Synthesis and biological evaluation of new potent and selective HCV NS5A inhibitors

Abstract

NS5A inhibitors are a new class of direct-acting antiviral agents which display very potent anti-HCV activity in vitro and in humans. Rationally designed modifications to the central biphenyl linkage of a known NS5A series led to selection of several compounds that were synthesized and evaluated in a HCV genotype 1b replicon. The straight triphenyl linked compound 11a showed similar anti-HCV activity to the clinical compound BMS-790052 and a superior cytotoxicity profile in three different cell lines, with an EC₅₀ value of 26 pM and a therapeutic index of over four million in an HCV replicon assay. This triphenyl analog warrants further preclinical evaluation as an anti-HCV agent.

Hepatitis C virus (HCV) is an important pathogen affecting nearly 170 million people worldwide.¹ HCV infections become chronic in about 50% of cases,² and about 20% of these chronic persons develop liver cirrhosis that can lead to hepatocellular carcinoma.³ Current therapies, interferon-alpha (IFN-α) and ribavirin, even when used with the newly approved HCV protease inhibitors Incivek and Victrelis, have limited efficacy and serious side-effects.⁴ Therefore, there is a need for more effective and safer small molecule anti-HCV agents.

In our continuing efforts to identify more effective direct-acting antiviral agents (DAA), a relatively new target, the non-structural protein NS5A, has emerged as an attractive objective.⁵ Currently in the clinic, the two most common targets for DAA with HCV are the non-structural proteins NS3 and NS5B.⁶ Among the recently discovered NS5A inhibitors, BMS-790052 showed a median effective antiviral concentration (EC₅₀) in vitro, in the picomolar range and demonstrated in clinical trials, a reduction in HCV RNA of over 3log₁₀ IU/mL at 24 h following a single dose of 10 mg (Fig. 1).⁷

Figure 1.

Chemical structure of BMS-790052.

We carefully studied the SAR of some known NS5A inhibitors,^7,8 including BMS-790052 and earlier hits BMS-858, BMS-824, and BMS-665. For the BMS-790052 and BMS-665 series of compounds we found that many of the compounds were symmetrical or almost symmetrical around a central core, this core had only/mainly pi electron interaction capabilities with the NS5A protein, and the length of these molecules is quite different, leaving room for us to modify and optimize these molecules. We hypothesized that the two phenyl rings of BMS-790052 act only as a core linker between the two substituted imidazolylpyrrolidine regions. Thus, we envisaged that changing the length and/or the geometry of this central biphenyl linkage might lead to a more potent compound showing, perhaps, less cytotoxicity. Therefore, a series of bis-imidazolylpyrrolidine compounds with phenyl, phenoxyphenyl, triphenyl, pyridinyldiphenyl, triazole containing, and tetraphenyl linkages have been synthesized and evaluated for their anti-HCV activity and cytotoxicity in different cell lines.

Phenyl, phenoxyphenyl, phenylthiopheneyl, and phenylbenzenesulfonamide linked compounds 5a–d were prepared in four steps by adapting a reported procedure⁹ as depicted in Scheme 1. Di bromoketones 1a and 1c,d were prepared by bromination of the corresponding diketone while 1b was prepared by Friedel-Crafts acylation of the corresponding diphenylether with bromoacetyl chloride. Reaction of 1a–d with N-Boc-l-proline to give the diesters 2a–d in good to excellent yield. The esters 2a–d were then refluxed in toluene with ammonium acetate to form imidazoles 3a–d. Following Boc deprotection with 6 N HCl and coupling with N-(methoxycarbonyl)-l-valine in presence of N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDAC), the target compounds 5a–d were obtained in fair to excellent overall yields.

Scheme 1.

Synthesis of 5a–d. Reagents and conditions: (a) N-Boc-l-proline, MeCN, Et₃N, rt, 2 h, 51–100%; (b) NH₄OAc, toluene, 95–100 °C, 14 h, 51–73%; (c) 6 N HCl, MeOH, 50 °C, 4 h, 85–99%; (d) HOBt, EDAC, N-(methoxycarbonyl)-l-valine, MeCN, DIPEA, rt, 14 h, 64–95%.

