Alkylsulfone-containing trisubstituted cyclohexanes as potent and bioavailable chemokine receptor 2 (CCR2) antagonists

Abstract

We describe novel alkylsulfones as potent CCR2 antagonists with reduced hERG channel activity and improved pharmacokinetics over our previously described antagonists. Several of these new alkylsulfones have a profile that includes functional antagonism of CCR2, in vitro microsomal stability, and oral bioavailability. With this improved profile, we demonstrate that two of these antagonists, 2 and 12, are orally efficacious in an animal model of inflammatory recruitment. CCR2

Monocyte chemoattractant protein-1 (MCP-1 or CCL2) is a CC chemokine overexpressed in many autoimmune and inflammatory conditions.¹ Its native receptor is CC chemokine receptor 2 (CCR2), which is a G protein-coupled receptor.² A primary function of this pair (MCP-1/CCR2) is the activation and migration of inflammatory cells to areas of inflammation. MCP-1 and CCR2 have been implicated in several diseases, including rheumatoid arthritis,³ atherosclerosis,⁴ multiple sclerosis⁵ and insulin resistance.⁶ This has resulted in a large effort focused on the design and synthesis of CCR2 antagonists.⁷ In this communication, we explore structural changes to a series of sulfone-containing CCR2 antagonists with the goal of reducing hERG channel activity and obtaining orally bioavailable compounds.

Recent reports from this laboratory have described the design and synthesis of cyclohexane-based CCR2 antagonists.⁸ The major focus of these early studies was to explore and define the SAR of this novel cyclohexyl template, so as to achieve maximum binding affinity and functional antagonism of CCR2. Unfortunately, the majority of our high affinity CCR2 antagonists suffered from hERG channel inhibition: a well known liability within the chemokine antagonist field.⁹ In an effort to moderate this hERG liability, we explored structural changes that would lower our overall lipophilicity. As shown in Table 1, when the starting phenyl sulfone 1 was modified to the methyl sulfone 2, the hERG inhibition was eliminated as observed via a hERG FLIPR assay. However, as is often the case in Log P lowering one can sacrifice binding affinity on the target, as was observed here for the methyl sulfone 2, which lost almost 2-fold in CCR2 binding affinity¹⁰ as compared to 1. Even though this transformation did eliminate the hERG activity, other channel binding issues were known to exist with our previous antagonists, and therefore, we employed a sodium channel binding assay¹¹ to help monitor this issue. In agreement with the hERG assessment, methylsulfone 2 also showed very little sodium channel binding (3% @ 10 µM). Continuing with the modifications, we added a nitrogen to the trifluoromethylbenzamide to give 3, which essentially retained the CCR2 binding affinty and channel profile of 2. Compounds 4 and 5 had a tert-butyl group substituted on the benzamide (instead of a trifluoromethyl), and although they had improved CCR2 binding affinity versus 2 and 3, respectively, compounds 4 and 5 did display an increase in sodium channel binding (52% and 46% @ 10 µM, respectively). For other substitutions of the benzamide, both 6 (4-methyl-3-trifluoromethyl) and 7 (3-trifluoromethoxy) had a deleterious effect on CCR2 binding. From here, we turned to the ethylsulfone, but 8 showed an increase in sodium channel activity, and 3-phenylbenzamide 9 revealed more hERG channel activity. The iso-propylsulfone 10 increased the CCR2 binding affinity 6-fold as compared to 2 without displaying hERG inhibition. However, the combination of 3-tert-butylbenzamide and iso-propylsulfone to give 11 increased CCR2 binding but also increased hERG channel binding. tert-Butylsulfones also proved to be compatible with CCR2 as 12, 13 and 14 all showed excellent CCR2 binding affinity with no hERG channel activity and only moderate sodium channel binding. From this data set, an increase in CCR2 binding affinity trended with the alkylsulfone group in this order: Me < Et < t-Bu < i-Pr.

Table 1.

