Gavestinel sodium salt

• Molecular Formula: C18H11Cl2N2NaO3
• Molecular Weight: 397.2 g/mol
• InChI: InChI=1S/C18H12Cl2N2O3.Na/c19-10-8-13(20)16-12(17(18(24)25)22-14(16)9-10)6-7-15(23)21-11-4-2-1-3-5-11;/h1-9,22H,(H,21,23)(H,24,25);/q;+1/p-1/b7-6+
• InChIKey: GRSDSTMFQHAESM-UHDJGPCESA-M

Gavestinel Sodium Salt (GV150526A): A High-Affinity Glycine-Site NMDA Receptor Antagonist for Stroke and Excitotoxicity Research

Gavestinel sodium salt (GV150526A; CAS 153436-38-5) is a non-competitive antagonist at the strychnine-insensitive glycine binding site of the N-methyl-D-aspartate (NMDA) receptor-channel complex [1]. It belongs to the 2-carboxyindole chemical series developed by GlaxoWellcome and was advanced to Phase III clinical trials as a neuroprotective agent for acute ischemic stroke [2]. Gavestinel binds the glycine site with nanomolar affinity (pKi = 8.5; Kd = 0.8 nM) and is distinguished from many earlier glycine-site ligands by its high >1,000-fold selectivity over glutamate recognition sites on NMDA, AMPA, and kainate receptors [1], as well as its oral bioavailability [3].

Research Model

Ischemic stroke and excitotoxicity model studies

Mechanism

Glycine-site NMDA receptor antagonist tool compound

Key Attribute

Reported high selectivity over glutamate recognition sites

Dosing Route

Supports oral and intravenous administration studies

• [1] NCATS Inxight Drugs – GAVESTINEL. National Center for Advancing Translational Sciences (NCATS). View Source
• [2] Di Fabio R, Cugola A, Donati D, et al. Identification and pharmacological characterization of GV150526, a novel glycine antagonist as potent neuroprotective agent. Drugs Fut 1998; 23: 61-69. View Source
• [3] BindingDB BDBM50010475 – GAVESTINEL (GV-150526X). Affinity Data: Ki = 8.5 nM. View Source

Why Gavestinel Sodium Salt Cannot Be Replaced by Other Glycine-Site NMDA Antagonists in Research Procurement

Glycine-site NMDA antagonists are not a homogeneous class; they divide into at least two pharmacological subclasses based on their binding characteristics at NR1 versus NR1/NR2 subunit combinations, with GV150526A and L-689,560 grouping in one class (two-site binding model at heteromeric receptors) and GV196,771A in another [1]. Beyond binding pharmacology, clinically meaningful differentiation arises from selectivity margins over non-NMDA glutamate receptors, oral bioavailability, CNS penetrance, and therapeutic window in ischemia models — parameters that vary by orders of magnitude even among close structural analogs within the indole-2-carboxylate series [2]. Substituting gavestinel with an alternative glycine-site antagonist without verifying these specific parameters risks introducing confounding off-target effects (e.g., AMPA/kainate antagonism at higher concentrations), inadequate brain exposure, or a truncated therapeutic window, any of which can invalidate experimental conclusions in neuroprotection or excitotoxicity studies.

Subclass Pharmacology

Glycine-site antagonists differ in NR1/NR2 binding profiles; GV150526A and L-689,560 cluster apart from GV196,771A, and interchange may shift receptor modulation context.

Selectivity Erosion

Quinoxalinedione-based antagonists such as licostinel lose selectivity at concentrations above 10 µM and block AMPA/kainate receptors, introducing off-target effects that are absent with gavestinel's indole-2-carboxylate scaffold.

Dosing and Exposure

Infusion-dependent analogs (e.g., ZD9379) require continuous delivery and have limited post-occlusion windows; oral bioavailability varies widely among indole-2-carboxylate analogs, making direct substitution without exposure verification risky.

• [1] Chopra B, Chazot PL, Stephenson FA. Characterization of the binding of two novel glycine site antagonists to cloned NMDA receptors: evidence for two pharmacological classes of antagonists. Br J Pharmacol 2000; 130: 65-72. DOI: 10.1038/sj.bjp.0703298. View Source
• [2] Di Fabio R, Conti N, De Magistris E, et al. Substituted analogues of GV150526 as potent glycine binding site antagonists in animal models of cerebral ischemia. J Med Chem 1999; 42(18): 3486-93. DOI: 10.1021/jm980576n. View Source

Quantitative Differentiation Evidence for Gavestinel Sodium Salt Versus Closest Glycine-Site NMDA Antagonist Comparators

