Preclinical Evaluation of 1H-Benzylindole Derivatives as Novel Human Immunodeficiency Virus Integrase Strand Transfer Inhibitors

Abstract

We have identified 1H-benzylindole analogues as a novel series of human immunodeficiency virus (HIV) integrase inhibitors with antiretroviral activities against different strains of HIV type 1 (HIV-1), HIV-2, and simian immunodeficiency virus strain MAC₂₅₁ [SIV(MAC₂₅₁)]. Molecular modeling and structure-activity relationship-based optimization resulted in the identification of CHI/1043 as the most potent congener. CHI/1043 inhibited the replication of HIV-1(III_B) in MT-4 cells at a 50% effective concentration (EC₅₀) of 0.60 μM, 70-fold below its cytotoxic concentration. Equal activities against HIV-1(NL4.3), HIV-2(ROD), HIV-2(EHO), and SIV(MAC₂₅₁) were observed. CHI/1043 was equally active against virus strains resistant against inhibitors of reverse transcriptase or protease. Replication of both X4 and R5 strains in peripheral blood mononuclear cells was sensitive to the inhibitory effect of CHI/1043 (EC₅₀, 0.30 to 0.38 μM). CHI/1043 inhibited integrase strand transfer activity in oligonucleotide-based enzymatic assays at low micromolar concentrations. Time-of-addition experiments confirmed CHI/1043 to interfere with the viral replication cycle at the time of retroviral integration. Quantitative Alu PCR corroborated that the anti-HIV activity is based upon the inhibition of proviral DNA integration. An HIV-1 strain selected for 70 passages in the presence of CHI/1043 was evaluated genotypically and phenotypically. The mutations T66I and Q146K were present in integrase. Cross-resistance to other integrase strand transfer inhibitors, such as L-708,906, the naphthyridine analogue L-870,810, and the clinical drugs GS/9137 and MK-0518, was observed. In adsorption, distribution, metabolism, excretion, and toxicity studies, antiviral activity was strongly reduced by protein binding, and metabolization in human liver microsomes was observed. Transport studies with Caco cells suggest a low oral bioavailability.

Although the development of combination regimens including protease and/or reverse transcriptase inhibitors has extended the length and quality of life for many human immunodeficiency virus (HIV)-infected individuals, antiviral resistance and toxicity problems of this lifelong therapy warrant the continuous need to develop new drugs preferentially targeting new targets. Integrase (IN), the third virally encoded enzyme required for HIV type 1 (HIV-1) replication, catalyzes the insertion of viral DNA into the host cell chromosome through a multistep process that includes two catalytic reactions: 3′ cleavage of the viral DNA ends and strand transfer of the viral DNA into the host DNA (11). After integration, the proviral DNA is replicated and genetically transmitted as part of the cellular genome. As such, integration defines a point of no return in the establishment of HIV infection and is therefore an attractive target for anti-HIV therapy.

A decade of research on effective inhibitors of IN yielded different mechanistic classes of compounds (34, 40, 41). However, most of these compounds did not exhibit antiviral activity or were toxic in cell culture. For most of the IN inhibitors with antiviral activity in cell culture, it was not unambiguously shown that integration was the sole target (12, 19, 33). The identification of a series of diketo acids (DKA) that specifically target strand transfer and prevent HIV-1 replication in cell culture provided the first proof of principle for HIV-1 IN inhibitors as antiviral agents (8, 17). L-731,988 is the prototype of these IN strand transfer inhibitors (INSTIs). In 2003, the Merck group characterized a series of metabolically stable heterocyclic compounds, represented by L-870,810, containing an 8-hydroxy-[1,6]-naphthyridine-7-carboxamide pharmacophore as a substitute for the 1,3-DKA moiety (Fig. 1) (42). For HIV-1-infected patients, administration of L-870,810 resulted in a 50-fold reduction in viral loads, but clinical studies were halted due to liver and kidney toxicity in dogs. In 2005, we initiated the development and preclinical evaluation of 1H-benzylindole diketo compounds based on a three-dimensional pharmacophore model for DKA-like derivatives acting as INSTIs (3, 10).

FIG. 1.

Structures of INSTIs.

Meanwhile, the pyrimidinone carboxamide compound from Merck, MK-0518 (raltegravir [Isentress]) (Fig. 1), has received FDA approval for treatment of HIV-1 infection in combination with other antiretroviral agents in treatment-experienced adult patients who have evidence of viral replication and HIV-1 strains resistant to multiple antiretroviral agents (25). MK-0518 reduced viral loads to undetectable levels (below 50 copies/ml) in nearly two-thirds of highly treatment-experienced patients infected with three-drug-class-resistant HIV and was generally well tolerated (6, 37). Gilead Sciences presented at the 14th Conference on Retroviruses and Opportunistic Infections its latest data on the antiviral activity of its experimental IN inhibitor GS-9137 (elvitegravir) (Fig. 1). An ongoing phase II clinical trial showed that GS-9137 at its highest dose level was able to significantly reduce HIV loads compared with a boosted protease inhibitor regimen (43).

