Prodrugs 2-(Phosphonomethyl)pentanedioic Acid (2-PMPA) as Orally Available Glutamate Carboxypeptidase II Inhibitors

Abstract

2-(Phosphonomethyl)pentanedioic acid (2-PMPA) is one of the most potent inhibitors of glutamate carboxypeptidase II (GCPII), a zinc metallopeptidase that cleaves glutamate from N-acetylaspartylglutamic acid and folylpoly-γ-glutamate. Due to the presence of multiple acidic groups, 2-PMPA exhibits poor oral bioavailability, limiting its therapeutic utility despite its potent GCPII inhibitory activity. One approach to address this challenge is to develop prodrugs of 2-PMPA with enhanced lipophilicity and improved oral absorption as demonstrated by tris-POC-2-PMPA and tetra-ODOL-2-PMPA. To expand the diversity of our prodrug strategy for 2-PMPA, we explored two promoieties for the phosphonate group of 2-PMPA, ProTide and cycloSal groups, while converting the two carboxylic acids to ester promoieties. The resulting prodrugs were assessed for their ability to deliver 2-PMPA in plasma in mice following oral administration. Among them, several cycloSal-based prodrugs delivered micromolar levels of 2-PMPA in plasma following oral administration, representing another effective prodrug strategy to orally deliver 2-PMPA.

Grphical Abstract

Glutamate carboxypeptidase II (GCPII), also known as prostate specific membrane antigen (PSMA), is a type II membrane-bound zinc metallopeptidase that cleaves the C-terminal glutamate from highly acidic peptide substrates such as N-acetylaspartylglutamic acid (NAAG) and folylpoly-γ-glutamate.¹ Given its ability to generate glutamate in the extracellular compartment of the brain, GCPII has been actively explored as a therapeutic target to treat neurological disorders in which excess glutamate is considered pathogenic.² It was later recognized that targeting GCPII outside of the brain may have therapeutic utility in non-neurological diseases³ such as prostate cancer⁴ and inflammatory bowel disease (IBD).⁵

Development of therapeutically effective GCPII inhibitors has been an active area of research. Although many potent GCPII inhibitors have been reported to date by multiple research groups, nearly all of them are competitive inhibitors mimicking the highly acidic nature of its substrates.⁶ As a result, many of these inhibitors suffer from poor membrane permeability and unfavorable pharmacokinetic properties. For instance, 2-(phosphonomethyl)pentanedioic acid 1 (2-PMPA, Figure 1), one of the most potent GCPII inhibitors with an IC₅₀ value of 0.3 nM,⁷ possesses two carboxylic acids and one phosphonic acid, forming at least triply, possibly quadruply, negatively charged species at physiological pH. Consequently, 2-PMPA displays negligible oral bioavailability and hence limited therapeutic utility despite its potent GCPII inhibitory activity.

Figure 1.

Chemical structures of 2-PMPA 1, tris-POC-2-PMPA 2, tetra-ODOL-2-PMPA 3, and protide- and cycloSal-based prodrugs of 2-PMPA.

One way to address this challenge is to develop prodrugs of 2-PMPA, in which its acidic moieties are masked by cleavable lipophilic promoieties (Figure 1). Our group discovered two distinct prodrugs of 2-PMPA, tris-POC-2-PMPA 2⁸ and tetra-ODOL-2-PMPA 3,⁹ both of which delivered high concentrations of 2-PMPA to plasma in mice following oral administration. These successful cases prompted us to further expand the diversity of 2-PMPA prodrugs by introducing ProTide and cycloSal groups as promoieties for the phosphonate group. Herein, we describe the design and synthesis of these new prodrugs as well as their ability to deliver 2-PMPA in plasma following oral administration.

