Suppression of Substance P Biosynthesis in Sensory Neurons of Dorsal Root Ganglion by Prodrug Esters of Potent Peptidylglycine (cid:97) -Amidating Monooxygenase Inhibitors*

,

The C-terminal amide is a prerequisite for full biological activity of many neuropeptides (1). These neuropeptides are typically synthesized as glycine-extended precursors and converted to the mature peptides by a family of enzymes involved in posttranslational modifications, e.g. O-glycosylation, phosphorylation, sulfation, and hydroxylation, as well as in proteolytic processing, e.g. endoproteolysis and exoproteolysis (2). The final step of posttranslational processing is catalyzed by an enzyme originally identified as peptidylglycine ␣-amidating monooxygenase (PAM) 1 (3)(4)(5). PAM is localized in secretory granules and requires copper, ascorbate, and molecular oxygen for activity (2,6,7). Recent studies have shown that PAM is actually a bifunctional enzyme that contains two distinct enzymatic activities and catalyzes the C-terminal amidation in a sequential manner. The first enzyme, peptidylglycine ␣-hydroxylating monooxygenase (PHM) (EC 1.14.17.3), requires the cofactors for activity and catalyzes the peptidylglycine ␣-hydroxylation reaction while the second enzyme, peptidyl ␣-hydroxyglycine ␣-amidating lyase (PAL) (EC 4.3.2.5), converts the intermediate into an ␣-amidated peptide and glyoxylate (8 -11).
Among the glycine-extended neuropeptides examined, glycine-extended substance P (substance P-Gly) has been demonstrated to possess the highest affinity for PAM partially purified from conditioned medium of cultured rat medullary thyroid carcinoma CA-77 cells (12). Substance P has been implicated in the pathogenesis of neurogenic inflammation (13,14). For example, elevated levels of substance P have been observed during inflammation, and depletion of substance P by chronic treatment of animals with capsaicin has been shown to lessen the severity of the inflammatory response (15)(16)(17). Furthermore, substance P may also play a role in rheumatoid arthritis (18). It has been shown that substance P stimulates the release of collagenase and prostaglandin E 2 from synoviocytes, resulting in a loss of cartilage, development of lesions in the adjacent bone, and perpetuation of the inflammatory process in the arthritic joint (19). The involvement of substance P in the pathophysiology of rheumatoid arthritis has been further supported by the observations that the release of substance P from the dorsal horn of polyarthritic rats is significantly accelerated and that the severity of arthritis in rats is increased upon infusion of substance P into the knee joints (20,21). Thus, suppression of substance P biosynthesis through PAM inhibition may be beneficial to diseases such as neurogenic inflammation and rheumatoid arthritis.
Several inhibitors of PAM have been identified previously, including acetopyruvate (4), [(4-methoxybenzoyl)oxy]acetic acid (22), trans-styrylthioacetic acid (23), benzylhydrazine (24), N-formyl amides (25), sulfite (26), and derivatives of organic acids (27). Most of these compounds inhibit PAM with IC 50 or K i values in the low micromolar to sub-millimolar range. Through mechanism-based inhibitor design, Zabriskie et al. (28) have found that D-phenylalanyl-L-phenylalanyl-D-vinylglycine, a substrate analog, inhibited PAM with an apparent K i of 20 M. Since substance P-Gly and other glycine-extended neuropeptides have displayed affinities in the low micromolar con-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  centrations for PAM (12,29), the compounds mentioned above are, therefore, relatively weak inhibitors and are not expected to alter the biosynthesis of the amidated neuropeptides significantly.
In a previous study, N-substituted homocysteine analogs were found to be potent inhibitors of PAM partially purified from conditioned medium of cultured rat medullary thyroid carcinoma CA-77 cells (30). Studies in cultured dorsal root ganglion (DRG) cells, however, showed only modest inhibition of substance P biosynthesis with these compounds (30). Since PAM is localized in secretory granules, these results suggest that the inhibitors were not accessible to the intracellular compartment of the cells. In the present study, several ester derivatives of hydrocinnamoyl-phenylalanylhomocysteine, one of the most potent PAM inhibitors, were synthesized to improve the intracellular accessibility of these compounds.
