Influence of a Novel Inhibitor (UM8190) of Prolylcarboxypeptidase (PRCP) on Appetite and Thrombosis

Abstract

Preclinical pharmacological characterization of a novel inhibitor (UM8190) of prolylcarboxypeptidase (PRCP) was investigated. We synthesized and evaluated a library of proline-based analogs as prospective recombinant PRCP (rPRCP) inhibitors and inhibitors of PRCP-dependent prekallikrein (PK) activation on human pulmonary artery endothelial cells (HPAEC). Among the newly synthesized compounds, UM8190 was further characterized in vivo using methods that encompassed a mouse carotid artery thrombosis model and animal model of food consumption. (S)-N-dodecyl-1-((S)-pyrrolidine-2-carbonyl) pyrrolidine-2-carboxamide [Compound 3 (UM8190)] was selected for further evaluation from the initial assessment of its PRCP inhibitory action (K_i= 43 µM) coupled with its ability to block PRCP-dependent PK activation on HPAEC (K_i= 34 µM). UM8190 demonstrated excellent selectivity against a panel of carboxypeptidases and serine proteases and blocked bradykinin (BK) generation and BK-induced permeability by 100%, suggesting that it may be useful in preventing the local production of large amounts of BK. Furthermore, UM8190 showed an anorexigenic effect when systemically administered to fasted mice, reducing food intake in a dose- and time-dependent manner. In a mouse carotid artery thrombosis model, it also demonstrated an antithrombotic effect. UM8190 is a selective PRCP inhibitor and it may represent a new anorexigenic, and antithrombotic drug, that works by inhibiting PRCP–mediated mechanisms.

INTRODUCTION

Obesity is a major risk factor for common disorders such as type 2 diabetes mellitus, cardiovascular disease, and cancer [1, 2]. α-Melanocyte-stimulating hormone (α-MSH or α-MSH_1–13), a thirteen amino acid peptide derived from proopiomelanocortin (POMC), is involved in regulation of food intake and energy homeostasis in mammals, [3] and is an inhibitor of inflammatory reactions.[4–6] α-MSH suppresses feeding behavior in mice and fish, [7] and POMC mRNA expression is increased in the pituitaries of genetically obese mice [8]. α-MSH appears to have a species-dependent regulatory function in energy homeostasis, with two tissue-dependant and opposing roles. In the CNS, it increases sensitivity to insulin, [9] while in peripheral tissues it seems to play a pivotal role in insulin resistance [9]. α-MSH is a potent agonist of the melanocortin 1 (MC1R) and melanocortin 4 (MC4R) receptors [10]. MC4R knockout mice develop a maturity onset obesity syndrome characterized by hyperphagia, hyperglycemia and hyperinsulinemia [11], and MC4R mutations have been linked to severe obesity in one study of French children [12]. MTII, a specific synthetic MC3R/MC4R agonist, inhibits food intake in rats [13]. The serine protease prolylcarboxypeptidase (PRCP) inactivates α-MSH by catalyzing cleavage at the carboxyl terminal Pro-Val bond, producing inactive α-MSH_1–12. suggests PRCP could have an appetite stimulating (orexigenic) effect [14, 15]. Therefore, PRCP inhibitors might be appetite suppressants.

PRCP also activates the plasma kallikrein-kinin system (KKS). The plasma KKS consists of high molecular weight kininogen (HK), prekallikrein (PK), and factor XII (FXII, Hageman factor). Activation of KKS is mediated through conversion of prekallikrein (PK) to α-kallikrein, resulting in release of bradykinin (BK) from HK [16]. BK is a potent proinflammatory peptide, which may induce vasodilation and vascular leak. Plasma kallikrein also activates FXII to activated factor XII (FXIIa). FXIIa plays a role in initiation of the intrinsic pathway of blood coagulation.

Thus, inhibition of PRCP-dependent PK activation could provide an anti-thrombotic effect.

We report the design, synthesize and biological and physiological characterization of a novel PRCP inhibitor, which has anti-thrombotic and appetite suppressant properties.

MATERIALS AND METHODS

Receptor Nomenclature

Nomenclature of receptors and other drug/molecular targets in this manuscript is in conformation with BJP’s Guide to Receptors and Channels [17].

General Methods

All commercially available chemicals and reagents used for synthesis were purchased from Sigma-Aldrich (St. Louis, MO) or Bachem (Torrence, CA) unless otherwise indicated. Analytical TLC plates (silica gel) were purchased from Sorbent Technologies (G Series-UV 254). Silica gel column chromatography used standard chromatography grade silica gel, 20–63 µM particle size, (Sorbent Technologies). ¹H NMR and C NMR spectra were recorded at either 400 or 500 MHz on a Bruker Avance DRX spectrometer. Elemental Analysis was obtained using a PerkinElmer 2400 Series II CHNS/O Analyzer. Low-resolution mass spectrometry analysis was performed on a Waters ZQ single-quadrupole system using either ESI positive (ESI+) or ESI negative (ESI-) electrospray ionization. High-resolution mass spectrometry (HRMS) was recorded using a Micromass Q-TOFmicro instrument. The purity of all target compounds was assessed using elemental combustion analysis (CHN) with the purity being assigned as ± 0.4% calculated values for the proposed formula, or LC-MS/ESI analysis coupled to UV-VIS diode array detector with a cutoff at 215 nm (95% peak area). All compounds used for subsequent studies had a minimum purity of ≥ 95%. Frozen human pulmonary artery endothelial cells (HPAEC), human endothelial cell growth medium (EGM), trypsin-EDTA (ethylenediaminetetraacetic acid), S2 cells, hygromycin, and primers for eNOS were purchased from Invitrogen (Carlsbad, CA). Trypsin-neutralizing solution (TNS) was purchased from Lonza (Walkersville, MD). SFX-insect serum free insect cell culture medium was purchased from Thermo Scientific (Logan, UT). Spectra/Por® Dialysis Membrane (MWCO 6–8 kDa), SP Sephadex 50–120, Angiotensin II (Ang II), des-Arg⁹ bradykinin (BK 1–8), bradykinin (1–7), and all other chemicals were purchased from Sigma (St. Louis, MO). Lactic dehydrogenase (LDH) assay was obtained using the “TOX 7” in vitro LDH assay also from Sigma (St. Louis, MO). Diethylaminoethyl (DEAE) cellulose was purchased from Whatman (Fairfield, NJ). HK, PK, corn trypsin inhibitor (CTI), plasma α-kallikrein, FXIIa, and antibody to human FXII were purchased from Enzyme Research Laboratory (South Bend, IN). HD-Pro-Phe-Arg-paranitroanilide (S2302) and Glu-Pro-Arg-pNA (S-2366) were purchased from DiaPharma (Franklin, OH). Ala-Pro-paranitroaniline (APpNA) and Z-Pro-Prolinal (z-Pro-Pro) were purchased from Bachem (Torrance, CA). Cell permeability, prostacyclin, and NO kits were all purchased from Cayman Chemical (Ann Arbor, MI).