The present study describes the antiretroviral activity, mechanism of action, resistance profile, and early adsorption, distribution, metabolism, excretion, and toxicity studies of the most potent congener of the 1H-benzylindole analogue series, CHI/1043 (Fig. 1; see Table 1).

TABLE 1.

Inhibition of HIV-1 IN enzymatic activity, replication of HIV-1(III_B), and cytotoxicity in MT-4 cells of 1H-benzylindole analogues

^aData represent average results ± standard deviations from at least three separate experiments.

^bConcentration required to inhibit the in vitro overall IN activity by 50%.

^cConcentration required to inhibit the in vitro strand transfer step by 50%.

^dEffective concentration required to reduce HIV-1-induced CPE by 50% in MT-4 cells.

^eCytotoxic concentration to reduce MT-4 cell viability by 50%.

^fSI, selectivity index (CC₅₀/EC₅₀ ratio).

^gCrystals were observed at a concentration of ≥62 μM.

^hEt, ethyl group.

MATERIALS AND METHODS

Compounds.

CHI/1043 and other 1H-benzylindole derivatives, such as CHI/1013, CHI/1014, CHI/1037, CHI/1038, CHI/1042, CHI/1047, CHI/1089, CHI/1096, CHI/1133, CHI/1202, and CHI/1203, were synthesized by the laboratory of A. Chimirri, Messina, Italy (see Table 1). Dextran sulfate (DS; average molecular weight, 5,000) was purchased from Sigma (Bornem, Belgium). Zidovudine (AZT) was synthesized according to the method described by Horwitz et al. (21). Efavirenz, saquinavir, nelfinavir, and ritonavir were obtained from the National Institutes of Health (Bethesda, MD). L-870,810, MK-0518, and GS-9137 were obtained from our institute (Katholieke Universiteit Leuven) (Fig. 1). The diketo derivative L-708,906 was synthesized at the National Cancer Institute. All compounds were dissolved in dimethyl sulfoxide at 10 mg/ml; DS was dissolved in Milli-Q-purified water.

Cells.

MT-4 cells (30) were grown in a humidified atmosphere with 5% CO₂ at 37°C and maintained in RPMI 1640 medium supplemented with 10% heat-inactivated fetal calf serum (FCS), 2 mM l-glutamine, 0.1% sodium bicarbonate, and 20 μg/ml of gentamicin. 293T cells were grown in Dulbecco's modified Eagle's medium (DMEM; Gibco BRL) supplemented with 10% FCS (Harlan Sera-Lab Ltd.), 2 mM glutamine (Gibco BRL), and 20 μg/ml gentamicin (Gibco BRL) at 37°C in a 5% CO₂ humidified atmosphere. Caco-2 cells were purchased from the European Collection of Cell Cultures. Liver microsomes were from In Vitro Technologies or Xenotech.

Virus strains.

The origins of HIV-1(III_B), HIV-1(NL4.3) (1, 35), HIV-2(ROD), HIV-2(EHO) (4, 36), and simian immunodeficiency virus strain MAC₂₅₁ [SIV(MAC₂₅₁)] (7, 35) have been described previously. The compounds were evaluated for their inhibitory effects against the replication of a variety of mutant HIV strains in MT-4 cells, such as nucleoside reverse transcriptase inhibitor (NRTI)-resistant strains (24, 27), nonnucleoside reverse transcriptase inhibitor (NNRTI)-resistant strains (15, 31), a protease-resistant strain (5), an L-708,906-resistant strain (14), and an L-870,810-resistant strain (20). The MK-0518-resistant strain harboring G140S and Q148H (37, 39) was generated through site-directed mutagenesis (26) of the pNL4.3 plasmid obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH, contributed by Malcolm Martin (Bethesda, MD). The HIV-1 X4 strain subtype B (syncytium-inducing) 92HT599 reagent (catalog no. 3301) from Neal Halsey and the BAL strain were also obtained through the NIH AIDS Research and Reference Reagent Program, Division of AIDS, NIAID.

Drug susceptibility assay.

The inhibitory effects of antiviral drugs on the HIV-induced cytopathic effect (CPE) in human lymphocyte MT-4 cell culture were determined by the MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay (32). This assay is based on the reduction of the yellow-colored MTT by mitochondrial dehydrogenase of metabolically active cells to a blue formazan derivative which can be measured spectrophotometrically. The 50% cell culture infective doses of the HIV strains were determined by titration of the virus stock by using MT-4 cells. For the drug susceptibility assays, MT-4 cells were infected with 100 to 300 50% cell culture infective doses of the HIV strains in the presence of fivefold serial dilutions of the antiviral drugs. The concentration of the compound achieving 50% protection against the CPE of HIV, which is defined as the 50% effective concentration (EC₅₀), was determined. The concentration of the compound killing 50% of the MT-4 cells, which is defined as the 50% cytotoxic concentration (CC₅₀), was determined as well.