Melting points were determined on a Thomas-Hoover Uni-Melt apparatus. 1 H-NMR spectra were recorded at 300 MHz on a Varian XL-300 or Bruker AM300 spectrometer with either chloroform or methanol as internal standard. DCI-mass spectra were measured on a Hewlett-Packard model 5985B single quadrupole mass spectrometer retrofitted with a Vacumetrics (California) desorption chemical ionization accessory. Thermospray mass spectra were measured on the Vestec 201 Thermospray single quadrupole mass spectrometer equipped with a HED (high energy dynode) detector. Electrospray mass spectra analysis was performed via loop injection using a Micromass Platform II mass spectrometer (Micromass, Manchester, UK).
Hydrocinnamoyl-L-phenylalanyl-D,L-homocysteine Thiolactone (Com-pound 2)-A suspension of 10 g of t-butoxycarbonyl (Boc)-L-phenylalanine (37.7 mmol), 5.75 g of D,L-homocysteine thiolactone hydrochloride (37.6 mmol), and 5.75 g of HOBT hydrate in 300 ml of methylene chloride was cooled in an ice bath. Triethylamine (5.3 ml, 38.1 mmol) was added in 1-ml aliquots followed by 7.25 g (37.9 mmol) of EDCI. The mixture was allowed to warm to room temperature and stirred for 17 h. The solvent was then evaporated in vacuo, and 300 ml each of ether and water were added to the residue. The phases were separated, and the organic layer was washed twice with 200 ml of 0.1 N HCl and once with saturated NaHCO 3 . The combined organic layers were dried over sodium sulfate/magnesium sulfate and filtered. The solvent was evaporated in vacuo to give 13.68 g of Boc-L-phenylalanyl-D,L-homocysteine thiolactone (99%) as a white foam. A solution of 2 g (5.5 mmol) of the above material and 5 ml (65 mmol) of trifluoroacetic acid was stirred for 45 min. The reaction was diluted with toluene before the solvents were evaporated in vacuo to give the product as white foam. This material was then dissolved in 40 ml of methylene chloride, and triethylamine (1.3 ml, 9.3 mmol) was added in portions. The pH was monitored to make sure that the reaction was basic. To this mixture was added 824 mg of hydrocinnamic acid (5.5 mmol), 840 mg of HOBT hydrate (5.5 mmol), and 1.04 g (5.5 mmol) of EDCI. The mixture was stirred for 90 min. The solvents were evaporated in vacuo, and the residue was taken up in ethyl acetate and water. The resultant emulsion was filtered through celite, and the phases were separated. The organic layer was washed once with water, twice with 0.1 N HCl, and once with saturated NaHCO 3 . It was then dried over sodium sulfate/magnesium sulfate and filtered, and the solvent was evaporated in vacuo. The resultant solid was dissolved in warm methylene chloride and purified by flash chromatography on silica gel (25% ethyl acetate/hexane) to give 1.1 g (51% overall, 3 steps) of hydrocinnamoyl-L-phenylalanyl-D,L-homocysteine thiolactone as a white solid.
Anal. Calcd for C 22  Hydrocinnamoyl-L-phenylalanyl-D,L-homocysteine n-Butyl Ester (Compound 4)-To a solution of 200 mg (0.5 mmol) of Compound 2 in 5 ml of tetrahydrofuran was added 320 mg of mercury (II) bistrifluoroacetate (0.75 mmol) and 37 mg (0.5 mmol) of n-butyl alcohol. The mixture was stirred for 3 h at room temperature and then cooled to 0°C. Sodium borohydride (57 mg, 1.5 mmol) was added to liberate the thiol. After 1 min, the reaction was quenched with 1 N HCl and extracted with ethyl acetate. The organic layer was washed with water and dried over magnesium sulfate. The solvents were evaporated under reduced pressure to yield a crude residue that was subsequently purified by recrystallization in ethyl acetate/hexane to give 67 mg (28%) of the product as a white solid (m.p. 85°C).
Anal. Calcd for C 26  Hydrocinnamoyl-L-phenylalanyl-D,L-homocysteine Benzyl Ester (Compound 5)-To 0.6 ml of dry benzyl alcohol was added 45 mg (2.0 mmol) of sodium. The mixture was heated to 130°C and stirred until the sodium dissolved and cooled to room temperature. A solution of 480 mg of Compound 2 in 6 ml of dry tetrahydrofuran was added and stirred for 90 s. The mixture was quenched with 2 N HCl, diluted with water, and extracted three times with methylene chloride. The combined organic layers were dried over sodium sulfate, filtered, and evaporated in vacuo. The resulting oil was purified by flash chromatography using 35% ethyl acetate/hexane until the benzyl alcohol was eluted. Further elution with 50% ethyl acetate gave 298 mg (47%) of the benzyl ester as SCHEME 1 a white solid (m.p. 129 -131°C).
