Peptide-derived mid-sized anti-HIV agents

The human immunodeficiency virus (HIV), the causative virus of acquired immunodeficiency syndrome (AIDS), was discovered by Montagnier et al. in 1983.¹  HIV infects human host cells to destroy their immune systems causing immunodeficiency. Currently the number of people in the world with HIV infection is certainly beyond 30 million and several anti-HIV drugs have been developed in the last 30 years (Fig. 1). HIV is classified as a retrovirus. DNA is produced from its RNA genome via the enzyme reverse transcriptase and is then incorporated into the host genome by an integrase enzyme. The first generation of anti-HIV drugs that were initially used for clinical treatment were reverse transcriptase inhibitors such as azidothymidine (AZT),²  which suppresses the enzyme action and blocks reverse transcription. The second generation of drugs that were clinically used consisted of protease inhibitors, which prevent the cleavage of HIV precursor proteins into active proteins. These drugs are usually administered in two- or three-drug cocktails in highly active anti-retroviral therapy (HAART), which has brought great success and hope in the clinical treatment of HIV infection and AIDS.²  HAART is capable of lowering the HIV level in the blood to below the detection level, but has side effects, the emergence of multi-drug resistant (MDR) HIV-1 strains and considerable expense. These serious drawbacks have encouraged the development of new drugs with different mechanisms of action. The molecular mechanism of HIV-1 replication involving the dynamic supramolecular mechanism at HIV entry and fusion steps has now been elucidated in more detail and is shown in Fig. 1. Initially gp120, an HIV envelope protein, interacts with a cellular surface protein, CD4. This leads to a conformational change in gp120 and its subsequent binding to the cellular coreceptors, chemokine receptors such as CCR5^3–7  and CXCR4.⁸  CCR5 and CXCR4 are the major coreceptors for the entry of macrophage-tropic (R5-) and T cell line-tropic (X4-) HIV-1, respectively. This binding triggers the exposure of another envelope protein gp41 and the penetration of its N-terminus into the cell membrane, followed by the formation of the trimer-of-hairpins structure of gp41, which leads to fusion of HIV to the cell membrane, completing the infection process.⁹  The clarification of this dynamic molecular machinery has encouraged us to develop inhibitors which block the HIV-entry/fusion steps targeting the receptors, CD4, CCR5 and CXCR4, and the viral proteins gp120 and gp41.

Enfuvirtide (fuzeon/T-20) from Roche/Trimeris, was approved by the Food and Drug Administration (FDA) in 2003 as the first “fusion inhibitor” to treat patients with advanced HIV in combination with other anti-HIV drugs.¹⁰  Subsequently, a CCR5 coreceptor antagonist, maraviroc (Pfizer), was approved by the FDA in 2007 for use in combination with other anti-HIV drugs as an entry inhibitor for the treatment of patients infected with R5-HIV-1.¹¹  In the same year, the FDA approved raltegravir (Isentress) (Merck Sharp & Dohme Corp.) as the first “integrase inhibitor”.^12,13  In 2012, elvitegravir (Gilead Sciences, Inc./JT) was approved by the FDA as an integrase inhibitor for use in patients starting treatment of HIV infection for the first time.^14,15  In 2013, dolutegravir (Shionogi/GSK) was also approved by the FDA as an integrase inhibitor, which is marketed as Tivicay.¹⁶  To date there have been many reviews describing development of reverse transcriptase and protease inhibitors. The present manuscript focuses on vaccines/fusion inhibitors, CCR5/CXCR4 antagonists, integrase inhibitors and CD4 mimics, and includes our research.

In general, use of antibodies and vaccines is an effective treatment for infectious diseases. In the case of HIV, immunization¹⁷  and de novo engineering of monoclonal antibodies (Abs), involving molecular evolution methods, have led to the development of HIV antibodies.¹⁸  However, very few monoclonal Abs show broad HIV-neutralizing activity. These include gp120 Abs, 2G12¹⁹  and b12²⁰  and gp41 Abs, 2F5^21–24  and 4E10.^23–25  An endodomain and an ectodomain of gp41 are separated by a transmembrane region. The gp41 ectodomain contains a hydrophobic amino-terminal fusion peptide, followed by NHR (HR1) and CHR (HR2) domains which have helical structures. In the membrane fusion process, NHR and CHR form a “six helical bundle” conformation, which consists of a central parallel trimer of NHR surrounded by three strands of CHR arranged in an antiparallel manner. According to the above mechanism, a useful strategy to design antigens that elicit broadly neutralizing antibodies is to produce artificial molecules that mimic the natural trimer on the viral surface, so that induced antibodies might recognize the NHR trimer and suppress formation of a natural “six helical bundle” structure. Such molecules, which are expressed in a recombinant form or on the surface of particles such as pseudovirions or proteoliposomes,^26–28  have been reported previously. Several synthetic antigens and inhibitors have also been developed using various templates connected with peptidomimetics corresponding to the native structure of gp41.^29–32  However, the templates, which are used for assembly of these helical peptides, have branched linkers with different lengths.³⁰  The distance between any two residues at the N-terminus of the N-region of gp41 is approximately 10 Å, according to X-ray crystallographic analysis.³³  Ideal mimetics of NHR might efficiently bind to neutralizing monoclonal Abs (mAb) but poorly to non-neutralizing mAbs.

