From Chemical Tools to Clinical Medicines: Nonimmunosuppressive

---

Perspective pubs.acs.org/jmc

From Chemical Tools to Clinical Medicines: Nonimmunosuppressive Cyclophilin Inhibitors Derived from the Cyclosporin and Sanglifehrin Scaffolds Zachary K. Sweeney,* Jiping Fu, and Brigitte Wiedmann Novartis Institutes for BioMedical Research, 4560 Horton Street, Emeryville, California 94608, United States ABSTRACT: The cyclophilins are widely expressed enzymes that catalyze the interconversion of the cis and trans peptide bonds of prolines. The immunosuppressive natural products cyclosporine A and sanglifehrin A inhibit the enzymatic activity of the cyclophilins. Chemical modification of both the cyclosporine and sanglifehrin scaffolds has produced many analogues that inhibit cyclophilins in vitro but have reduced immunosuppressive properties. Three nonimmunosuppressive cyclophilin inhibitors (alisporivir, SCY-635, and NIM811) have demonstrated clinical efficacy for the treatment of hepatitis C infection. Additional candidates are in various stages of preclinical development for the treatment of hepatitis C or myocardial reperfusion injury. Recent publications suggest that cyclophilin inhibitors may have utility for the treatment of diverse viral infections, inflammatory indications, and cancer. In this review, we document the structure−activity relationships of the nonimmunosuppressive cyclosporins and sanglifehrins in clinical and preclinical development. Aspects of the pharmacokinetic behavior and chemical biology of these drug candidates are also described.

■

INTRODUCTION The cyclosporins and the sanglifehrins are peptidic macrocycles that inhibit a class of prolyl isomerase enzymes known as the cyclophilins.1 These compounds, which generally have immunosuppressive properties, are natural products made by fungi.2,3 Synthetic modification of the cyclosporin and sanglifehrin scaffolds can provide molecules that are less immunosuppressive but retain cyclophilin binding activity.4 Such “nonimmunosuppressive cyclophilin inhibitors” have provided critical tools for the elucidation of the mechanism of action of the cyclosporins and the biological properties of the cyclophilins. Furthermore, three compounds from this class have demonstrated clinical efficacy for the treatment of hepatitis C infection.5 Additional classes of cyclophilin inhibitors, including the sanglifehrins, have been discovered, and these molecules have also served as starting points for the discovery of structurally diverse drug candidates. Many publications suggest that nonimmunosuppressive cyclophilin inhibitors may be useful medicines for a wide spectrum of diseases. In particular, cyclophilin inhibitors have been proposed to have potential utility for the treatment of diverse viral infections,6 inflammation,7 cardiac failure,8 and cancer.9 Recently, there has been a resurgence in the application of cyclic peptides as medicinally active agents.10 In this context, the unique medicinal and pharmaceutical properties of the cyclosporins and the sanglifehrin family of macrocyclic peptides are also instructive. For example, the remarkable structure− activity relationships of these natural products illustrate correlation between the ligand conformational energy landscape and protein−ligand affinity.11 Studies of the relationship between the in vitro cyclophilin affinity and cellular potency of © XXXX American Chemical Society

these molecules also provide insight into factors that contribute to the cellular permeability of macrocyclic peptides.12,13 In this review, we document the discovery, pharmacokinetic behavior, and structure−activity relationships of the nonimmunosuppressive cyclosporins and sanglifehrins. Clinical and preclinical drug candidates are described together with the efficacy of these nonimmunosuppressive cyclophilin inhibitors in cellular and animal disease models.

■

DISCOVERY OF CYCLOSPORIN A AND THE CYCLOPHILINS Cyclosporin A14 (1, CsA, Figure 1A) was originally discovered in a screening program at Sandoz directed toward the identification of noncytotoxic immunosuppressive agents. The development of this compound revolutionized organ transplant treatment, and cyclosporine A has also found application for the treatment of psoriasis, rheumatoid arthritis, and uveitis.2 Mechanistic studies gradually revealed that many of the immunosuppressive effects of 1 can be attributed to the formation of a cytoplasmic ternary complex between the abundant cellular protein cyclophilin A (PPIA) and the two subunits of the phosphatase calcineurin (Figure 2).15,16 Formation of this ternary complex prevents calcineurin from dephosphorylating the transcriptional regulator nuclear factor of activated T-cells (NFAT), thereby preventing production of proteins associated with the immune response.17,18 The discovery of the cyclophilin proteins occurred in parallel with the investigation of the cyclosporin family of natural Received: February 10, 2014

A

dx.doi.org/10.1021/jm500223x | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Perspective

Figure 1. Left: Cyclosporin A (1, CsA). Right: Solid state structure of 1 (DEKSAN).