Anal. Calcd for C 29  Hydrocinnamoyl-L-phenylalanyl S-nicotinoylhomocysteine Benzyl Ester (Compound 6)-Nicotinoyl chloride (25 mg, 0.18 mmol) was added to 86 mg (0.18 mmol) of Compound 5 in 1 ml of pyridine. The reaction was stirred at room temperature for 1 h before quenching with water. The solvent was evaporated under reduced pressure and then taken up in methylene chloride and water. The resulting emulsion was filtered through celite, and the layers were separated. The aqueous layer was extracted with methylene chloride. The combined organic layers were dried over sodium sulfate, filtered, and evaporated to give a residue that was subsequently purified by filtration through silica gel (25% ethyl acetate/ether followed by 50% ethyl acetate/ether) to give 101 mg (100%) of product as a white solid (m.p. 118 -121°C). A sample was recrystallized from ethyl acetate/hexane for elemental analysis.
Using the same procedure as the preparation of Compound 5, 167 mg of 2-naphthoyl-L-phenylalanyl-D,L-homocysteine thiolactone was converted to the benzyl ester that was recrystallized from ethyl acetate/ hexane to give 20 mg (10%) of the benzyl ester as a white solid (m.p. 144 -150°C).
Anal. Calcd for C 31  PAM Enzyme Assay-PAM was partially purified from conditioned medium of cultured rat medullary thyroid carcinoma CA-77 cells by DEAE and Sephacryl 300 SF column chromatography (12). The enzyme assay was performed according to the method published previously (30). Briefly, PAM (13.5 milliunits) was pre-incubated with various concentrations of inhibitors in 150 mM Tes, pH 7.0, and 0.001% Triton X-100 for 20 min at room temperature in a total volume of 50 l. An equal volume of a solution containing 4 M substrate, N-dansyl-D-Tyr-Phe-Gly, and 6 mM ascorbate was added and incubated further for 20 min. The reaction was terminated by adding 10 l of 100 mM EDTA. The product, N-dansyl-D-Tyr-Phe-NH 2 , was separated from substrate by a C 18 cartridge column on high pressure liquid chromatography and detected by a fluorimeter.
DRG Cell Assay-The sensory neurons of DRG from 1-to 3-day-old Sprague-Dawley rats [Tac:N(SD)fBR] were cultured as described previously (31). To suppress the growth of non-neuronal cells, 10 M cytosine arabinoside was added during the second and third days of culture (32). After 6 days in culture, cells were treated with or without PAM inhibitors for 1 day followed by incubation with 500 M ascorbate overnight. Subsequently, the cells were washed with Ham's F-12 medium, extracted with 2 M acetic acid, and centrifuged. The supernatant was lyophilized, redissolved in a buffer containing 0.001% Triton X-100 in 150 mM Tes, pH 7.0, and divided in two. One-half of the extract was used directly to quantitate substance P by a radioimmunoassay using substance P antiserum that recognizes substance P with an affinity 10,000-fold higher than that of substance P-Gly (29). The other half was treated with 40 g of partially purified PAM, 0.5 mg/ml catalase, 1 M CuSO 4 , and 3 mM ascorbate in 150 mM Tes, pH 7.0, in a total volume of 35 l for 2 h at room temperature to convert substance P-Gly to substance P. The amount of substance P in the latter sample, again measured by radioimmunoassay, was the sum of substance P and substance P-Gly in the cells. The substance P-Gly levels in sensory neurons were calculated by subtracting the amount of substance P in samples without PAM treatment from the corresponding samples treated with the enzyme.

RESULTS
When the DRG cells were dissociated and grown one day in culture, both the neuronal cells and non-neuronal cells such as glial cells and fibroblasts were seen (results not shown). The growth of the non-neuronal cells could be suppressed almost entirely after treatment of the cells with cytosine arabinoside during the second and third days of culture, resulting in a nearly homogenous population of neuronal cells that were inter-connected with neurites (results not shown). The effects of PAM inhibitors on substance P biosynthesis were assessed using these cells.
