Preclinical disposition of MGS0274 besylate, a prodrug of a potent group II metabotropic glutamate receptor agonist MGS0008 for the treatment of schizophrenia

Abstract MGS0274 besylate is an ester‐based lipophilic prodrug of a metabotropic glutamate (mGlu) 2 and mGlu3 receptor agonist MGS0008 and being developed for the treatment of schizophrenia. We investigated the disposition of these compounds in rats and monkeys and in vitro metabolism in humans to evaluate whether MGS0274 besylate could be useful as a prodrug in humans. After the oral administration of MGS0274 besylate to monkeys (2.89 mg/kg), MGS0008 was immediately found in plasma, reached a maximum concentration at 4 hours postdose, and decreased with a terminal half‐life of 16.7 hours; MGS0274 was barely detectable. The oral bioavailability as MGS0008 was 83.7%, which was approximately 20‐fold greater than that after oral dosing of MGS0008 (3.8%). In rats, MGS0008 penetrated the cerebrospinal fluid and was eliminated slower than from plasma. The in vitro metabolism study indicated that MGS0274 was rapidly hydrolyzed to MGS0008, which was not further metabolized. After the intravenous administration of MGS0008 to rats and monkeys, almost all the dose was excreted unchanged in urine. These results suggested that MGS0274 was, as expected, presystemically hydrolyzed to MGS0008 after gastrointestinal absorption and that MGS0008 was distributed throughout the body without further metabolism and ultimately excreted in urine in the animals. Furthermore, the hydrolytic activity against MGS0274 in the human liver S9 fraction was comparable to that in monkeys, suggesting the possibility of the rapid presystemic hydrolysis of MGS0274 to MGS0008 in humans, as it is in monkeys. Consequently, MGS0274 besylate is expected to function as a preferable prodrug in humans.


| INTRODUC TI ON
Group II mGlu (mGlu2/3) receptors play a role in regulating glutamatergic tone in the forebrain and limbic area, where glutamatergic dysregulation has been reported in patients with schizophrenia, and agonists for mGlu2/3 receptors are thought to be beneficial for the treatment of schizophrenia. 6,28,29 LY2140023 discovered by Eli Lilly and Company, which is an orally bioavailable prodrug of the mGlu2/3 receptor agonist LY404039 (Figure 1), significantly improved positive and negative symptoms of schizophrenia in a phase 2 proof-of-concept trial, compared with placebo. 31 However, LY2140023 failed to demonstrate efficacy in a phase 3 clinical trial. 1 This raised questions as to whether mGlu2/3 receptors are actually therapeutic targets for schizophrenia. Thereafter, a pharmacogenetic analysis of the patients treated with LY2140023 showed that the single nucleotide polymorphisms located in the serotonin 2A receptor were associated with a change in the positive and negative syndrome scale, suggesting a pharmacogenetic relationship between the single nucleotide polymorphisms and the response to LY2140023 treatment. 21 In addition, a post hoc analysis of data from phase 2 and phase 3 clinical trials for LY2140023 indicated that the patients with schizophrenia who had been ill for 3 years or less or previously treated with a dopamine D2 receptor antagonist exhibited the therapeutic response to LY2140023. 19 Thus, the potential of mGlu2/3 receptors and their agonists as therapeutic targets and medications for schizophrenia is still worth investigating.
The mGlu2/3 receptor agonists, LY404039 and LY354740 ( Figure 1), are rigid glutamate analogs and exhibited poor gastrointestinal absorption in humans (oral bioavailability of 3-6%), 2,8 probably because of poor membrane permeability arising from their hydrophilic properties. Accordingly, a prodrug approach was inevitable to improve their oral bioavailability. 4,31 F I G U R E 1 Chemical structures of the 14 C or stable-isotope labeled MGS0008 and MGS0274 besylate. Asterisk donates the position of the 14 C label. For comparison, the chemical structures of LY354740, LY544344 (a prodrug of LY354740), LY404039, and LY2140023 (a prodrug of LY404039) are also shown 4,31 13 2 15 2 4 To improve the gastrointestinal absorption of hydrophilic compounds, transportable prodrugs or ester-based lipophilic prodrugs are generally worth being developed. 32 Both LY544344 (a prodrug of LY354740) and LY2140023 ( Figure 1) were transportable prodrugs designed to be absorbed via intestinal peptide transporter 1 4,30 and have indeed improved oral bioavailability in humans, compared with their parent compounds. 2,18 Although the transportable prodrug approach was successful, systemic exposure to the prodrug was still observed for LY2140023, accounting for roughly 30% of its parent compound in humans 31 and 15% of circulating radioactivity in monkeys. 2 In general, higher exposures to pharmacologically inactive compounds, such as prodrugs themselves, should be avoided because these compounds can trigger toxicities and/or cause insufficient exposure to active components. 37 MGS0008, (1S,2S,3S,5R,6S)-2-amino-3-fluorobicyclo[3.1.0]hexane -2,6-dicarboxylic acid (Figure 1), is a rigid glutamate analog with a potent agonist activity for mGlu2/3 receptors. In rat in vivo models known to be indicative of antipsychotic potential, treatment with MGS0008 significantly decreased phencyclidine-induced locomotor hyperactivity and reduced conditioned avoidance responses, demonstrating that MGS0008 has antipsychotic potential. 27