Overall integration assay using an ELISA.

To determine the susceptibility of the HIV-1 IN enzyme to different compounds, we used an enzyme-linked immunosorbent assay (ELISA). The overall integration assay uses an oligonucleotide substrate for which one oligonucleotide (5′-ACTGCTAGAGATTTTCCACACTGACTAAAAGGGTC-3′) is labeled with biotin at the 3′ end and the other oligonucleotide (5′-GACCCTTTTAGTCAGTGTGGAAAATCTCTAGCAGT-3′) is labeled with digoxigenin at the 5′ end. For the strand transfer assay, a precleaved oligonucleotide substrate (the second oligonucleotide lacks GT [underlined] at the 3′ end) was used. The IN enzyme was diluted in 750 mM NaCl, 10 mM Tris (pH 7.6), 10% glycerol, and 1 mM β-mercaptoethanol. To perform the reaction, 4 μl of diluted IN (corresponding to a concentration of 1.6 μM) and 4 μl of annealed oligonucleotides (7 nM) were added in a final reaction volume of 40 μl containing 10 mM MgCl₂, 5 mM dithiothreitol, 20 mM HEPES (pH 7.5), 5% polyethylene glycol, and 15% dimethyl sulfoxide. As such, the final concentration of IN in this assay was 160 nM. The reaction was carried out for 1 h at 37°C. Reaction products were denatured with 30 mM NaOH and detected by an ELISA on avidin-coated plates (22).

Time-of-addition experiments.

MT-4 cells were infected with HIV-1(III_B) at a multiplicity of infection (MOI) of 0.5, and the compounds were added at different time points after infection (0, 1, 2, 3, 8, 9, 10, 11, 12, 13, 15, 25, and 26 h). Viral p24 antigen production was determined at 31 h postinfection (HIV-1 p24 core profile ELISA; duPont, Dreieich, Germany). The reference compounds (DS, AZT, and ritonavir) were added at standardized concentrations corresponding to 100 times the EC₅₀ value, determined at an MOI of 0.01. MK-0518 and CHI/1043 were added at 0.05 and 27 μM (35 to 45 times their EC₅₀ values), respectively.

Quantification of different HIV-1 DNA species during HIV infection by real-time PCR.

MT-4 cells (1.5 × 10⁶ cells per tube) were incubated with HIV-1 NL4.3 (corresponding to 150 ng of p24) in the absence or presence of the test compounds. Inhibitors were added to the cells 1 h prior to infection. After a 2-h incubation at 37°C, the cells were washed with phosphate-buffered saline, transferred to new medium, and seeded in a 24-well plate (ca. 300,000 infected cells/well). When infection medium was replaced with new medium, fresh inhibitors were added. In each 24-well plate, uninfected MT-4 cells were incubated in parallel. Each time a sample was prepared for quantitative Alu PCR (Q-PCR) analysis, an aliquot of uninfected cells was prepared as well. DNA extractions and quantification of late reverse transcripts, two-long-terminal-repeat (2-LTR) circles, and integrants were done as described earlier (38).

Lentiviral vector transduction assay.

HIV-1-derived vector particles, pseudotyped with the envelope of vesicular stomatitis virus (VSV), were produced by transfecting 293T cells with three plasmids as described previously (38). The transfer plasmid used was the pCHGFPWS plasmid, as described previously (38). The day prior to transduction, 293T cells were seeded in 24-well plates at ca. 1.5 × 10⁵ cells per well. Transductions with the lentiviral vectors were carried out at an MOI of 10. Vector was added to the cells in the presence of DMEM-1% FCS. After 4 h of incubation, medium was replaced with DMEM containing 10% FCS. In each 24-well plate, 293T cells that were not transduced were incubated in parallel. Inhibitors of lentiviral transduction were added to the cells 1 h prior to transduction. When transduction medium was replaced by DMEM-10% FCS, fresh inhibitors were added. DNA extractions and quantification of late reverse transcripts, 2-LTR circles, and integrants were done as described earlier (38).

Selection of antiviral resistance.

The resistance selection of HIV-1(III_B) against CHI/1043 was initiated at a low MOI (MOI = 0.01) in MT-4 cells and a drug concentration of 0.15 μM. Every 3 to 4 days the MT-4 cell culture was monitored for the appearance of an HIV-induced CPE. When a CPE was observed, the cell-free culture supernatant was used to infect fresh, uninfected MT-4 cells in the presence of an equal or higher concentration of the compound. When no virus breakthrough was observed, the infected cell culture was subcultivated in the presence of the same concentration of the compound. The compound concentration was gradually increased.

PCR amplification and sequencing of the coding region of the IN gene.