The tricyclic and tetracyclic linked analogs were prepared as described in Scheme 2. The key intermediate boronate 8 was prepared from the bromo derivative 7a by a palladium catalyzed cross-coupling reaction with bis(pinacolato)diboron,¹⁰ while the bromide 7a was prepared from commercially available 2,4′-dibro-moacetophenone, 6a (Fig. 2) via esterification and cyclization, analogous to the preparation of compounds 3 in Scheme 1. Suzuki coupling of the boronate 8 with various dihalogenated arenes (1,4-diiodobenzene, 2,5-dibromopyridine, 1,3-dibromobenzene, 2,5-dibromothiophene, or 4,4′-dibromobiphenyl) resulted in the formation of tricyclic and tetracyclic linked compounds 9a–e in excellent yields. After Boc deprotection with HCl and coupling with N-methoxycarbonyl-l-valine, compounds 11a–e were obtained in fair overall yields (Scheme 2). Tricyclic analog 13 (Table 1) with the pyridine ring shifted to the terminal core position relative to 11b was prepared by the sequence outlined in Scheme 2 utilizing boronate 7d with pyridyl bromide 12 (Fig. 2) in place of the dibromide starting at step b.¹¹

Scheme 2.

Synthesis of 11a–e. Reagents and conditions: (a) Pd(PPh₃)₄, bis(pinacolato) diboron, KOAc, 1,4-dioxane, 80 °C, 16 h, 84%; (b) Pd(PPh₃)₄, NaHCO₃, 1,2-dimethoxyethane, H₂O, dihalide (1,4-diiodobenzene, 2,5-dibromopyridine, 1,3-dibromobenzene, 2,5-dibromothiophene, or 4,4′-diiodobiphenyl), 80 °C, 14 h, 80–99%; (c) 6 N HCl, MeOH, 50 °C, 4 h, 86–95%; (d) HOBt, EDAC, N-(methoxycarbonyl)-l-valine, MeCN, DIPEA, rt, 14 h, 28–59%.

Figure 2.

Chemical structures for 6a–c, 7a–d and 12.

Table 1.

In vitro anti-HCV activity and cytotoxicity data for compounds 5a–d, 11a–e, 13, 17and 18

Compd	X	Anti-HCVa (pM)		Cytotoxicity CC50 (μM)
		EC50	EC90	PBM	CEM	Vero
5a		>1000	>1000	>100	>100	>100
5b		30,000	90,000	16	16	19
5c		100,000	>100,000	15	13	>100
5d		90,600	>100,000	93	48	99
11a		26	90	>100	>100	>100
11b		26	68	3.9	2.3	6.5
11c		800	2900	18	3.0	>100
11d		100	300	26	12	>100
11e		240	1000	>100	>100	>100
13		200	700	5.5	31	29
17		200	900	18	5.0	35
18		600	1633	46	>100	>100
BMS-790052		6.6	23	19	9.6	21

^aAll dose–response determinations were calculated from at least four concentrations. All concentrations were run in triplicate and reported as mean values.

As shown in Scheme 3, the triazolodiphenyl linked compound 17 was prepared from azido 7c and acetylene 14 by using Cu(I)-catalyzed Huisgen azide–alkyne 1,3-dipolar cycloaddition (CuAAC).¹² The iodo derivative 7b, prepared from the corresponding bromoketone 6b (Fig. 2), was reacted with trimethylsilylacetylene under Sonogashira coupling conditions¹³ then desilylated under basic conditions to give the acetylene derivative 14 (Scheme 3). The cycloaddition reaction of azido compound 7c (prepared analogously to Scheme 1) and acetylene 14 resulted in the formation of triazole intermediate 15 in 67% yield. After N-Boc deprotection with 6 N HCl and coupling with N-methoxycarbonyl-l-valine, the triazolo-diphenyl compound 17 was obtained in good overall yield (Scheme 3). The related triazole 18 (Table 1) was prepared by Scheme 3 and started from the Boc’ed azide 7c and involved Boc removal (70%) and acylation with N-(methoxycarbonyl)-l-valine (95%). A CuAAC reaction of 1,4-diethynylbenzene with the resulting azide gave 18 in a 40% isolated yield.

