JAC Advance Access originally published online on January 23, 2006
Journal of Antimicrobial Chemotherapy 2006 57(3):411-423; doi:10.1093/jac/dki464
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Review |
Dawn of non-nucleoside inhibitor-based anti-HIV microbicides
1 Drug Discovery Program, Parker Hughes Institute, 2657 Patton Road, St Paul, MN 55113, USA; 2 Paradigm Pharmaceuticals, LLC, 2685 Patton Road, St Paul, MN 55113, USA
* Corresponding author. Tel: +1-651-628-9988; Fax: +1-651-628-9891; E-mail: ODCRUZ{at}IH.ORG
 |
Abstract |
|---|
Top
Abstract
IntroductionThe emergence of HIV/AIDS as a disease spread through sexual intercourse has prompted the search for safe and effective vaginal and rectal microbicides for curbing mucosal viral transmission via semen. Since endogenous reverse transcription is implicated in augmenting the sexual transmission of HIV-1 infection, potential microbicides should have the inherent ability to optimally inhibit both wild-type and drug-escape mutants. The non-nucleoside reverse transcriptase inhibitors (NNRTIs), which bind to an allosteric site on RT, are an important arsenal of drugs against HIV-1. The clinical success of NNRTI-based HIV/AIDS therapies has led to extensive structural and molecular modelling studies of enzyme complexes and chemical synthesis of second- and third-generation NNRTIs. Rationally designed NNRTIs deduced from changes in binding pocket size, shape and residue character that result from clinically observed NNRTI resistance-associated mutations exhibit high binding affinity for HIV-1 RT and robust anti-HIV activity against the wild-type and drug-escape mutants without cytotoxicity. Notably, membrane permeable tight binding NNRTIs have the ability to inactivate cell-free as well as cell-associated HIV-1 in semen without metabolic activation. Consequently, NNRTIs currently under development as experimental microbicides include thiourea-PETT (where PETT stands for phenethylthiazolylthiourea) derivatives (PHI-236, PHI-346 and PHI-443), urea-PETT derivatives (MIV-150), oxypyrimidines (S-DABOs), thiocarboxanilides (UC-781) and diarylpyrimidines (TMC-120). Mucoadhesive formulations of these NNRTIs have been studied for safety and efficacy in animal models and some have entered Phase I safety testing in humans. This review focuses on the structural, biological and preclinical studies relevant to the clinical development of these NNRTIs as molecular virucides intended to prevent the sexual transmission of HIV-1.
Keywords: AIDS/HIV , reverse transcriptase , non-nucleoside reverse transcriptase inhibitors , NNRTIs
|
Introduction |
|---|
Newly acquired HIV-1 infections are largely the result of heterosexual transmission.1,2 Currently, about half of the 42 million people living with HIV/AIDS are women.3,4 In the United States, the proportion of women living with HIV/AIDS has increased from 8% of all cases during 1981–87 to 17% during 1993–95, and 28% in 2003.5 Worldwide, more than 90% of all adolescent and adult HIV-1 infections have resulted from heterosexual intercourse.6 Women are particularly vulnerable to heterosexual transmission of HIV-1 due to substantial mucosal exposure to seminal fluids.7 Also, male-to-female transmission of HIV-1 is more efficient than female-to-male transmission.8
|
Microbicides for curbing mucosal viral transmission |
|---|
The emergence of HIV/AIDS as a disease spread through sexual intercourse has prompted the search for safe and effective vaginal and rectal microbicides for curbing mucosal viral transmission via semen.9 Both cell-free and cell-associated viruses are present in semen,10,11 and cervical mucus from HIV-1-infected persons.12 Leucocytes are the predominant HIV-1-infected cell types in genital tract secretions of both men and women.13,14 Because semen usually has >106 leucocytes per ejaculate, HIV-1-infected cells would be a more important source of transmission than free virus.
Microbicides can provide protection by inactivating viruses or preventing viruses from replicating either in semen or in the infected host cells that line the vaginal and rectal wall. Both contraceptive and non-contraceptive microbicides are currently under investigation.9,15–22 More than 60 products or compounds are in various pre-clinical stages of development; 18 of these products have entered Phase I safety trials, 4 are in Phase II expanded safety and preliminary efficacy trials and 6 are in Phase II/IIB or Phase III efficacy trials.23 The majority of the first-generation products are being tested for preventing pregnancy as well as protection against sexually transmissible HIV-1.9 These microbicides under development work in a variety of ways that include the following:
- Disrupting the viral membrane or envelope;
- Blocking the receptor–ligand interactions or post-fusion events essential for infectivity;
- Inhibiting the intracellular replication of the virus; or
- Altering the vaginal microenvironment, reducing the susceptibility to infection or enhancing the local immune response.