Compound 1 was previously found to be a potent PAM inhibitor in vitro, with an IC 50 of 10 nM. However, this compound was not effective in suppressing the biosynthesis of substance P in DRG cells; only 25% inhibition was observed at 10 M (Table  I). The homocysteine moiety was absolutely required for inhibition of substance P production in these cells; no significant inhibition was obtained when it was replaced by cysteine (results not shown). Since compounds that contain a charged group, such as a carboxylic acid, frequently exhibit poor cell penetration, we attributed the large difference in the potencies between the in vitro isolated enzyme and the cell assays to the inability of Compound 1 to enter DRG cells. One approach to circumvent this problem was to increase the hydrophobicity of the compound by cyclizing the sulfhydryl group with the Cterminal carboxyl group. As expected, the resulting thiolactone (Compound 2) was inactive in the in vitro PAM assay since a free sulfhydryl group was necessary to coordinate the active site copper ion. Unfortunately, this compound only showed a slight improvement in the DRG cell assay; substance P biosynthesis was inhibited by 37% at 10 M (Table I). Therefore, a number of prodrug esters of Compound 1 were synthesized. The methyl (Compound 3), butyl (Compound 4), and benzyl (Compound 5) esters were 200-to 500-fold weaker in potencies in the in vitro PAM assay when compared with the parent compound; their respective IC 50 values were 2.3, 1.9, and 5.4 M. In DRG cells, the prodrug esters inhibited substance P biosynthesis with increased potency. The results also indicate that increased hydrophobicity of the esterified group led to better cellular activity. For example, Compound 5 inhibited the production of substance P in DRG cells by 74% at 10 M (Table  I). The esters were expected to inhibit substance P biosynthesis after intracellular conversion to the corresponding carboxylates by esterase. Indeed, incubation of the prodrug ester 3 with rabbit liver esterase (10 units/ml) for 10 min at 37°C produced the parent compound based on its potency in the in vitro PAM inhibition with an IC 50 of 8 nM (results not shown), similar to that observed with Compound 1.
Compound 5 was selected for further modification. Its sulfhydryl group was replaced with thionicotinate (Compound 6), thioacetate (Compound 7), or thiobenzoate (Compound 8). These compounds were inactive in the in vitro PAM assay as expected. Surprisingly, Compound 6 showed a decreased activity in the DRG cell assay when compared with Compound 5 (Table I). Also, replacement with a thioacetate group did not result in an improved inhibition of substance P biosynthesis when compared with the results obtained with Compound 5. Compound 8 was the most potent in this series; it inhibited the production of substance P in DRG cells with an estimated IC 50 of 2 M (Fig. 1).
In a previous study, Compound 9 was also identified to be a potent inhibitor of PAM in vitro, and it was more effective than Compound 1 in the DRG cell assay (Tables I and II). Using the same strategy as described in Table I, the benzyl ester derivative of Compound 9 (Compound 10) was synthesized. Unfortunately, this compound was less potent than its parent compound in the DRG cell assay (Table II). Therefore, optimization of this series of compounds was not pursued further.
The effect of PAM inhibition on the relative abundance of substance P and its precursor substance P-Gly in DRG cells was investigated using Compound 8. As described previously (31), cultured DRG cells produce more substance P-Gly than substance P in the absence of exogenous ascorbate. Under this condition, PAM is not fully activated since ascorbate is required for reoxidation of the active site copper during substrate turnover. Addition of 500 M ascorbate and incubation of the DRG cells overnight led to a change in the substance P-Gly to substance P ratio from 3.2 to 0.3 (Table III). Compound 8 at 3 M significantly inhibited the conversion of substance P-Gly to substance P even in the presence of the reducing agent. The ratio of substance P-Gly to substance P in the DRG cells was 1.5, but the total amount of the two peptides was not significantly different from that obtained in control cells or cells treated with ascorbate overnight (Table III).

