Discovery of Novel N-Substituted Prolinamido Indazoles as Potent Rho Kinase Inhibitors and Vasorelaxation Agents

Inhibitors of Rho kinase (ROCK) have potential therapeutic applicability in a wide range of diseases, such as hypertension, stroke, asthma and glaucoma. In a previous article, we described the lead discovery of DL0805, a new ROCK I inhibitor, showing potent inhibitory activity (IC50 6.7 μM). Herein, we present the lead optimization of compound DL0805, resulting in the discovery of 24- and 39-fold more-active analogues 4a (IC50 0.27 μM) and 4b (IC50 0.17 μM), among other active analogues. Moreover, ex-vivo studies demonstrated that 4a and 4b exhibited comparable vasorelaxant activity to the approved drug fasudil in rat aortic rings. The research of a preliminary structure–activity relationship (SAR) indicated that the target compounds containing a β-proline moiety have improved activity against ROCK I relative to analogues bearing an α-proline moiety, and among the series of the derivatives with a β-proline-derived indazole scaffold, the inhibitory activity of the target compounds with a benzyl substituent is superior to those with a benzoyl substituent.


Introduction
Arterial hypertension is a long-term medical condition in which the blood pressure in the arteries is persistently elevated. Hypertension already affects 1.1 billion people worldwide, leading to coronary artery disease, stroke, heart failure, peripheral vascular disease, vision loss and chronic kidney disease. Despite the development of approaches for the prevention and control of raised blood pressure, there remains a significant need for the discovery of novel antihypertensive agents.
Rho kinase (Rho-associated coiled-coil-containing protein kinase (ROCK)) belongs to a family of serine/threonine protein kinases and acts as a major downstream effector of Rho A [1]. It plays an important role in regulating a variety of cellular functions such as smooth-muscle-cell contraction, cell adhesion, migration and proliferation [2]. ROCK contains two isoforms, ROCK I and ROCK II, which share approximately 60% overall amino acid sequence identity and 92% identity within the kinase domain [3]. However, the expression of the two isoforms has different tissue specificity. kinase domain [3]. However, the expression of the two isoforms has different tissue specificity. ROCK II shows high expression in brain tissue, while ROCK I is preferentially expressed in the tissues of the pancreas, liver, lung and heart [3]. Accumulating evidence shows that the dysregulation of ROCK activity or expression is associated with various diseases, including hypertension [4], stroke [5], asthma [6,7], cancer [8,9] and glaucoma [10][11][12], which indicates that ROCK is a potential novel target for drug development.
Over the past two decades, ROCK has aroused extensive interest, and numerous ROCK inhibitors have been developed from a variety of distinct scaffolds including isoquinoline [13], quinazoline [14], indazole [15][16][17][18][19][20], benzimidazole [21,22], benzothiazole [23], quinazolinone [24,25], diaminopyrazine [26], benzamide [27], chroman-3-amide [28] and urea [29]. Usually, ROCK II inhibitors were widely applied in the central nervous system (CNS) to cure stroke, Alzheimer's disease and other diseases, and ROCK I inhibitors showed therapeutic potential for the treatment of hypertension [30]. To date, only two ROCK inhibitors, fasudil (approved in 1995 for the treatment of cerebral vasopasm in Japan) and ripasudil (approved in 2014 for treatment of glaucoma in Japan) have been approved for clinical use. The narrow therapeutic window restricts the application of the existing ROCK inhibitors, leading to the development of novel ROCK inhibitors. Our group has been focused on the discovery of novel Rho kinase inhibitors for the treatment of hypertension.
5-Nitro-1H-indazole-3-carbonitrile (DL0805, 1) ( Figure 1) is a new ROCK inhibitor with an IC50 value of 6.7 μM against ROCK I, which was discovered by high-throughput screening (HTS) [31]. Previous literature had disclosed the development of ROCK I inhibitors with an indazole scaffold attached to a rigid aromatic heterocycle linking structure (I, II, III) ( Figure 1) [32][33][34]. These studies highlighted the importance of the indazole for ROCK I activity. Herein, we report the design, synthesis and the structure-activity relationship (SAR) research of a series of novel N-substituted prolinamido indazoles as ROCK I inhibitors, and their vasorelaxant activity evaluation based on the DL0805 template.