Proviral DNA extraction of MT-4 cells, infected with HIV-1(III_B) selected in the presence of CHI/1043, was performed using the QIAamp blood kit (Qiagen, Hilden, Germany). PCR amplification and sequencing of the IN-encoding sequence were done as described previously (19). Mutations present in more than 25% of the global virus population can be detected in a mixture of the subject protein with the wild-type protein by means of population sequencing.

Plasma protein binding.

The plasma protein binding was estimated by determining the EC₅₀ of CHI/1043 in parallel with reference compounds in the presence of 50% human serum. The increase in EC₅₀ was calculated in comparison to the EC₅₀ determined under standard conditions (10% FCS).

Permeability experiments using Caco-2 cells.

Apical-to-basolateral transport experiments with Caco-2 cells were performed as described previously (2, 28). Briefly, prior to the experiment, the transport plates were precoated with culture medium containing 10% serum to avoid nonspecific binding. After 21 to 28 days of culture, the cells were ready for permeability experiments.

The culture medium from the apical wells was removed, and the inserts were transferred to a wash row in a transport plate without inserts, which was previously prepared with 1.5 ml transport buffer (Hanks balanced salt solution, 25 mM HEPES, pH 7.4). The transport buffer in the basolateral well contained 1% bovine serum albumin. Transport buffer (Hanks balanced salt solution, 25 mM MES [morpholineethanesulfonic acid], pH 6.5) was added to the inserts, and the cell monolayers were equilibrated in the transport buffer system for 30 min at 37°C in a polymix shaker. The transepithelial electrical resistance value in each well was measured just before the experiment in the culture plate by an EVOM chopstick instrument (World Precision Instruments). The transport buffer (pH 6.5) was removed from the apical side, and 425 μl fresh transport buffer (pH 6.5) including the test substance was added to the apical (donor) well. The plates were incubated in a polymix shaker at 37°C with a low shaking velocity of approximately 150 to 300 rpm.

After a 30-min incubation, the inserts were moved to new prewarmed basolateral (receiver) wells every 30 min. Twenty-five-microliter samples were taken from the apical solution after approximately 2 min and at the end of the experiment. Three hundred microliters was taken from the basolateral (receiver) wells at each scheduled time point, and the post value of transepithelial electrical resistance value was measured at the end of the experiment. For all collected samples, acetonitrile was added to a final concentration of 50% and stored at −20°C until analysis by liquid chromatography-mass spectrometry.

The cumulative fraction absorbed (FA_cum) was calculated as [summ](C_Ri/C_Di), where C_Ri is the receiver concentration at the end of interval i and C_Di is the donor concentration at the beginning of interval i. A linear relationship was obtained.

The determination of permeability coefficients (P_app, in cm/s) was calculated by the formula (k·V_R)/(A·60), where k is the transport rate (min⁻¹), defined as the slope obtained by linear regression of FA_cum as a function of time (min), V_R is the volume in the receiver chamber (ml), and A is the area of the filter (cm²).

Metabolic stability screening in the presence of liver microsomes and the determination of in vitro intrinsic clearance (CL_int).

CHI/1043 was incubated at 37°C with human liver microsomes (pooled, 20 mg/ml; In Vitro Technologies) in order to determine the metabolic stability. The incubation mixture consisted of human liver microsomes (0.5 mg/ml), NADPH (1 mM), and CHI/1043 (2 μM) in a total volume of 0.3 ml potassium phosphate buffer, 100 mM, pH 7.4. The reaction was started by the addition of NADPH. After 0, 10, 20, and 30 min, aliquots (50 μl) were removed and acetonitrile (150 μl) was added in order to stop the reaction. The samples were centrifuged (10 min at 20,800 × g at 10°C), and 5 μl of the supernatant was injected into the liquid chromatography-mass spectrometry system for determination of CHI/1043 concentrations.

The concentrations of CHI/1043 in the incubation samples were calculated from a curve obtained by linear regression of calibration standard concentrations. The natural logarithm (residual concentration) was plotted against time. The slope of the line is the elimination rate constant (k), from which the elimination half-life (t_1/2) was calculated.

The equation for determining the t_1/2 was −0.693/k. The equation for determining the CL_int was 0.693/t_1/2 × (ml incubation volume/mg microsomes). The mean value of the CL_int was calculated and converted to μl/min/mg.

RESULTS

Inhibition of HIV and SIV replication in cell culture.