N36 and C34 are respectively, NHR and CHR-derived helical peptides. We have designed and synthesized a three-helical bundle mimetic corresponding to the equivalent trimeric form of N36, and investigated whether mice immunized with this N36 trimer mimetic can induce antibodies with high binding affinity for the N36 trimer.³⁴  Since N36 has relatively high hydrophobicity and low aqueous solubility, our design of an N36-derived peptide called for the triplet repeat of arginine and glutamic acid fused to the N-terminus to increase the aqueous solubility of the peptide, which was designated as N36RE (Fig. 2A). A C₃-symmetric template was designed to form a triple helix, which mimics the gp41 pre-fusion form precisely (Fig. 2B). The linker tethered to this template²⁸  has three same branches of equal length and possesses a hydrophilic structure and a ligation site required for coupling with N36RE. A template with its three-armed aldehyde scaffold (Fig. 2B) was conjugated with Cys-containing unprotected N36RE (N36REGC) by thiazolidine ligation^35–39  to produce the trimer triN36e (Fig. 2C). According to the CD analysis of the synthetic peptides, the helical content of the trimer triN36e is higher than that of the monomer N36RE, and the mixture of triN36e and a C34-derived monomer peptide, C34RE, has high helicity compared with triN36e alone³⁴  indicating that the interaction of C34RE with triN36e induces a higher helical form.⁴⁰  In the experiments of antibody induction, mice were immunized with the trimer triN36e, and antibody production was then evaluated by serum titer ELISA against coated synthetic monomer N36RE and trimer triN36e antigens. The triN36e-induced antisera showed approximately 30 times higher affinity for the triN36e antigen than for the N36RE antigen, proving that the triN36e-induced antisera have a structural preference for binding with triN36e. In anti-HIV assays antisera from the trimer triN36e immunization showed an approximately 4-fold higher neutralizing activity than those from the monomer N36RE immunization. As a result, the N36 trimeric form can induce antibodies with higher neutralization activity than the monomer form.

The ectodomain of gp41 is composed of two peptides, a 51-mer from the NHR region and a 43-mer from the CHR region, designated as N51 and C43, respectively.⁴¹  To date, there have been many reports that several CHR region peptides prevent formation of a natural “six helical bundle” structure by binding to the central parallel trimer of NHR thereby inhibiting membrane fusion.⁴²  An NHR region peptide C34, with 34 residues, shorter than C43, has potent inhibitory activity against HIV-1 fusion.³³  As described in the Introduction (above), a 36-residue peptide, Enfuvirtide (fuzeon/T-20, Roche/Trimeris), which is tilted slightly away from the helical region of CHR toward the C-terminal side from C34 and which has 24 residues in common with C34, was approved by the FDA in 2003 for its clinical use in HIV/AIDS treatment as the first fusion inhibitor.¹⁰  Such peptides bind to the NHR region preventing formation of the six helical bundle structure.⁴³  C34 contains the amino acid residues required for docking into the hydrophobic pocket of the central parallel trimer of NHR and potently inhibits HIV-1 fusion.³³  We have therefore designed and synthesized a three-helical bundle mimetic corresponding to the equivalent trimeric form of C34. In the C34-derived peptide, the triplet repeat of arginine and glutamic acid was fused to the C-terminus to increase its aqueous solubility, and a glycine thioester was added to the C-terminus, and the product was designated as C34REG-thioester (Fig. 3A). To form a triple helix, which mimics the gp41 pre-fusion form, the C₃-symmetric template depicted in Fig. 3B was designed with a linker having three branches of equal length, a hydrophilic structure and a ligation site for coupling with the C34REG-thioester. An unprotected C34REG-thioester was coupled with a template possessing a three-armed cysteine scaffold to yield the trimer triC34e (Fig. 3B and 3C).⁴⁴  CD analysis showed that both the C34-derived monomer C34REG and the trimer triC34e form random structures, which are different from N36-derived peptides,⁴⁴  and that the mixture of C34REG and N36RE and that of triC34e and N36RE form α-helix structures while the helical content of the latter mixture is significantly lower than that of the former mixture. This indicates that the assembly of three peptide strands in triC34e by covalent bonds might cause some difficulty in formation of a six-helical bundle structure by the trimer triC34e with three N36 peptides. It is noteworthy that the HIV-1 inhibitory potency of triC34e is one hundred times higher than that of C34REG (Table 1), suggesting that a trimeric form is critical to an inhibitory structure. In addition, triC34e has no significant cytotoxicity. For comparison, the HIV-1 inhibitory activity of the N36 trimer mimetic triN36e is three times higher than that of the monomer N36RE, suggesting that N36 peptide content is crucial, but both have modest inhibitory activity.³⁴  Immunogenicity of the C34 trimer mimetic triC34e was also investigated⁴⁵  as was that of the N36 trimer mimetic triN36e. The antisera produced by immunization of triC34e showed 23-fold higher binding affinity for the trimer triC34e than for the monomer C34REG. This result is consistent with that seen in immunization of the N36 trimer mimetic triN36e. However, the neutralization activity of the triC34e-induced antibodies is not sufficiently high, being nearly equal to that of the monomer C34REG-induced antibodies. Taken together, these facts suggest that the NHR region is more suitable as a vaccine target than the CHR region.