| Animals
Male Wistar rats (7 weeks old) were purchased from Charles River  (see Table S1-B). Aliquots of the supernatant from unlabeled incubation samples and authentic standards were also injected into a liquid chromatography-mass spectrometry (LC-MS) system under the same chromatographic conditions to compare their chromatographic retention times and mass spectral data (see Table S1-B).

| Hydrolytic activity against MGS0274 in rat, monkey, and human sera and tissue S9 fractions
Rat, monkey, or human serum that had been pre-incubated for 5 minutes at 37°C was spiked with MGS0274 besylate at 10 μmol/L.  Table S1-C). The formation rate of MGS0008 from MGS0274 in the reaction mixture was calculated using the following equation:

| Inhibition of cytochrome P450s
The inhibition potential of MGS0008 (0.3-100 µmol/L) and MGS0274 besylate (0.03-10 µmol/L) on the activity of human cytochrome P450 (CYP) isoforms (CYP1A2, 2B6, 2C8, 2C9, 2C19, 2D6, and 3A) was investigated using probe substrates for each CYP isoform and human liver microsomes, of which the final protein concentration was set at 0.1 mg/mL (see Table S2). The assay was started by spiking the reaction mixture, which had been pre-incubated at 37°C for 5 minutes, with NADP + . After incubation for 5 to 10 minutes, the reaction was terminated by adding acetonitrile containing an internal standard for each probe substrate, and the resultant mixture was centrifuged. Aliquots of the supernatant were subjected to bioanalysis using LC-MS/MS (see Table S1-D).

| Induction of cytochrome P450s
The primary cultured cryopreserved human hepatocytes were

| Pharmacokinetics in rats and monkeys
The pharmacokinetic profiles of MGS0008 and MGS0274 were investigated in fasted male Wistar rats (n = 3) and fed male cynomolgus monkeys (n = 4). Monkeys had been given 60 g of a pri-  Tables S1-E Table S1-E).

| Pharmacokinetic analysis
The plasma concentration-time profiles of MGS0008 and MGS0274 were analyzed using a non-compartmental analysis method with pharmacokinetic analysis software Phoenix ® WinNonlin ® 6.2 or later (Certara, Princeton, NJ).

| Prediction of human pharmacokinetic parameters
The prediction of CL total , Vd ss , and t 1/2 in humans was performed using single-species allometric scaling methods 14 with the pharmacokinetic parameters obtained from the intravenous administration of MGS0008 to rats and monkeys, unbound fraction of MGS0008 in rat, monkey, and human plasma (1.0), and body weight of rats (0.3 kg), monkeys (5 kg), and humans (60 kg).

| Plasma protein binding
The unbound fractions of MGS0008 in rat, monkey, and human plasma were found to be 103.

| In vitro metabolite profiling of MGS0274 in rats, monkeys, and humans
The metabolite profiling of MGS0274 was conducted with rat, monkey, and human cryopreserved hepatocytes and besylate. In all the species tested, a predominant peak was detected in RI-HPLC analyses without other metabolite peaks after 1 hour incubation ( Figure 2) and identified as MGS0008 based on the chromatographic retention time and mass spectral data (Table S3).

| Hydrolytic activity against MGS0274 in rat, monkey, and human sera and tissue S9 fractions
The results are summarized in Table 1. In rats, all the tissue S9 fractions exhibited almost equal activity levels, followed by the serum.
In monkeys, the hydrolytic activity was equally high in S9 fractions from the intestine, liver, and kidney, followed by the lung S9 fraction.
In contrast to rats, monkey serum had no apparent hydrolytic activity against MGS0274. In humans, the liver S9 fraction exhibited the highest hydrolytic activity against MGS0274, followed by S9 fractions from the lung and intestine, while no apparent hydrolytic activity was observed in the serum or kidney S9 fraction.