Molecular Design
In order to discover potent ROCK I inhibitors, we further optimized the structure of the DL0805 template. The attractive indazole core of DL0805 was preserved, owing to the frequent presence of this scaffold in ROCK I inhibitors. Considering the potential genotoxic hazard of the NO2 group, we conceived of the novel N-substituted prolinamido indazole ROCK I inhibitors, wherein the NO2 group can be replaced with a flexible N-substituted prolinamido group instead of the substituted

Molecular Design
In order to discover potent ROCK I inhibitors, we further optimized the structure of the DL0805 template. The attractive indazole core of DL0805 was preserved, owing to the frequent presence of this scaffold in ROCK I inhibitors. Considering the potential genotoxic hazard of the NO 2 group, we conceived of the novel N-substituted prolinamido indazole ROCK I inhibitors, wherein the NO 2 group can be replaced with a flexible N-substituted prolinamido group instead of the substituted rigid heteroaromatic ring described in the literature (e.g., I, II, III) ( Figure 1). Four related series of N-substituted prolinamido indazoles were designed and are listed in Figure 2

Chemistry
A concise route was developed to synthesize compounds 2a-2f and 3a-3c using racemic or D/L-proline as the starting material (Scheme 1). Alkylation of proline provided the intermediates 6a-6f. Initially, amidation of 6a with 5-aminoindazole employing EDCI in CH2Cl2 at room temperature gave the target compound 2a only in low yield because of 6a's solubility. When we changed to use DMF as reaction solvent and raised the reaction temperature to 80 °C, target compound 2a was obtained in moderate yield. Thus, this reaction condition was used to synthesize the target compounds 2b-2f. Analogues 3a-3c were similarly prepared by coupling the intermediates 6a-6c with 6-aminoindazole.
Next, we embarked on the synthesis of analogue 4 following a similar synthetic route used for preparing compounds 2 and 3. Unfortunately, an inseparable mixture was observed during the direct alkylation of β-proline. Therefore, new synthetic strategies to access analogues 4 and 5 were performed and described in Scheme 2. The target compounds 4a-4i were synthesized using racemic or D/L-β-proline as the starting material. Esterification of β-proline with SOCl2 and ethyl alcohol gave the ester 7. Subsequently, alkylation of ester 7 with substituted benzyl bromides provided compounds 8a-8i, which were hydrolyzed and then reacted with 5-aminoindazole to produce compounds 4a-4i. Intermediate 7 was also used to prepare the target compounds 5a-5f in an analogous fashion. Acylation of ester 7 with substituted benzoyl chloride afforded compounds 10a-10f that were saponified to 11a-11f and coupled to 5-aminoindazole to furnish the other set of target compounds 5a-5f.

Chemistry
A concise route was developed to synthesize compounds 2a-2f and 3a-3c using racemic or D/L-proline as the starting material (Scheme 1). Alkylation of proline provided the intermediates 6a-6f. Initially, amidation of 6a with 5-aminoindazole employing EDCI in CH 2 Cl 2 at room temperature gave the target compound 2a only in low yield because of 6a's solubility. When we changed to use DMF as reaction solvent and raised the reaction temperature to 80 • C, target compound 2a was obtained in moderate yield. Thus, this reaction condition was used to synthesize the target compounds 2b-2f. Analogues 3a-3c were similarly prepared by coupling the intermediates 6a-6c with 6-aminoindazole.
Next, we embarked on the synthesis of analogue 4 following a similar synthetic route used for preparing compounds 2 and 3. Unfortunately, an inseparable mixture was observed during the direct alkylation of β-proline. Therefore, new synthetic strategies to access analogues 4 and 5 were performed and described in Scheme 2. The target compounds 4a-4i were synthesized using racemic or D/L-β-proline as the starting material. Esterification of β-proline with SOCl 2 and ethyl alcohol gave the ester 7. Subsequently, alkylation of ester 7 with substituted benzyl bromides provided compounds 8a-8i, which were hydrolyzed and then reacted with 5-aminoindazole to produce compounds 4a-4i. Intermediate 7 was also used to prepare the target compounds 5a-5f in an analogous fashion. Acylation of ester 7 with substituted benzoyl chloride afforded compounds 10a-10f that were saponified to 11a-11f and coupled to 5-aminoindazole to furnish the other set of target compounds 5a-5f.