Animal Husbandry

Protocols were approved by Yale University and the University of Mississippi Institutional Review Boards. Animals were cared for as reported previously. Briefly, mice were fed a low-fat diet and housed 5 per cage.

Purification of Recombinant PRCP (rPRCP) from Schneider 2 Cells

The induction and purification of rPRCP from PRCP-transfected S2 cells was performed according to a previously published protocol [18].

The Effect of PRCP Inhibitors on PK Activation on HPAEC

Human pulmonary artery endothelial cells (HPAEC) were purchased from Invitrogen (Carlsbad, CA) and were cultured according to the vendor’s method. The cells were grown in Dulbecco’s Modified Eagle’s Medium (DMEM) overnight in 96 well plates (Costar). Cells were washed gently three times with HEPES-carbonated buffer (137 mM NaCl, 3 mM KCl, 12 mM NaHCO₃, 14.7 mM HEPES, 5.5 mM Glucose, 0.1% Gelatin, 2 mM CaCl₂, 1 mM MgCl₂, 7.1 pH, 37°C) between each incubation step. Gelatin blocking buffer (1% Gelatin) was prepared by adding appropriate amount of 5% gelatin stock to HEPES-carbonated buffer as described above. In the first step, gelatin blocking buffer was used to reduce the non specific binding. As part of this step, the cells were incubated with 1% gelatin buffer for 1 hour. In second and third steps, the cells were incubated with 20 nM high molecular weight kininogen (HK), and with 20 nM of PK and different concentrations of PRCP inhibitors (1 µM to 3 mM), respectively.

The Effect of PRCP Inhibitors on rPRCP Activity

The effects of the inhibitors of PRCP were determined using a previously published method [19]. Briefly, rPRCP was incubated in the presence or absence of various PRCP inhibitors in HEPES-carbonated buffer containing 1 mM Ala-Pro-paranitroaniline (APpNA, a PRCP chromogenic substrate). The final volume was 100 µL and the time of incubation was 60 min. Negative control had only 1 mM APpNA in HEPES-Carbonated buffer. Generation of free p-nitroaniline from APpNA was determined by monitoring changes in absorbance at 405 nm. Assays were done a minimum of 3 times.

Determination of the Specificity of Compound UM8190

After finding the optimal concentration of α-kallikrein, FXIIa, FXIa, or trypsin to produce a high signal-to-noise ratio, the effects of various concentrations of UM8190 on these serine proteases were determined according to published methods. All values were calculated as percentage of positive control after subtraction of background. Briefly, purified human serum was added to 200 µL of substrate buffer solution containing hippuryl-L-arginin in triplicated test tubes. Commercially available pooled normal human serum was used. The tubes were tightly covered, vortexed for 15 seconds and incubated for 60 min at 37 °C in a shaker water bath. The enzymatic reaction was stopped by addition of 250 µL of 1 M HCl solution and vortexed for 15 seconds. The blanks were stopped immediately after the addition of serum by 250 µL of 1M HCl before incubation. All samples were incubated on ice for 5 min. 1500 µL ethyl acetate was added, vortexed for 30 s and centrifuged for 10 min at 4000 g. 1 mL of the upper layer (ethyl acetate) was transferred into a 5 mL test tube and placed in a boiling water bath for 45–60 min to evaporate the ethyl acetate thoroughly, 3 mL of 1M NaCl solution was added to each tube and vortexed for 30 s. All samples were placed in water bath at 70 °C for 5 min to redissolve all the residual hippuric acid, then vortexed for 30 s. After 15 min of incubation at room temperature, the absorbance of hippuric acid was read at λ=228 nm.

The effect of UM8190 on the major membrane and plasma carboxypeptidases was determined. The effect of UM8190 on subtilisin activity in the reaction was examined with the specific substrate ZAALpNA according to the manufacturer’s suggestion (Sigma-Aldrich). The reaction mixture composed of 0.1 mM ZAALpNA in the 50 mM Tris-HCl/10 mM CaCl₂ buffer, pH 8.5. The reaction mixture was incubated at 37 °C for 1 h. The reaction was stopped by the addition of 0.5 M HCl. Studies determined the effect of UM8190 on CPA, according to the manufacturer’s recommendation (Sigma-Aldrich).

Determination of the Effect of UM8190 on PRCP-Induced Bradykinin Generation on HPAEC

HPAEC were incubated with 100 nM HK for 1 h at 37°C, as previously described [20]. After incubation, cells were washed and treated with 100 nM PK in the absence or presence of UM8190. Supernatants were collected and either frozen at −70 °C or immediately deproteinized with trichloroacetic acid. BK in the samples was determined using a commercial kit (MARKIT BK, Dainippon Pharmaceutical; Osaka, Japan), performed according to the manufacturer instructions. The metabolites of BK were determined by LC-MS.