Evaluation of alkylsulfone derivatives

Compd #	R	R1	R2	X	CCR2 bindinga IC50 (nM)	hERGb IC50 (µM)	Na+ bindingc %Inh @10 µM
1	Ph	CF3	—	C	1.0 ± 0.2 (7)	30	NT
2	Me	CF3	—	C	1.8 ± 0.8 (12)	>80	3
3	Me	CF3	—	N	2.2 ± 0.1 (2)	>80	0
4	Me	t-Bu	—	C	1.2 (1)	>80	52
5	Me	t-Bu	—	N	0.7 (1)	>80	46
6	Me	3-CF3	4-CH3	C	5.2 (1)	>80	36
7	Me	OCF3	—	C	5.0 (1)	>80	32
8	Et	CF3	—	C	1.3 ± 0.02 (2)	>80	26
9	Et	Ph	—	C	1.1 (1)	60	NT
10	i-Pr	CF3	—	C	0.3 (1)	>80	NT
11	i-Pr	t-Bu	—	C	0.5 ± 0.2 (2)	77	NT
12	t-Bu	CF3	—	C	0.96 ± 0.26 (19)	>80	48
13	t-Bu	OCF3	—	C	1.2 (1)	>80	61
14	t-Bu	3-CF3	5-F	C	1.3 (1)	>80	48

^aIC₅₀ values (n) are displayed as mean ± SD (n = 2) and mean ± SEM (n > 2).

^bhERG FLIPR assay (n = 1).

^cRat Na⁺ channel binding assay (n = 1). NT = not tested.

As shown in Table 2, four alkylsulfones (2, 8, 10, and 12) were selected for chemotaxis, in vitro microsomal incubation, Caco-2, hERG patch-clamp¹² and sodium patch-clamp¹³ evaluation. All four alkylsulfones displayed excellent chemotaxis values, hence confirming their ability to operate as potent functional antagonists. Another benefit of the alkylsulfone was observed in the in vitro microsomal stability assay, as all four alkylsulfones were extremely stable as compared to phenylsulfone 1. In addition, although permeability was universally poor, as measured by Caco-2, the tert-butylsulfone 12 did show a measurable value. Three of these compounds (2, 8, and 12) were also taken into hERG and sodium patch-clamp assays. The patch-clamp values were in-line with the in vitro assessment and indicated a low to moderate liability.

Table 2.

Evaluation of alkylsulfone derivatives

		IC50a (nM)		Human microsomal stabilityc (% remaining)	hERGd IC50 (µM)	Caco-2 PAP-BL (nm/s)	hERG patch clamp (4 Hz) (% Inh)	Na+ patch clamp (4 Hz) (% Inh @10 µM)
#	R	CCR2 binding	Chemotaxisb
1	Ph	1.0 ± 0.2 (7)	0.5	88	30	<15e	NT	NT
2	Me	1.8 ± 0.8 (12)	2.6	100	>80	<15	4% @10 µM	22%
8	Et	1.2 ± 0.07 (4)	0.6	100	>80	<15	3.6% @3 µM	4.2%
10	i-Pr	0.34 (1)	0.2	100	>80	<15	NT	NT
12	t-Bu	0.96 ± 0.26 (19)	0.2	100	>80	30	4.2% @10 µM	19%

^aIC₅₀ values (n) are displayed as mean ± SD (n = 2) and mean ± SEM (n > 2).

^bChemotaxis in human monocytes (n = 1) with 0.1 M BSA.

^cPercent remaining after 10 min incubation in human hepatic microsomes.

^dhERG FLIPR assay (n = 1). NT = not tested.

^e<15 is limit of detection for this assay.

With promising antagonists in hand, we selected two compounds for further evaluation in four species pharmacokinetic (PK) studies. As shown in Table 3, compound 2 displayed some oral exposure across the four species with dog being the best (F% = 51). Compound 2 showed a wide disparity in clearance (iv) with high clearance recorded in mouse (CL = 70 mL/min/kg) and low clearance recorded in dog (CL = 3 mL/min/kg). As shown in Table 4, the tert-butylsulfone 12 had more consistent oral bioavailability across the four species than the methylsulfone 2. This improved bioavailability for 12 may be a reflection of improved permeability as noted in the Caco-2 value. Compound 12 also had low clearance values across three species with rat being the outlier.