ROCK I Inhibitory Activity Evaluation
The twenty-four target compounds were initially evaluated for their percentage inhibition against ROCK I with a ROCK I assay kit (CY-1160, Cyclex, Nagoya, Japan) at 20 μM ( Table 1). The nine compounds with significant inhibition against ROCK I were further evaluated in full concentration-response plots to determine their IC50 values ( Table 2). The data illustrate some clear SAR trends. Firstly, the target compounds of series III (4) and IV (5) containing a β-proline moiety

ROCK I Inhibitory Activity Evaluation
The twenty-four target compounds were initially evaluated for their percentage inhibition against ROCK I with a ROCK I assay kit (CY-1160, Cyclex, Nagoya, Japan) at 20 μM ( Table 1). The nine compounds with significant inhibition against ROCK I were further evaluated in full concentration-response plots to determine their IC50 values ( Table 2). The data illustrate some clear SAR trends. Firstly, the target compounds of series III (4) and IV (5) containing a β-proline moiety

ROCK I Inhibitory Activity Evaluation
The twenty-four target compounds were initially evaluated for their percentage inhibition against ROCK I with a ROCK I assay kit (CY-1160, Cyclex, Nagoya, Japan) at 20 µM (Table 1). The nine compounds with significant inhibition against ROCK I were further evaluated in full concentration-response plots to determine their IC 50 values ( Table 2). The data illustrate some clear SAR trends. Firstly, the target compounds of series III (4) and IV (5) containing a β-proline moiety have improved activity against ROCK I relative to analogue series I (2) and II (3) bearing an α-proline moiety. This implied that a near-linear molecule probably had better combination with ROCK I than an "angular"-shaped molecule. Secondly, among the series of the derivatives with a β-proline-derived indazole scaffold, the inhibitory activity of the target compounds (series III) with a benzyl substituent (4) was superior to those with a benzoyl substituent (series IV 5). This suggested that the free rotation of a single bond was probably beneficial to the ROCK I inhibitory activity. Thirdly, the data of series III compounds showed that the substituent group had some influence on the activity (CH 3 > H > Br > OCH 3 > F > NO 2 , CN). This implied that the activity was probably affected by multiple factors, and the volume factor may be important for the activity. In addition, the data also revealed the (S)-enantiomer had a higher activity than the (R)-enantiomer (4i versus 4h, respectively) with the (1H-indazole-5-yl)-pyrrolidine-3-carboxamide scaffold. In contrast, the (R)-and (S)-enantiomers had a similar activity (2b versus 2c; 3b versus 3c) in the (1H-indazole-5-yl)-pyrrolidine-2-carboxamide and the (1H-indazole-6-yl)-pyrrolidine-3-carboxamide scaffolds.
0.51 b 5b 11.5 b a These data were means of multiple experiments (N = 4) with errors within 30% of the mean; b these data were obtained by single determinations.

Vasorelaxant Activity Evaluation
Abnormalities in the Rho/ROCK signaling pathway are associated with various cardiovascular diseases, especially hypertension. The inhibition of the Rho/ROCK pathway can cause the vessels to relax [4]. The norepinephrine (NE)-or potassium chloride (KCl)-induced model of the rat aortic ring can trigger sustained vessel contraction and is usually used to evaluate vasorelaxant activity. DL0805 has shown vasorelaxant activity [31]. Therefore, the nine compounds with significant inhibition against ROCK I were further tested for their vasorelaxant activity in rat aortic rings in both the high-potassium and NE models [35] (Table 3). Compounds 4a, 4b and 4c showed low micromolar EC 50 values in both vasorelaxant assays. The potent active compound 4b with excellent activity has been further evaluated [35], and further pharmacokinetic and safety evaluations are in progress.