Glycine-Site Binding Affinity: Gavestinel pKi 8.5 vs. Licostinel IC50 59 nM vs. MDL 100,748 Ki 140 nM

Gavestinel (GV150526A) displays a glycine-site affinity of pKi = 8.5 (Ki ≈ 3.2 nM) in [3H]glycine displacement assays on rat cortical membranes, with an independently determined Kd of 0.8 nM [1][2]. By comparison, the quinoxalinedione licostinel (ACEA 1021) exhibits an IC50 of 59 nM at the glycine receptor , representing an ~18-fold lower potency. MDL 100,748, another glycine-site antagonist, has a Ki of 140 nM (0.14 μM) , approximately 44-fold weaker than gavestinel. 5,7-Dichlorokynurenic acid (5,7-DCKA) shows a Ki of 79 nM [4], approximately 25-fold weaker. The parent scaffold kynurenic acid is substantially weaker still, with affinity in the micromolar range. All measurements were obtained under comparable [3H]glycine or [3H]MDL-105,519 displacement conditions using rodent brain membrane preparations.

Glycine-site Affinity

Reported

pKi 8.5

Ki ≈ 3.2 nM, Kd 0.8 nM

Supports glycine-site selective blockade at low concentrations

vs. licostinel IC50 59 nM, MDL 100,748 Ki 140 nM; rat cortical membranes

Evidence Dimension	Binding affinity at the strychnine-insensitive glycine site of the NMDA receptor
Target Compound Data	pKi = 8.5 (Ki ≈ 3.2 nM); Kd = 0.8 nM
Comparator Or Baseline	Licostinel (ACEA 1021) IC50 = 59 nM; MDL 100,748 Ki = 140 nM; 5,7-DCKA Ki = 79 nM; Kynurenic acid > 1 μM
Quantified Difference	Gavestinel ~18-fold higher affinity than licostinel; ~44-fold higher than MDL 100,748; ~25-fold higher than 5,7-DCKA
Conditions	[3H]glycine or [3H]MDL-105,519 displacement in rat cortical/hippocampal membrane preparations

Why This Matters

Higher target affinity enables lower experimental concentrations, reducing the risk of off-target binding at the glutamate recognition site or other receptors, which is critical for mechanistic studies where selective glycine-site blockade must be isolated from AMPA/kainate antagonism.

• [1] NCATS Inxight Drugs – GAVESTINEL. National Center for Advancing Translational Sciences (NCATS). View Source
• [2] BindingDB BDBM50010475 – GAVESTINEL (GV-150526X). Affinity Data: Ki = 8.5 nM. View Source
• [3] Bertin Bioreagent – Gavestinel (sodium salt) CAT N°: 32845. Kis = 0.003, 10, 7.94, and 3.98 μM for glycine site, NMDA glutamate, AMPA, and kainate, respectively. View Source
• [4] McNamara D, Smith E, Calligaro D, et al. Activity of 5,7-dichlorokynurenic acid, a potent antagonist at the N-methyl-D-aspartate receptor-associated glycine binding site. Mol Pharmacol 1990; 38(4): 506-11. PMID: 2147422. View Source

Glutamate Receptor Selectivity: Gavestinel >1,000-Fold Discrimination vs. ACEA 1021 Concentration-Dependent Loss of Selectivity

Gavestinel binds with Kis of 0.003 μM (glycine site), 10 μM (NMDA glutamate site), 7.94 μM (AMPA), and 3.98 μM (kainate), yielding selectivity ratios of ~3,333-fold over the NMDA glutamate site, ~2,647-fold over AMPA, and ~1,327-fold over kainate [1]. It is also inactive (IC50 > 10 μM) against a panel of 70 additional receptors, transporters, and ion channels [1]. In head-to-head comparative studies using rat neocortical slice electrophysiology, the widely used glycine-site antagonist ACEA 1021 (licostinel) showed a markedly narrower selectivity window: at 10 μM, it reduced NMDA responses to 27.9% of control and kainate responses to 79.9% of control, but at 30 μM all NMDA-, AMPA-, and kainate-induced responses were inhibited [3]. In vivo, ACEA 1021 at 10 mg/kg i.v. failed to discriminate between NMDA- and AMPA-mediated excitations in the CA1 region [3]. This concentration-dependent erosion of selectivity is a documented liability of the quinoxalinedione scaffold that is absent from the indole-2-carboxylate chemotype of gavestinel.