The Effect of UM8190 on PRCP-Induced Nitric Oxide NO Formation on HPAEC

HPAEC were treated with 100 nM HK and incubated for 1 h at 37 °C. After washing three times with HEPES buffer, cells were then incubated with 100 nM PK in the absence and presence of 100 µM of UM8190 for 1 h at 37 °C. The solution was collected to measure the amount of nitrate + nitrite (the final products of nitric oxide metabolism) in each sample using a fluorometric assay (Cayman Chemicals, Ann Arbor, MI) according to the manufacturer’s protocol. The fluorescence was read at an excitation wavelength of 360 nm and an emission wavelength of 460 nm using BioTek Synergy HT Multi-Mode Microplate Reader. Nitrate + nitrite levels in each sample were normalized to that for the buffer alone.

The Effect of UM8190 on PRCP-Induced 6-keto Prostaglandin F_1α Release from HPAEC

HPAEC were treated with 100 nM HK and incubated for 1 h at 37 °C. Cells were then incubated with 100 nM PK ± 30 µM of compound UM8190 for 1 h at 37 °C. The solution was collected to measure the amount of 6-keto prostaglandin F_1α (a stable analog of prostacyclin) in each sample using a competitive acetylcholinesterase (AChE) enzyme immunoassay (Cayman Chemicals, Ann Arbor, MI) according to the manufacturer’s protocol. The absorbance was measured spectrophotometrically at 405 nm. The data was analyzed using a computer spreadsheet provided on the manufacturer’s website. 6-keto prostaglandin F_1α level in each sample was normalized to that for the buffer alone.

The Effect of UM8190 on HPAEC Permeability via PRCP-Dependent PK Activation

The effect of the inhibitors of PRCP on vascular permeability was assessed using an in vitro vascular permeability assay kit (Chemicon/Millipore, MA) according to the manufacturer’s protocol. Briefly, collagen coating solution in 0.2X PBS, pH 7.1 was added to the inserts. After incubating for 1 h at room temperature, the inserts were hydrated with cell growth medium for 15 min and seeded with 200 µL of cell suspension (1.0 × 10 HPAEC/mL). The plate was incubated at 37 °C for 24 h until a cell monolayer was formed. The inserts were then treated with cell basal medium (negative control); 1 µg/mL LPS (positive control); 0.1 µM HK/PK complex in the absence or presence of HOE-140 (1 µM) and lisinopril (1 µM, an angiotensin converting enzyme) and incubated at 37 °C for 18 h. The effect of the UM8190 (100 µM) alone and in combination with HK/PK complex (0.1 µM) on endothelial permeability was also studied. The fluorescence was read at an excitation wavelength of 485 nm and an emission wavelength of 528 nm using BioTek Synergy 2 Multi-Mode Microplate Reader.

Effects of UM8190 on the Metabolism of Angiotensin II and Bradykinin by rPRCP

The specificity of UM8190 was assessed by LC/MS analysis of the metabolism of angiotensin III (Ang III, Ang _2–8) to angiotensin 2–7 (Ang_1–7) and des-Arg bradykinin (BK_1–8) to BK_1–7 by rPRCP, as previously described [18].

Effect of UM8190 on Food Intake

Mice were single housed one week before the experiment. Male mice (n=6 per group; 4 months old) were then food deprived for 24 hours. Thirty minutes before food was re-introduced, mice were injected ip (100 µl total volume) with either saline (vehicle control) or 1, 10, or 100 mg/kg of UM8190. Food intake was measured at 1 hour, 2 hours, 4 hours 8 hours and 24 hours.

Mouse Carotid Artery Thrombosis Models

C57Bl/6 WT mice and C57Bl/6 mice deficient in factor fXII (FXII^−/−), were used in these studies. The procedures were approved by the Institutional Animal Care and Use Committee of Vanderbilt University. After anesthesia with pentobarbital (50 mg/kg IP), the right common carotid artery was exposed and fitted with a Doppler flow probe (Model 0.5 VB, Transonic System, Ithaca, NY). PBS with or without Compound UM8190 (0.8 mg/kg) was infused into the internal jugular vein 15 min before vascular injury in 50 µl volume. Thrombus formation was induced by applying two 1 × 1.5 mm filter papers (GB003, Schleicher & Schuell, Keene, NH) saturated with FeCl₃ (3.5% for wild type mice, 12.5% for FXII^−/−mice) to opposite sides of the artery for three min. Flow was monitored for 30 min. Mice were sacrificed by pentobarbital overdose after conclusion of the experiment, while under anesthesia.

Statistical Analysis

Results are expressed as mean ± standard error of mean (SEM) of at least three independent experiments each performed in triplicates. One-way ANOVA was performed and differences between groups were considered significant when p< 0.05 as verified by Newman-Keul’s or Dunnett’s post hoc test. Two representative concentrations (K_i and absolute inhibition) of UM8190 were chosen for statistical analysis of the inhibition studies. UM8190 was compared with z-Pro-Pro. For all comparisons, statistical significance was defined as p < 0.05.