The ProTide technology¹⁰ was originally developed as a prodrug strategy for nucleotide-based drugs though it can be extended to non-nucleotide phosphorus-containing molecules such as 2-PMPA. We chose to synthesize two ProTide-based prodrugs of 2-PMPA. We used a combination of phenyl and L-alanine isopropyl ester to mask the phosphonate group of 2-PMPA in both prodrugs while the two carboxylic acids were masked with either ethyl or isopropyloxycarbonyloxymethyl (POC) ester. Synthesis of protide-based prodrug with diethyl ester 9 is outlined in Scheme 1. The orthogonally protected 2-PMPA tetraester 5 was prepared by addition of dibenzyl phosphite to diethyl 2-methylenepentanedioate 4. After removing the two benzyl groups by catalytic hydrogenation, compound 6 was converted to phosphonic acid monophenyl ester 7 via a two-step procedure. Compound 7 was subsequently reacted with thionyl chloride to form phenyl phosphonochloridate 8, which was then coupled with L-alanine isopropyl ester to yield the desired protide-based prodrug 9.

Scheme 1.

Synthesis of ProTide-Based 2-PMPA Prodrug 9.^a

^aReagents and conditions: (a) HP(O)(OBn)₂, DBU, rt; (b) H₂, Pd/C, EtOH; (c) (i) SOCl₂/DMF, sulfolane, 70 °C; (ii) PhO-SiMe₃, 90 °C; (d) SOCl₂, ACN, 70 °C; (e) L-alanine isopropyl ester, DIEA, DCM, between −10 to −30 °C.

Synthesis of di-POC ester ProTide 16 is outlined in Scheme 2. The orthogonally protected 2-PMPA tetraester 10⁹ was converted to the corresponding diester 11 by catalytic hydrogenation. After esterifying the two carboxylic acid groups of 11 with POC-Cl, the resulting di-POC ester 12 was converted to phosphonic acid 13. Following a strategy analogous to the synthesis of prodrug 9 from compound 6, compound 13 was converted into prodrug 16.

Scheme 2.

Synthesis of ProTide-Based 2-PMPA prodrug 16.^a

^aReagents and conditions: (a) H₂, Pd/C, EtOAc; (b) POC-Cl, Et₃N, DMF, 60 °C; (c) TMSBr, CHCl₃, rt; (d) P(OPh)₃, pyridine, 115 °C; (e) SOCl₂, ACN, 70 °C; (f) L-alanine isopropyl ester, DIEA, DCM, between −30 °C and −10 °C.

CycloSal-based prodrugs have been applied to various nucleotides in which their phosphate group formed cyclic diester with 2-hydroxybenzyl alcohol (saligenol) derivatives.¹¹ Like the ProTide technology, the cycloSal approach can be extended to phosphonic acids such as 2-PMPA.¹² As an extension of this prodrug approach, 2-aminobenzyl alcohol derivatives were also employed as promoieties to mask the phosphate group, resulting in cyclic phosphoramidate-based prodrugs. As shown in Scheme 3, we synthesized four cycloSal-based prodrugs 18a-c and 20 with different combinations of promoieties on the phosphonate and carboxylate groups of 2-PMPA. Phosphonic acid 13 was converted to dichloridate 17 by treatment with thionyl chloride, followed by salicyl alcohol or 2-aminobenzyl alcohol derivatives, to give the desired cycloSal-based prodrugs 18a-c as mixtures of diastereomers in variable ratios. Similar chemical transformations were carried out on compound 19 to synthesize cycloSal-based prodrugs 21, which possess a dibenzyl ester.

Scheme 3.

Synthesis of cycloSal-based 2-PMPA prodrugs 18a-c and 21.^a

^aReagents and conditions: (a) SOCl₂, CHCl₃, 60 °C; (b) Salicyl alcohol or 2-aminobenzyl alcohol derivatives, DIEA (or Et₃N), DCM, 0 °C or rt; (c) SOCl₂, CHCl₃, 60 °C; (d) (2-Amino-3-methylphenyl)methanol, DIEA/pyr, DCM, −78 °C to 0 °C.