| Induction of cytochrome P450s
The induction potential of MGS0008  (Tables S5 and S6).

| Pharmacokinetics of MMGS0008 in rats
The plasma concentration-time profiles of MGS0008 after a single intravenous or oral administration of MGS0008 (3 mg/kg) to male rats under fasted conditions are shown in Figure 3, and the associated pharmacokinetic parameters are summarized in Table 2. After intravenous administration, the plasma MGS0008 level rapidly decreased, with a t 1/2 of 1.54 hours, and the CL total was calculated to be 548 mL/h/ kg. The Vd ss was estimated to be 545 mL/kg, indicating the extravascular distribution of MGS0008. In urine, 93.2% of the administered dose was excreted as the unchanged form within 48 hours, demonstrating metabolically stable properties in vivo and urinary clearance as the major route of elimination for MGS0008. The CL renal was calculated to be 511 mL/h/kg from the amount of MGS0008 recovered in urine and the plasma AUC, which was approximately equal to the CL total and glomerular filtration rate (GFR) in rats (516 ± 114 mL/h/kg), 9 suggested that MGS0008 is less likely to undergo tubular secretion or reabsorption. After oral administration, MGS0008 reached a C max of 1990 ng/mL at 1.17 hours postdose and rapidly declined, with a t 1/2 of 0.968 hours. The oral bioavailability was estimated to be 78.1%.
TA B L E 1 Hydrolytic activity against MGS0274 in sera and tissue S9 fractions of rats, monkeys, and humans Values are presented as the mean of triplicate determinations, with SD in parentheses. a Formation rate was not calculated because MGS0008 was not detected in the reaction mixture.
F I G U R E 3 Plasma concentration-time profiles of MGS0008 after a single intravenous or oral administration of MGS0008 to fasted male rats at a dose of 3 mg/kg. After intravenous administration, the MGS0008 level rapidly decreased, with a t 1/2 of 1.54 h. After oral administration, the MGS0008 level reached a C max of 1990 ng/mL and rapidly decreased, with a t 1/2 of 0.968 h. The oral bioavailability was estimated to be 78.1%. Data are represented as the mean ± SD (n = 3). The plasma concentrations declined below the lower limit of quantification (3 ng/mL) at 24 h TA B L E 2 Pharmacokinetic parameters of MGS0008 after a single intravenous or oral administration of MGS0008 to fasted male rats at a dose of 3 mg/kg The extent of CSF penetration of MGS0008 after a single oral administration of MGS0008 was also evaluated in fasted male rats at a dose of 3 mg/kg ( Figure 4 and Table 3); at this dose, phencyclidine-induced locomotor hyperactivity was reportedly inhibited. At the first sampling time point (1.5 hours), MGS0008 in the CSF had already reached a C max value and then decreased at a slower rate than that observed in plasma. The AUC values in plasma and CSF were estimated to be 4330 h·ng/mL and 123 h·ng/mL, respectively.
The AUC ratio of CSF to plasma was calculated to be 2.8%.

| Pharmacokinetics of MGS0008 in monkeys
The plasma concentration-time profiles of MGS0008 after a single intravenous or oral administration of MGS0008 (1 mg/kg) to male monkeys under fed conditions are shown in Figure 5, and the associated pharmacokinetic parameters are summarized in Table 4