In order to identify the most potent congener of the 1H-benzylindole series (10), several derivatives were tested for their inhibitory effect on IN enzymatic activity and their effect against HIV-1 replication in MT-4 cells. The biological results showed that the DKA (CHI/1043, CHI/1014, CHI/1089, CHI/1047, and CHI/1203) were more active than the corresponding esters (CHI/1042, CHI/1013, CHI/1133, CHI/1096, and CHI/1202) in both assays (Table 1) . Moreover, all methoxy-substituted DKA exhibited higher potency than the corresponding unsubstituted derivative CHI/1038. In particular, CHI/1043, characterized by the presence of one methoxy group at position 4, proved to be the most potent compound in the enzyme assays at nanomolar concentrations (Table 1). Taking into account these results, we held the 4-methoxy group constant and introduced an additional methoxy moiety at position 7 of the indole nucleus, obtaining the disubstituted derivative CHI/1203. This modification provided an INSTI active at nanomolar concentrations but did not improve the potency of the most active derivative, CHI/1043. Moreover, when the synthesized compounds were tested against HIV-1 replication in cell cultures, all DKA showed an EC₅₀ lower than 10 μM and proved to be better than corresponding esters. Furthermore, also in this test, the 4-methoxy-substituted CHI/1043 and 4,7-dimethoxy analogue CHI/1203 presented the best anti-HIV activities and high selectivity indices, thus holding promise for further development. These findings showed that the additional methoxy group at position 7 produced no change either against IN or in a cell-based assay, thus suggesting that the improvement of activity is mainly connected with the presence of the 4-methoxy substitution. In the rest of the experiments, we evaluated only the most potent congener, CHI/1043, which inhibits overall and strand transfer enzymatic activities of HIV-1 IN, with 50% inhibitory concentrations ranging from 0.11 to 0.14 μM. CHI/1043 was found to inhibit HIV-1(III_B) replication in MT-4 cells at an EC₅₀ value of 0.60 μM (Table 1). Fifty percent cytotoxicity was observed at 41 μM in mock-infected MT-4 cells, resulting in a selectivity index of 70 (Table 1). The anti-HIV activity of CHI/1043 was also evaluated against HIV-1(NL4.3), HIV-2(ROD), HIV-2(EHO), and SIV(MAC₂₅₁) in MT-4 cells (Table 2). Several reference compounds, i.e., the NRTI AZT, the NNRTI efavirenz, and the protease inhibitor nelfinavir, were evaluated in parallel. CHI/1043 was threefold less active against SIV(MAC₂₅₁) than it was against HIV(III_B), but HIV-2(ROD) and HIV-2(EHO) remained susceptible to the inhibitory effect, whereas efavirenz was inactive against HIV-2(ROD), HIV-2(EHO) and SIV(MAC₂₅₁). Replication of X4 (IIIB and 92HT599) and R5 (BAL) strains in peripheral blood mononuclear cells was twofold more sensitive to the inhibitory effect of CHI/1043 (EC₅₀, 0.30 to 0.38 μM) (data not shown) than HIV-1(III_B) replication in MT-4 cells. In conclusion, CHI/1043 showed higher potency than the DKA L-708,906. Also, the compound proved to be active against the replication of HIV-1 and HIV-2.

TABLE 2.

Antiretroviral activity of CHI/1043 and reference compounds

Compound	Mean EC50 ± SD (μM)a
	HIV-1		HIV-2		SIV(MAC251)
	IIIB	NL4.3	ROD	EHO
RT inhibitors
AZT	0.027 ± 0.003	0.021 ± 0.002	0.067 ± 0.037 (2.4)	0.056 ± 0.019 (2)	0.011 ± 0.005 (0.4)
Efavirenz	0.015 ± 0.001	0.0019 ± 0.0013	>31.68 (>2,128)	>31.68 (>2,128)	>31.68 (>2,128)
IN inhibitors
L-708,906	12.4 ± 6.6	24.5 ± 6.4	62.1 ± 34.9 (5)	ND	22.6 ± 20.8 (1.8)
CHI/1043	0.60 ± 0.38	0.73 ± 0.22	2.11 ± 0.70 (3.5)	1.76 ± 0.76 (1.2)	4.76 ± 3.98 (3.1)
Protease inhibitor nelfinavir	0.028 ± 0.001	0.023 ± 0.014	0.050 ± 0.002 (1.8)	0.049 ± 0.017 (1.3)	0.050 ± 0.008 (1.8)

^aConcentrations of each compound required to inhibit the CPE of retroviruses in MT-4 cells by 50%. In parentheses are the increases (n-fold) in EC₅₀ value compared to activity against HIV-1(III_B). All data represent mean values ± standard deviations for at least three separate experiments. ND, not determined.

Inhibitory activity of CHI/1043 against mutant HIV strains.

Next, CHI/1043 was evaluated for its inhibitory effects against the replication of a variety of mutant HIV strains (Table 3). We observed no cross-resistance of CHI/1043 against the NRTI-, NNRTI-, and protease inhibitor-resistant virus strains (Table 3).

TABLE 3.