To identify a key structure required for the expression of the inhibitory activity of the CHR-derived trimer mimetic triC34e against HIV-1 fusion, the C34 dimer mimetic was chemically synthesized. The same unprotected C34REG-thioester was stoichiometrically coupled with the same C₃-symmetric template with a three-armed cysteine scaffold, followed by carboxymethylation of the free thiol groups with iodoacetamide to produce the C34 dimer mimetic diC34e (Fig. 3C).⁴⁶  The HIV-1 inhibitory activity of diC34e was found to be nearly equal to that of the trimer mimetic triC34e (Table 1). This indicates that two units of the C34 peptide in the dimer form can bind to the NHR region in a cooperative manner. In the C34 dimer and trimer mimetics, dimerization or trimerization of the C34 peptide fails to increase the α-helicity of the peptides judging by CD analysis.⁴⁶  In the mixture with the N36 monomer N36RE, all the C34-derived peptides, monomer C34REG, dimer diC34e and trimer triC34e show similar α-helicity in spite of the difference in the number of units in the C34 peptide.

Interaction of CD4 with the HIV envelope protein gp120 causes a conformational change which is followed by its subsequent binding to the cellular coreceptors, CCR5^3–7  and CXCR4⁸  as described in the Introduction. CXCR4 is the major coreceptor for the entry of T cell line-tropic (X4-) HIV-1 strains, which are the major species in the late stage of HIV infection and AIDS. It is conceivable that CXCR4 antagonists can block entry of X4-HIV-1 into cells. On the other hand, stromal cell-derived factor 1 (SDF-1)/CXCL12 and CXCR4, an important endogenous ligand/receptor pair, play physiological roles in embryogenesis of cardiovascular, hematopoietic and central nervous systems. CXCL12 and CXCR4 have also been relevant to various pathological conditions such as cancer,^47–49  leukemia^50,51  and rheumatoid arthritis.^52,53  Thus the SDF-1/CXCR4 axis is an excellent drug target, and CXCR4 antagonists might overcome these diseases. To date, several peptidic and non-peptidic CXCR4 antagonists, have been developed.

The 14-mer peptide T140 (1), a polyphemusin II analog has been found by Fujii and Tamamura to be a potent CXCR4 antagonist (Fig. 4A).⁵⁴  Tachyplesins and polyphemusins, which are naturally contained in the hemocyte debris of the Japanese horseshoe crab (Tachypleus tridentatus) and the American horseshoe crab (Limulus polyphemus), respectively, are 17-mer and 18-mer self-defense peptides that show broad spectrum antimicrobial activity against several strains of bacteria and viruses.^55,56  Our continuing structure-activity relationship studies on these peptides have led to a polyphemusin analog, T22 ([Tyr^5,12, Lys⁷]-polyphemusin II)^57,58  and a shortened 14-mer peptide, T140, as anti-HIV peptides.⁵⁹  T22 and T140 strongly block X4-HIV-1 entry through their competitive binding to CXCR4.^60–62  Four amino acid residues contained in T140, Arg², l-3-(2-naphthyl)alanine (Nal)³, Tyr⁵ and Arg¹⁴, are particularly important for high potency.⁶³  Since T140 is not sufficiently stable in mouse/feline serum or in rat liver homogenate,^64,65  it was modified at the N-/C-terminus to suppress the biodegradation. This led to development of more effective compounds, having high CXCR4-antagonistic activity and increased biostability. The biostable T140 analogues^65,66  so obtained have significant inhibitory activity against HIV infection as well as against cancer/leukemia:^48,49  4F-benzoyl-TN14003/BL-8040/BKT-140 (BioLineRx Ltd.) is a Phase II drug candidate for the treatment of acute myeloid leukemia (AML), and other types of hematological cancer (http://www.biolinerx.com). In addition, BL-8040 mobilizes hematopoietic stem cells from the bone marrow into peripheral blood, and also induces the mobilization of cancer cells from the bone marrow and other sites thereby exposing these cells to anticancer therapy inducing apoptosis. Pre-clinical studies have shown that BL-8040 is efficient, both alone and in combination with the anticancer drug rituximab, in reducing bone marrow metastasis of lymphoma cells and stimulating lymphoma cell death (http://www.biolinerx.com). To develop low molecular weight CXCR4 antagonists, a pharmacophore-guided approach was performed based on four indispensable residues of T140,⁶⁷  Arg², Nal³, Tyr⁵ and Arg¹⁴, and adopting cyclic pentapeptides as conformationally restricted templates with functional groups used in the efficient discovery of bioactive lead compounds in medicinal chemistry.^68–73  From the library of cyclic pentapeptides using two l/d-Arg, l/d-Nal, l/d-Tyr and Gly, FC131 (2) was found to be a potent CXCR4 antagonist comparable to T140⁷⁴  (Fig. 4). The pharmacophore-guided approach using cyclic pentapeptide templates proved to be useful for downsizing of T140 (1) into FC131 (2).⁷⁴  FC131 derivatives such as compound 3, containing amidine type peptide bond isosteres have been developed. Replacement of peptide bonds in FC131, except for the d-Tyr-Arg position, with an amidine moiety improved inhibitory activity against SDF-1 binding and HIV-1 infection by X4 strains. Furthermore, these analogues showed selectivity for CXCR4 and not for CXCR7 and CCR5, which are the targets shared by SDF-1 and HIV-1, respectively.⁷⁵  Based on the β-hairpin structure of polyphemusin II, several protein epitope mimetic (PEM)⁷⁶  molecules, such as POL3026 (4)⁷⁷  and POL6326 (5),⁷⁸  possessing potent and selective antagonistic activity against CXCR4 were designed and optimized in biological assays. POL6326 has moved into a Phase II clinical trial for autologous stem cell transplantation in newly diagnosed multiple myeloma patients. FPI-X4 (6), with a molecular weight of 1830, corresponds to amino acid residues 408–423 of human serum albumin (HSA). The peptide has moderate inhibitory activity against HIV-1 infection by the X4-HIV-1 strain (IC₅₀=15.8 μM) and also moderate binding affinity (IC₅₀=8.6 μM) for CXCR4 but not for CCR5. FPI-X4 does not induce Ca²⁺ mobilization or receptor internalization, and it thus acts as an inverse agonist for CXCR4.⁷⁹  LY2510924 (7) was identified by the combinational method of a medium throughput screen with a rational design approach and specifically blocks SDF-1 binding to CXCR4 (IC₅₀=0.079 nM).⁸⁰  LY2510924 has now moved into Phase II clinical studies for cancer.