DISCUSSION
Since PAM requires both ascorbate and copper for activity (33,34), several investigators have demonstrated changes in the production of amidated peptides by altering the levels of these cofactors in vivo. For example, a diet deficient in vitamin C has been shown to cause a 30-fold increase in the levels of glycine-extended gastrin with a concomitant 2-fold decrease in gastrin in extracts of guinea pig antra (35). Likewise, chronic treatment with a copper chelator N,N-diethyldithiocarbamate or its disulfide dimer disulfiram in rats produced a dose-dependent increase in glycine-extended ␣-melanotropin in the pituitary (36). Unexpectedly, treatment with disulfiram, but not with N,N-diethyldithiocarbamate, was effective in the reduction of ␣-melanotropin in rat pituitary neurointermediate lobe and cholecystokinin octapeptide in cerebral cortex (37). Similar treatment with disulfiram in rats also resulted in increased levels of substance P-Gly in various areas of the brain, including preoptic area, medial basal hypothalamus, pons, medulla, and spinal cord (38). However, these treatments may cause other nonspecific effects, since other enzymes such as dopamine ␤-hydroxylase, superoxide dismutase, and lysyl oxidase also require ascorbate and/or copper for activity (39 -42). Therefore, specific, potent inhibitors of PAM would be desirable to confirm these observations.
The most potent PAM inhibitors discovered to date are Nsubstituted homocysteine analogs (30). These compounds were shown to inhibit PAM partially purified from conditioned medium of cultured rat medullary thyroid carcinoma CA-77 cells with IC 50 values in the low nanomolar range. Despite their potent activity in the in vitro enzyme assay, the homocysteine analogs were only moderately effective in regulating the biosynthesis of substance P in cultured DRG cells. Since the poor inhibitory activity would be due to poor cell penetration, a number of prodrug esters of hydrocinnamoyl-phenylalanylhomocysteine were prepared to increase the intracellular accessibility of these compounds. Hydrocinnamoyl-phenylalanyl-(S-benzoyl-homocysteine) benzyl ester (Compound 8) was identified as the most potent compound, inhibiting substance P biosynthesis in DRG cells with an estimated IC 50 about 2 M or 1.2 g/ml (Fig. 1). Although this potency is sufficient to modulate the production of amidated peptides, it is surprising to note that the IC 50 value obtained in the DRG cell assay is somewhat higher than expected, considering the parent Compound 1 inhibited PAM with an IC 50 of 10 nM in vitro. Several reasons may explain this observed discrepancy in the potencies obtained from the in vitro enzyme and cellular assays. It is possible that Compound 8 is still not optimal for penetration through the DRG cell membrane. Alternatively, the intracellular esterases in the DRG cells might not be effective in converting the prodrug ester to the active PAM inhibitor, or the compound may not be stable throughout the incubation period since dimerization of the compound did occur upon storage (results not shown). Third, since PAM is localized in secretory granules inside the cells, inhibitors would need to cross two membranes to access PAM. If the prodrug esters are converted to the active Compound 1 by intracellular esterases immediately after crossing the plasma membrane, the resulting compound would not be as effective in penetrating through the membrane of the secretory granules to inhibit PAM.
Inhibition of an enzyme frequently results in an excessive accumulation of the precursor and/or up-regulation of the enzyme (43). Under these conditions, the precursor itself may exert biological effects similar to that of the mature product if a sufficient concentration of the precursor is reached to overcome its weak potency. Likewise, up-regulation of enzyme may normalize the biosynthesis of the mature product. Therefore, neither an excessive accumulation of the precursor nor upregulation of the enzyme would provide the desired pharmacological effects expected for drug intervention. In the present study, a limited quantity of PAM in the scarce DRG cells did not permit an easy assessment of whether PAM was up-regulated in these cells after inhibition. Nevertheless, a significant decrease in the level of substance P in DRG cells after treatment with Compound 8 suggests indirectly that PAM up-regulation was not likely to have occurred. Furthermore, the fact that the total amount of substance P and its precursor, substance P-Gly, remained the same in the presence or absence of ascorbate or PAM inhibitor (Table III) supports the notion that substance P-Gly was not excessively accumulated during PAM inhibition. Thus, inhibition of the ␣-amidation process appears to be an effective strategy for suppression of substance P biosynthesis. The prodrug esters of PAM inhibitors described here may also be useful to assess the pathological role of other amidated peptides such as neuropeptide Y, a powerful stimulant of food intake (44), whose precursor, glycine-extended neuropeptide Y, has also been shown to be recognized by PAM with high affinity (12).