Discovery of Imidazopyridine Derivatives as Highly Potent Respiratory Syncytial Virus Fusion Inhibitors

Abstract

A series of imidazolepyridine derivatives were designed and synthesized according to the established docking studies. The imidazopyridine derivatives were found to have good potency and physical-chemical properties. Several highly potent compounds such as 8ji, 8jl, and 8jm were identified with single nanomolar activities. The most potent compound 8jm showed an IC₅₀ of 3 nM, lower microsome clearance and no CYP inhibition. The profile of 8jm appeared to be superior to BMS433771, and supported further optimization.

Human respiratory syncytial virus (RSV) is one of the most important respiratory pathogens that cause lower respiratory tract infections such as bronchiolitis and pneumonia in infants and young children, resulting in up to 125,000 hospitalizations annually in the United States.¹ The mortality rate among children admitted to hospital is approximately 3% for those with heart and lung problems and up to 1% for those without these risk factors.^2,3 In adults and the elderly, RSV pneumonia is increasingly recognized as a significant cause of morbidity and mortality, being associated with more than 17,000 deaths annually between 1991 and 1998.⁴⁻⁶

Although the research into the prevention and treatment of RSV infection has been ongoing for almost 40 years, vaccine development is difficult,^7,8 and to date, there is no clinically approved vaccine. Passive immunization with the monoclonal antibody Palivizumab (Synagiss) has provided about 50% protection to high-risk children.⁹ Ribavirin, the only approved small molecule for treatment of RSV, has to be given by a prolonged aerosol, and there are certain doubts as to its safety versus efficacy in treatment of RSV infection. As such, the unmet medical need for additional effective and safe treatments for RSV is paramount.

Fusion of the viral envelope or infected cell membranes with uninfected cell membranes is an essential step in the virus life cycle. The disruption of viral attachment and entry to cells is a common strategy in the design of antiviral therapies, as evidenced by successful approaches to influenza and HIV inhibition.¹⁰ In the case of the Pneumovirinae subfamily, which includes RSV, a distinctive feature is the absence of hemagglutinin/neuraminidase fusion proteins and the dependence on the F protein alone for virus binding and cell entry.¹¹ The associated G glycoprotein appears to facilitate viral attachment. This pivotal role for the fusion protein has inspired a number of approaches over the years and led to compounds suitable for clinical evaluation.¹²

Several chemical series have been reported as RSV fusion inhibitors and two apparently progressed into early clinical development. One leading series is a class of benzimidazole pyridines.¹³⁻¹⁶ The lead JNJ-2408068 (1) demonstrated nanomolar potency in RSV in vitro assays. TMC-353121 (2) was reported to access an additional binding pocket in the binding space¹³⁻¹⁶ and displayed subnanomolar RSV in vitro activity with an unusual lung tissue residence half-life of 25 h. Another example of tricyclic imidazolines 3 was reported to inhibit RSV in cell culture at 100–250 ng/mL.^15,16 Recently, GS-5806 (4) was described as a novel RSV fusion inhibitor achieved proof-of-concept in human RSV challenge studies.¹⁷ In addition, VP-14637 (6) was in Phase I trials prior to a decision not to develop it further, in part due to development costs.¹⁸ RFI-641(7) was in Phase II clinical trials in 2000–2001 for the secondary prevention and therapeutic treatment of RSV infections in adults demonstrating both large therapeutic window in vitro and safety in vivo. Nevertheless, no further development information has been reported recently.^19,20

BMS-433771 (5) is an orally bioavailable and modestly potent RSV inhibitor showing an EC₅₀ of 24 nM over a range of both laboratory and clinically relevant RSV strains.²¹ BMS-433771 was progressed into preclinical evaluation.²² Although the above advantages endorsed BMS-433771 for preclinical development, some potential concerns caught our attention according to our internal assessment, including modest activity, high in vivo clearance, short T_1/2 duration time (36 min), etc. Recognizing the areas for improvement, our efforts were focused on the optimization of potency and PK properties.

BMS-433771 was speculated to bind in the hydrophobic pocket created by the trimerization of the N-terminal heptad repeats of the RSV F1 protein.^23,24 According to the reported binding mode and the internal docking results (Figure 1), there remains some space around the BMS-433771 benzimidazole portion in the hydrophobic pocket close to LYS191 and Val192, as well as the side chain part, which is exposed to the surface of the pocket around Asp194 and Leu195. This observation prompted us to explore both benzimidazole region and side chain in order to improve potency and physicochemical properties.

Figure 1.

RSV fusion inhibitors under development.

Figure 2.

Docking model of RSV fusion inhibition with BMS-433771 (right)¹⁶ and 8h (left).

Several bioisosteres of the benzimidazole were examined to identify the optimal replacement (such as pyrazole, indazole, or isoxazole). Only imidazolepyridine analogue 8a exhibited good potency presumably due to its nice fitting the binding pocket. Inspired by the promising results, some hydrophobic substitutions including methyl, fluoro, trifluor, etc., were introduced to the R1 and R2 positions of imidazole[1,2-a]pyridine core. The general synthesis of these derivatives 8a–i is summarized in Scheme 1.