We anticipated that in vitro to in vivo extrapolation of ADME properties of these prodrugs could be challenging due to the multiple promoieties incorporated onto the 2-PMPA scaffold. Therefore, we conducted two-time point pharmacokinetic studies in mice to confirm delivery of 2-PMPA into plasma following oral administration.

ProTide-based prodrugs 9 and 16 were orally administered at a dose equivalent to 10 mg/kg of 2-PMPA in mice. Plasma samples were collected at 30 min and 2 h post-administration. As shown in Figure 2, prodrug 9 delivered low concentrations (≤100 nM) of 2-PMPA to plasma at both time points. No enhancement in bioavailability was observed compared to oral administration of 2-PMPA itself, which was detected in plasma at a concentration of ~250 nM 30 min post-oral administration.⁹ Prodrug 16 delivered approximately 15-fold higher plasma levels of 2-PMPA compared to prodrug 9, but the overall plasma exposure remained low (1.58 ± 0.34 μM and 1.12 ± 0.04 μM at 30 min and 2 h, respectively) when compared to previously described prodrugs of 2-PMPA.^{8, 9}

Figure 2.

Plasma concentrations of 2-PMPA following oral administration of prodrug 9 (A) and prodrug 16 (B) in mice at 10 mg/kg 2-PMPA equivalent dose. Data expressed as a mean ± standard error (SEM) (n=3).

CycloSal-based prodrugs 18a-c and 21 were also evaluated in a two-time point in vivo pharmacokinetic studies in mice. Similar to the ProTide-based prodrugs, cycloSal-based prodrugs 18a-c and 21 were orally administered at a dose equivalent to 10 mg/kg of 2-PMPA. Plasma levels of released 2-PMPA were measured at 30 min and 2 h post-administration (Figure 3). Interestingly, all the cycloSal-based prodrugs delivered higher plasma levels of 2-PMPA at both time points compared to the ProTide-based prodrugs 9 and 16. The highest levels of 2-PMPA (23.7 ± 5.07 μM) were observed at 30 min post-administration of prodrug 18a. However, 2-PMPA plasma levels quickly declined to 4.00 ± 1.41 μM at 2 h post-administration, suggesting rapid clearance. Prodrugs 18b and 21 delivered nearly 4-fold lower concentrations of 2-PMPA at 30 min compared to prodrug 18a. Interestingly, these prodrugs exhibited sustained plasma levels over a period of 2 h. Prodrug 18c displayed 2-fold lower levels of 2-PMPA (10.4 ± 1.44 μM) at 30 min compared to prodrug 18a but slightly higher levels at 2 h (4.91 ± 0.91 μM), suggesting lower clearance of 2-PMPA. It should be noted that the plasma levels of 2-PMPA achieved by these prodrugs, especially prodrug 18a, are comparable to those reported for tetra-ODOL-2-PMPA orally administered at the same equivalent dose of 2-PMPA (10 mg/kg),⁹ and nearly 5-fold higher (dose-normalized) compared to tris-POC 2-PMPA (15.1 ± 0.35 μM at 30 min following 30 mg/kg equivalent dose of 2-PMPA).⁸

Figure 3.

Plasma concentrations of 2-PMPA following oral administration of cycloSal-based prodrugs 18a (A), 18b (B), 18c (C), and 21 (D). The prodrugs were orally administered at a dose equivalent to 10 mg/kg of 2-PMPA in mice. Data (μmol/L) expressed as a mean ± standard error (SEM) (n = 3).

It should be noted that intravenously administered 2-PMPA was found to be quickly cleared from the plasma⁹ and that the sustained plasma levels of 2-PMPA achieved by prodrug 18b and 21 may suggest the presence of a low-clearance depot of 2-PMPA in the plasma. To better understand the in vivo bioconversion of cycloSal-based prodrugs, we conducted metabolite identification (MET-ID) studies in plasma samples following oral administration of 18b (Figure 4). Interestingly, prodrug 18b itself was not detected in plasma samples. Analysis of extracted ion chromatograms for all possible intermediates derived from prodrug 18b identified dicarboxylic acid intermediate 22 in the plasma samples collected at 30 min and 2 h post-administration. Although we were not able to quantify the concentrations of 22 in the absence of the reference standard, it is conceivable that this intermediate serves as a depot for 2-PMPA in the plasma, enabling sustained levels. The same intermediate 22 was also detected in the plasma of mice treated with prodrug 21 (Figure S1), which may explain the nearly superimposable 2-PMPA plasma pharmacokinetics shown by prodrugs 18b and 21.