Table 3.

Pharmacokinetic data for compound 2

Species	Dose (mpk) iv/po	F%a	CLiva (mL/min/kg)
Mouse	3/54	16	70
Rat	6/72	1	42
Cyno	1/14	9	14
Dog	1/14	51	3

^aValues are an average from two animals.

Table 4.

Pharmacokinetic data for compound 12

Species	Dose (mpk) iv/po	F%a	CLiva (mL/min/kg)
Mouse	5/100	13	25
Rat	4/43	14	54
Cyno	1/10	26	12
Dog	1/10	74	5

^aValues are an average from two animals.

With both 2 and 12 showing oral bioavailability, it was our desire to test these CCR2 antagonists in a mouse model of inflammatory cellular recruitment. The MCP-1/CCR2 pair plays a major role in mediating the egress of inflammatory monocytes (defined as Ly6C+F4/80+) from bone marrow to blood,¹⁴ and this can be emulated with the thioglycollate (TG)-induced peritonitis model.¹⁵ However, 2 and 12 have poor activity versus mouse CCR2, hence our TG model had to be performed in a human-CCR2 knock-in mouse (we did dose 12 in a TG-induced peritonitis model using wild-type mice, however 12 did not show any activity—data not shown). As shown in Table 5 with human-CCR2 knock-in mice, compounds 2 and 12 were orally dosed in separate experiments one hour before thioglycollate challenge, and both 2 and 12 showed a significant reduction of inflammatory monocytes in blood as compared to vehicle (similar findings were observed with these monocytes in the peritoneal cavity—data not shown). Hence, these results validate the in vivo activity of compounds 2 and 12.

Table 5.

6-h TG model in human-CCR2 knock-in mice with 2 and 12

^a% of Ly6C+F4/80+ cells vs. TG control in peripheral blood.

The synthesis of compound 12, shown in Scheme 1, is used as a representative example of these alkylsulfone antagonists. The synthesis commenced with mesylation of the homochiral alcohol 15.¹⁶ The resulting mesylate was used, without purification, in a displacement reaction to give 16, which was subsequently oxidized to sulfone 17. The carbamate of 17 was then removed prior to coupling with a methionine derivative to yield 18. The lactam was formed under our modified Freidinger¹⁷ conditions (MeI and then Cs₂CO₃ in DMF) to give 19. Final elaboration was performed by way of benzamide installation followed by tertiary amine formation to afford 12.

Scheme 1.

Reagents and conditions: (a) (i) MsCl, TEA, DCM, 0 °C, quant; (ii) NaS-t-Bu, DMF, 78%; (b) oxone, IPA, H₂O, 79%; (c) H₂, Pd/C, MeOH; (ii) BOP, NMM, N-Cbz-l-Met-OH, DMF, 95% (two steps); (d) (i) Mel; (ii) Cs₂CO₃, DMF, 76%; (e) (i) H₂, Pd/C, MeOH; (ii) BOP, NMM, 3-trifluoromethylbenzoic acid, DMF, 90% (2 steps); (f) (i) TFA, DCM, quant; (ii) acetone, NaBH(OAc)₃, DCM; (iii) 37% HCHO, NaBH(OAc)₃, DCM, 96% (two steps).

In summary, we have demonstrated that trisubstituted cyclohexanes containing alkylsulfones are potent functional antagonists of CCR2 that have an improved hERG channel profile as compared to our previously described antagonists. Two of these alkylsulfone antagonists, 2 and 12, also displayed in vitro microsomal stability and oral bioavailability. With this improved profile, we established that these CCR2 antagonists are orally efficacious in an animal model of monocyte recruitment, one of the hallmarks of autoimmune disease.