Scheme 1. Synthesis of 8a–i Imidazopyridine Derivatives.

Reagents and conditions: (a) ethyl bromopyruvate, ethanol; (b) LiAlH₄, THF, 0 °C; (c) (i) NBS, CH₃CN, reflux; (ii) SOCl₂, DCM; (d) 1-cyclopropyl-1,3-dihydroimidazo[4,5-c] pyridin-2-one(13), DMF, NaH; (e) 4-(TBDMSO)-butyne, CuI, Pd(PPh₃)₂Cl₂, Et₃N, CH₃CN, microwave, 100 °C; (f) TBAF/THF 0 °C to rt; (g) Pd/C, H₂, 3 h, 70–85%

The synthesis of imidazole[1,2-a]pyridine analogue 8 commenced with commercially available 2-amino-pyridine (9), which was cyclized with ethyl bromopyruvate via heating in ethanol to generate the imidazole[1,2-a]pyridine-2-ethyl ester 10, followed by reduction with LAH, bromination, chlorination, and then alkylation with the right part: 1-cyclopropyl-1,3-dihydroimidazo[4,5-c]pyridin-2-one(13)²³ to yield intermediate 3-bromo-imidazole[1,2-a]pyridine 14. This intermediate was then subjected to a Sonogashira coupling with 4-(t-butyldimethylsiloxy) butyne under microwave conditions. After removal of TBDMS, the intermediate 16 was quickly reduced with palladium on activated carbon to afford imidazole[1,2-a]pyridine derivatives 8a–i in good yield (70–80%).

The antiviral activity of these compounds was evaluated by the reduction of the cytopathic effect (CPE) induced by the long (A) strain of virus replicating in HEp-2 human lung epithelial carcinoma cells⁴ (Table 1). The introduction of fluoro at 6 position of 8b resulted in the reduction of activity about 4-fold with an EC₅₀ of 96 nM in comparison to BMS-433771. It was found that the substitutions at 6-position (R1) of imidazolepyridine were very rigid and could not tolerate imidazolepyridine derivatives. Some of the results are presented below. The 6- methyl and chloro analogues (8a and 8c) were much less potent with an EC₅₀ of 0.217 and 0.422 μM, respectively. Replacing the chloro with trifluoro substitution at 6-position (R1) resulted in a nearly inactive analogue 8d with an EC₅₀ of 2.89 μM. Surprisingly, the substitutions at the 7-position (R2) can tolerate more broad modification and exhibited a trend toward potency improvement. Lastly, unprecedented single digital nanomolar potency was obtained when methyl (8e), ethyl (8f) or chloro (8j) were introduced as R2 of 8, these compounds demonstrated 3–5 folds activity improvement in comparison to 5 with an EC₅₀ below 10 nM and CC₅₀ greater than 100 μM. The introduction of 7-substitution chloro or methyl had a significant impact on retaining potency. These promising results encouraged us to further explore the appropriate substitution pattern as shown in Table 1.

Table 1. Antiviral Activity of Imidazole[1,2-a]pyridine Analogues 8a–j.

compds	R1	R2	EC50 (μM)a A strain	CC50 (μM)a
5			0.028	>100
8a	Me	H	0.217	>100
8b	F	H	0.096	>100
8c	Cl	H	0.422	>100
8d	CF3	H	2.89	>100
8e	H	Me	0.006	>100
8f	H	Et	0.007	>100
8g	H	F	0.023	>100
8h	H	H	0.017	>100
8i	H	CF3	0.217	>100
8j	H	Cl	0.007	>100

^aEC₅₀: the concentration of test compound that protects 50% of infected cells. CC₅₀: the concentration of drug that manifests cytotoxicity toward 50% of uninfected HEp-2 cells in the absence of virus. Values are means of two or more experiments performed on consecutive weeks.²⁵

According to the new insight of the distinct impact of 7-substitution on potency improvement, a number of 7-chloroimidazopyridine analogues 8ja–r bearing various side chain were synthesized as shown in Table 2 to identify more potent compounds with favorite MDO properties. The synthesis of 8jc–jf was described in Scheme 2 below, whereas the preparation of 8jg–jm followed the coupling and reduction steps in Scheme 1 and condensation step in Scheme 2.

Table 2. Antiviral Activity of Imidazole[1,2-a]pyridine Analogues 8j–jr.

compds	R3	EC50 (μM)	CC50 (μM)
8j	(CH2)4OH	0.007	>100
8ja	H	3.06	22.27
8jb	CN	0.086	22.49
8jc	CH2OH	1.93	>100
8jd	COOH	25.5	>100
8je	CONH(CH2)2OH	2.23	>100
8jf	CONH(CH2)2SO2Me	0.082	>100
8jg	CH=CHCOOH	0.207	>100
8jh	(CH2)2COOH	0.053	>100
8ji	(CH2)2CONH2	0.005	>100
8jj	(CH2)2CONH(CH2)2OH	0.010	>100
8jk	(CH2)2CONHSO2Me	0.046	>100
8jl	(CH2)2SO2Me	0.006	>100
8jm	(CH2)2SO2Et	0.003	>100
8jn	NH-n-butyl	0.041	>100
8jo	NH-cyclohexane	0.072	>100
8jp	NH(CH2)2–morpholine	0.083	>100
8jq	CH2NH(CH2)2OH	0.683	>100
8jr	Ph(p-SO2Me)	0.089	>100

Scheme 2. Synthesis of 7-Substituted Imidazopyridine Derivatives.