Figure 2. Left: Complex between cyclophilin A, cyclosporin A, and calcineurin (PDB code 1MF8). Right: Structure of cyclosporin A bound in cyclophilin A conformation (PDB code 1CWA).

products. The first cyclophilin was purified from porcine kidney cortex by Fischer et al. in the early 1980s,19 although it was only realized 4 years later that this protein was identical to the cyclosporine A binding proteins identified in mammalian thymocytes20,21 and in human T-cells.22,23 Cyclophilin proteins have been identified in all phyla from E. coli to H. sapiens. For example, the yeast S. cerevisiae has eight cyclophilins. Fruit flies have 14 cyclophilins and human cells at least 19 cyclophilins.24 The archetype cyclophilin is the 18 kDa cyclophilin A (PPIA) which consists almost only of the catalytic domain. Other paralogs have additional domains that can raise the molecular weight to 354 kDa. These domains are thought to regulate protein−protein interactions and intracellular location, thereby controlling the participation of these proteins in many biological processes.25 The cyclophilins form a subset of the peptidylprolyl cis/trans isomerases (rotamase, PPIase, EC 5.1.2.8.). These enzymes catalyze the interconversion of the cis and trans peptide bonds of prolines (Figure 3). Proteins are synthesized in a trans proline conformation, but at least 30% of peptidylprolyl bonds are found in a cis conformation, suggesting that most proteins interact at some point with a prolyl isomerase enzyme. However, it is not clear if all biological functions observed for cyclophilin proteins, such as the chaperone function (see below), require isomerase activity. Several PPIases, including

PPIL2, PPIL4, PPIL6, and SDCCAG-10, lack an essential and conserved tryptophan/histidine and do not have detectable isomerase activity or bind 1.1 Depletion of single cyclophilins is not strictly lethal in mammals,26 although homozygotic mice lacking cyclophilin A have a lower survival rate than wild-type animals. Cyclophilin B (PPIB) knockout mice die prematurely and show decreased size and body weight, skeletal deformations, and reduced bone density. Cyclophilin F (PPIF) knockdown mice have abnormal mitochondrial physiology and decreased cerebral infarction size. However, the manageable hypertension and renal toxicities associated with administration of 1 suggest that cyclophilin function in humans can be inhibited without causing serious adverse events. Defects in cyclophilin expression level and activity have been associated with a wide range of diseases, although in many cases, an improved understanding of these enzymes is required to more firmly establish a causative role. In addition to their enzymatic activity, cyclophilins act as chaperones for a wide range of biomolecules. Cyclophilin B, for example, serves as a chaperone for procollagen, and human diseases related to collagen have been linked to alterations in this enzyme.27 This cyclophilin also transports apolipoprotein B, which is the major component of low density lipoproteins involved in cholesterol transport.28 The expression level of cyclophilin B modulates the prolactin-mediated growth and migration of breast cancer cells.29 Another proline isomerase, cyclophilin 60 (PPIL2), transports the cyclophilin receptor CD147/EMMPRIN to the cell surface, and inhibition of this transporter causes less accumulation of the immunomodulator CD147 at the cell surface without affecting synthesis.30 Cyclophilin A and cyclophilin B can be secreted from cells stimulated by oxidative stress and act as chemokines for the leukocytes with CD147 on their surface.7,31 These cyclophilins therefore play a positive