|
Genital inflammation and sexual transmission of HIV-1 |
|---|
Genital tract inflammation and immune defence mechanisms play a key role in sexual transmission of HIV-1.24,25 The ‘first-generation’ anti-HIV microbicides currently undergoing clinical evaluation are non-specific surfactants or non-specific anionic polymers that variably interfere with virus binding/fusion/entry.9 A major challenge has been to design ‘mechanism-based’ microbicides that are highly effective against HIV-1 while lacking adverse effects on in vivo sperm functions as well as the integrity of vaginal mucosa. This is especially important because currently available over-the-counter detergent-based spermicidal microbicides have been shown to damage the cervicovaginal epithelium,26–30 cause an acute inflammatory tissue response31,32 and enhance the risk of promoting opportunistic infections in the genitourinary tract.33–37
In addition, genital ulcerative diseases increase the risk of HIV-1 acquisition by up to 13-fold, both by disrupting the integrity of the genital mucosa and by attracting and activating HIV target cells to the area.38 They also contribute to HIV-1 transmission by increasing the viral load in the genital tract of infected persons, mainly due to genital ulcers, which are likely to bleed, with concomitant shedding of HIV-1.39 Activation of leucocytes that accumulate at the site of inflammation results in the production of cytokines.39 These cytokines have different effects on HIV-1 replication. Proinflammatory cytokines stimulate HIV-1 replication through activation of NF-
B, a cellular factor that activates HIV-1 transcription.40,41 Women exposed to detergent-type spermicides such as nonoxynol-9 (N-9) develop high levels of inflammation in the vagina and increased levels of proinflammatory cytokines in cervicovaginal lavage fluid.32 Thus, women with vaginal inflammatory conditions induced by potential vaginal irritants may be at higher risk for HIV-1 infection due to multiple mechanisms of barrier dysfunction.
|
NNRTIs as anti-HIV microbicides |
|---|
Design of potent inhibitors of HIV-1 reverse transcriptase, an enzyme responsible for the reverse transcription of the viral RNA to proviral DNA, has been a focal point in translational AIDS research.42,43 The success of non-nucleoside inhibitors of HIV-1 reverse transcriptase (NNRTIs) for the clinical treatment of AIDS has led to the computer-aided design and chemical synthesis of second- and third-generation potent NNRTIs.9,18,21,22,44–46 Several NNRTIs, due to their unique virucidal property attributed to their tight binding to HIV-1 RT, are currently under development as candidate microbicides to inhibit cell-free as well as cell-associated HIV-1 replication in genital tract secretions.18,21,22,47–59
|
Advantages of tight binding NNRTIs |
|---|
The rationale for the development of tight binding NNRTIs as microbicides is that unlike nucleoside analogue RT inhibitors (NRTIs) they do not require metabolic activation to achieve antiviral activity. Therefore, NNRTIs can directly exert their antiviral action against cell-free and cell-associated HIV-1 within the vaginal or rectal cavity. Although NNRTIs belonging to diverse classes exhibit potent anti-HIV-1 activity, only tight binding lipophilic NNRTIs are endowed with microbicidal activity. Additional criteria for an NNRTI to be an ideal microbicide include the following:
- Ability to rapidly cross membrane barriers.
- Prolonged or irreversible inhibition of HIV-1 RT activity.
- Rapid virucidal activity without metabolic activation.
- Retain antiviral activity under acidic conditions.
- Stable under various climatological temperatures.
- Minimal binding to genital tract components.
- Retain virucidal activity following drug removal.
- Lack systemic absorption to prevent drug resistance.
- Should not perturb the vaginal mucosa and normal vaginal flora.
Considerable progress has been achieved in regard to such NNRTIs that act as tight binding inhibitors to the HIV-1 RT enzyme.44,46 These NNRTIs are expected to inactivate RT enzymic activity in a highly efficient manner, thus eliminating the ability of the virus to initiate a new round of infection. Because of the possibility for transmission of HIV-1 from serodiscordant couples via assisted reproductive technology procedures (e.g. artificial insemination, in vitro fertilization), it is also desirable to provide a non-contraceptive prophylactic agent capable of suppressing HIV-1 infectivity. Thus, the availability of a non-spermicidal microbicide is equally important for: (i) sexually active women to allow pregnancy while protecting both mother and her fetus or infant from HIV-1; and (ii) as a prophylactic antiviral agent to curb the transmission of cell-free and cell-associated HIV-1 via semen during assisted reproductive procedures. Examples of tight binding NNRTIs currently under development as experimental microbicides include thiourea-PETT (where PETT stands for phenethylthiazolylthiourea) derivatives (PHI-236, PHI-346 and PHI-443), urea-PETT derivatives (MIV-150), oxypyrimidines (S-DABO), thiocarboxanilides (UC-781) and diarylpyrimidines (TMC-120).
|
NNRTI binding pocket |
|---|
NNRTIs are structurally diverse hydrophobic compounds that bind to an allosteric site of HIV-1 RT,60–62 which is
10 Å away from the polymerase active site in the ‘palm’ domain of the p66 subunit (Figure 1a). NNRTI binding induces rotamer conformation changes in amino acid residues (Y181 and Y188) and makes the ‘thumb’ region of the enzyme more rigid.63,64 Both events alter the substrate-binding mode and/or affect the translocation of the double strand, which are critical for the polymerase function, thereby leading to non-competitive inhibition of the enzyme.63,64
|
Drug-resistant RT mutants have been proposed to obstruct the binding of NNRTI to HIV-1 RT. NNRTI resistance is primarily associated with mutations of the amino acids lining the lipophilic NNRTI binding pocket in the p66 subunit of the RT.46,64 Dozens of mutant strains have been characterized as resistant to first-generation NNRTIs, including L100I, K103N, V106A, E138K, Y188I/C and Y188H.65–67 In particular, the K103N and Y181C mutants are the most difficult to treat, because they are resistant to most of the NNRTIs that have been examined.65 For example, primary mutations associated with resistance to nevirapine involve residues K103, V106, V108, Y181, Y188 and G190, which have van der Waals contact with the NNRTI. The mutations of these residues lead to the weakening of the NNRTI binding to HIV-1 RT. Currently used NNRTIs, delavirdine, efavirenz and nevirapine, are less active against RT with primary mutations K103N or Y181C.66–68
|
Rational design of tight binding thiourea NNRTIs |
|---|
A novel computer model of the NNRTI binding pocket of HIV-1 RT was constructed by superimposing nine individual RT–NNRTI crystal structures and generating a van der Waals surface that encompassed all the overlaid ligands.44,46,64 This ‘composite binding pocket’ revealed a different and unexpectedly larger NNRTI binding site than shown in or predictable from any of the individual structures and served as a probe to more accurately define the potentially usable space in the binding site (Figure 1b). This composite NNRTI binding pocket model was used to design potent NNRTIs against wild-type RT and drug-resistant RT mutants.45,69–76 Molecular modelling and score functions were used to analyse how drug-resistance mutations would change the RT binding pocket shape, volume and chemical make-up of PETT NNRTIs, and how these changes could affect NNRTI binding.21,22,44–46,64 The inhibitory constants (Ludi Ki values) of docked NNRTIs were calculated using a calibration procedure which involved calculation of the Ki values of 45 protein–ligand complexes having known Ki values and known crystal structures.