RESULTS

PRCP Inhibitor Optimization

We utilized two assays to evaluate candidate PRCP inhibitors. One involved continuous monitoring of free p-nitroaniline generation derived from rPRCP-catalyzed cleavage of the chromogenic substrate Ala-Pro-p-nitroaniline (APpNA). The other was a cell-based assay measuring PRCP-dependent PK activation on human pulmonary artery endothelial cells (HPAEC). Initially, simple Z-Pro-Pro-NH-amide derivatives were evaluated. Analogs containing N-benzyl substituents (2a - 2j) demonstrated modest PRCP inhibitory activity , while the analog containing a phenethyl group (2k) was slightly advantageous (Table 1). The most dramatic inhibitory effects were observed in the homologous alkyl series (2l − 2o), with analog 2m inhibiting rPRCP with an K_i= 43 µM. The dodecyl (C12) substitution appeared to provide the optimal carbon group length, and requirement for a saturated hydrocarbon was not apparent. To test this, we substituted an ethoxy-3-propyl repeating unit (2o) for the dodecane group in 2m. Analog 2o was ineffective in inhibiting rPRCP and blocking PK activation. Calculated logP values for each analog (Table 1) also suggested a requirement for lipophilic substituents. However, aqueous solubility was limited for a number of analogs, particularly those with ClogP values > 4.5.

Table 1.

rPRCP Inhibition Data (Z-Pro-Pro-NH Amides)

Compd	R	CLogP	rPRCP Ki (µM)	PRCP-Dependent PK Activation on HPAEC Ki (µM)
2a		2.44	338.46 ± 21	NE
2b		3.00	467.7 ± 18	900 ± 75
2c		3.00	203.08 ± 17	NE
2d		3.36	135.4 ± 12.5	488.6 ± 32
2e		2.32	289.2 ± 23.6	771.4 ± 26
2f		2.32	393.8 ± 12.4	450 ± 46
2g		3.56	80 ± 6.2	NE
2h		3.56	NE	192.86 ± 15.6
2i		2.76	541.5 ± 38.5	835.7 ± 41.6
2j		3.55	110.8 ± 8.6	282.86 ± 21.4
2k		2.37	289.2 ± 32.2	565.7 ± 28.8
2l		3.20	80 ± 4.3	437.14 ± 16.1
2m		5.29	61.5 ± 5.7	50.8 ± 8.2
2n		6.54	61.54 ± 4.6	64.3 ± 4.3
2o		6.96	61 ± 5.3	141.4 ± 17.6
2p		0.69	NE	NE

CLogP (octanol-water partition coefficient) values were calculated using ChemBioDraw calculator (v. 12.0; CambridgeSoft). NE; No Effect.

Since the PRCP inhibitory activity of 2m was modest, further structural modifications were carried out by replacing the N-carboxybenzyl group with ring isosteres to generate proline A-ring N-substituted amines that are suitable precursors for producing HCl salts. Unfortunately, none of these derivatives inhibited rPRCP (Table 2); however, a compound lacking the proline A-ring carbamate [3 (UM8190)] proved to be an effective inhibitor (Ki=43 µM) and blocked PRCP-dependent PK activation on HPAEC’s (K_i=34 µM). Replacement of the proline A-ring with heterocyclic isosteres afforded derivatives (5 and 7a-7c) devoid of PRCP inhibitory activity, with the exception of azetidine derivative 7d, which was equivalent in its ability to inhibit PRCP (Table 3). Construction of analogs represented by B-ring azetidine isosteres (10, 12, and 14) further established the requirement for an unsubstituted N-group (Table 4).

Table 2.

PRCP Inhibition Data (N-Substituted-Pro-Pro-NH dodecylamides)

Cpnd	R	rPRCP Ki (µM)	PRCP-Dependent PK Activation on HPAEC Ki (µM)
3(UM8190)	H	43.08 ± 3.6	34.07 ± 4.3
4a		NE	NE
4b		NE	NE
4c		NE	NE
4d		NE	NE
4e		NE	NE

Table 3.

PRCP Inhibition Data (N-Substituted-Pro-NH Dodecylamides)

Cpnd	R	rPRCP Ki (µM)	PRCP-dependent PK activation on HPAEC Ki (µM)
5		NE	NE
6	H	61.54 ± 8.2	477 ± 28.7
7a		NE	NE
7b		NE	NE
7c		NE	NE
7d		43.7 ± 5.2	17.36 ± 3.2

NE; No effect

Table 4.

PRCP Inhibition Data (Azetidine-NH Dodecylamides)

Cpnd	R	rPRCP Ki (µM)	PRCP-dependent PK activation on HPAEC Ki (µM)
10	H	195.1 ± 15.4	50.14 ± 8.1
12		108.92 ± 12.5	29.6 ± 3.6
14		84.92 ± 7.6	60.43 ± 7.2

Compound UM8190 is a Selective PRCP Inhibitor

We chose compound UM8190 from the library of analogs based on results with the initial chromogenic- and cell-based screening assays, and evaluated its effect on other plasma serine proteases. Compound UM8190 inhibited rPRCP in a dose-dependent manner with a K_i value of 43 µM Fig. (1A), but failed to inhibit α-kallikrein, factor XIIa (FXIIa), factor XIa (FXIa) or trypsin at concentrations > 1.0 mM. z-Pro-Prolinal (our lead compound) had a little effect on rPRCP activity at the tested concentrations, Fig. (1A). This finding is consistent with a previously described observation and suggests that z-Pro-Prolinal is a very weak inhibitor of PRCP [19]. For comparison, soybean trypsin inhibitor (SBTI) inhibited α-kallikrein, FXIa, and trypsin with Ki values of 6.15, 0.62, and 0.12 µM, respectively (Table 5), while corn trypsin inhibitor blocked FXIIa with a Ki of 0.12 µM under our experimental conditions. Next, investigations were performed to determine whether UM8190 inhibited PRCP-dependent PK activation on endothelium. As shown in Fig. (1B), compound UM8190 inhibited the activation of PK to kallikrein in a dose-dependent manner. While z-Pro-Prolinal had a little effect on the production of kallikrein, UM8190 markedly reduced the hydrolysis of S2302 by kallikrein produced on HPAEC with K_i values of 34 µM.

Fig. 1. Inhibition of PRCP by compound UM8190.