A non-peptidic small molecule, bicyclam AMD3100 (8) (Genzyme Corp.), was reported as the first CXCR4 antagonist to enter clinical trials for the treatment of HIV-1-infected patients (Fig. 4A),^81,82  but its application as an anti-HIV drug was discontinued because of its adverse cardiovascular effects. Subsequently, several small CXCR4 antagonists with potent anti-HIV activity have been reported based on the structure of AMD3100 (8) containing at least two nitrogen atoms (e.g. pyridine groups) on each side of the p-xylene template.^83,84  However, these compounds are subject to rapid oxidative metabolism and have poor biostability. Based on AMD3100 (8), several non-cyclam CXCR4 antagonists have been developed.^85–87  AMD3100 (8), designated as plerixafor/mozobil (Genzyme Corp.) was used as an immunostimulant to mobilize hemopoietic stem cells into the blood in patients with cancer. The stem cells can be extracted from the blood in patients for transplantation using granulocyte colony-stimulating factor (G-CSF). Combination of G-CSF with plerixafor can increase the number of patients that succeed with stem cell transplantation.⁸⁸  A tetrahydroquinoline compound AMD11070 (AMD070) (9) (Genzyme Corp.) has been found to be a CXCR4 antagonist by recent antiviral evaluation and pharmacokinetic analysis.^89,90  Phase I/II studies of AMD11070 are being conducted to assess its effect on X4-HIV-1 infection. MSX-122 (10), which has been identified by rational design and analysis of emerging structural and pharmacologic data is a partial CXCR4 antagonist which fails to mobilize stem cells, which can reduce the risk of long-term blocking of metastasis caused by other CXCR4 antagonists.⁹¹  TG-0054 (burixafor) (11) is a selective CXCR4 antagonist provided by TaiGen Biotechnology Co., Ltd. (Taipei, Taiwan) and is currently in Phase II clinical trials to assess the therapeutic effect of HSC mobilization alone or in combination with granulocyte colony-stimulating factor (G-CSF) in patients with multiple myeloma, non-Hodgkin's lymphoma, and Hodgkin's disease.⁹²  Chemical modification of the cyclic pentapeptide FC131 (2) has led to the development several non-peptidic CXCR4 antagonists. An indole template modified the peptide backbone of FC131 (2) and the disposition of the original pharmacophore moieties (Fig. 4B). Small CXCR4 micromolar level antagonists linked with three pharmacophore moieties as in compound 13 have been found.⁹³  Nonpeptide compounds with a dipicolylamine (Dpa)–zinc(ii) complex structure have been developed as potent and selective antagonists against CXCR4.⁹⁴  A Dpa–Zn complex with a xylene scaffold (14) binds to CXCR4 at the 50 nM level. Combination of alkylamino and pyridiyl moieties, which are contained as common structural features in the Dpa–Zn complex (14) and AMD3100 (8), has led to the development of compounds 15 and 16 with 30 nM and 10 nM activity levels for binding to CXCR4, respectively.⁹⁵  A small CXCR4 antagonist, KRH-1636 (17) (Kureha Chemical & Daiichi Sankyo Co. Ltd.), which was derived by the intensive modification of the N-terminal tripeptide of T140, Arg-Arg-Nal, was found to be an orally bioavailable.⁹⁶  Continuous efforts have led to identification of several derivatives which may be promising as novel inhibitory drugs for treatment of cancer patients and include KRH-2731.⁹⁷  These small compounds are attractive and useful leads for the future development of non-peptidic CXCR4 antagonists.