Figure 3. Interconversion of proline rotamers catalyzed by the cyclophilins. B

dx.doi.org/10.1021/jm500223x | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Perspective

trans-position. Furthermore, the hydroxyl group of the [BMT]1 (BMT = (4R)-4-[(E)-2-butenyl]-4-methyl-L-threonine) residue forms a hydrogen bond with the carbonyl group of [MeLeu]4 and holds these groups in proximity to one another (Figure 2, right).49 The binding of 1 to cyclophilin is directly mediated by interactions with P9, P10, P1, and P2 groups (see Figure 1 for numbering), although substitutions at other positions of the undecapeptide can impact cyclophilin binding affinity.53 The interface between 1 and calcineurin consists primarily of the P4, P5, and P6 amino acids of the macrocycle.54 A general strategy for the design of nonimmunosuppressive cyclophilin inhibitors has been to reduce calcineurin binding by modifying one of the [Leu]4-[Val]5-[Leu]6 groups while substituting the [NMeGly]3 group with a substituent that favors the cyclosporin-bound conformation. Cyclosporin A is very lipophilic and is widely distributed to tissue compartments in vivo. Plasma protein binding is reported to be approximately 95%, although determined values can vary dramatically.55 Analysis of the pharmacokinetic profile of cyclosporin A and other cyclophilin inhibitors is complicated by the high expression of cyclophilin A in blood cells, especially erythrocytes.56 This leads to a dose and exposure-dependent blood/plasma partitioning of cyclophilin inhibitors. For example, following oral administration, the exposure of the nonimmunosuppressive cyclophilin 8 in blood is not dose proportional while plasma exposure increases in a linear fashion.44 The immunosuppressive effect of cyclosporin A is associated with the unbound concentration in plasma,55 and it should be expected that the pharmacological efficacy of nonimmunosuppressive cyclophilin inhibitors would be proportional to free drug exposure in plasma. Therefore, relatively high total doses of cyclophilin inhibitors may be required to saturate binding of cyclosporin A in erythrocytes and provide free drug concentrations required for efficacy. The pharmacology of 1 is further complicated by its significant interaction with endogenous transporters. This undecapeptide is a potent inhibitor of organic anion transporting proteins (i.e., OATP1B1 and OATP1B3) in vitro and in vivo.57,58 Compound 1 is also a strong inhibitor of the bilesalt export pump (BSEP),59 the multidrug resistance protein 1 (MDR1, ABCB1, P-gp),59 and MDR2.60 The structure−activity relationships regarding the inhibition of MDR1 by a number of cyclosporins have been reported in detail,61 and valspodar (PSC-833, 3′-oxo-4-butenyl-4-methyl-Thr 1 ][Val 2 ]cyclosporine),62 a nonimmunosuppressive cyclosporin inhibitor of MDR1 that does not have affinity for cyclophilins, was studied clinically for its potential to improve response to chemotherapy.63 Alisporivir (8,64 Figure 6), a hydrophobic nonimmunosuppressive cyclophilin inhibitor closely related to 1, has also been reported to be an inhibitor of OATP1B1 and MDR2 activity. These transporters are involved with the excretion of bilirubin and/or the bilirubin−glucuronide conjugate into the bile.65 Inhibition of this pathway can cause hyperbilirubinemia, and some cases of hyperbilirubinemia following administration of 8 have been reported. Recent studies have found that large, hydrophobic peptides are often OATP inhibitors.66 Interestingly, as described below, addition of hydrophilic groups to the cyclosporin template seems to significantly reduce the inhibition of certain transporters.

role as part of the innate immunity against infections. Dysregulation of these cyclophilins can also cause inflammation, as is seen in myocardial reperfusion injury32 and chronic asthma.33 These biological functions of the cyclophilins suggest that cyclophilin inhibitors may modulate the immune response without interfering with calcineurin function. Cyclophilin inhibitors have been investigated in a variety of disease models and in clinical studies for their ability to reduce cellular apoptosis by inhibiting the formation of the mitochondrial transition pore (MTP).34−38 Assembly of the mitochondrial transition pore is part of the cellular stress response. Prolonged, excessive activation of MTPs has been implicated in the increased mitochondrial membrane permeability and cell death associated with traumatic brain injuries and strokes.39,40 Numerous in vitro and in vivo studies have been performed to test how cyclophilin F activity affects the outcome of disease models involving the MTP, and nonimmunosuppressive cyclophilin inhibitors have proven to be effective inhibitors of the MTP in in vitro and in vivo models.40 Advanced clinical trials using 1 as a protective agent for reperfusion injury following myocardial infarction are reported to be ongoing.41 Viruses rely heavily on host proteins, including the cyclophilins, as their genome does not encode all of the proteins essential for their life cycle. The most compelling evidence for cyclophilin participation in viral amplification, including efficacy of compound treatment in clinical trials, exists for the treatment of human immunodeficiency virus (HIV) and hepatitis C virus (HCV). Although cyclophilin inhibitors potently suppress HIV infection, naturally preexisting coat protein mutants are resistant to treatment and preclude the broad therapeutic use of cyclophilin inhibitors against HIV.42 Treatment of HCV with cyclophilin inhibitors has been very promising however. Three compounds, NIM811 (7,84 Figure 5), alisporivir (8,64 Figure 6), SCY-635 (9,46 Figure 7), have been investigated in clinical trials, and each has demonstrated therapeutic efficacy.43 Several cyclophilins contribute to the HCV viral life cycle,44,45 and strong evidence indicates that cyclophilin A interacts with the viral NS5A protein to stimulate the replication of the RNA genome.44,46 In contrast to most direct antivirals, significant resistance to cyclophilin inhibitors is rarely observed, and cyclophilin inhibitors are potent across all HCV genotypes.47,48