Scheme 3.

Synthesis of triazole 17. Reagents and conditions: (a) (i) trimethylsilyl-acetylene, CuI, Pd(PPh₃)₄, Et₃N, THF, rt, 2 h; (ii) K₂CO₃, THF, MeOH, rt, 3 h, 72%; (b) 7c, CuSO₄, sodium ascorbate, t-BuOH, H₂O, 95–100 °C, 14 h, 67%; (c) 6 N HCl, MeOH, 50 °C, 4 h, 85%; (d) HOBt, EDAC, N-(methoxycarbonyl)-l-valine, MeCN, DIPEA, rt, 14 h, 80%.

All synthesized NS5A inhibitors were evaluated for inhibition of HCV RNA replication in Huh7 cells containing a subgenomic HCV 1b replicon. Cytotoxicity in Huh7 cells was determined simultaneously with anti-HCV activity by extraction and amplification of both HCV RNA and ribosomal RNA (rRNA).¹⁴ Cytotoxicity was determined in peripheral blood mononuclear (PBM), CEM (a human-T-cell-derived cell line), and Vero (kidney epithelial cells from the African green monkey) cells.¹⁵ The results are summarized in Table 1.

First, it is noteworthy that compound 5a which possess only one phenyl ring as a linker showed a complete loss of cytotoxicity, but also significantly less potency against HCV replication with a EC₅₀ >1 nM. This suggested that the length of the linker could have a marked effect on both anti-HCV activity as well as cytotoxicity. Furthermore, increase of the linker’s size to three straight aromatic rings (compounds 11a and 11b) did not severely impact the anti-HCV activity, while for compound 11a, no cytotoxicity was observed up to 100 μM in PBM, CEM and Vero cells. It is interesting that 13 which has the pyridine ring shifted to the outside aryl ring of the core has similar cyctotoxicity but 10-fold less anti-HCV activity when compared to its similar pyridyl analog 11b. However, the four straight aromatic ring (11e) seems to be problematic with a 37-fold loss of activity at the EC₅₀ level when compared to the biphenyl derivative BMS-790052.

Introduction of torsion angle by incorporation of an oxygen atom (5b), a sulfonamide (5c), a 1,3-substituted phenyl (11c), a 1,3-triazole (17) or a thiophene ring (11d) between the two phenyl ring of the linker led to a substantial decrease of activity (16- to 15,000-fold). With a 2-phenylthiophene core (5d) a torsion angle was also introduced and accompanied by an almost complete loss of anti-HCV activity. Compound 18 which contains five aromatic rings in its core was less potent but gave a similar profile to its shorter triazole analog 17 but with significantly reduced cytotoxicity.

From this SAR study we can conclude that: (i) the length of the straight central core linkage is optimal at 2–3 phenyl or pyridyl rings (11a, 11b and BMS-790052), shorter (5a) or longer (11e) than this range caused markedly reduced antiviral activity; (ii) for the three aryl core linkage, pyridyl is tolerated as the middle aryl ring (11b) while a ten fold loss of anti-HCV activity is observed with pyridyl at the outer aryl position (13); (iii) straight linked compounds (11a, 11b and BMS-790052) have a superior anti-HCV effect relative to the non-linear linked compounds (5b–d, 11c, 11d, 17 and 18); and (iv) for the three aryl core linkage, the linkage composed completely of phenyl rings rather than containing a pyridyl ring results in an improved cytotoxicity profile (11a vs 11b and 13).

In summary, 12 HCV inhibitors with different lengths and angles at the central aromatic linkage were synthesized and evaluated for anti-HCV activity and cytotoxicity in cell based assays. This study revealed that the straight tricyclic aromatic linked compounds 11a and 11b retain potent anti-HCV activity in the low picomolar range and, in the case of compound 11a, have significantly less cytotoxicity in vitro versus BMS-790052, with a therapeutic index CC₅₀/EC₅₀ over four million. Thus, this compound warrants further preclinical investigation as a potential anti-HCV agent.