The chemokine receptor CXCR4 belongs to the seven transmembrane G protein-coupled receptor (GPCR) family, and many of GPCRs, including the chemokine receptors, exist as dimers and/or higher order oligomers and express physiological functions. Chemokine receptors such as CXCR4 form homodimers and/or heterodimers with other chemokine receptors.^98,99  We designed and synthesized several CXCR4 bivalent ligands consisting of two molecules of an FC131 derivative, [cyclo(-d-Tyr-Arg-Arg-Nal-d-Cys-)], connected by poly(l-proline) or PEGylated poly(l-proline) linkers of various lengths (18, 19)¹⁰⁰  (Fig. 4C). A maximum binding affinity for CXCR4 was observed for bivalent ligands with two types of the linkers of suitable lengths (5.5–6.5 nm). Our experimental results have shown that the native state of the CXCR4 dimer has the distance between the ligand binding sites (5.5–6.5 nm), and that fluorescent-labeled bivalent ligands are useful tools for cancer diagnosis that can assess the density of CXCR4 on the surface of cancer cells. Thus, we synthesized bivalent CXCR4 ligands with near infrared (NIR) dyes at the terminus or the center of the poly-l-proline linker. These are valuable probes which are useful in studies of the behavior of cells expressing CXCR4 (Fig. 5).¹⁰¹ 

To investigate the possible multimeric forms of CXCR4, trivalent ligands with rigid poly-l-proline linkers were designed and synthesized. Judging by the CXCR4 binding affinity of the trivalent ligands, the ligands recognize the dimeric form of CXCR4 on the cellular surface. In particular, the ligand with 9-l-proline linkers binds to CXCR4 with remarkable specificity judged by the fluorescent imaging and analysis using flow cytometry. In comparison with the corresponding monomer ligand, the dimer and the trimer ligands showed 17- and 47-fold increases in binding activity, respectively, suggesting a synergistic effect in the binding of the ligand units. However, the IC₅₀ of the trimer is approximately 3-fold higher than that of the dimer, suggesting that three patterns exist for the dimer recognition in the trivalent ligand (Fig. 6).¹⁰²  Since it has been proven that CXCR4 does not exist as the trimer, it is presumed that the dimer units are oligomerized. In future, the multimeric form should be investigated by the design of ligands with rigid linkers.

The development of CCR5 antagonists was validated by the finding that people with the CCR5Δ32 deletion are not only healthy but highly resistant to HIV-1 infection.¹⁰³  Several pharmaceutical companies have investigated novel CCR5 antagonists with suitable pharmaceutical properties. A solitary CCR5-selective antagonist, maraviroc (20) (Pfizer Inc.) has been approved by the FDA (Fig. 7),¹¹  and is used for the treatment of patients infected with R5-HIV-1. Takeda Pharmaceutical Co. Ltd. developed TAK-779 (21)^104,105  but its clinical trials were terminated because of local reactions at s.c. injection sites. Subsequent development produced TAK-220 with a piperidine-4-carboxamide structure (22) (Takeda Pharmaceutical Co. Ltd./Tobira Therapeutics, Inc.), which has high CCR5 binding activity and resistance to metabolic modification.¹⁰⁶  Vicriviroc (SCH-D/SCH417690) with a piperidinopiperazine unit (23) was developed by Merck & Co.¹⁰⁷  The safety and efficacy of vicriviroc have been established, but Merck & Co. failed to prove that the current regimens using vicriviroc are more effective than preceding regimens. A CCR5 antagonist, ONO-4128/873140 with a spirodiketopiperazine scaffold (24) (GSK/Ono Pharmaceutical Co., Ltd.), has been developed.¹⁰⁸  The spirodiketopiperazine is an attractive scaffold and leads to more diverse derivatives through combinatorial chemistry. Phase III studies with this compound were discontinued on account of its hepatotoxicity. Taken together, several other CCR5 antagonists have progressed to clinical trials, but to date, no drug other than maraviroc has been approved by the FDA.

HIV-1 integrase (HIV-IN) is a critical enzyme for stable infection of host cells because it catalyzes the insertion of viral double-stranded DNA into the chromosomal genome of human host cells through 3′-end processing and strand transfer reactions. HIV-IN, a 32 kDa protein, has 288 amino acid residues, and is divided into an N-terminal, catalytic core and C-terminal domains.^109,110  The catalytic core domain has three acid residues, Asp64, Asp116 and Glu152, which are indispensable for coordination with two magnesium ions to catalyze the 3′-end processing and strand transfer reactions that are correlated with cleavage and formation of DNA phosphodiester bonds, respectively (Fig. 8A).^111–113  Thus, IN strand transfer inhibitors possessing two-magnesium-binding pharmacophores, which target the three carboxylate residues, have been developed. Initially, diketoacids (DKAs), which have a two-magnesium-binding pharmacophore, have been reported as first generation IN inhibitors (25, 26) based on an interactive model of the binding of these inhibitors to IN through coordination with two magnesium ions.¹¹⁴ 