|
PETT NNRTIs can be viewed as two chemical moieties linked together by a thiourea group (Figure 2). The left side of the molecule is either a 2-aminothiazole group (PETT) or a 5-bromopyridyl group (trovirdine) capable of forming an intramolecular H-bonded heterocyclic ring.77,78 The right side of the molecule is a pyridyl or phenyl ring separated from the thiocarbonyl group by an ethyl linker. Trovirdine docked into the NNRTI binding site of HIV-1 RT had a higher binding score than the parent compound (PETT) and fits into the butterfly-shaped binding region with one part residing in Wing 1 and the other in Wing 2. Docking studies with PETT and trovirdine revealed that the Wing 2 region defining space of the butterfly-shaped binding pocket has a substantial molecular volume (
160 Å) surrounding the phenyl ring of PETT compounds that can be more efficiently occupied by a larger functional group to achieve a high binding affinity even against the problematic Y181C and Y188 CRT mutants. The binding of NNRTI forces RT residue W229 to change its position slightly and causes residues Y181 and Y188 to rotate into another rotamer conformation. Consequently, the binding pocket would be substantially larger than it was before NNRTI binding, forcing the primer-template into an inactive binding conformation and rendering the protein inactive. This volume change is a direct consequence of the different positions of the Y181, Y188 and W229 side chains before and after the NNRTI binding. When Y181 and Y188 are mutated to cysteine residues, the volume change due to NNRTI binding is smaller and the impact of NNRTI inhibiting the RT mutants would be attenuated. This idea fits well with adding a larger functional group at the Wing 2 region. Therefore, NNRTI, which has a maximum occupancy at the Wing 2 region, was predicted to have an advantage against Wing 2 mutants, such as the Y181C and Y188C mutants. Thus, an ideal NNRTI should: (i) be highly potent against wild-type RT and therefore afford a considerable activity loss against mutants (i.e. a picomolar-level inhibitor against wild-type RT may still be effective against RT mutants at nanomolar levels); (ii) maximize the occupancy at the Wing 2 region, which will have an advantage against the Wing 2 mutants; and (iii) the substitutions should match with the chemical nature of the residue that is potentially mutated in the RT mutant.
|
Thiourea NNRTIs as anti-HIV microbicides |
|---|
Using the composite binding pocket, a series of phenyl, heterocyclic and alicyclic ring-containing thiourea NNRTIs were synthesized that yielded high binding affinity for HIV-1 RT and robust anti-HIV activity against wild-type and drug-escape mutants.21,44,64,69–76 The thiourea NNRTIs were more potent against drug-sensitive and multidrug-resistant strains of HIV-1 than the three classes of NNRTIs currently in clinical use to treat HIV infections.69–76 Representative thiourea NNRTIs under development as molecular virucidal microbicides include:
PHI-236 (N-[2-(2,5-dimethoxy phenylethyl)]-N'- [2-(5-bromopyridyl)]-thiourea)
Replacing the 2-pyridyl ring of trovirdine with a 2,5-dimethoxy-substituted phenyl ring was predicted to yield a potentially more active PETT-NNRTI with a larger molecular surface area, higher binding score and lower Ludi Ki value (Figure 2). As expected, PHI-236 displayed high affinity (Ludi Ki = 0.07 µM) for the NNRTI binding pocket of HIV-1 RT and abrogated HIV replication at nanomolar concentrations (IC50[p24] = <1 nM) with a selectivity index of >100 000.69,79 PHI-236 displayed robust anti-HIV activity against drug-resistant HIV-1 strains and genotypic and/or phenotypically NRTI/NNRTI-resistant primary clinical non-subtype B isolates originating from South America, Asia and sub-Saharan Africa (Table 1). PHI-236 was 4 times more potent than trovirdine, 40 times more potent than zidovudine, 60 times more potent than emivirine, 80 times more potent than delavirdine and 1000 times more potent than nevirapine against the multidrug-resistant HIV-1 strain, RT-MDR, with mutations involving RT residues M41L, L74V, V106A and T215Y.69 Similarly, PHI-236 was 500–1000 times more effective than delavirdine and nevirapine against the problematic NNRTI-resistant HIV-1 strain A17 with a Y181C mutation. The anti-HIV-1 activity of PHI-236 against the A17 variant with Y181C and K103N mutations was also superior. Exposure of human sperm to PHI-236 at doses 106-times its anti-HIV IC50 value had no adverse effects on human sperm functions or the viability of human female genital tract epithelial cells.22
|
PHI-236 was remarkably more potent than the standard NNRTI drugs against the problematic multidrug-resistant HIV-1 strains with mutations involving RT residues M41L, L74V, K103N, V106A, Y181C or T215Y. PHI-236 (IC50 = 0.009 µM) was 270-fold more potent than zidovudine (IC50 = 2.44 µM) against genotypically and phenotypically NRTI-resistant non-subtype B HIV-1 isolates originating from South America, Asia and sub-Saharan Africa carrying 2–5 thymidine analogue mutations in amino acid sequence 20–219 (M41L, E44D, D67N, T69D, K70R, L74V, K103N, F116S, M184V, Y181C, L210W, T215Y or K219Q).22 This is particularly relevant because a high percentage of newly infected individuals harbour NRTI/NNRTI-resistant mutants with increased incidence of HIV subtypes.80,81 Subtype B predominates in North America and Europe.82 However, HIV-1 subtype B currently accounts for only 12% of the estimated 42 million HIV-infected individuals worldwide and the vast majority of new infections are caused by non-subtype B HIV-strains.83