■

CYCLOSPORINS Compound 1 (Figure 1) is a cyclic undecapeptide with a molecular weight of approximately 1200 Da. A variety of spectroscopic and structural studies indicate that 1 and related molecules are highly flexible. The affinity to cyclophilins can be modulated by substitution of the cyclosporin in a manner that reduces the energy of the cyclophilin-bound conformation.49−51 Conformational flexibility is also thought to contribute to the remarkable cellular permeability and oral bioavailability of this high molecular weight peptide.52 In nonpolar solvents and in the solid (unbound) state, all of the amide hydrogens of 1 are engaged in intramolecular hydrogen bonds with amide carbonyl groups (Figure 1, right). In this configuration, the hydrophobic side chains and the N-alkyl groups form a hydrophobic layer of alkyl chains that sit above and below the polar macrocyclic amide linkages. Presumably, cyclosporins adopt this conformation when passing through cellular membranes. In polar solvents and when bound to cyclophilins, the hydrophilic amide groups are revealed and all peptide bonds are in the C

dx.doi.org/10.1021/jm500223x | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

■

Perspective

NONIMMUNOSUPPRESSIVE CYCLOSPORINS The discovery of 1 and its mysterious, novel mechanism of action generated a tremendous amount of interest in the industrial and academic medicinal chemistry community. At Sandoz, a team of medicinal chemists in collaboration with Dieter Seebach (ETH Zurich) developed synthetic methods that led to the preparation of hundreds of cyclosporin analogues.67−71 The discovery of additional naturally occurring cyclosporins also allowed the structure−activity relationships of the cyclosporins, with respect to suppression of the cellular immune response, to be elucidated.72 Following the discovery of cyclophilin A, an ELISA assay was developed that allowed the affinity of cyclosporins for cyclophilin A to be determined.73 This assay employed a conjugate of [D-Lys]8-CsA appended to bovine serum albumin through the lysine functionality. While the authors determined that the immunosuppressive activity of most of the cyclosporins correlated with their affinity for cyclophilin A, it was observed that two analogues that featured bulky groups in the P3 position including 273 (Figure 4) retained affinity for

on the interaction between the CsA/cyclophilin complex and calcineurin. Replacement of the [Leu]4 in 1 to produce [Val]4CsA (4,76 Figure 5) or [4′-OHLeu]4-CsA (5,76 Figure 5) retained or improved cyclophilin A binding while sharply reducing activity in the immunosuppression assays. From these results, the authors concluded that “calcineurin has a very “tight-binding” pocket” for [Leu]4, and this supposition was later confirmed by cocrystallization experiments.54

■

P3,P4-MODIFIED CYCLOSPORINS Previous studies with 1 had revealed that appropriate substitution of the [Gly]3-group49,79 or the addition of small hydrophobic elements to the [Leu]4-residue of 180 improved cyclophilin binding by increasing the population of the cyclosporin binding conformer. This information led to the investigation of the anti-HIV activity of a series of nonimmunosuppressive cyclosporins that had greater affinity for cyclophilin A than 1 itself.81,82 Representative analogues [(D)MeSer]3-[(4-OH)-MeLeu]4-CsA (6,81 Figure 5) and [isoleu-

Figure 4. Early nonimmunosuppressive cyclophilin inhibitors [(D)NMePhe]3-CsA and [MeAla]6-CsA.