Figure 4.

Metabolite identification (MET-ID) studies in plasma samples collected at 30 min (black) and 2 h (red) following oral administration of 18b at a 10-mg/kg equivalent dose of 2-PMPA, compared with a 10 μM standard solution of 18b (green). At 30 min, 18b [m/z = 560.1891; retention time (RT) = 6.11 min] was not detected and dicarboxylic acid intermediate 21 [m/z = 328.0944, RT = 4.04 min] was observed. Intermediate 22 was also detected in plasma samples collected 2 h post oral administration of 18b, with a 65% decline in the peak area compared to the 30 min plasma samples.

Encouraged by the favorable plasma pharmacokinetics of 2-PMPA observed by these prodrugs, we analyzed brain tissues for 2-PMPA levels to evaluate their ability to deliver 2-PMPA to the central nervous system. As shown in Figure 5, brain levels of 2-PMPA were substantially lower than plasma levels for all cycloSal-based prodrugs, with the brain-to-plasma ratio ranging from 0.005 to 0.05. The limited brain distribution of 2-PMPA was observed in other orally available 2-PMPA prodrugs, including tris-POC-2-PMPA⁸ and tetra-ODOL-2-PMPA,⁹ underscoring the challenge of developing prodrugs to enhance the brain distribution of the parent molecules.¹³

Figure 5.

Brain concentrations of 2-PMPA following oral administration of cycloSal-based prodrugs 18a (A), 18b (B), 18c (C), and 21 (D). The prodrugs were orally administered at a dose equivalent to 10 mg/kg of 2-PMPA in mice. Data (nmol/g) expressed as a mean ± standard error (SEM) (n = 3).

Among the two types of promoieties explored for the phosphonate group of 2-PMPA, cycloSal approach showed promise in enhancing the oral bioavailability of 2-PMPA. It appears that the diesters of these prodrugs are quickly hydrolyzed and that the resulting dicarboxylic acids are responsible for sustained release of 2-PMPA in the plasma. This may explain the low brain-to-plasma ratio of 2-PMPA because the dicarboxylic acid intermediates are unlikely to penetrate the blood-brain barrier. Although these prodrugs have limited therapeutic utility in neurological disorders, they certainly hold therapeutic potential for non-neurological diseases that would benefit from chronic oral treatment with a GCPII inhibitor. Future efforts will be directed at more in-depth investigation of the bioconversion pathway by integrating in vitro ADME profiling (permeability and liver and plasma stability), which should help us strategize further structural optimization. It should be noted that these prodrugs consist of multiple stereoisomers, which may contribute differently to plasma 2-PMPA levels. A synthetic strategy based on a derivative of (S)-2-PMPA,¹⁴ the more potent enantiomer, could mitigate complications associated with these stereoisomers in future studies. Given the three distinct plasma pharmacokinetic profiles shown by the four prodrugs 18a (high C_max and high clearance), 18b and 21 (low C_max and slow clearance), and 18c (balanced C_max and clearance), cycloSal-based prodrugs could be tailor-designed to achieve optimal plasma pharmacokinetics of 2-PMPA unique to specific disease conditions.

Supplementary Material

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.5c00384.

Synthetic procedures, ¹H, ¹³C, and ³¹P NMR spectra, and pharmacological experimental methods (PDF)

ABBREVIATIONS

2-PMPA

2-(phosphonomethyl)pentanedioic acid

GCPII

glutamate carboxypeptidase II

cycloSal

cycloSaligenyl