Docking studies revealed that replacement of a pyridylethyl group of trovirdine with a 2,5-dimethoxy-substituted phenyl ring provides favourable contacts with binding site residues and stronger binding to RT.79 The addition of 2,5-dimethoxy groups in PHI-236 increased the molecular volume in the Wing 2 region of the binding site by 18 Å.69,79 Thus, PHI-236, which has a maximum occupancy at the Wing 2 region and is more closely in contact with residues L100 and L234 has an advantage against Wing 2 mutants, such as the Y181C and Y188C mutants (Figure 3a). An energy-minimized model of PHI-236 in the RT binding site revealed the largest molecular surface area in contact with the protein and thus achieved the highest lipophilicity score. The larger surface area and favourable chemical properties of PHI-236 contributed to a better lipophilic score and better Ludi Ki value than trovirdine. The 2-methoxy group situated beneath the ethyl linker fits favourably into a cavity of the binding pocket, providing additional contact with protein residues. Likewise, the 5-methoxy group provides close contact with residues P95 and W229 and is near Y181. Inhibitor interactions with P95 and W229 are especially desirable because no RT mutations have been reported for these residues.
|
The improved anti-HIV activity of PHI-236 against RT mutants is also consistent with the structural analysis of PHI-236 binding to RT based on X-ray crystallographic data.79 The hydrogen (H) bonds involved in the intramolecular interaction between a thiourea NH and the pyridyl N [N-H ... N = 2.671] locks the molecule into a relatively planar conformation. A second H bond between a thiourea N atom and the thiocarbonyl-S atom [N-H2 ... S = 3.403] allows the formation of inversion-related H bonded dimers. This compact conformation would allow PHI-236 to more easily fit into the NNRTI binding site (Figure 3a).
PHI-236 due to its lipophilic as well as tight binding attributes could provide protection by directly inactivating HIV-1 or preventing HIV-1 replication either in semen or the infected host cells that line the vaginal wall. PHI-236 prevented the vaginal transmission of HIV-1 monotropic R5 strain BaL in the humanized severe combined immunodeficient (Hu-SCID) mouse model of vaginally transmitted HIV-1.22 PHI-236 had no effect on the viability of normal human female genital tract epithelial cells and did not affect human sperm functions at a concentration 106-times its in vitro anti-HIV IC50.22
PHI-346 (N-[2-(1-cyclohexenyl)ethyl]-N'- [2-(5-bromopyridyl)]-thiourea)
Docking studies using the computer-generated model of the NNRTI binding pocket suggested that the replacement of the planar pyridyl ring of trovirdine with a non-planar cyclohexenyl ring, which occupies a larger volume, would better fit the spacious Wing 2 region of the NNRTI binding pocket. The calculated Ludi Ki value of 0.16 µM for the cyclohexenyl-substituted thiourea NNRTI, PHI-346, was better than the Ki value of 0.64 µM for trovirdine.72 Functionalization at the 5'-position of the pyridyl ring of cyclohexenyl ring-containing thioureas with a Br atom led to a significant increase in anti-HIV activity as well as gain of spermicidal function.18,45,72
PHI-346 was 20 times more potent than trovirdine, 200 times more potent than zidovudine, 300 times more potent than emivirine, 400 times more potent than delavirdine and 5000 times more potent than nevirapine against the multidrug-resistant HIV-1 strain RT-MDR.72 PHI-346 was 3 times more effective against the multidrug-resistant HIV-1 strain RT-MDR than it was against HTLVIIIB with wild-type RT. PHI-346 due to its lipophilic (log P = 4.01) as well as tight binding (Ludi Ki = 0.16 µM) attributes exhibited remarkable microbicidal activity against a multidrug-resistant HIV-1 strain in the Hu-SCID mouse model of vaginally transmitted HIV-1 with a high selectivity index against normal human female genital tract epithelial cells.18,45
When PHI-346 docks into the NNI binding site of RT, it fits into the butterfly-shaped binding region with one part of the molecule residing in Wing 1 and the other in Wing 2 (Figure 3b). The docking results indicated that the cyclohexenyl group of PHI-346 is situated in the Wing 2 region of the NNRTI binding pocket, providing contact with RT residues including Y181 (Figure 4a). In addition, the cyclohexenyl group contains more ring hydrogens than the heterocyclic pyridyl ring and therefore has more hydrogen-atom-mediated contacts and fewer carbon-atom-mediated contacts with RT residues than trovirdine. Because the cyclohexenyl rings are conformationally more flexible than aromatic-ring-containing thiourea derivatives, they are likely to have an added advantage by being able to fit an uncompromising binding pocket more effectively than conventional NNRTIs.