Antiviral activity of CHI/1043 and reference compounds against HIV-1 strains resistant to reverse transcriptase and protease inhibitors

Compound	Mean EC50 ± SD (μM) for strain typea
	Reverse transcriptase inhibitor resistant		Protease inhibitor resistant
	NRTI resistant	NNRTI resistant
RT inhibitors
AZT	0.20 ± 0.02 (7.3)	0.020 ± 0.005 (0.7)	0.026 ± 0.003 (1.2)
Efavirenz	0.00095 ± 0.00032 (0.6)	0.086 ± 0.025 (57.2)	0.0016 ± 0.0003 (0.8)
IN inhibitors
L-708,906	7.2 ± 1.2 (0.6)	19.3 ± 14.7 (0.8)	7.4 ± 0.5 (0.6)
CHI/1043	0.86 ± 0.50 (1.4)	0.84 ± 0.24 (1.4)	0.68 ± 0.30 (1.1)
Protease inhibitor nelfinavir	0.029 ± 0.017 (1.0)	0.035 ± 0.006 (1.3)	0.17 ± 0.07 (7.4)

^aConcentration required to inhibit the viral CPE by 50% in MT-4 cells. In parentheses are the increases (n-fold) in EC₅₀ value, compared to activity against the HIV-1(III_B) strain. All data represent mean values ± standard deviations for at least three separate experiments.

Mechanism-of-action studies. (i) Time of intervention.

A time-of-addition experiment was carried out to confirm HIV IN as a target of inhibition. This experiment determines how long the addition of an anti-HIV compound can be postponed within the viral replication cycle before losing its antiviral activity. Virus was added at a high MOI (0.5) to synchronize all steps of viral replication. Reference compounds with a known mode of action were included. DS, a polyanion, is known to interfere with binding of the virus to the cell. The nucleoside analogue AZT inhibits the reverse transcription process. Ritonavir is an inhibitor of proteolytic cleavage. The addition of DS cannot be delayed after infection, and the addition of AZT and ritonavir can be delayed 3 and >10 h after infection, respectively (Fig. 2). The known strand transfer inhibitor, MK-0518, lost its activity when added more than 7 h beyond infection. The same profile was seen with CHI/1043 (Fig. 2). These data confirmed IN as a target of interaction of CHI/1043.

FIG. 2.

Time-of-addition assay. MT-4 cells were infected with HIV-1(III_B) at an MOI of 0.5, and the test compounds were added at different times postinfection. Viral p24 antigen production was determined at 31 h postinfection. Compounds used were DS (20 μM), AZT (1.9 μM), ritonavir (2.8 μM), MK-0518 (0.05 μM), and CHI/1043 (27 μM). The results are from a representative experiment that was repeated at least once.

(ii) Effect of CHI/1043 on HIV-1 infection and HIV-1 vector transduction kinetics.

We carried out Q-PCRs on DNA extracted from cells transduced with HIV vectors or infected with HIV. We transduced 293T cells with VSV G protein-pseudotyped HIV-1 vectors encoding enhanced green fluorescent protein, and we infected MT-4 cells with HIV-1(NL4.3), in the absence or presence of 12.2 μM CHI/1043 (16 times its EC₅₀ as determined in the MT-4/MTT assay). Results of DNA quantification are presented in Fig. 3. Neither during lentiviral transduction nor during virus infection was inhibition of early cDNA synthesis observed in the presence of CHI/1043 (Fig. 3A and D, respectively). However, no integrated proviral DNA was detected by Q-PCR (Fig. 3B and E), whereas the amount of 2-LTR circles clearly increased in the presence of CHI/1043 (Fig. 3C and F). These data confirm that integration of viral cDNA is blocked by CHI/1043.

FIG. 3.

Kinetics of VSV G protein-pseudotyped HIV-1 vector transduction in the presence of CHI/1043. 293T cells (∼1.5 × 10⁵ cells) were transduced with VSV G protein-pseudotyped HIV-1 vector at an MOI of 10 in the absence (⋄) or presence (▴) of 12.2 μM CHI/1043. DNA extracts were made from samples taken at different time points, and formations of total HIV-1 DNA (A), integrated proviral DNA (B) and 2-LTR circles (C) were quantified using Q-PCR. Kinetics of HIV-1 infection in the presence of CHI/1043. MT-4 cells (1.5 × 10⁶ cells) were infected with HIV-1(NL4.3) (150 ng p24) in the absence or presence of 12.2 μM CHI/1043. DNA extracts were made from samples taken at different time points, and formations of total HIV-1 DNA (D), integrated proviral DNA (E), and 2-LTR circles (F) were quantified using Q-PCR. All data represent mean values ± standard deviations from at least two separate experiments.

(iii) Selection and analysis of HIV-1 strains resistant to CHI/1043.