Figure 5. P4-modified nonimmunosuppressive cyclophilin inhibitors 4 and 5, 6, and NIM811 (7).

cyclophilin A but were not immunosuppressive. To the best of our knowledge these P3-modified cyclosporins were the first molecules specifically identified as nonimmunosuppressive cyclophilin inhibitors. Scientists at Merck also profiled the cyclophilin binding and immunosuppressive activity of a large collection of synthetic cyclosporins.74 They noted that “a number of (these) compounds can interact with cyclophilins but are much less immunosuppressive than expected”. [MeAla]6-CsA (3,74 Figure 4), in which the [MeLeu]6 of 1 has been replaced with a NMealanine residue, was one of the analogues highlighted in their published report. This compound had approximately 50% of the cyclophilin binding affinity of 1 but only 0.4% of the immunosuppressive activity. [MeAla]6-CsA was utilized by the Schreiber group as part of their classic investigations into the mechanism of 1 and FK506.75 In particular, it was observed that the [cyclophilin A-[MeAla]6-CsA] complex had negligible affinity for calcineurin. This observation provided strong evidence for the relevance of the ternary cyclophilin/CsA/ calcineurin complex in the immunosuppressive pharmacology of 1. A detailed structure−activity relationship study describing analogues of 1 that retained cyclophilin binding activity while having reduced affinity for calcineurin was published in 1994 by Sandoz.76 Although the crystal structure of the ternary CsA/ cyclophilin A/calcineurin complex was not available at the time, the initial binding,73 NMR,77 and crystallographic78 studies interrogating the structure of the CsA/cyclophilin A complex had confirmed that positions 4−8 were remote from the cyclophilin binding interface. Consequently, modification of the P4 group was explored in order to probe the potential impact

cine]4-CsA (7,84 NIM-811, Figure 5) were found to be inhibitors of HIV viral amplification in cellular assays and essentially devoid of immunosuppressive activity. Interestingly, while 6 was 4-fold more potent than 7 in binding assays, it was less potent in the anti-HIV cellular assay. This result indicated that modifications of the side chains of 1 with polar residues must be carefully considered in order to avoid compromising cellular permeability.83 Compounds substituted at P3 (such as 6) could be prepared using methodology developed by Seebach and Sandoz, in which treatment of 1 with excess lithium diisopropylamine (6.0−14.0 equiv) generates an enolate at P3 that can be reacted with various electrophiles (Scheme 1). The P3 Re-stereoselectivity ratio obtained in this alkylation step can be as high as 7/1. The lead compound from these initial studies, 7,84 has higher affinity for cyclophilin A than 1. This inhibitor has been studied clinically as an HCV medicine, and it has also been employed in a number of animal and cellular models of disease. The discovery of 7 appears to have predated the detailed SAR studies described above, as it was originally isolated as a minor component from a threonine-enriched fermentation with a strain of Tolypocladium niveum that produced 1. Genetic manipulation of the producing strain allows 7 to be produced on large scale directly from fermentation using microbial broths enriched with N(Me)-isoleucine. The anti-HIV activity of 7 was attributed to its ability to interfere with the binding of cyclophilin A to the HIV-p24gag protein.82,42 Following the discovery of the antiviral effect of 1 on HCV replication,85,86 development of 7 focused on its potential as a treatment for HCV infection. Importantly, despite D

dx.doi.org/10.1021/jm500223x | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Perspective

Scheme 1. P3 Enolate Generation and Modificationa

a

(a) Lithium diisopropylamide, THF. (b) Electrophile (i.e., CH3I, CH2O, CH3SSCH3).