A pyrimidinone derivative, raltegravir (Isentress) (27) (Merck Sharp & Dohme Corp.), was the first IN inhibitor to be approved by the FDA.^12,13  Initially, raltegravir was approved only for patients with resistance to other HAART drugs in 2007, but later the FDA expanded its approval of raltegravir to its use in combination with other anti-HIV agents in 2009. Elvitegravir (28) (Gilead Sciences, Inc./JT) with a quinolone template, was the second IN inhibitor to be approved by the FDA.¹¹⁵  This compound has properties of nanomolar levels of IN inhibitory and anti-HIV activities as well as moderate bioavailability and low clearance. A CYP3A inhibitor, cobicistat can protect raltegravir from metabolism, and these drugs can be used in a combinational regimen. The compounded drug Stribild contains elvitegravir, cobicistat and two reverse transcriptase inhibitors, tenofovir and emtricitabine.¹⁵  Taking this tablet once per day causes effective and sufficient results in patients. In addition, elvitegravir has cross-resistance with raltegravir-resistant strains.¹¹⁶  Dolutegravir (Tivicay) (29) (Shionogi/GSK) is a potent anti-HIV agent with a low clearance and good oral bioavailability, and was approved by the FDA in 2013 as the third IN inhibitor.^16,117,118  Even once daily monotherapy in patients with dolutegravir causes effective reduction in RNA levels, high retention of blood concentrations and suitable pharmacokinetic profiles. Development of these three HIV-1 IN inhibitors has recently advanced AIDS chemotherapy although combinational regimens are necessary because emergence of resistant mutants has been reported.

Recently, we have developed different types of HIV-1 IN inhibitors.^119,120  Screening an overlapping peptide library derived from HIV-1 gene products led to finding three Vpr-derived fragment peptides with significant IN inhibitory activity (Fig. 9). These three inhibitory peptides are sequentially overlapping. Twelve- and eighteen-mer peptides derived from the above original Vpr sequence conjugated by an octa-arginyl group, a cell membrane permeable unit¹²¹  at the C-terminus have IN inhibitory activity and anti-HIV activity in cell-base assays. The details of the mechanism of action of these peptides is unclear although it is possible that they bind to the cleft between the amino-terminal domain and the core domain of HIV-1 IN and inhibit the function of IN. This cleft region is distinct from the active site which is the nucleic acid interacting surface. This suggests that the Vpr-derived peptides inhibit the IN function in an allosteric manner. The above original Vpr sequences are originally located in an α-helical region of the parent Vpr protein. Conjugation of an octa-arginyl group to the inhibitory peptides led to not only significant inhibition of HIV replication associated with an increase in cell-permeability but also caused relatively high cytotoxicity. To develop new generation inhibitors, stapled peptides, stabilized α-helical peptidomimetics, were adopted in place of octa-arginine conjugation to increase the cell-permeability of the above peptides. A stapling strategy is an alternative means to boost cell penetration (Fig. 10).¹²²  A series of stapled peptides, which have a covalent hydrocarbon link formed by a ruthenium-catalyzed ring-closing metathesis reaction between sequential turn pitches of the α-helix, were designed and synthesized (Fig. 11).¹²³  According to CD analysis, stapled peptides form α-helical structures while the corresponding linear peptides form β-sheet structures. Cell-based assays showed that some of the stapled peptides have potent anti-HIV activity comparable to that of the original octa-arginine-conjugated peptide and lower cytotoxicity as an advantage.¹²⁴  In fluorescent imaging experiments, these stapled peptides were found to be significantly cell permeable. The application of this stapling strategy to Vpr-derived IN inhibitory peptides led to a remarkable increase in their potency in cells and a significant reduction in their cytotoxicity. Subsequently, it was found that the functional role of the octa-arginine sequence might be correlated to binding affinity for the target DNA and thus the IN inhibitory activities in vitro as well as cell membrane permeability. Oligo-arginine plays a critical role in DNA binding because the arginine guanidino groups can bind to phosphate groups of DNA. According to effects of the length of oligo-arginine sequences on DNA binding and IN inhibitory activities, the addition of tetra-/hepta-arginine is sufficient to produce an increase in IN inhibitory activities without a significant increase in cytotoxicity. Compounds 30 and 31 with tetra- and penta-arginine sequences have potent IN inhibitory and DNA binding activities and relatively low cytotoxicities, thus stapled peptidic IN inhibitors are useful lead compounds (Fig. 12).