CHI/1043-resistant HIV-1 strains were selected by serial passage of HIV-1(III_B) in the presence of increasing concentrations of the compound. After 70 passages, the selected strain was able to grow at a compound concentration of 25 μM, a concentration that is >40-fold higher than the concentration required to inhibit the replication of wild-type HIV-1(III_B) by 50% (EC₅₀) (0.60 μM). The mutations T66I and Q146K were detected in the IN of the selected strain. Both IN mutations are located in close proximity to each other and to the active site of IN. This clustering of resistant substitutions near the active site suggests that CHI/1043, like other DKA, binds at or near the enzyme active site and thereby directly disrupt IN catalysis. The mutant strain carrying the IN mutations T66I and Q146K showed reductions in susceptibility to inhibition by several INSTIs, with increases in strain resistance ranging from >5.9-fold to >55.2-fold (Table 4). In addition, CHI/1043 showed reductions in activity against strains resistant to L-708,906, L-870,810, or MK-0518 (Table 4), with increases in strain resistance ranging from 5.9- to >39-fold. These results were in agreement with previous resistance profiles obtained with IN inhibitors (13, 14, 16, 18, 20).

TABLE 4.

Reduced susceptibilities of resistant HIV strains to CHI/1043 and reference compounds

Compound	Fold resistance of strain resistant toa:
	L-708,906 (T66I, L74M, and S230R)	L-870,810 (L74M, E92Q, and S230N)	MK-0518 (G140S and Q148H)	CHI/1043 (T66I and Q146K)
IN inhibitors
CHI/1037	>14.2	8.9	>13	>13
CHI/1043	12.5	5.9	>39.1	>55.2
L-708,906	>20.5	ND	>5.9	>5.9
L-870,810	17.9	47.5	>954.7	23.5
GS/9137	149.6	83.8	3,185.6	154.0
MK-0518	20.4	11.3	412.5	14.4
RT inhibitors
AZT	1.1	2.0	1.1	1.8
Efavirenz	1.2	0.9	0.8	1.9
Protease inhibitor ritonavir	0.7	0.8	0.7	1.8

^aIncreases (n-fold) in EC₅₀ value, compared to activity against HIV-1(III_B) for L-708,906-resistant and L-870,810-resistant strains and HIV-1(NL4.3) for MK-0518-resistant and CHI/1043-resistant strains. Mutations of the indicated strains are given in parentheses. The results are from a representative experiment that was repeated at least once. ND, not determined.

Evaluation of the physicochemical and pharmacokinetic properties of CHI/1043.

We initiated early preclinical adsorption, distribution, metabolism, excretion, and toxicity studies with CHI/1043. Analysis of protein binding of CHI/1043 revealed that the presence of 50% human serum has a significant impact on the antiviral activity of CHI/1043 (19-fold increase in comparison to standard conditions [10% FCS]) (Table 5). Similar effects were observed for other INSTIs (L-731,988 and L-870,810), the binding inhibitor DS, and the NNRTI efavirenz. Only the antiviral effect of AZT was not influenced by the presence of human serum.

TABLE 5.

Effect of protein binding on antiviral effect of HIV inhibitors

Compound	Fold increase in EC50a
Adsorption inhibitor DS	18.8
RT inhibitors
AZT	0.7
Efavirenz	15.3
IN inhibitors
L-708,906	14.6
CHI/1043	18.9
L-870,810	17.6

^aIncrease in EC₅₀ in the presence of 50% human serum was calculated in comparison to the EC₅₀ determined in the presence of standard conditions (10% FCS). The results are from a representative experiment that was repeated at least once.

The results of the metabolization experiments using human liver microsomes are expressed as CL_int. CHI/1043 showed a CL_int of 11 μl/min/mg. The metabolization data for CHI/1043 are still acceptable for the continuance of drug development, since CL_int for a candidate drug should be lower than 30 μl/min/mg. However, permeability coefficients determined in Caco transport experiments (P_app, 5.5 × 10⁻⁶ cm/s) proved that CHI/1043 has no satisfactory features for uptake after oral administration, since the desired P_app for a candidate drug should be >15 × 10⁻⁶ cm/s.

DISCUSSION

Development of new antiretroviral drugs against novel targets, with different resistance profiles and improved safety and tolerability, is urgently needed for both treatment-naive and treatment-experienced patients, especially those with multidrug-resistant HIV. Besides reverse transcriptase and protease, the two viral enzymes targeted by current antiretroviral therapy, HIV encodes a third enzyme, IN. DKA were the first compounds reported to interfere with HIV replication through specific inhibition of the integration step (17). When we embarked on this drug development project in 2005, there was a need for the development of first-generation inhibitors of HIV IN for the clinic. The goal of this study was to identify more-potent HIV IN inhibitors with interesting pharmacological profiles. We started the development of 1H-benzylindole diketo compounds (10) based on a three-dimensional pharmacophore model for DKA-like derivatives acting as INSTIs (3, 9).

