Discovery of Heteroaryl Urea Isosteres for Formyl Peptide Receptor 2 Agonists

Abstract

Formyl peptide receptor 2 (FPR2) agonists have shown efficacy in inflammatory-driven animal disease models and have the potential to treat a range of diseases. Many reported synthetic agonists contain a phenylurea, which appears to be necessary for activity in the reported chemotypes. We set out to find isosteres for the phenylurea and focused our efforts on heteroaryl rings. The wide range of potencies with heterocyclic isosteres demonstrates how electronic effects of the heteroatom placement impact molecular recognition. Herein, we report our discovery of benzimidazole and aminophenyloxadiazole FPR2 agonists with low nanomolar activity.

Formyl peptide receptors (FPRs) are a family of G-protein-coupled receptor (GPCR) pattern recognition receptors with important roles initiating the innate immune response.¹ Three family members, FPR1, FPR2, and FPR3, have been identified with high sequence homology but different specificity for endogenous ligands. When exposed to bacterial-derived peptides or damage-associated peptides, the receptors recruit leukocytes to the site of infection or injury.^2,3 In addition to initiating the immune response, FPR2 has been identified for its role in resolution of inflammation. FPR2 agonists have been shown to decrease neutrophil recruitment to an active site of inflammation and differentiate macrophages to a pro-wound-healing phenotype, promoting the clearance of apoptotic cells. Chronic inflammation plays a role in many diseases, so the pro-resolution characteristics of FPR2 agonists have potential as a treatment for many of these conditions, by attenuating the associated inflammation.⁴⁻⁶ FPR2 agonists have been shown to have strong efficacy in pre-clinical animal disease models, including arthritis, sepsis, fibrosis, heart disease, and Alzheimer’s disease.⁷⁻¹⁰ Recently, FPR2 agonist 1 showed significant improvements in survival, improved cardiac function, and preservation of heart tissue in multiple rodent heart failure models.^11,12

A number of potent, synthetic FPR2 agonists have been reported in the literature, and a common structural feature of many of these molecules is a phenylurea.¹³⁻¹⁵ We used phenylurea 1 as a template to look for urea isosteres that maintain the FPR2 potency while potentially changing pharmaceutical properties and developed a pharmacophore model for FPR2 activity.

Compound activity was assessed by measuring inhibition of cAMP in cells stimulated with forskolin that over-express human FPR2 or FPR1 (details in Supporting Information). Following isostere strategies in the literature,¹⁶ many of our initial attempts to modify or replace the urea were unsuccessful, until three heteroaryl motifs were identified (Figure 1, Table 1). N-Linked benzimidazole 2 matches the hydrogen-bonding pattern of the urea but loses significant activity. The C-linked benzimidazole 3, as well as the N-linked phenyloxadiazole 4, were also moderately active. Overlays of the structures show that all three maintain a hydrogen bond acceptor in a similar region as the carbonyl in the urea but display different patterns of hydrogen bond acceptors and donors in the region of the two urea NH bonds. All three compounds displayed selectivity over FPR1. Initial analogues of compound 2 did not result in improved potency, so we report herein structure–activity relationship (SAR) studies and optimization of the C-linked benzimidazole 3 and N-linked phenyloxadiazole 4 initial leads.

Figure 1.

Heterocyclic phenylurea replacements.

Table 1. cAMP Data for Initial Heterocyclic Phenylurea Replacements.

	EC50 (nM)
compd	FPR2	FPR1
1	5	400
2	1 300	>10 000
3	680	>10 000
4	250	5 500

Compounds could be synthesized from intermediates 5 and 6 used for the synthesis of phenylurea 1 (Scheme 1).¹¹C-Linked benzimidazoles were readily synthesized from acid intermediate 5, first by coupling the appropriate benzenediamine followed by dehydration under acidic conditions to produce the desired products in good yield. The N-linked heterocycles were accessed via amine intermediate 6, which was synthesized via a Curtius rearrangement of acid 5, followed by in situ formation of the benzyl carbamate with benzyl alcohol and then reductive cleavage of the carbamate. N-Linked benzimidazole 2 was synthesized via displacement of 2-chloro-1H-benzo[d]imidazole with amine 6. Phenyloxadiazole 4 was synthesized via BOP coupling of the phenyl oxadiazolones to yield the desired products. Other N-linked heterocycles reported in this Letter were accessed from intermediate 6 using published procedures, which are shown in the Supporting Information.¹⁷

Scheme 1. Synthesis of Target Molecules.