The binding of gp120 to CD4 causes a conformational change in gp120, followed by the binding to the co-receptor CCR5 or CXCR4, as described in the Introduction (Fig. 13). Although many attempts to develop soluble CD4 molecules as anti-HIV drugs have not been successful, CD4-related molecules are known to inhibit the interaction of gp120 with CD4 and could be inhibitors of HIV entry. To data, several small-sized CD4 mimics have been developed in this and other laboratories, including NBD-556 (32),^125,126  YYA-021 (33),^127–129  JRC-II-191 (34)¹³⁰  and BMS806 (35)¹³¹  (Fig. 14). NBD-556, YYA-021, and JRC-II-191 bind to gp120, and suppress binding of HIV to CCR5 or CXCR4 blocking an HIV entry. BMS806 binds to gp120 with no significant effect on CD4 binding, but blocks the CD4-induced exposure of gp41. We have developed several CD4 mimics^132–136  based on NBD-556 and YYA-021. These compounds interact with a conserved pocket in gp120, the “Phe43 cavity”, and cause a conformational change of gp120, opening the envelope,^137,138  as is observed in the binding of soluble CD4 to gp120. As a result, some neutralizing antibodies such as an anti-V3 monoclonal antibody KD-247 can bind to gp120 (Fig. 13).¹³⁹  Very recently, CD4 mimics such as YIR-819 (36) and YIR-821 (37) with a monocyclohexyl group and a guanidino group have been developed, and were found to interact effectively with Val430 and either Asp368 or Asp474 on the surface of the Phe43 cavity. These compounds show a remarkable synergistic anti-HIV activity with KD-247, thus might be useful envelope protein openers and are desirable drug candidates for the combinational use with neutralizing antibodies.

To date, several anti-HIV drugs have been developed and used clinically for treatment of AIDS and HIV-infected patients. During the late 20th century, a combinational use of reverse transcriptase inhibitors and protease inhibitors, designated HAART, provided great success in clinical treatments. Recently, novel drugs including entry inhibitors and integrase inhibitors, which belong to different categories, have been developed successively and approved by the FDA for clinical use. However, serious clinical problems including side effects, the emergence of MDR strains, and high costs, have not disappeared and brand-new drugs with novel mechanisms of action are still required. This review article has focused on newly developed vaccines and fusion inhibitors, coreceptor antagonists, integrase inhibitors and CD4 mimics.

The design of HIV vaccines and fusion inhibitors based on the native structural mimic of proteins involved in the dynamic supramolecular mechanisms of HIV fusion is a effective strategy. The N36 trimer mimic antigen with complete equivalency induces antibodies with structural preference for the antigen as well as significant neutralizing activity.

Trimer and dimer mimics of C34 with complete equivalency have 100-fold higher anti-HIV-1 activity than the corresponding monomers. Effective inhibitors, such as six-helical bundle formation in the gp41 assembly, which target protein–protein interactions have attracted broad attention as mid-size drugs, and further development in this area is anticipated.

The SDF-1/CXCR4 axis is a significant drug target and several CXCR4 antagonists, peptidic and non-peptidic, have been developed. Based on horseshoe crab peptides, a 14-mer peptide, T140 has been found to be a potent CXCR4 antagonist. T140 analogs block X4-HIV-1 entry through competitive binding to CXCR4 and have significant inhibitory activity against cancer/leukemia as well as against HIV infection. A T140 analog, 4F-benzoyl-TN14003/BL-8040/-BKT-140 (BioLineRx Ltd.), is in Phase II clinical trials for the treatment of acute myeloid leukemia (AML). Several small CXCR4 antagonists, which are attractive and useful leads for the development of non-peptidic antagonists, have been found to date. In addition, CXCR4 bivalent ligands containing two molecules of an FC131 derivative, [cyclo(-d-Tyr-Arg-Arg-Nal-d-Cys-)], connected by poly(l-proline) or PEGylated poly(l-proline) linkers, have been synthesized. Bivalent ligands having the linkers with suitable lengths (5.5–6.5 nm) accurately recognize the native state of the CXCR4 dimer, suggesting a 5.5–6.5 nm separation of the ligand binding sites. Fluorescent-labeled bivalent ligands are also useful tools for cancer diagnosis.

A single CCR5-selective antagonist, maraviroc (Pfizer Inc.) has been approved by the FDA while several other CCR5 antagonists have progressed to clinical trials.

Development of three small HIV-1 IN inhibitors has recently been advanced in AIDS chemotherapy although they require combinational regimens. These include raltegravir (Isentress) (Merck Sharp & Dohme Corp.), elvitegravir (Gilead Sciences, Inc./JT) and dolutegravir (Tivicay) (Shionogi/GSK). Allosteric type HIV-1 IN inhibitors distinct from the above drugs have been developed and are Vpr-derived fragment peptides. The subsequent application of a stapling strategy to Vpr-derived IN inhibitory peptides led to a remarkable increase in their potency in cells.

CD4 has been a conventional target for AIDS chemotherapy, and several small-sized CD4 mimics have been found. These compounds cause a conformational change of gp120 and envelope opening which is observed in the binding of soluble CD4 to gp120. Thus, neutralizing antibodies such as an anti-V3 monoclonal antibody KD-247 can bind to gp120. These compounds show a highly remarkable synergistic anti-HIV activity with KD-247, and are desirable drug candidates for the use in combination with neutralizing antibodies as well as for entry inhibition.

In the face of the loss of efficacy of HAART due to the emergence of MDR viruses, a change of regimens of the drug combination in HAART is effective if the amounts of the virus and CD4 in blood are monitored. Taken together, the number of available potent drugs becomes a key for treatment of AIDS and HIV-infected patients. Entry inhibitors such as CCR5/CXCR4 antagonists and CD4 mimics, fusion inhibitors, and IN inhibitors might be important optional agents for an increase in the available drug repertoire.