Scheme 2. Method for Modification of P3 and P4 of Cyclosporinsa

(a) Ring-opening ((i) Ac2O, DMAP, CH2Cl2; (ii) Me3O+BF4−; (iii) NaOCH3; (iv) H2SO4). (b) P4 removal by Edman degradation ((i) PhNCS, DMAP; (ii) TFA). (c) Addition of new P3−P4 groups using peptide coupling methodologies. (d) Removal of P3 group ((i) NaBH4; (ii) MsOH; (iii) NaOMe). (e) Removal of Boc group. (f) Cyclization using high dilution peptide coupling conditions. a

pore formation in hepatic cells.91 Other investigations suggested that 7 is neuroprotective in vitro.92,93 Finally, 7 was found to be able to modify the pathology of leukocyte reactivation in a murine asthma model.33 Such anti-inflammatory activity is consistent with other studies showing that extracellular cyclophilins play a role in leukocyte trafficking.94,30 Wenger and co-workers developed a synthetic sequence that allows for the modification of both the P3 and P4 groups of 1 (Scheme 2).95 Thus, the macrocycle is opened at P3−P4 and the P4 residue is excised by Edmund degradation. A new P3− P4 sequence can be appended to the resulting linear peptide. Through a clever reduction and rearrangement sequence, the original NMeGly P3 group is removed. Finally, macrocyclization in the presence of coupling reagent under high dilution conditions provides P3−P4 modified analogues of 1. Using this technology and published structure−activity relationships, scientists at Debiopharm SA and the University of Lausanne introduced a new lead nonimmunosuppressive cyclophilin inhibitor, Debio-025 (8,96 since renamed alisporivir, Figure 6). Compound 8 contains the [Val]4 group shown to mitigate calcineurin binding in the early Sandoz studies as well as a [(D)-Ala]3 residue. It also contains an N-Et (in place of NMe) functionality at P4. The P4 N-Me group is a major site of CYP3A4-mediated metabolism for some cyclophilin inhibitors, including 1,97 and the N-Me for N-Et substitution may reduce the rate of this undesired biotransformation.

the high genetic variability of HCV, 7 was found to be able to effectively block rapid resistance development. In vitro, resistance emerged slowly and amounted to only 2-5 fold lower susceptibility to 7 compared to wild-type replicon.87 These studies demonstrated that nonimmunosuppressive cyclophilin inhibitors have a high genetic barrier to viral resistance development, as expected for host-targeting agents. In a phase Ib monotherapy study, 7 was investigated for antiviral activity in HCV-infected patients.88 Doses up to 600 mg b.i.d. were employed; however, no significant antiviral response was achieved after 14 days. In combination with interferon-α, however, this dosing regimen was found to increase the reduction in HCV RNA from 0.6 log 10 (interferon only) to 2.8 log 10 (7 + interferon) following 7 days of treatment. Compound 7 has also been employed as an inhibitor of cyclophilin function in a number of animal models of disease. The results highlight the wide range of disease processes associated with cyclophilin function. In a CCl4 model of liver fibrosis, administration of 7 reduced the extent of liver necrosis.89 This activity was attributed to the inhibition of cyclophilin B and F activity. The compound also reduced cholestatic necrosis and apoptosis in a murine induced cholestasis model and reduced liver injury following hepatectomy,90,34 results ascribed to the ability of 7 to block mitochondrial apoptosis by inhibiting mitochondrial transition E

dx.doi.org/10.1021/jm500223x | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Perspective

Figure 7. Structure of 9. Figure 6. Structure of alisporivir (8).