|
PHI-443 (N'-[2-(2-thiophene)ethyl]-N'- [2-(5-bromopyridyl)] thiourea)
PHI-443, a rationally designed thiophene thiourea NNRTI is a potent inhibitor of NRTI-resistant, NNRTI-resistant and multidrug-resistant HIV-1.21,46,71 As predicted, PHI-443 was 5 times more potent than trovirdine, 50 times more potent than zidovudine, 75 times more potent than efavirenz, 100 times more potent than delavirdine and 1250 times more potent than nevirapine against the multidrug-resistant HIV-1 strain RT-MDR. It was 12 times more potent than trovirdine, 1250 times more potent than delavirdine and 2500 times more potent than nevirapine against strain A17 with a Y181C mutation. Similarly, PHI-443 was more effective than trovirdine, delavirdine and nevirapine against the problematic NNRTI-resistant HIV-1 strain A17-variant with both Y181C and K103N mutations in RT. Notably, PHI-443 was active against 23 genotypically and/or phenotypically NRTI/NNRTI-resistant primary clinical HIV-1 isolates (subtypes A, B, F and G carrying 2–7 thymidine analogue mutations) with a mean IC50 value of 0.02 µM (Table 1).21
PHI-443 formulated via a non-emulsifying gel lacked mucosal toxicity in the rabbit and porcine models of vaginal inflammation.21,84 PHI-443-treated HIV-1-infected blood lymphocytes in the presence of semen led to complete loss of their ability to transmit a clinical HIV-1 isolate with genotypic (D67N, L214F, T215D, K219Q) and phenotypic (zidovudine IC50 = 0.2 µM) NRTI-resistance in the vaginally inoculated Hu-PBL-SCID mouse model of genitally transmitted AIDS.21 PHI-443 is lipophilic (log P = 4.39) and therefore can readily enter membrane milieu such as the plasma membrane or the membrane envelope surrounding the HIV-1 core. Exposure of human sperm to PHI-443 at doses 30 000 times its anti-HIV IC50 had no effect on sperm motility, kinematics, cervical mucus penetrability or the viability of genital tract epithelial cells.21 Consequently, PHI-443 will be a tremendous advantage for the development of a non-spermicidal broad-spectrum prophylactic anti-HIV microbicide.
Docking studies with PHI-443 indicated that the thiophene ring situated in the Wing 2 region of the NNRTI binding pocket provides better contact with RT residues including Y181.71 The thiophene group of PHI-443 was found to be located in close proximity of the Y181 residue (Figure 4b). In this docked position, the sulphur (S) atom of the thiophene ring is only 4.4 Å away from the C atom of the Y181 residue, which is mutated to an S atom in the RT Y181C mutant strains (A17 and A17 variant). The S atom of the thiophene group is more compatible with the sulphur-containing cysteine-181 residue than the pyridyl group of trovirdine.
The solubility, stability, pharmacokinetic features, bioavailability, toxicity and efficacy of thiourea NNRTIs have been examined in test animal species.84–86 At the dose ranges tested (10–400 mg/kg) thiourea NNRTIs were non-toxic to mice when administered intravenously or intraperitoneally.85,87 Thiourea NNRTIs showed favourable pharmacokinetics and did not cause acute or subacute toxicity. Pharmacokinetic studies following oral administration of thiourea NNRTIs (250–400 mg/kg bolus dose) indicated a very low capacity (<2%) to be absorbed through the mucosal epithelium.85,87 The estimated oral bioavailability of PHI-236 was only 2.2% with a predicted maximum concentration of 1.5 µM.85 PHI-236 showed a rapid absorption with time to reach a maximum plasma PHI-236 concentration of 5.8 min. The poor oral bioavailability of these NNRTIs was also confirmed by the lack of systemic absorption via the vaginal route, which is advantageous for topical application as microbicides.
Other tight binding NNRTIs that doomed them as systemic antiviral drugs are currently being pursued as intravaginal/intrarectal microbicides. Such NNRTIs include MIV-150, UC-781 and TMC-120 (Figure 5).
|
MIV-150 [(1S;2S)-N-(cis-6-fluoro-2-hydroxy-3-propionyl-phenyl)cyclopropyl]-N'-(5-cyanopyrid-2-yl)urea]
In pre-clinical studies, MIV-150 blocked in vitro infectivity of free virus with a reported IC50 value of 0.01 µM and inhibited viral replication at low nanomolar concentrations.88 The rate of in vitro resistance development to MIV-150 was three times slower than that of efavirenz.
In microbicide efficacy studies, MIV-150 showed potent activity against simian-human immunodeficiency virus (SHIV) in monkeys and prevented infection when dosed after SHIV inoculation and showed a good profile in pre-clinical safety and toxicology.47 A Phase I clinical study has revealed that the compound has low oral bioavailability, is well tolerated and not easily systemically absorbed.