Reagents and conditions: (a) POCl₃, DMF, 70 °C; (b) NaH, DMF, 70 °C; (c) NaClO₂, tBuOH/THF; (d) HATU, Et₃N.

It was speculated from the docking results that the R3 position might be exposed to an open surface area of the pocket, which could be flexible. Various modifications in this region may be tolerable, although there is potential hydrogen bond interaction between the terminal hydroxyl with Lys191. On the basis of this model, new analogues 8ja–r with a variety of terminal groups at 3-position (R3) were investigated.

Table 2 revealed that many 7-chloro derivatives including 8jb, 8jf, 8jh–jp, and 8jr were very potent RSV inhibitors, with EC₅₀ of below 100 nM. Interestingly, even without a side chain, a cyano analogue 8jb still exhibited prominent antiviral activity compared to 8ja, although mild cytotoxicity was observed. Hydroxylmethyl, carboxylic acid, or amide replacement at R3 led to remarkable activity loss (8jc, 8jd, and 8je). The antiviral potency of methyl sulfone analogue 8jf still remained with an IC₅₀ below 100 nM. For 8jf, acid elongation derivatives 8jg and 8jh were not very potent. In contrast, several 7-chloro derivatives 8ji, 8jj, 8jl, and 8jm containing amide and alkyl sulfonyl as terminal groups of side chain demonstrated high antiviral acidity with single digital nanomolar IC₅₀s. The most potent compound 8jm has an IC₅₀ of 3 nM, which is around 9-fold more potent than BMS-433771 (5). A few substituted amino-linked analogues like 8jn and 8jp exhibited some potency decline, and 8jq had much weak activity with an IC₅₀ of 0.68 μM. Replacing the hydroxyl butyl with more hinderous para-methyl sulfonyl phenyl at R3 position resulted in more than 10-fold activity loss (0.089 μM).

The physicochemical and metabolic properties of several potent 7-chloroimidazolepyridine analogues and BMS compound 5 were profiled below (Table 3). Compound 8jh exhibited highly enhanced solubility and metabolic stability; however, its activity and permeability were not optimal. The solubility of the amide 8ji is slightly better than BMS compound 5; 8jj is more soluble with up to 3-fold LYSA increase. The PAMPA value of 8ji is subtly increased; however, 8jj has a little lower one (Table 3). Compounds 8jl and 8jm appear equally soluble as BMS compound 5 and much more permeable. All five compounds significantly improved microsome stabilities. These significant advantages suggested superior PK properties with higher bioavailability and lower clearance could be achieved for these several leading compounds toward further development. More DMPK and in vivo data will be reported in due course.

Table 3. Physicochemical and Metabolic Properties of Leading Compounds.

compds	solubility LYSAa (μg/mL)	permeability PAMPAa (10–6 cm/s)	microsome clearance humanb/mousec (mL/min/kg)	CYP inhibition 3A4/2C9/2D6 (IC50, μM)
5	39	2.2	16.3/82	27.5/>50/>50
8j	20	3.1	11.5/80	NA
8jh	541	0.30	0.0/9.1	>50/>50/>50
8ji	87	2.5	13.5/59	NA
8jj	107	0.96	2.9/47	NA
8jl	59	3.6	2.5/35	>50/>50/>50
8jm	46	2.6	4.6/62	>50/>50/>50

^aValues are means of two or more experiments performed on consecutive weeks.

^bHuman: high (>15), medium (5–15), low (<5).

^cMouse: high (>68), medium (23–68), low (<23).

In summary, a series of imidazolepyridine derivatives were conceived and synthesized according to our docking studies. It was found 7-substitution is crucial for potency maintaining and improvement. Several highly potent compounds, 8ji, 8jl, and 8jm, were identified with single digital nanomolar activities of which the most potent compound 8jm showed an IC₅₀ of 3 nM with up to 9-fold activity improvement relative to 5, low to medium microsome clearance, and no CYP inhibition. The promising profile of 8jm supported further optimization efforts of this series, which will be reported in due course.

Glossary

Abbreviations

RSV

respiratory syncytial virus

LYSA

lyophilization solubility assay

PAMPA

partial permeability assay

MDO

multiple dimensional optimization

CYP

Cytochrome P450 enzymes

Supporting Information Available

Biological assays, synthetic procedures, and analytical data for selected compounds. This material is available free of charge via the Internet at http://pubs.acs.org.

The authors declare no competing financial interest.