The HIV replication cycle involves many protein–protein interactions such as the dynamic supramolecular mechanisms of entry/fusion steps. Thus the mid-size drugs described above might become effective drug candidates because these compounds can recognize spacious interfaces, and their further development is valuable.

HIV

human immunodeficiency virus

AIDS

acquired immunodeficiency syndrome

AZT

azidothymidine

HAART

highly active anti-retroviral therapy

MDR HIV-1

multi-drug resistant HIV-1

R5-HIV-1

macrophage-tropic HIV-1

X4-HIV-1

T cell line-tropic HIV-1

FDA

the Food and Drug Administration

Ab

monoclonal antibody

NHR

N-terminal heptad repeat

HR

heptad repeat

CHR

C-terminal heptad repeat

IC₅₀

50% inhibitory concentration

CC₅₀

50% cytotoxic concentration

SDF-1

stromal cell-derived factor 1

Nal

l-3-(2-naphthyl)alanine

AML

acute myeloid leukemia

PEM

protein epitope mimetic

HAS

human serum albumin

G-CSF

granulocyte colony-stimulating factor

Dpa

dipicolylamine

GPCR

G protein-coupled receptor

NIR

near infrared

IN

integrase

DKA

diketoacid

This work was supported in part by the “Research Program on HIV/AIDS” from Japan Agency for Medical Research and Development (AMED), a “Grant-in-Aid for Scientific Research” from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan, and a “Health and Labor Sciences Research Grant” from the Japanese Ministry of Health, Labor, and Welfare. It was also supported in part by JSPS Core-to-Core Program, A and the Platform for Drug Discovery, Informatics, and Structural Life Science of MEXT, Japan. The authors wish to acknowledge their collaborators: Drs Nobutaka Fujii (Kyoto University), Naoki Yamamoto (National University of Singapore), Tsutomu Murakami (National Institute of Infectious Diseases), Hideki Nakashima (St. Marianna University), Hiroaki Mitsuya (Kumamoto University), Toshio Hattori (Tohoku University), Michinori Waki (Kyushu University), Akira Otaka (The University of Tokushima), Itaru Hamachi (Kyoto University), Masao Matsuoka (Kyoto University), Shuzo Matsushita (Kumamoto University), Kazuhisa Yoshimura (National Institute of Infectious Diseases), Shigeyoshi Harada (National Institute of Infectious Diseases), John O. Trent (University of Louisville), Stephen C. Peiper (Thomas Jefferson University), Zixuan Wang (Thomas Jefferson University), Huangui Xiong (University of Nebraska Medical Center), Shuichi Kusano (St. Marianna University), Shigemi Terakubo (St. Marianna University), Akio Ojida (Kyushu University), Shinya Oishi (Kyoto University), Satoshi Ueda (Kyoto University), Jun Komano (National Institute of Infectious Diseases), Eiichi Kodama (Tohoku University), Kenji Ohba (National University of Singapore), Emiko Urano (National Institute of Infectious Diseases), Kasthuraiah Maddali (National Cancer Institute), Yves Pommier (National Cancer Institute), John A. Beutler (National Cancer Institute), Aikichi Iwamoto (The University of Tokyo), Tomohiro Tanaka (Tokyo Medical and Dental University), Chie Hashimoto (Tokyo Medical and Dental University), Hiroshi Tsutsumi (Tokyo Medical and Dental University), Tetsuo Narumi (Tokyo Medical and Dental University), Haruo Aikawa (Osaka University), Takaaki Mizuguchi (Tokyo Medical and Dental University), and Wataru Nomura (Tokyo Medical and Dental University), Mr Kenichi Hiramatsu (Kyoto University), Takanobu Araki (Kyoto University), Teppei Ogawa (Kyoto University), Hiroki Nishikawa (Kyoto University), Yasuaki Tanabe (Tokyo Medical and Dental University), Toru Nakahara (Tokyo Medical and Dental University), Hiroshi Arai (Tokyo Medical and Dental University), Taro Ozaki (Tokyo Medical and Dental University), Akira Sohma (Tokyo Medical and Dental University), Yu Irahara (Tokyo Medical and Dental University), Takaharu Suzuki (Tokyo Medical and Dental University), Taisuke Koseki (Tokyo Medical and Dental University), Shohei Taketomi (Tokyo Medical and Dental University) and Barry Evans (Medical College of Georgia), and Ms Akane Omagari (Kyoto University), Ai Esaka (Kyoto University), Miki Nakamura (Kyoto University), Yuko Yamada (Tokyo Medical and Dental University), Aki Ohya (Tokyo Medical and Dental University), Chihiro Ochiai (Tokyo Medical and Dental University), Aiko Ogawa (Tokyo Medical and Dental University), Ami Nozue (Tokyo Medical and Dental University), Kyoko Itotani (Tokyo Medical and Dental University) and Miho Tanabe (Tokyo Medical and Dental University).