Panel A. Effects of UM8190 on recombinant PRCP (rPRCP). 100 ng rPRCP was incubated with increasing concentrations of UM8190 or z-Pro-Prolinal (z-Pro-Pro). Panel B. Effects of UM8190 or z-Pro-Pro on PRCP-dependent PK activation in human pulmonary artery endothelial cells (HPAEC). Panel C. UM8190 inhibits the metabolism of BK_1–8 to BK_1–7 by rPRCP. Panel D. UM8190 blocks the metabolism of Ang III (angiotensin 2–8) to Ang _2–7 by rPRCP. Panel E. Compound UM8190 is a competitive inhibitor of PRCP. For panel A and E, generation of paranitroanilide from APpNA was measured. For panel B the release of paranitroaniline from the kallikrein substrate S2302 was measured. Data are expressed as mean ± SEM. ^**P < 0.01 versus z-Pro-Pro.

Table 5.

Effect of UM8190 on Recombinant Prolylcarboxypeptidase (rPRCP), Serine Proteases and Carboxypeptidases

Blocking Agent	rPRCP	Kallikrein	FXIIa	FXIa	Ki (µM) Trypsin	CPA	CPB	CPN	CPM
UM8190	43	NE	NE	NE	NE	NE	NE	NE	NE
SBTI	NE	6.15	NT	0.62	0.12	NT	NT	NT	NT
CTI	NT	NT	0.12	NT	NT	NT	NT	NT	NT
1,10- Phenanthroline	NE	NT	NT	NT	NT	486	738.5	615.4	393.85

Inhibitor concentrations required to produce 50% inhibition of recombinant prolylcarboxypeptidase, various abundant serine proteases, and carboxypeptidases. rPRCP; recombinant prolylcarboxypeptidase, FXIIa; activated factor XII, FXIa; activated factor XI, CPA; carboxypeptidase A; CPB; carboxypeptidase B, CPN; carboxypeptidase N, CPM; carboxypeptidase M, SBTI; soy bean trypsin inhibitor, CTI; corn trypsin inhibitor.

^*NE : denotes no effect

^**NT : indicates not tested

Since PRCP is a serine carboxypeptidase, we tested the effect of UM8190 on other carboxypeptidases. The serine carboxypeptidases CPN [21] and CPM [22] regulate kinin activity through proteolysis of BK, and are mainly involved in regulating chronic inflammation. Compound UM8190 did not inhibit metabolism of hippuryl-lysine by CPN (hCPN, partially purified from plasma) or CPM [23, 24], while 1,10-phenanthroline blocked CPN and CPM with K_i values of 615.4 and 393.8 µM respectively (Table 5). Carboxypeptidase A (CPA) is a highly conserved protease found in pancreas and mast cell granules. Although its substrate selectivity is different from that of PRCP, both CPA and PRCP cleave the C-terminal aromatic or aliphatic amino acids of proteins or peptides. Carboxypeptidase A [CPA, EC 3.4.17.1] and carboxypeptidase B [CPB, EC 3.4.17.2] were not inhibited by UM8190. 1,10-phenanthroline inhibited CPA and CPB with K_i values of 486 and 738.5 µM, respectively (Table 5).

Ang II [25] and BK [18] are well-established proteolytic targets of PRCP. Previously, we described a LC/MS-based method for characterizing Ang II metabolism by rPRCP. This method was used to evaluate the effects of UM8190 on PRCP-catalyzed cleavage of Ang II to Ang_1–7 and BK_1–8 to BK_1–7. Compound UM8190 (100 µM) effectively blocked both enzymatic reactions Fig. (1C and 1D). Further studies with purified rPRCP determined that UM8190 was a competitive inhibitor of PRCP Fig. (1E). These findings suggested that UM8190 is a selective inhibitor of PRCP.

Effects of Compound UM8190 on PRCP-Mediated Processes In Vitro

PRCP-dependent PK activation stimulates vascular endothelial cells to produce NO and PGI₂ via bradykinin-mediated B₂ receptor activation [26]. We initially determined the generation of bradykinin (BK) in the absence or presence of UM8190 using a previously described method [20]. PRCP-induced increases in kallikrein and kallikrein-dependent bradykinin were measured in HPAECs. Incubation of HPAECs with the complex of HK/PK caused an increase in BK generation Fig. (2A). Conversely, UM8190 (100 µM) reduced PK activation by PRCP and subsequent BK generation downstream Fig. (2A). Next, the effect of UM8190 on BK-induced release of NO and PGI₂ was determined. Compound UM8190 (100 µM) reduced NO and PGI₂ production by 80% Fig. (2B, 2C) for cells on which HK/PK complexes were allowed to assemble. PK activated by PRCP also facilitates BK generation on endothelial cells, with a subsequent increase in cell permeability [26]. Consistent with this, treatment of HPAEC with 0.1 µM HK alone did not alter cell permeability, while addition of the HK/PK complex (0.1 µM each) resulted in significantly increase permeability Fig. (2D). In the presence of HK/PK, compound UM8190 (100 µM) reduced cell permeability by 90% Fig. (2D). These results demonstrate the efficacy of UM8190 in inhibiting known PRCP-mediated reactions.

Fig. 2. Effect of compound UM8190 on PRCP-mediated pathways.

Panel A. Compound UM8190 blocks bradykinin generation on HPAEC. Panel B. Compound UM8190 blocks bradykinin-induced nitric oxide generation on HPAEC. Panel C. Compound UM8190 blocks bradykinin-induced prostacyclin generation on HPAEC. Panel D. Compound UM8190 blocks bradykinin-induced HPAEC monolayer permeability. Data are expressed as mean ± SEM. ^**P < 0.01 versus control.