| Pharmacokinetics of MGS0274 in monkeys
We administered MGS0274 besylate orally to fed male monkeys at a dose of 2.89 mg/kg (1-mg equivalent of MGS0008/kg) and examined the plasma concentration profiles of MGS0274 and MGS0008 ( Figure 6 and Table 5). Although MGS0274 was barely detected, MGS0008 was detected at the first sampling time point (30 minutes postdose) and reached a C max of 688 ng/mL at 4 hours postdose, accounting for approximately 30-fold over the C max after the oral administration of MGS0008 (1 mg/kg). The oral bioavailability as MGS0008 was 83.7%, which was approximately 20-fold higher than that after the oral administration of MGS0008 F I G U R E 4 Plasma and cerebrospinal fluid (CSF) concentrationtime profiles of MGS0008 after a single oral administration of MGS0008 to fasted male rats at a dose of 3 mg/kg. At the first sampling time point (1.5 h), the concentration of MGS0008 in the CSF already reached a C max value and then decreased at a slower rate than that observed in plasma. Data are represented as the mean ± SD (n = 3). The plasma concentration declined below the lower limit of quantification (3 ng/mL) at 24 h TA B L E 3 Plasma and cerebrospinal fluid (CSF) concentrations of MGS0008, and pharmacokinetic parameters after a single oral administration of MGS0008 to fasted male rats at a dose of 3 mg/ kg Data are presented as the mean ± SD (n = 3). The concentration in the plasma at 24 h postdose was less than the lower limit of quantification (3 ng/mL) and was treated as zero for calculation of AUC. -, not applicable; NC, not calculated.
F I G U R E 5 Plasma concentration-time profiles of MGS0008 after a single intravenous or oral administration of MGS0008 to fed male monkeys at a dose of 1 mg/kg. After intravenous administration, the MGS0008 level rapidly decreased, with a t 1/2 of 1.48 h, to be below the lower limit of quantification (3 ng/mL) at 24 h. After oral administration, the MGS0008 level reached a C max of 20.5 ng/mL and showed sustained concentrations at low levels. The oral bioavailability was estimated to be 3.8%. Data are represented as the mean ± SD (n = 4) itself (3.8%). Moreover, the t 1/2 of MGS0008 (16.7 hours) was apparently prolonged, compared with that observed after the in-

| Prediction of human pharmacokinetic parameters for MGS0008
The prediction of human pharmacokinetic parameters was performed using single-species allometric scaling methods with the animal CL total , animal Vd ss , and unbound fraction. Since MGS0008 was unlikely to bind to plasma proteins in all the species tested, the unbound fraction value was set at 1. The CL total , Vd ss , and t 1/2 in humans were predicted to be 146 mL/h/kg, 545 mL/kg, and 2.6 hours from the rat data, and 69.3 mL/h/kg, 181 mL/kg, and 1.8 hours from the monkey data, respectively.