S-DABO (5-isopropyl-2-[(methylthiomethyl)thio]-6-(benzyl)-pyrimidin-4-(1H)-one)
NNRTIs belonging to the dihydro-alkoxy-benzyl-oxopyrimidine (DABO) family are highly potent and specific inhibitors of HIV-1 RT.49,50,89,90 Structure–activity relationship profiles of DABOs together with molecular modelling studies on their putative binding mode have shown that the presence of a C2-alkoxy (DABOs) or C2-alkylthio (S-DABOs) side chain is a structural determinant for the anti-HIV activity (Figure 5).89,90–95 Modelling studies predicted that the addition of a methyl, ethyl or isopropyl group at the fifth position of the thymine ring would lead to higher affinity for the NNRTI binding pocket (Ludi Ki values 0.05–0.56 µM) as well as conformational rotation of the 6-benzyl ring, which affects the positions of nearby Tyr-181, Tyr-183 and Tyr-188 residues. Biological evaluation indicated that S-DABO showed highest potency and selectivity index among the DABO derivatives tested.89 As expected, S-DABO derivatives are highly active against both wild-type and the Y181C HIV-1 strains. S-DABO derivatives are also endowed with potent virucidal activity. DABO compounds have been shown to suppress HIV-1 multiplication in cultures acutely infected with a very high multiplicity of infection for up to 40 days. Consequently, S-DABO derivatives are currently being explored as microbicides to prevent mucosal HIV transmission.48,51
UC-781 (N-[4-chloro-3-(3-methyl-2-butenyloxy)phenyl]-2-methyl-3-furancarbothio amide)
UC-781, a tight binding thiocarboxanilide NNRTI originally developed as a crop protection agent, is currently being developed as a vaginal and rectal microbicide. UC-781 is 5- to 10-fold more potent than conventional NNRTIs.96–99 Its binding mode resembles that of the structurally dissimilar NNRTI Cl-TIBO, with a common hydrogen bond between each carboxanilide NH- group and the main-chain carbonyl oxygen of Lys-101.99 Brief exposure of HIV-1 virions to UC-781 led to persistent endogenous RT inhibition leading to viral inactivation.52,53 Although UC-781 due to its ability to rearrange and adapt to a mutated NNRTI pocket has greater resilience to mutations in RT, prolonged exposure of wild-type strain to UC-781 resulted in emergence of RT mutations L100I, K101I, V106A, Y181C and F227L that conferred resistance.100 UC-781 was markedly less effective against mutant strains with RT mutations L100I, K101I, V106A, Y181C and F227L.
The highly conserved W229 is thought to be vital for correct protein folding or for stabilizing the complex between RT and the template-primer.62,101 Most mutations of W229 eliminate in vitro polymerase activity and viral infectivity.102–104 Consequently, W229 is a prime amino acid candidate within the HIV-1 RT for targeted design of NNRTIs. Interestingly, the interaction between the thiocarboxanilides and W229 in HIV-1 RT is crucial for antiviral potency of UC-781. Site-directed mutagenesis studies evaluating the sensitivity of residue 229 mutations have revealed substantial resistance to UC-781 (21-fold).104–106 Furthermore, modelling studies and crystallographic analysis have confirmed that the pentenyl ether group of UC-781 interacts in an optimal way with regard to distance and positioning of the methyl groups of UC-781 with the aromatic group of Trp.107 Structure–activity relationship studies have revealed that UC-781 contains the optimum 5-atom length pentenyl ether moiety at the 3-position of the thiocarboxanilide ring required for its antiviral potency.105,106
The abrogation of infectivity of virus produced by UC-781-treated cells is useful as a microbicide targeted to inhibit HIV-1 in the cervicovaginal and rectal milieus. Consequently, the virucide property of UC-781 has been extensively studied in in vitro and in vivo models of mucosal HIV-1 transmission.52–56,108–111 UC-781 was active against HIV-1 variants (R5SF162 and X4LAV) of different co-receptor tropisms in human lymphoid tissue ex vivo. R5 HIV-1 which is known to transmit infection, is more sensitive to UC-781 inhibition than is X4. A 24 h treatment with UC-781 was shown to block cell-free or cell-associated HIV-1 infection of monocyte-derived dendritic cells and CD4+ T cell co-cultures, representing the primary targets during sexual transmission.56 However, secondary culture of UC-781-exposed cells revealed latent infection.56 Only continuous treatment of cell-free or cell-associated virus with micromolar concentrations of UC-781 resulted in inhibition of proviral integration. Thus, UC-781 may not be efficient in vivo against some primary HIV-1 isolates with pre-existing resistance to NNRTIs.
