation as described previously [1113]. All calculations were donewithcomputer software using IBM PC[II].

Downloaded from http://jid.oxfordjournals.org/ at University of Birmingham on June 8, 2015

of uninterrupted zidovudine monotherapy for generation of PBMC-derived virus stocks as described previously [8]. These isolates, designated 14a-4/87 (early zidovudine therapy) and I 4a-6/89 (late zidovudine therapy), were confirmed previously to be zidovudine-sensitive and -resistant, respectively [8). Compounds. Protease inhibitor Ro 3 1-8959 (previously called compound XVII) was provided by Roche Products (Welwyn Garden City, UK) [I -3]. The compound was dissolved in dimethyl sulfoxide (OMSO) and stored at -20°C until used. The highest final concentration of OM SO tested in these experiments was ~0.001% when stocks ofRo 31-8959 were diluted in medium. Zidovudine, ddC, and rIFN-aA were prepared as described previously [8, 9). Viral replication assay. Cell-free culture supernatants were assayed by an HIV- 1 p24 antigen ELISA (NEN Research Products, Boston) as described previously [l0]. Experimental design. Eight experiments were done. In each, uninfected 4-day phytohemagglutinin-stimulated PBMC (5 X 106 cells) were suspended in a 5-mL final volume as described previously [8]. Multiply diluted fixed-ratio combinations of the drugs or single drugs were added to each flask. Simultaneously, cells were exposed to HIV -1 inocula (1000 TCIOso/10 6 cells) without a subsequent wash. In experiments 1-4, and 5-8, isolates 14a-4/87 and 14a-6/89 were used, respectively. In addition, uninfected drug-treated toxicity controls were maintained at the highest concentrations of each agent studied (either alone or in two-drug regimens). In all experiments, culture medium was changed twice weekly as described previously [8], and cellfree culture supernatants were harvested for HIV-l p24 antigen production on day 7 in culture. Cell proliferation and viability were assessed by the trypan blue dye exclusion method. To evaluate Ro 31-8959 susceptibility of the HIV-l isolates, experiment 1 (with isolate 14a-4/87) and experiment 5 (with isolate 14a-6/89) were done. The ranges ofRo 31-8959 concentrations tested were 0.001-100.0 nM (experiment 1) and 1.2520.0 nM (experiment 5). Experiments 2-4 and 6-8 were combined therapy studies. The concentration ranges ofagents tested were 0.005-0.02 u.M zidovudine (for the zidovudine-sensitive HIV -1 isolate), 0.25-1.0 /.lM zidovudine (for the zidovudine-resistant HIV-l isolate), 0.025-0.1IlM ddC, 8-32 units/rnl. rIFNaA, and 7.5-30 nMRo 31-8959. Mathematical analysis ofsingle-agent ED50 and combined drug interactions. The EO so of each single-agent therapy in experiments 1 and 5 was determined with the computer software program of Chou, Chou, and Talalay [11-13]. The multiple-drugeffect analysis of Chou and Talalay [12], based on the median-effect principle and the isobologram technique, was used to analyze combined drug effects as described previously [11-13). Combination index (CI) values of < I, 1, and> I indicate synergism (i.e., greater than the expected additive effect when two agents are combined), additive effects, and antagonism (i.e., less than the expected additive effect), respectively.

lID 1992;166 (November)

Concise Communications 200r --

-

-

-

-

-

-

-

Inleoled

Control

200r--

-

-

-

-

-

-

-

-

-

Ro 15

Azr

Ro

Ro AZl

Ro

.5

+

30 nM

AZ

-.,

-

-

Ro AZT Ro 7.5 25 + nM ~M AZT - - - - -

nM

-

-

~M

-

AZl - -

-

-

1 ~M

-

-

-

---,

Ro 31-8959 ond ddC

'€'

!

i:i; (.:l

;::

150

100