Reagents and conditions: (a) benzene-1,2-diamine, HBTU, Et₃N, DMF, 99%; (b) acetic acid/acetonitrile, microwave, 100 °C, 49%; (c) DPPA, Et₃N, toluene then BnOH; (d) Pd/C, 1 atm H₂, 45% over two steps; (e) 2-chloro-1H-benzo[d]imidazole, DIEA, acetonitrile, microwave, 160 °C, 58%; (f) 5-phenyl-1,3,4-oxadiazol-2(3H)-one, BOP, DIEA, DMF, 29%.

In an effort to optimize C-linked benzimidazole 3, substitution on the benzimidazole was explored (Table 2). Chloro substituents in the 5-position provided the most potent benzimidazoles, as demonstrated by compound 7. Electron-donating substitutions such the methyl ether in compound 8 or substitutions at the 6-position, such as chloro in compound 9, resulted in a significant loss in cAMP activity. Introduction of a pyridyl ring increased activity, and azabenzimidazole 10 was the most active C-linked heterocycle identified. The additional nitrogen could make a specific protein contact, because regioisomer 11 resulted in a loss in potency. An alternate explanation is that the regioisomers have difference tautomeric preferences, which influence the H-bond strength of the NH. The narrow SAR in this series led us to focus more on the N-linked phenyloxadiazole lead 4.

Table 2. cAMP Data for the Benzimidazole Analogues.

Similar to the C-linked benzimidazoles, optimization of phenyloxadiazole showed that chloro substituents on the phenyl ring were optimal for activity (Table 3). Para-chloro substituted compound 12 was the most potent compound identified in this series, whereas both meta- and ortho-substituted chlorophenyloxadiazoles 13 and 14 were less active. Combining other substituents as small as fluoro with the para-chloro substitution in compound 15 also resulted in loss of activity. Replacing the para-chloro with a methyl ether resulted in a weakly active compound 16, but switching to the more electronegative trifluoromethyl ether in 17 regained potency, again demonstrating the importance of electron withdrawing groups on the phenyl ring for potency. The effect of ortho fluoro substitution on the methoxyphenyl was probed in analogues 18 and 19. Similar to previous results, 11 one ortho fluoro is important for potency. After SAR around the phenyl ring was explored resulting in compound 12 as a nanomolar agonist, we next examined the effect of changing the heterocyclic ring.

Table 3. cAMP Data for Optimization of Phenyloxadiazole 4.

Although the first cryo-EM structure of a small molecule bound to FPR2 was recently reported,¹⁸ no structures were available at the time of these studies. In the absence of a protein structure of these ligands bound to FPR2, we used small-molecule crystallography and molecular modeling to guide our understanding of FPR2 agonism. Overlay of small-molecule structures of compounds 1 and 12 shows a close overlap (Figure 2). If the 3-amino-4-phenylpyrrolidin-2-ones cores are superimposed, then the urea oxygen and the N3 of the oxadiazole align as well. However, the terminal phenyl in compound 12 is shifted relative to the phenyl in compound 1. It remained to be determined if the urea and oxadiazole were only important as spacers for placement of the phenyl rings to maintain activity or if the heteroaryl rings made important interactions with the protein. With the ed terminal chlorophenyl, a series of five-membered ring heterocycles were screened to test the effect of heteroatom placement on activity.

Figure 2.

Overlay of small-molecule crystal structures of compound 1 (green) and compound 12 (tan).

Synthesis of an array of 1,3-disubstituted five-membered heterocycles in place of the oxadiazole in compound 12 resulted in a set of compounds which were compared to the oxadiazole as urea isosteres (Table 4). When compared to 1,3,4-oxadiazole 12, the 1,2,4-oxadiazole isomers 20 and 21 had a striking loss of potency, whereas the analogous 1,3,4-thiadiazole 22 maintained potency. Modeling studies indicated there is little difference in the rotamer preferences of the pendant groups (Figure 3), so electronics likely drive the different in potency. One hypothesis is that the bridging nitrogen in the 4-position of the 1,2,4-oxadiazoles has a repulsive interaction with FPR2 that disfavors binding.