A human cervical explant culture model has been shown to be effective in transmitting infectious cell-free and cell-associated HIV-1 across the mucosal barrier. In vitro exposure of human cervical tissue to micromolar concentrations of UC-781 was shown to protect the cervical tissue from both T- (IIIB) and M-tropic (Ba-L and ADA-M) HIV-1 isolates, clinical virus isolates from different HIV-1 clades (B, C, G and O) and from HIV-1 resistant to NRTIs.55,56,111 In this explant model, UC-781 was able to prevent direct tissue infection as well as transinfection by migratory cells.111 UC-781 demonstrated a memory effect, in which cells pre-treated in vitro with the NNRTI were protected from HIV-1 replication for 6–12 days post-compound treatment.111,112 UC-781 treatment of infected cells resulted in the release of attenuated virus.56 UC-781 alone and formulated in a vaginal gel was active in the presence of semen in neutralizing cell-associated HIV-1 transmission in this model.55 Gel formulations of UC-781 (0.1 and 1%) effectively blocked the infection by the three subtype strains (A, C and CRF01_AE) representing predominant circulating strains in southern Africa and southeast Asia. UC-781 exhibits marked temperature and pH stability,54 but is prone to oxidation and photolysis, thus requiring protection from oxygen and light. Repeated intravaginal exposure to 5% UC-781 incorporated in replens gel did not cause local inflammation or damage of the vaginal mucosa and epithelia.54 A Phase I study of the safety and acceptability of three different doses of UC-781 is currently underway.113
TMC-120 [(1S;2S)-N-(cis-6-fluoro-2-hydroxy-3-propionyl-phenyl)cyclopropyl]-N'-(5-cyanopyrid-2-yl)urea]
Diarylpyrimidine analogues such as TMC-120 (dapivirine) were predicted to bind HIV-1 RT in multiple conformations thereby escaping the effects of drug-resistance mutations.114,115 Structural studies showed that TMC-120 could adapt to changes in the NNRTI binding pocket by conformationally variable binding modes.115 TMC-120 has favourable in vitro activity profile against NNRTI-resistant variants. However, in vitro data suggest that, over time, some degree of resistance to TMC-120 does occur in HIV with the L100I and K103N mutations. Consequently, TMC-120 was less effective against HIV-1 strains with L100I, K103N and Y181L mutations.
In vitro studies showed that both viral infection as well as integration could be blocked by prior exposure to TMC-120. TMC-120 was 10-fold more potent than UC-781. However, unlike UC-781, TMC-120 showed significant degradation and binding to serum albumin. In co-cultures of dendritic cells and mitogen-activated CD4+ T cells, the primary targets in sexual transmission, TMC-120 prevented infection of both cell-free as well as cell-associated virus at nanomolar concentrations.57,58 However, a 10-fold higher concentration was required to completely prevent proviral integration. Furthermore, TMC-120 had a low selectivity index of 2400 and emergence of drug-resistant strains was apparent after two mutations. Additional studies showed that TMC-120 inhibits HIV-1 infection by X4 and R5 strains of virus in cervical explant models. TMC-120-exposed cervical tissues resisted viral challenge up to 6 days suggesting an anti-HIV memory effect. The diarylpyrimidine NNRTIs showed comparable activity to a range of clinically derived recombinant viruses representing HIV-1 group M subtypes A through H, as well as circulating recombinant forms (CRFs) CRF01_AE, CRF02_AG, CRF05_DF and HIV-1 group O from different geographic groups. TMC-120 gel formulated either in carbopol 940 or hydroxyethyl cellulose showed 70–100% efficacy in the vaginal Hu-SCID mouse model of HIV transmission of cell-associated R5 (CCR5-tropic) and R4 (CXCR4-tropic) HIV-1 strains, the phenotypes most commonly transmitted sexually.58 A gel incorporating 0.3% TMC-120 was non-irritating to the rabbit vaginal mucosa following a 10 day daily application.116
|
Lipophilicity of tight binding NNRTIs |
|---|
Tight binding hydrophobic NNRTIs can readily enter the lipid bilayer of the cell plasma membrane and are sequestered in cellular compartment(s) that enables access to HIV during subsequent virus exposure. Fusion of incoming virus with the drug-treated cell membrane is thought to allow diffusion of the membrane-resident NNRTI into the hydrophobic capsid core of the virus, thereby allowing binding to HIV-1 RT within.117 Because of their rapid association rate and a very slow dissociation rate, reverse transcription would be inhibited for prolonged periods following fusion and entry, thereby preventing the cell from becoming infected as well as resulting in attenuated infectivity of virions. In contrast, orally bioavailable NNRTIs are able to transit cells, and thus they would not remain entirely within the plasma membrane lipid bilayer. Consequently, thiourea NNRTIs with poor aqueous solubility (<0.001 mg/mL), high octanol–water partition coefficient (log P values >4) and poor oral bioavailability (<2%) may accumulate to a greater extent in specific subcellular compartments providing the maximal protection against HIV-1 infection by inhibiting endogenous and intracellular reverse transcription.
|
Lipophilic NNRTI formulations |
|---|
The effectiveness of a vaginal/rectal microbicide is dependent on the bioadhesion of the formulation and the bioavailability of the drug. The vaginal cavity exhibits an aqueous environment containing glandular secretions that create an acidic pH in the range of 4.0–5.5. The pharmaceutical formulation for a microbicide should include acceptable as well as compatible non-contraceptive or contraceptive carriers and optionally other therapeutic ingredients. Vaginal or rectal formulations can be administered as a suppository or pessary. Other vaginal or rectal preparations can include creams, films, foams, gels, pastes, rings, sponges, spray formulations or tampons containing the active agent(s) and acceptable carriers.
Many of the current anti-HIV agents intended for vaginal/rectal microbicide development are poorly soluble in aqueous medium.118 Poor solubility can lead to poor dissolution kinetics, suboptimal bioavailability, ineffective protection and the need for higher dosing, thereby contributing to mucosal toxicity. To be effective as microbicides, the vaginal and rectal concentrations of virucidal NNRTIs should be in the range of several orders of magnitude in excess of their reported in vitro activity against wild-type and drug-escape mutants prior to and immediately following exposure to HIV-1. Additionally, it is advantageous to co-administer NNRTIs with other antivirals to provide synergistic response, to minimize toxicity and to ensure complementary resistance patterns.