Effects of Compound UM8190 on Food Intake in Mice

Recent studies indicated that PRCP plays an important physiological role in the regulation of food intake. Investigations were performed to extend these findings by demonstrating that PRCP inhibition reduces food intake. First, the effects of UM8190 on 14 metabolic markers in mice were determined (Table 6). Compound UM8190 did not cause significant changes in electrolyte levels, or markers of kidney and liver function within 8 hours of administration. There was a slight increase (20%) in blood amylase compared to control (n=3, P = 0.2), and blood glucose levels were decreased by 15% in treated mice (n=3, P = 0.26). Although not statistically significant, the reduction in blood glucose could be due to inhibition of hypothalamic release of POMC, as suggested previously [27].

Table 6.

The Effect of UM8190 on the Function of Kidney and Liver, Electrolyte, and Fluid Balance

Metabolic Panel	Vehicle	Compound 3 (400 µM)
ALB	3.9	3.9
ALP	80	93
ALT	34	33
AMY	822	1041
TBIL	0.4	0.4
BUN	16	19
Ca2+	9.9	9.9
PHOS	10	8.0
CRE	0.2	0.3
GLU	195	165
Na+	152	151
K+	5.8	5.4
TP	5.4	5.3
GLOB	1.5	1.5

Next, investigations were performed to assess the effectiveness and minimum time required for UM8190 to reduce food intake. Mice were deprived of food for 24 h prior to being randomly assigned to receive intraperitoneal infusions of UM8190 at 1, 10, or 100 mg/kg. Food intake was determined at fixed intervals (1, 2, 4, 8, or 24 h) after drug administration. As shown in Fig. (3), Panel A, administration of UM8190 started to attenuate food intake within 1 h. Significant declines in food intake were observed when mice were treated with UM8190 at 10 (6 mice per group, P < 0.05 vs saline) or 100 mg/kg (6 mice/grop, P < 0.05 vs saline, P < 0.05 vs 10 mg/kg UM8190) at all time intervals Fig. (3, Panels A through E), with a dose effect clearly apparent. Mice treated with the lowest UM8190 dose (1 mg/kg) also had lower food intake, although the reduction was not significant at 2, 4, 8, or 24 h. Zhou and colleagues [28] from Merck Research Laboratories recently reported on the orexigenic effect of a substituted prolyl-2-benzimidazole PRCP inhibitor in mice. Although the affinity of UM8190 for rPRCP is significantly lower than that of the prolyl-2-benzimidazole inhibitor, Compound UM8190 (100 mg/kg) was five times more potent at decreasing food intake at a similar dose. Our results are consistent with published studies showing that PRCP inhibition or disruption of the PRCP gene results in reduced food intake and decreased body weight in mice [15]. Cumulatively, the in vitro and in vivo studies demonstrate that UM8190 is a potent anorexigenic agent that inhibits PRCP-dependent pathways.

Fig. 3. Effect of compound UM8190 on food intake in mice.

Mice were given a single dose (i.p.) of compound UM8190 (1, 10, and 100 mg/kg) and food intake was measured over 1, 2, 4, 8, and 24 hours. The numbers of animals per group was 6. *, P <0.05 vs saline, P < 0.05 vs 10mg/kg UM8190. Data are expressed as mean ± SEM. T denotes each time unit. Data are expressed as mean ± SEM. T denotes each time unit.

The Antithrombotic Effect of Compound UM8190 in Normotensive Mice

Previously, we proposed that PRCP could contribute to thrombin generation and blood clot formation by converting PK to α-kallikrein, promoting the sequential conversion of the protease zymogens factor XII (FXII), factor XI (FXI), and factor IX (FIX) of the intrinsic pathway of blood coagulation to their active forms (FXIIa, FXIa, and FIXa, respectively). As FXII-deficient (FXII^−/−) mice are resistant to thrombus formation in arterial injury and cerebral ischemia-reperfusion models, [29, 30] we postulated that UM8190 might have an antithrombotic effect. Compound UM8190 was tested in mice with a model in which carotid artery thrombosis is induced by exposing the vessel to varying concentrations of ferric chloride (FeCl₃). In wild type C57BL/6 mice (n=5), UM8190 (10 mg/kg IV) did not prevent vessel occlusion, nor prolong the time to arterial occlusion, compared to vehicle (PBS) when thrombosis was induced with 3.5% FeCl₃ (the lowest concentration that reproducibly causes arterial occlusion in wild type mice) [29, 31]. Thrombus formation in this model is dependent on FXII, and it is possible that PK activation through FXIIa, or PK-independent activation of FXII could have overwhelmed any effect of UM8190 on PK activation. To address this concern, we tested UM8190 on thrombus formation in FXII^−/− mice, using 12.5% FeCl₃, which causes a high rate of vessel occlusion in these animals [29]. While all FXII^−/− control mice treated with PBS (n=5) experienced rapid vessel occlusion with 12.5% FeCl₃, 4 of 6 FXII^−/− mice treated with UM8190 (10 mg/kg IV) did not develop occlusion. These results suggest UM8190 has an antithrombotic effect, although the potency is difficult to determine at this point.

DISCUSSION

PRCP (also referred to as lysosomal carboxypeptidase) is a serine protease found on the surface of a variety of cells, including vascular endothelial cells [32–34]. It was first identified in lysosomal fractions of kidney as an activity that cleaves Ang II and BK, and was subsequently determined to be a potent activator of PK in the presence of HK [35]. As expected for a carboxypeptidase, PRCP cleaves after the penultimate C-terminal proline residue on Ang II, and after the C-terminal most proline on des-Arg⁹-bradykinin (BK_1–8) [18]. While human PK also has a penultimate proline residue, activation of this protease zymogen ultimately requires cleavage of the internal Arg ³⁷¹ -Ile ³⁷² peptide bond by a protease with endopeptidase activity [36]. Although it has been suggested that PRCP has an endopeptidase activity, it is not clear if PRCP has such activity, or renders PK more susceptible to activation by other mechanisms [37]. Observations linking PRCP to hypertension, preeclampsia and the metabolic syndrome are consistent with its reported effects on the renin-angiotensin and kallikrein-kinin systems [38–40], and indicate that PRCP might be an attractive target for treating cardiovascular or inflammatory disorders. More recently, PRCP has been implicated in the regulation of food intake in mice [15, 41]. PRCP activity appears to lead to increased food intake by converting α-MSH (also referred to as α-MSH_1–13) to α-MSH_1–12.