elongation of the P3 substituent from [-SCH3] to the [-SCH2CH2N(CH3)2] group found in 9 reduced in vitro immunosuppressive activity while improving anti-HIV potency. The preclinical profile of 9 has been extensively documented.105 The compound is a slightly more potent inhibitor of cyclophilin A isomerase activity than 1. Reduction of HCV replication in vitro was reported to be time-dependent, with full replication inhibition not reached until 72 h. Compound 9 appears to be a weaker inhibitor of transporter activity than many other cyclophilin inhibitors. Although complete inhibition of P-gp efflux activity was observed at moderate (15 μM) concentrations, inhibition of MRP2-mediated transport was very weak (approximately 150 μM).106 In a phase Ib study, patients were treated with 9 3 times daily at individual doses ranging from 100 to 300 mg. No compound related toxicities were reported, and at the highest dose there was a mean 2.24 log 10 unit decline in plasma HCV RNA. Minimum plasma concentrations similar to the in vitro EC90 were required in order for an appreciable antiviral effect to be observed.46 The antiviral activity of 9 in this clinical study also correlated with the apparent induction of endogenous components of the interferon-mediated immune response. This observation, while not yet reproduced by other groups, is supported by in vitro studies in PBMC cells. It is possible that antiviral activity of 9 may result from both modulation of the innate immune system and the disruption of cyclophilin A− HCV NS5A interactions.107,108 Enanta Pharmaceuticals has described the results of their preclinical profiling with EDP-546 (10,109 structure not disclosed), a nonimmunosuppressive cyclophilin inhibitor derived from cyclosporin A. The company has reported plans to begin phase I studies with this molecule in early 2014. The structure of 10 has not been released, but patent applications have focused on structures similar to 11 (Figure 8). In cellular passaging experiments with 10 in genotype 1a and 1b replicon systems, moderate (2- to 5-fold) levels of resistance arose by selection of the D320E variant within the NS5A protein. Antiviral activity of combinations with HCV protease inhibitors and HCV NS5A inhibitors was generally additive, and co-incubation of 10 with these antiviral agents reduced the emergence of viral resistance.110 The cellular activity of the inhibitor was not impacted by the addition of 40% human serum to the growth media, and 10 was more stable than 8 in human microsomes. In preclinical pharmacokinetic studies, 10 had low blood clearance and significant distribution into the liver. Good solubility was observed in aqueous buffer solutions, particularly at pH values less than 7. In addition, it was reported that 10 was a much weaker inhibitor of the bilirubin and drug transporters MRP2 (approximately 100-fold) and OATP1B1 (approximately 3-fold) than 8.111

The cyclosporin 8 is approximately 5- to 10-fold more potent than 1 in the cellular subgenomic HCV replicon systems.98 This inhibitor has also been reported to be more potent in antiviral cellular assays than 7. The improved antiviral activity of 8 relative to the earlier generations of cyclophilin inhibitors may be due to the increased cyclophilin affinity provided by the [(D)-Ala]3 group.96 HCV resistance development to 8 was a slow process, and when the compound was combined with direct-acting antiviral agents, cells were rapidly cleared of replicon.99 The modified cyclosporin 8 was originally developed as a treatment for HIV infection. Initial 10-day studies in infected patients showed that once-daily administration could cause modest reductions in HIV-1 plasma RNA. A subsequent study in HIV-HCV co-infected patients, using significantly higher doses than were explored in the initial efficacy trial, demonstrated that drug exposure unexpectedly increased in a nonlinear fashion at higher doses.100 Furthermore, the compound caused a significant reduction (average of −3.6 log 10 units) in plasma HCV RNA but only a modest and more variable decrease in HIV RNA (average of −0.6 log 10 units). The improved clinical activity of 8 against HCV relative to 7 can be attributed to a combination of higher drug exposure and improved potency. The anti-HCV efficacy of 8 as monotherapy or in combination with interferon and ribavirin has since been studied in several large clinical studies.44,101 Viral breakthrough due to resistance development has been rare and was most likely related to low exposure, confirming the high barrier to resistance development observed in vitro. Treatment associated hyperbilirubinemia, possibly associated with inhibition of the bilirubin transporters MDR2 and OATP1B, was transient and reversible. Other clearly drug-related side effects have not been described. This cyclophilin inhibitor appears to be particularly effective for treating patients infected with genotype 2 or genotype 3 HCV, and phase II studies in this patient population are ongoing. Inhibitor 8 has also been studied in several murine models of muscular dystrophy, based on its ability to slow cellular necrosis by desensitizing the mitochondrial permeability pore.102 In these studies, administration of 8 was consistently associated with a decrease in muscle wasting. Cyclosporin derivative 946 (Figure 7) is a nonimmunosuppressive cyclophilin inhibitor currently in phase II development for treatment of HCV infection. This compound was originally discovered at Aventis in a program targeting novel cyclophilin inhibitors for the treatment of HIV.103 Enzymatic hydroxylation of 1 by Sebekia benihana provided [4′-(OH)MeLeu]4-CsA. Selective thioalkylation of the [MeGly]3 residue of [4′(OH)MeLeu]4-CsA104 gave a series of thioalkyl compounds that feature amine groups appended to P3. In this series, F

dx.doi.org/10.1021/jm500223x | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Perspective

Figure 8. Structures of Enanta (11), S&T Global (12), and Allergan (13) cyclosporins.