A novel submicron (30–80 nm) particle microemulsion formulation (namely GM-144) that offers rapid dispersion and an enhanced drug absorption profile was developed as a drug delivery vehicle for spermicidal NNRTIs.119 Microemulsions as thermodynamically stable, isotropically clear dispersions of water, oil and surfactants, can deliver larger amounts of topically applied agents into the mucosa than do traditional lotions and creams because of their capacity for enhanced solubilization.118,120,121 A drug that is dissolved rather than suspended in a vehicle is in a form immediately available for absorption and is therefore generally more rapidly and more effectively absorbed. The components used for the gel-microemulsion (GM-144) formulation are non-toxic solubilizers for lipophilic drugs used in the preparation of a variety of topical, oral and injectable medications. The microemulsion-based lipophilic formulation offers several benefits for vaginal delivery, including increased absorption, potent contraceptive activity and decreased toxicity.122
A novel non-toxic, non-spermicidal, self-emulsifying gel (namely Conceival) with improved solubility of lipophilic anti-HIV drugs was developed as a non-contraceptive drug delivery vehicle.123 The solubility values for thiourea NNRTIs with admixtures of varying amounts of water in hydroxylic excipients showed an exponential rather than a linear relationship with increasing amounts of excipients. These solubility studies indicated that formulations containing 1–2% active pharmaceutical ingredients are feasible by adjusting the ratios of the excipients for pre-clinical and clinical studies. The characteristic feature of Conceival is its unique ability to disperse rapidly on contact with the aqueous vaginal environment to bring about self-emulsification or dispersion. The formulated anti-HIV drugs in Conceival are rapidly released in aqueous vaginal fluid, thereby facilitating good dispersibility of the drug(s) especially during coitus. Nearly any water-insoluble drug may be formulated in Conceival so as to increase its solubility, and hence its bioavailability as a vaginal or rectal microbicide.
|
Drug resistance to NNRTI microbicides |
|---|
The extensive variability of HIV-1 has a potential impact on epidemiology, diagnosis and therapy, as well as the prevention of infection. Viral diversity appears to radiate out of sub-Saharan Africa, where over 28 million of the total 42 million infected persons live. Although subtype B is still predominant in Europe, the USA and Australia, an increasing prevalence of non-clade B subtypes and circulating recombinant forms has been reported by several surveys in previously homogeneous clade B countries. Greater than 40% of new infections in Europe are presently non-B African and Asian variants. Consequently an ideal microbicide must be able to neutralize a wide range of circulating HIV-1 subtypes to prevent new infections especially in the third world.
In addition, recent studies suggest that the male or female genital tract represents a distinct replication compartment for HIV-1 and that such compartments may serve as a virus reservoir.124 Genital secretions of NNRTI-experienced women harbour HIV-1 with K103N and K238N mutations and these variants can persist for years in the absence of drug selection.125 These findings highlight the potential for further increases in sexual transmission of HIV-1 resistant to NNRTI and have particular relevance to design of NNRTIs active against these mutations. The repeated application of an NNRTI as a vaginal virucide may select for NNRTI-characteristic resistance in the drug-treated individuals. Consequently, poor oral bioavailability of thiourea NNRTI is advantageous for topical application as it may limit systemic absorption following mucosal application. An NNRTI that can enter and inactivate free virus is likely to enhance the effectiveness of a microbicide, as it could inactivate HIV-1 in seminal plasma before the virus reaches a target cell. The lack of detection of thiourea NNRTIs in the plasma of test animal species following intravaginal administration of 2% thiourea NNRTI-containing gel argues against potential selection of thiourea NNRTI-resistant virus.21,84 In general, although RT mutations can confer some resistance to second-generation NNRTIs, the concentration of these NNRTIs readily attainable in lipophilic gel formulations is expected to be much higher than their IC90 values for even the most resistant mutants.
|
Transparency declarations |
|---|
None to declare.
|
Acknowledgements |
|---|
NNRTI microbicide research and development in the authors' laboratories was supported in part by National Institute of Health Grants (AI052594, AI052633, AI054352 and HD043683) as well as Institutional Funds at the Parker Hughes Institute, which is a non-profit 501 (C) 3 tax exempt research organization with a public charity status. We thank the Crystallography Group, especially Dr Chen Mao, for the computer models of thiourea NNRTIs.
|
References |
|---|
1. Centers for Disease Control. Diagnosis of HIV/AIDS—32 states. 2000–3. MMWR 2004; 53: 1106–10.[Medline]
2.
Mujeeb SA, Altaf A. The AIDS pandemic. N Engl J Med 2003; 349: 814–5.
3. UNAIDS. Report on the Global HIV/AIDS Epidemic: December 2003. http://www.unaids.org/wad2004/report.html (24 November 2005, date last accessed).
4. National Institute of Allergy and Infectious Diseases Fact Sheet. HIV Infection in Women 2004. http://www.niaid.nih.gov/factsheets/womenhiv.htm (24 November 2005, date last accessed).
5. Centers for Disease Control (2003) Fact Sheet: HIV/AIDS Among US Women: Minority and Young Women at Continuing Risk. Atlanta, GA 2003. www.cdc.gov/hiv/pubs/facts/women.htm (24 November 2005, date last accessed).
6. Newman MC. Women and HIV. UCSF center for HIV information. March 2004. http://hivinsite.ucsf.edu/InSite.jsp?