Molecular Docking Studies
To identify the possible binding modes of our inhibitors, molecular docking of compound 4a was performed to elucidate key interactions within the active site of ROCK I. As shown in Figure 3, docking of compound 4a into the binding site of Rho kinase indicates two key hydrogen-bond interactions between the N and NH in the indazole ring and Met 156, respectively. The amide NH is predicted to form a hydrogen bond with Ala 215. Furthermore, a pi-cation interaction between the terminal phenyl ring and Lys 105 was also observed. In addition, molecular docking of compound 2a was also performed. In contrast to compound 4a, two key hydrogen-bond interactions between the N and NH in the indazole ring and Met 156 were still maintained. However, the hydrogen-bond interaction between the amide NH with Ala 215 was not observed. The result is in agreement with the activity result, that the activities of series III and IV were superior to those of series I and II.  19 .65 a --No activity; a these data were means of multiple experiments (N = 6) with errors within 30% of the mean; b these data were obtained by single determinations.

Molecular Docking Studies
To identify the possible binding modes of our inhibitors, molecular docking of compound 4a was performed to elucidate key interactions within the active site of ROCK I. As shown in Figure 3, docking of compound 4a into the binding site of Rho kinase indicates two key hydrogen-bond interactions between the N and NH in the indazole ring and Met 156, respectively. The amide NH is predicted to form a hydrogen bond with Ala 215. Furthermore, a pi-cation interaction between the terminal phenyl ring and Lys 105 was also observed. In addition, molecular docking of compound 2a was also performed. In contrast to compound 4a, two key hydrogen-bond interactions between the N and NH in the indazole ring and Met 156 were still maintained. However, the hydrogen-bond interaction between the amide NH with Ala 215 was not observed. The result is in agreement with the activity result, that the activities of series III and IV were superior to those of series I and II.

Materials and Methods
All melting points were obtained on a Yanaco melting-point apparatus and were uncorrected. ESI mass spectra were performed on the Thermo Exactive Plus LC-MS spectrometer. Nuclear magnetic resonance (NMR) spectra were recorded on a Bruker 300 or 400 MHz NMR spectrometer with TMS as an internal standard. Chemical shifts (δ values) and coupling constants (J values) were given in ppm and Hz, respectively. Column chromatography was performed with silica gel (160-200 mesh, Qingdao Haiyang Chemical, Qingdao, China). Unless otherwise noted, reagents and solvents were purchased from Acros Chemical Co, (Geel, Belgium) or other commercial providers and used without further purification.

General Procedure for the Preparation of Compounds 6
To a solution of proline (2.0 g, 17 mmol) and KOH (2.85 g, 51 mmol) in i-PrOH (50 mL), 4-substituted benzyl derivative (21 mmol) was added at 40 °C. Afterwards, the mixture was stirred for 8 h at 40 °C and then the reaction mixture was cooled to room temperature. 6 M HCl was added to adjust the pH value of the mixture to 4-5 and then CHCl3 (50 mL) was added. The mixture was stirred for 12 h, followed by filtration and evaporation in vacuo. The residue was purified by recrystallization in acetone at 0 °C to give 6 (about 90% yield).

Materials and Methods
All melting points were obtained on a Yanaco melting-point apparatus and were uncorrected. ESI mass spectra were performed on the Thermo Exactive Plus LC-MS spectrometer. Nuclear magnetic resonance (NMR) spectra were recorded on a Bruker 300 or 400 MHz NMR spectrometer with TMS as an internal standard. Chemical shifts (δ values) and coupling constants (J values) were given in ppm and Hz, respectively. Column chromatography was performed with silica gel (160-200 mesh, Qingdao Haiyang Chemical, Qingdao, China). Unless otherwise noted, reagents and solvents were purchased from Acros Chemical Co, (Geel, Belgium) or other commercial providers and used without further purification.

General Procedure for the Preparation of Compounds 6
To a solution of proline (2.0 g, 17 mmol) and KOH (2.85 g, 51 mmol) in i-PrOH (50 mL), 4-substituted benzyl derivative (21 mmol) was added at 40 • C. Afterwards, the mixture was stirred for 8 h at 40 • C and then the reaction mixture was cooled to room temperature. 6 M HCl was added to adjust the pH value of the mixture to 4-5 and then CHCl 3 (50 mL) was added. The mixture was stirred for 12 h, followed by filtration and evaporation in vacuo. The residue was purified by recrystallization in acetone at 0 • C to give 6 (about 90% yield).