Molecular modeling and structure-activity relationship of this series enabled us to develop the most potent congener encoded, CHI/1043 (10). CHI/1043 inhibited the replication of HIV-1(III_B) in MT-4 cells at an EC₅₀ of 0.60 μM, which is 70-fold below its cytotoxic concentration (Table 1). This compound is 14-fold more selective than the DKA L-708,906. Similar activities against HIV-1(NL4.3), HIV-2(ROD), HIV-2(EHO), and SIV(MAC₂₅₁) were observed (Table 2). Replication of X4 and R5 strains in PBMCs was twofold more sensitive to the inhibitory effect of CHI/1043 (EC₅₀, 0.30 to 0.38 μM; data not shown). CHI/1043 inhibited IN activity in oligonucleotide-based enzymatic assays at low micromolar concentrations (Table 1). CHI/1043 proved at least sixfold more active in inhibiting HIV IN enzymatic activity than L-708,906, whereas L-870,810, MK-0518, and GS-9137 showed at least 10-fold-higher levels of activity (Table 1). Time-of-addition experiments (Fig. 2) and Q-PCR (Fig. 3) corroborated that the anti-HIV activity of CHI/1043 in cell culture is based upon the selective inhibition of proviral DNA integration. It should be noted that CHI/1043 retained activity against strains resistant to reverse transcriptase and protease inhibitors (Table 3). These data further substantiate the specificity of the compound.

In order to compare CHI/1043 to known classes of INSTIs in more detail, we profiled the activity of CHI/1043 on IN carrying resistance mutations for DKA, L-870,810, and MK-0518 in an MT-4/MTT susceptibility assay. In addition, we selected HIV-1 strains in the presence of increasing concentrations of CHI/1043 and evaluated the genotypic and phenotypic (cross-)-resistance profile of these strains (Table 4). The T66I and Q146K mutations were present in the IN of the CHI/1043-selected strain. These mutations are located in the catalytic core domain of IN, close to the active site. Both IN mutations have been previously identified to be associated with in vitro resistance against DKA (14, 17) and/or the DKA analogue S-1360 (13), and the T66I mutation has also been selected by the clinical trial drug GS-9137 both in vitro (23, 43) and in vivo (29). On the phenotypic level, the CHI/1043-selected HIV(III_B) strains were more than 55-fold less susceptible to the inhibitory effect of CHI/1043, the compound used for selection. Cross-resistance against the DKA L-708,906 (>5.9-fold), the naphthyridine analogue L-870,810 (23.5-fold), and the clinical drugs GS-9137 (154.0-fold) and MK-0518 (14.4-fold) was observed as well. A similar phenotypic profile was observed for the L-708,906-resistant strain (i.e., lower-level cross-resistance in the cases of MK-0518 and L-870,810 and higher-level cross-resistance in the case of GS-9137). These genotypic and phenotypic data suggest a CHI/1043-to-IN binding mode comparable with binding modes of other DKA and IN. Previously, it has been suggested that DKA and L-870,810 have discordant resistance profiles (16); however, from our profiling data, it became obvious that cross-resistance between the INSTIs CHI/1043, L-708,906, L-870,810, MK-0518, and GS-9137 is apparent although incomplete, supporting the idea of partly overlapping binding modes between the INSTIs developed to date. The highest-level cross-resistance was observed with the G140S/Q148H IN variant (Table 4), an HIV-1 strain associated with clinical raltegravir resistance (18, 39).

Next, we determined the pharmacological properties of CHI/1043. Antiviral activity of CHI/1043 was strongly reduced by protein binding (Table 5). Comparable decreases in antiviral activity were seen in the presence of the DKA L-708,906, the naphthyridine analogue L-870,810, the adsorption inhibitor DS, and the NNRTI efavirenz. Metabolization data obtained with CHI/1043 (CL_int of 11 μl/min/mg) are on the borderline of continuing drug development, since CL_int for a candidate drug should be lower than 30 μl/min/mg. However, permeability coefficients determined in Caco transport experiments (P_app, 5.5 × 10⁻⁶ cm/s) proved that CHI/1043 has low oral bioavailability.

In summary, we have presented a class of 1H-benzylindole analogues as a novel series of selective HIV INSTIs. The antiviral activity of the most potent congener, CHI/1043, surpasses that of the early DKA; however, during our preclinical evaluation of CHI/1043, the more potent pyrimidinone carboxamide compound of MK-0518 (Merck) has received FDA approval for treatment of HIV-1 infection (25). In addition, an ongoing phase II clinical trial showed that GS-9137 (elvitegravir; Gilead Sciences) was able to significantly reduce HIV loads (43). Based on its potent biological activity and mode of action, CHI/1043 is an interesting new anti-HIV compound, but the poor in vitro pharmacological properties of CHI/1043 in comparison with those of presently available IN inhibitors confound further clinical development. Furthermore, because of substantial cross-resistance among currently developed INSTIs, efforts to develop new IN inhibitors should not only focus on the discovery of new strand transfer inhibitors. Rather, we should also aim to develop novel inhibitors of other functions of IN, like 3′ processing and binding of IN to its cellular cofactors, i.e., LEDGF/p75.