| D ISCUSS I ON
MGS0008 is a rigid glutamate analog with a potent agonist activity for mGlu2/3 receptors and has been shown to exert antipsychotic actions in animal models of psychiatric disorders. 27,36 Based on the similarities in the chemical structures between MGS0008 and the preceding compounds, such as LY404039, the possibility that MGS0008 might exhibit poor oral bioavailability in humans was predicted. Furthermore, a certain level of a transportable prodrug of LY404039 (LY2140023) remained in plasma. 31 Therefore, we decided to develop an ester-based lipophilic prodrug to further reduce the systemic exposure to the prodrug itself. Currently, MGS0274 besylate, a prodrug of MGS0008, is being developed for the treatment TA B L E 4 Pharmacokinetic parameters of MGS0008 after a single intravenous or oral administration of MGS0008 to fed male monkeys at a dose of 1 mg/kg Data are presented as the mean of two animals. First, the disposition of MGS0008 after intravenous administration in rats and monkeys was investigated, and we confirmed that almost all the administered compound was excreted as the unchanged form in urine. Moreover, no metabolites of MGS0008 were generated in the hepatocytes of rats and monkeys. These results indicate that the CL total of MGS0008 is equivalent to the CL renal . The CL renal values were almost equal to each GFR, suggesting that MGS0008 is less likely to undergo tubular secretion or reabsorption in the animal species that were tested. Although MGS0008 did not bind to plasma proteins in rats and monkeys, the Vd ss were limited and roughly corresponded to the extracellular volumes. Thus, MGS0008 could conceivably penetrate into the extravascular spaces, but its hydrophilic property (cLogP = −1.18) makes it difficult to cross cell membranes easily, and this limitation is thought to influence its tissue distribution in rats and monkeys. Also, in human in vitro studies, MGS0008 was not metabolized in hepatocytes and did not bind to plasma proteins.
Based on these results, similar distribution and excretion profiles of MGS0008 to those observed in the preclinical species seem possible in humans. In addition, a single-species allometric scaling of the animal pharmacokinetic parameters demonstrated that the t 1/2 of MGS0008 in human plasma is likely to be relatively short (1.8-2.6 hours); this suggests that managing medication adherence could be difficult, since a reduced dosing frequency, such as once a day, is optimal, especially for the treatment of psychiatric disorders. 7,25 Of note, a single-species allometric scaling method was used for the prediction of human pharmacokinetic parameters for the following reasons: it had reportedly performed better than in vitro-in vivo extrapolation methods, 23,33 and the CL total predicted from in vitro data had been underestimated in in-house examination with a variety of compounds (data not shown).
In a previous study, the oral administration of MGS0008 to rats at a dose of 3 mg/kg exerted antipsychotic actions in an animal model of schizophrenia. 27 Hence, the distribution of MGS0008 in the CSF as a surrogate for the central nervous system (CNS) exposure at this dose level in rats was evaluated. The AUC ratio of MGS0008 in the CSF relative to plasma was 2.8%, and MGS0008 exhibited a slower elimination from the CSF, compared with that from plasma. These characteristics, such as the limited penetration into the CNS and the slow elimination from the CNS, likely arise from the hydrophilic property of MGS0008. A similar pharmacokinetic profile was also reported in a study examining LY2812223 in humans, the chemical structure of which resembles that of MGS0008. 24 The bioavailability after the oral administration of MGS0008 in monkeys was only 3.8%, whereas it was 78.1% in rats. Since MGS0008 was not metabolized in the preclinical species, the large species differences in oral bioavailability might be due to differences in the rates of its gastrointestinal absorption, such as transporter-mediated uptake in rats. Species differences in oral bioavailability have also been reported for LY354740 (high in dogs and low in humans) 8,17 and LY404039 (high in rats and low in humans). 2,26 MGS0008, which has a fluorine-substituted structure of LY354740 at the C3 position, is also likely to exhibit low oral bioavailability in humans.
MGS0274 is designed to liberate MGS0008, l-menthol, acetaldehyde, and carbon dioxide (Figure 7), and this reaction is considered to be rapidly initiated by enzymatic hydrolysis and followed by spontaneous cleavage. Indeed, in vitro enzymatic hydrolysis of MGS0274 to MGS0008 was confirmed in all the species tested. As for l-menthol and acetaldehyde, they were not directly detected in this study; however, there seems to be no remarkable species differences in their metabolism. In rats, monkeys, and humans, l-menthol and acetaldehyde are thought to be metabolized to mainly l-menthol glucuronide 10,13,40 and to acetate, 5,20,35 respectively. Metabolites liberated from ester promoiety should be safe or low toxic. Indeed, acetaldehyde is a low toxic fragment which is rapidly detoxified, and F I G U R E 7 Hydrolytic conversion profile of MGS0274. MGS0274 is designed to liberate MGS0008, l-menthol, acetaldehyde, and carbon dioxide, and this reaction is considered to be rapidly initiated by enzymatic hydrolysis and followed by spontaneous cleavage. This concept was supported by the in vitro metabolite profiling study as shown in Figure 2 l 3 2 there are many marketed prodrugs which release acetaldehyde via hydrolysis of their ester promoiety. 3 On the other hand, l-menthol has not been used as ester promoiety before. However, the systemic exposure level of l-menthol after dosing of MGS0274 besylate is expected to be much lower than that after dosing of l-menthol at 4 mg/kg, which is an acceptable daily intake 39  Consequently, we confirmed that MGS0274 was rapidly and extensively hydrolyzed to the parent compound MGS0008 after gastrointestinal absorption, and MGS0008 was distributed throughout the body without further metabolism and then excreted renally in the animal species that were tested. In monkeys, compared with MGS0008, an oral dose of MGS0274 besylate successfully improved the bioavailability of MGS0008 by up to approximately 20-fold, with the prolonged plasma t 1/2 of MGS0008.
Furthermore, the hydrolytic activities against MGS0274 in the liver S9 fraction were comparable between monkeys and humans. Thus, MGS0274 besylate is expected to function properly as a prodrug and to exhibit a preferable pharmacokinetic profile for MGS0008 in humans, as it does in monkeys. Based on the preclinical data obtained, MGS0274 besylate has entered phase 1 clinical trials.

ACK N OWLED G EM ENTS
We acknowledge Dr Takashi Hashihayata and Mr Hiroki Urabe in Taisho Pharmaceutical Co., Ltd. for synthesizing MGS0274 besylate, MGS0008, their stable-isotope labeled compounds, and MGS0039.
In addition, we thank Mr Shigeji Jingu for critical reading of the manuscript.

D I SCLOS U R E S
The authors declare that there is no conflict of interest.

S U PP O RTI N G I N FO R M ATI O N
Additional supporting information may be found online in the Supporting Information section at the end of the article.