General Procedure for the Preparation of Compounds 7
To a solution of ethanol (150 mL), SOCl 2 (10.34 g, 86.86 mmol) was added at 0 • C. Afterwards, β-proline (5.0 g, 43.43 mmol) was added in several separated portions at 0 • C. The mixture was stirred for 10 h at 40 • C and then evaporated in vacuo. The residue was used immediately without further purification.

General Procedure for the Preparation of Compounds 8 or 10
To a solution of 7 (400 mg) in dry CH 2 Cl 2 (8 mL), Et 3 N (849 mg, 8.39 mmol) and 4-substituted benzyl or benzoyl derivative (3.36 mmol) were added at room temperature. Afterwards, the mixture was refluxed for 4 h and then evaporated in vacuo. The residue was purified by column chromatography on silica gel (10-20% EtOAc in CH 2 Cl 2 ) to give 8 or 10 (about 80% yield).

General Procedure for the Preparation of Compounds 9 or 11
To a solution of 8 or 10 (0.775 mmol) in EtOH (3 mL), 4 M NaOH (0.7 mL, 2.8 mmol) was added at room temperature. Afterwards, the mixture was stirred for 1 h and then evaporated in vacuo. After being diluted with water and extracted with EtOAc, the pH of the aqueous phase was adjusted to 1-2 with 6 M HCl. The mixture was extracted with n-BuOH and then the organic phase was washed with saturated NaCl. The organic phase was dried over Na 2 SO 4 and evaporated in vacuo. The residue was used immediately without further purification. Compounds 2a-2f; 3a-3c; 4a-4i; 5a-5f To a solution of intermediate 9, 11 or 6 (0.39 mmol) and 5-or 6-aminoindazole (0.47 mmol) in dry DMF (3 mL), EDCI (90 mg, 0.47 mmol) was added at r.t. Afterwards, the mixture was stirred for 7 h at 80 • C and then the reaction mixture was cooled to room temperature. The mixture was evaporated in vacuo and the residue was purified by column chromatography on silica gel (2%-10% CH 3 OH in CH 2 Cl 2 ) to give the target compounds.   bubbled with 95% O 2 /5% CO 2 at 37 • C. Rings were allowed to equilibrate for 60 min at a resting tension of 1.2 g with changes of buffer every 20 min.

General Procedure for the Preparation of Target
After the equilibration period, the aortic rings were constricted with a high K + (60 mM) K-H solution to stimulate the tissue and to test its availability. Then, the rings were washed with normal K-H buffer to restore the basic tension of 1.2 g. The aortic rings were then stimulated with a high K + (60 mmol/L) K-H buffer ornorepinephrine (1 µM) to evaluate the vasorelaxant effects of the compounds. When the contraction reached the platform, compounds were added to the bath in batches at 5-min intervals to reach the accumulative concentrations of 1, 2, 5, 10, 25 and 50 µM. The tension of vessels was recorded and nonlinear regression was used to calculate the EC 50 values.

Molecular Docking
Molecular docking was performed with Discovery Studio 2016 software package (version 2016, BIOVIA, San Diego: Dassault Systèmes, CA, USA, 2016). The crystal structure of ROCK I, with co-crystal ligand obtained from the Protein Data Bank (PDB code: 3NDM) [36], was used to simulate the binding mode between our compounds and ROCK I protein. The original water molecules were removed from the coordinates set. The co-crystal ligand was used to determine the binding site and was then removed prior to docking. The docking calculation was carried out by using the LibDock protocol. Smart Minimizer algorithm was used to minimize docked poses with CHARMm force field. The default parameter settings were used. The obtained docking poses were ranked by LibDock score and the bestscored pose was chosen.

Conclusions
In summary, a series of potent ROCK I inhibitors based on a hybrid prolinamido indazole scaffold were designed and synthesized. Six compounds 4a, 4b, 4c, 4d, 4g and 4h showed activity profiles superior to DL0805, validating our design strategy. These results have established preliminary SAR trends, and molecular docking studies on the active compound 4a provide a foundation for future molecular optimization of this promising scaffold.