Table 4. cAMP Data Showing the Effect of Heterocyclic Urea Replacement.

Figure 3.

Overlay of compounds 20–27.¹⁹ Individual compounds were optimized in the QM software package Jaguar²⁰ using M06-2X/6-31G** in the gas phase, constraining the lactam–amino torsion to a value similar to those observed in small-molecule crystallographic structures of compounds 1 and 12.

In order to test this theory, heterocycles 23–27 with oxygen and methine at this bridging position were synthesized. Oxazole 23 loses ∼8-fold in cAMP activity, indicating that the diazole of compound 12 is important for activity. Isoxazole 24 provided an increase in activity, which supported the hypothesis that placing a methine at the corresponding O1 position of oxadiazole 12 creates a more favorable interaction with FPR2. An overlay of isoxazole 24 with molecule 1 places the isoxazole nitrogen in a region similar to that of the urea oxygen, likely making a similar interaction with the protein. Surprisingly, the activity of isoxazole 25 was 160-fold less than that of the oxadiazole 12, which is potentially due to the increased hydrogen-bonding potential of nitrogen over oxygen in these heterocyclic rings. Calculations comparing the hydrogen bond potential of the urea carbonyl, isoxozoyl nitrogen, and oxygen give their values as −6.0, −4.5, and −1.4 kcal, respectively. Pyrazole 26 and methylpyrazole 27 were significantly less active, demonstrating that hydrogen bond donors or additional steric bulk is not tolerated in this region.

With two heterocyclic urea isosteres identified, we analyzed the effect on other efficacy end points and pharmaceutical properties (Table 5). In comparison to the phenylurea 1, compounds 12 and 24 had a marginal loss of FPR2 cAMP activity and maintained similar selectivity against FPR1. The chemotaxis assay measures the inhibition of movement by HL-60 cells, a neutrophil-like lineage, and the phagocytosis assay evaluates the stimulation of macrophages to engulf particles.¹² Both compounds inhibit chemotaxis and stimulate phagocytosis at low nanomolar activity, which are in vitro measures of pro-resolution activity. The heterocyclic replacements had significantly higher protein binding than the phenylurea, which is likely due to the increased lipophilicity, as predicted by the corresponding increase in clogP values. The increase in permeability by the heterocycles translates into a moderate increase in oral exposure for both molecules.

Table 5. In Vitro Potency, Selectivity, and Pharmaceutical Properties of Compounds 12 and 24 Compared to 1.

	compd
	1	12	24
hFPR2 EC50, nM	5	32	10
hFPR1 EC50, nM	400	1 400	470
chemotaxis IC50, nM	7.5	28	19
phagocytosis EC50, nM	<1	<1	<1
protein binding (h, %free)	12.7	1.5	0.5
clogP	1.75	2.46	3.23
PAMPA, pH = 5.5, nm/s	440	970	1 100
mouse PK (AUC0-inf), nM·h, 3 mpk dose	5 600	8 400	15 000

In summary, several five-membered ring heterocyclic urea isosteres were identified in the context of FPR2 agonists with oxadiazole 12 and isoxazole 24 both being potent and displaying favorable pharmaceutical properties. Distal chloro substitution was optimal in all cases, and the arrangement of heteroatoms was important for potency. The wide range of results is best shown by the almost 200-fold difference in potency for isoxazole isomers. This study reinforces the differential electronic effects of heteroatom placement in heterocycles, as demonstrated by a dramatic potency difference between oxadiazole and isoxazole isomers when used as urea replacements.

Glossary

Abbreviations

BOP

((1H-benzo[d][1,2,3]triazol-1-yl)oxy)tris(dimethylamino)phosphonium hexafluorophosphate(V)

cAMP

cyclic adenosine monophosphate

DPPA

diphenyl phosphorylazide

FPR

formyl peptide receptor

HBTU

2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate, or hexafluorophosphate benzotriazole tetramethyl uronium

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.2c00079.

• Additional information about biological assays, synthesis procedures and small-molecule crystal structures (PDF)

• X-ray crystallographic data for 1 (CIF)

• X-ray crystallographic data for 12 (CIF)

The authors declare no competing financial interest.