A series of 3- and 4-modified cyclosporin analogues were prepared by S&T Global.112−115 The majority of these compounds were prepared starting from [4′-(OH)-MeLeu]4CsA, the product of oxidative biotransformation of 1 used for the production of 9. Data provided indicate that the preferred compounds, unlike 9, do not epimerize at the P3 position significantly in MeOH at slightly elevated temperatures. Many examples such as 12 (Figure 8) showed potent anti-HCV in the cellular replicon assay.113 Allergan has also recently reported the synthesis of a series of analogues that contain modifications to the cyclosporin scaffold in the P3 position.116 In this work, [MeLeu]4 group of 1 was maintained. Interestingly, substitution with a thioether group was generally found to maintain the inhibition of cyclophilin A isomerase activity while reducing the ability of the compound to reduce calcineurin phosphatase activity in the presence of cyclophilin A. Furthermore, the compounds were determined to be significantly weaker inhibitors of immune response in a mouse mixed-lymphocyte assay. Compound 13 (Figure 8), for example, is nearly equipotent to 1 in an in vitro enzymatic assay, but this analogue was a much weaker inhibitor of a mixed lymphocyte reaction measuring immunosuppressive potential (EC50 1 = 60 nM, EC50 13 = 3000 nM). As the [Leu]4 group of 1 is maintained in these analogues, these data, together with results originally reported by the Aventis and Sandoz groups,103,73 indicate that certain P3 substitutions may reduce calcineurin inhibition.

Figure 9. Structures of the Aurinia P1-modified example 14 and poorly permeable example 15.

Traditionally, the immunosuppressive potential of cyclosporin derivatives was tested in isolated T-cell assays. In this context, the immunosuppressive activity of cyclosporins largely correlates with the ability of cyclophilin A−cyclosporin complexes to inhibit the activity of calcineurin. However, as described above, cyclophilins appear to be involved in other immune response processes such as innate immunity and inflammation. In order to explore the potential of cyclophilin inhibitors to act as inhibitors of leukocyte trafficking by inhibiting extracellular cyclophilins, a series of impermeable cyclosporins were prepared.83 The benzimidazole derivative 15 (Figure 9) was determined to have a similar affinity to cyclophilin A as 1 in an in vitro proline isomerase binding assay. A cellular cyclophilin binding assay developed using a fluorescently labeled cyclophilin derivative revealed that 15 is at least 50-fold less permeable than 1. As expected, despite its significant restriction to the extracellular space, 15 inhibits leukocyte migration toward cyclophilin A. This inhibitor also reduces leukocyte activation in a mouse model of allergic contact hypersensitivity.

■

P1-MODIFIED CYCLOSPORINS The [BMT]1 residue characteristic of the cyclosporins is unique in that in the cyclophilin-bound conformation, this group extends from the cyclophilin binding interface to a region very near the calcineurin-binding region (Figure 2). Early work had suggested that P1-modified cyclosporins can retain cyclophilin binding affinity without maintaining the immunosuppressive activity of 1.73,117 In particular, the Merck group highlighted the cyclophilin binding activity of [CH(OH)CH(CH 3 )CH2SCH3]1-CsA, which has 178% of the cyclophilin binding activity of 1 but only 10% of the immunosuppressive potency.74 Aurinia has reported that extension of the BMT group of the cyclosporins and addition of a D-Ala group in the P3 position provides a series of nonimmunosuppressive cyclophilin analogue molecules such as 14 (Figure 9).118 This molecule has more than 14-fold greater inhibition of cyclophilin A isomerase activity relative to 1. Furthermore, while specific assay methodology was not provided, it was asserted that this example has only 3% of the immunosuppressive activity of 1. Although the structure has not been revealed, in July 2013 it was reported that Aurinia had identified a lead candidate and was in the process of fully evaluating the scope of the anti-HCV activity of this analogue.

■

P5-MODIFIED CYCLOSPORINS Although detailed data are not available, in a patent filing Scynexis explored the potential for alkylation at P5 to influence the immunosuppressive activity and antiviral potency of 1.119 Earlier, chemists at Sandoz had demonstrated that alkylation of the free NH at this position can influence the conformation and immunsuppressive activity of cyclosporin derivatives.120 Compound 16 (Figure 10) was prepared via alkylation of [(D)MeAla]3-CsA with dimethylallyl bromide. This analogue was reported to bind to cyclophilins A and D with good (