Although α-MSH_1–13 is an anorexigenic peptide involved in energy homeostasis [3], central injection of α-MSH_1–13 reduces release of proinflammatory mediators such as TNF-alpha and nitric oxide (NO) in endotoxin treated mice [42]. The implication of these studies is that the consumption of PRCP inhibitors may have both anorexigenic and anti-inflammatory effects. α-MSH_1–13 -mediated MC4R activation results in a decrease in LPS/IFN-gamma-induced inflammation through reducing iNOS and COX-2 expression in astrocytes [43]. α-MSH_1–13 can exert anti-inflammatory effects via MC1R and MC4R. However, evidence suggests that MC1R may produce anti-inflammatory effects [44]. In addition, it is widely expressed in endothelial cells, neutrophils, monocytes, macrophages, fibroblasts and astrocytes [45–48]. Conversely, Mc4r knockout mice have hyperphagia, hyperglycemia and hyperinsulinemia, reflecting development of the maturity onset obesity syndrome [11]. Thus, α-MSH_1–13 in the brain plays an important role in the development of anorexia by activating MC4R, while it plays an anti-inflammatory role in neural and non-neural cells through the activation of MC1R, reflecting its pleiotropic activities. These studies suggest that inhibitors of PRCP may not only influence energy homeostasis but also induce the release of anti-inflammatory mediators. Surprisingly, chronic infusion of α-MSH_1–13 results in a decrease in mean arterial pressure and physical activity in rats [49]. Apparently the mechanism does not involve α-MSH_1–13 - MC1R or MC4R – mediated pathways, suggesting a previously unknown function for α-MSH_1–13. Although the hypotensive property of α-MSH_1–13 could be species specific, the finding suggests that we should be cautious when assessing drugs that cause a robust elevation of α-MSH_1–13 as a potent inhibitor of PRCP may cause a sudden unsafe drop of blood pressure. Further investigations are required to determine the effect of UM8190 on expression patterns of α-MSH_1–13, MC1R, and MC4R in in vitro cultured neurons and in vivo assays.

Previous studies on the effects of PRCP inhibition used inhibitors that were either modified peptides [50] or small molecules (aryl imidazoles or benzimidazoles) [28]. These drugs did not penetrate the blood-brain barrier well. We synthesized a class of reversible, potent and selective PRCP inhibitors to address this limitation. Several laboratories, including ours, previously showed that 1-[1-(benzyloxycarbonyl)-L-prolyl] prolinal (Z-Pro-Pro-OH; 1) inhibits PRCP, albeit with a high IC₅₀ (>2 mM) [19]. We synthesized a series of Pro-Pro B-ring amide analogs (Z-Pro-Pro-NH amides) and explored several classical isosteric replacements for the proline ring. Of the derivatives generated in this study, UM8190, compound 6 and compound 7d were found to have the highest affinity and best selectivity for PRCP in vitro, (Tables 2 and 3). In the present study, since UM8190, compound 6 and compound 7d had a similar affinity for PRCP, UM8190 was further characterized in vivo. However, our future studies will focus on characterizing Compounds 6 and 7d. The structure-activity relationship analysis has revealed a requirement for an unsubstituted nitrogen proline ring-A, further optimized with a dodecyl amide group on proline ring B. We are currently investigating conformationally constrained derivatives to improve PRCP inhibitory activity while retaining enzyme selectivity and in vivo potency.

We found that the administration of UM8190 induced suppression of appetite, indicating appetite is partially dependent on PRCP in fasted mice. Compound UM8190 induced appetite loss that was dose- and time-dependent. Next, investigations were performed to determine the effect of UM8190 on glucose homeostasis (α-MSH can reduce IL-6) [51]. Since IL-6 causes insulin resistance in hepatocytes [52], we proposed that compound UM8190-induced increase in α-MSH concentration would decrease glucose levels. Our study indicates that glucose levels in compound UM8190-treated mice were relatively low but not statistically significant compared with control mice (Table 6). Compound UM8190 induced a small increase in the level of serum amylase, suggesting the drug induced mild pancreatitis. Although intrapancreatic activation of pancreatic proteases has been considered as the common mechanism underlying pancreatitis [53, 54], we cannot readily explain mechanism (s) by which the actions of UM8190 induce increases in amylase. However, Mayerle [53] suggests that the ratio of proinflammatory/anti-inflammatory responses could be a potential triggering factor in affecting pancreatic proteases.

Our results with UM8190 in thrombosis models are somewhat at odds with those recently published by Adams and co-workers using a gene-trap technique to generate mice with significantly reduced, but not absent PRCP expression (>75% reduction in PRCP antigen in kidney, the organ with the greatest level of PRCP expression) [55]. The hypomorphic (PRCP gt/gt) mice had significantly elevated blood pressure. In the FeCl3-induced carotid artery thrombosis model PRCP gt g mice displayed a shortened time to arterial occlusion, and enhanced plasma thrombin generation, consistent with a hypercoagulable state. Since BK causes an induction of tissue plasminogen activator (tPA, a clot dissolving enzyme) [56], it is also tempting to speculate that α-kallikrein through BK generation protects normal endothelial cell morphology against thrombotic disorders by promoting tPA generation. However, whether PRCP, an upstream regulator of kallikrein is involved in the generation of tPA under pathologic conditions in in vitro and animal remains unknown. This phenotype was attributed partly to reduced PK activation, as protease inhibitors with activity directed against α-kallikrein produced a similar phenotype. We, in contrast, did not observe shortening of carotid artery occlusion times in wild type C57Bl/6 mice treated with UM8190, and observed evidence for an antithrombotic effect in FXII deficient mice. The reasons for the discrepancies between our work and the study by Adams et al. are not clear although differences in levels of PRCP activity expression and subtle variations in thrombosis models may be contributing. Indeed, while Adams et al. observed a shortening of time to thrombotic occlusion in wild type mice treated with the α-kallikrein/FXIIa inhibitor Pro-Phe-Arg-CMK, Renne et al. [30] reported that the same inhibitor produced an antithrombotic effect in mice. Furthermore, Revenko et al. [57] recently reported that mice treated with anti-sense oligonucleotides to specifically reduce plasma PK are resistant to formation of arterial and venous thrombosis, consistent with PK contributing to thrombus formation through contact activation. One additional comment on our results with UM8190 in the mouse thrombosis model is required. We observed an effect of PRCP inhibition in animals lacking FXII. If PRCP is contributing to thrombosis through PK, then α-kallikrein may be mediating this effect independent of its role as an activator of FXII. Alternatively, the data may be interpreted as indicating that the prothrombotic role of PRCP is not mediated through PK. One could speculate that UM8190 reduces Ang III catabolism by PRCP, leading to the activation of AT-1 receptors. Evidence indicates that activation of AT1 results in generation of plasminogen activator inhibitor-1 (PAI-1), promoting thrombosis and fibrosis [58]. Similarly, angiotensin IV (Ang IV, a metabolite of Ang III) also is linked to induction of PAI-1 acting through angiotensin type 4 (AT4) receptors [58]. Clearly, additionally, work needs to be done to clarify the roles of PRCP and PK in thrombus formation to determine if these proteins are suitable targets for antithrombotic therapy.

CHEMICAL STRUCTURES

PRCP inhibitor synthesis

Several laboratories including ours have shown that 1-[1-(benzyloxycarbonyl)-L-prolyl] prolinal inhibits PRCP with an IC₅₀ > 1 mM [19]. We initially synthesized a series of proline B-ring amides by condensing 1 with various primary amines under standard conditions [59] to yield analogs 2a −2p (Scheme 1). ¹H NMR analysis of the resulting products were complex, as anticipated, due to rotational cis/trans isomerism about the N-carbamoyl and NH-amide bonds generating multiple resonances [60, 61]. PRCP assay-guided evaluation resulted in selection of 2m (containing a dodecyl group) for further optimization.

Scheme 1.

Synthesis of Z-Pro-Pro-Amides^a.

^aReagents and conditions: (a) R-NH₂,1-(3-dimethylaminopropyl)-3-ethylcarbodiimide-HCl, HOBT, N,N-diethylisopropylamine, CH₂Cl₂, 25 °C, 18 h.

Variations in the proline A-ring N-substituent were produced by synthesis of a series of N-alkyl and N-benzyl amines. The N-CBz group of 2m was deprotected under standard conditions (H₂, Pd/C) yielding compound 3 (UM8190), which was subsequently subjected to reductive amination reactions [59] [RCHO, NaBH(OAc)₃] to produce analogs 4a-4e. The products were isolated and characterized as their respective HCl salts (Scheme 2).

Scheme 2.

Synthesis of N-Alkyl-Pro-Pro-Amides^a

^a Reagents and conditions: (i.) H₂, Pd/C, MeOH, HCl (aq.), 3 h; (ii.) R-CHO, CHCl₃, NaBH(OAc)₃, 2h.

A series of proline A-ring isosteres (7a-7d) were also synthesized from intermediate 6 (derived from catalytic hydrogenation of N-Cbz analog 5) using standard condensation reactions with carboxylic acids or their suitably protected analogs (Scheme 3).

Scheme 3.

Synthesis of N-Alkyl-Pro-Amides^a

^a Reagents and conditions: (i.) H₂, Pd/C, MeOH, HCl (aq.), 3 h; (ii.) R-COOH, 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide-HCl, HOBT, N,N-diethylisopropylamine, CH₂Cl₂, 25 °C, 18 h.

Azetidine bioisosteres (12 and 14) were prepared from commercially available N-Boc azetidine-2-carboxylic acid (8). Reaction of azetidine 9 (derived from the condensation reaction of 8 with dodecylamine under standard conditions) with trifluoroacetic acid afforded the deprotected analog 10, which was reacted with 8 or Z-Pro-OH yielding N-protected analogs 11 and 13, respectively (Scheme 4). Standard methods for N-CBz and N-Boc deprotection were used to generate the final products 12 and 14. All of the synthesized compounds were characterized by using a combination of analytical methods of analysis including ¹H- and ¹³C-NMR, HRMS (ESI), and elemental combustion analysis (CHN).

Scheme 4.

Synthesis of Azetidine bioisosteres^a

^aReagents and Conditions (i.) CF₃COOH:CH₂Cl₂ (1:1), 25 °C; (ii.) 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide-HCl, N,N-diethylisopropylamine, CH₂Cl₂, 25 °C, 18 h; (iii.) H₂, Pd/C, MeOH, HCl (aq.), 25 °C, 3 h.

CONCLUSION

The current study supports the hypothesis that PRCP plays roles in food-intake and thrombosis. Compound UM8190 is a novel synthetic PRCP inhibitor, which has antithrombotic and food suppressant properties. Compound UM8190, and similar compounds may be of great importance in addressing the human epidemic of obesity and diabetes, and associated chronic inflammation.

ABBREVIATIONS

PRCP

Prolylcarboxypeptidase

FXII

Factor XII (Hageman Factor)

Factor FXII^−/−

Factor XII Knockout

BK

Bradykinin

α-MSH

Alpha-Melanocyte-Stimulating Hormone

AT1

Angiotensin Type 1