Glutamate Receptor Selectivity

Reported

>1,000-fold selectivity

Gly site vs. NMDA/AMPA/kainate

Supports selectivity profiling over non-NMDA receptors

Licostinel loses selectivity above 10 µM; gavestinel selectivity context maintained

Evidence Dimension	Selectivity ratio: glycine site Ki vs. glutamate recognition site Ki at NMDA, AMPA, and kainate receptors
Target Compound Data	Kis: glycine site 0.003 μM; NMDA 10 μM; AMPA 7.94 μM; kainate 3.98 μM. Selectivity ratios: 3,333× (NMDA), 2,647× (AMPA), 1,327× (kainate). Inactive against 70 receptors (IC50 > 10 μM)
Comparator Or Baseline	ACEA 1021 (licostinel): at 10 μM, NMDA response reduced to 27.9%, kainate to 79.9%, AMPA increased by 2.4%. At 30 μM, all NMDA-, AMPA-, and kainate-induced responses were inhibited. In vivo at 10 mg/kg i.v., no clear discrimination between NMDA and AMPA excitation
Quantified Difference	Gavestinel maintains >1,000-fold selectivity across all tested concentrations; ACEA 1021 loses functional selectivity above 10 μM in vitro and at behaviorally active doses in vivo
Conditions	[3H]ligand displacement binding assays (rat brain membranes) for target; neocortical slice electrophysiology and in vivo CA1 recordings for comparator

Why This Matters

For researchers requiring selective glycine-site blockade without confounding AMPA/kainate receptor modulation, gavestinel provides an experimentally cleaner tool than quinoxalinedione-based antagonists, particularly in studies using concentrations above 10 μM or requiring in vivo dosing.

• [1] Bertin Bioreagent – Gavestinel (sodium salt) CAT N°: 32845. Kis = 0.003, 10, 7.94, and 3.98 μM for glycine site, NMDA glutamate, AMPA, and kainate, respectively. View Source
• [2] NCATS Inxight Drugs – GAVESTINEL. Displays >1,000-fold selectivity over NMDA, AMPA and kainate binding sites. View Source
• [3] Lingenhöhl K, Pozza MF, Olpe HR. Reevaluation of ACEA 1021 as an antagonist at the strychnine-insensitive glycine site of the N-methyl-D-aspartate receptor. Neuropharmacology 1998; 37(6): 729-37. DOI: 10.1016/s0028-3908(98)00062-8. View Source

Therapeutic Window in Focal Ischemia: Gavestinel Single Bolus at 6 h vs. ZD9379 Infusion Protocol Required for Efficacy

In the rat middle cerebral artery occlusion (MCAO) model of focal ischemia, a single intravenous bolus of gavestinel at 3 mg/kg administered up to 6 hours post-occlusion produced a statistically significant reduction in infarct volume measured histologically at 24 hours (P < 0.05 vs. MCAO control), with concomitant protection of somatosensory evoked potential (SEP) responses [1][2]. By contrast, the glycine-site antagonist ZD9379 required a more complex dosing regimen to achieve efficacy in the same model: a 5 mg/kg i.v. bolus initiated either 30 minutes before MCAO or 30 minutes after MCAO, followed by a continuous 5 mg/kg/h infusion for 4 hours, to significantly reduce infarct volume (P < 0.001 vs. vehicle) [3]. No ZD9379 data support efficacy with dosing delayed to 6 hours post-occlusion. This distinction — a single bolus with 6-hour flexibility versus an infusion-dependent protocol with a 30-minute window — represents a practical and mechanistic advantage for gavestinel in experimental stroke research.

Delayed-Dosing Window

Reported

Single bolus up to 6 h post-MCAO effective

Rat permanent MCAO model; infarct volume reduction

Supports delayed-intervention model designs

Infusion-dependent ZD9379 effective only ≤30 min post-occlusion; cross-study review

Evidence Dimension	Therapeutic time window in the rat permanent MCAO model: dosing delay from occlusion to effective treatment
Target Compound Data	Gavestinel 3 mg/kg i.v. single bolus, effective when administered up to 6 h post-MCAO; significant infarct volume reduction (P < 0.05)
Comparator Or Baseline	ZD9379: 5 mg/kg i.v. bolus + 5 mg/kg/h infusion for 4 h, effective when bolus started 30 min pre-MCAO or 30 min post-MCAO only; no data for delayed monotherapy
Quantified Difference	Gavestinel: single injection, 6 h window. ZD9379: infusion-requiring, ≤30 min window (pre- or early post-occlusion). Approximate 12-fold longer therapeutic window for gavestinel
Conditions	Rat permanent MCAO model; infarct volume measured by histology (TTC staining or equivalent) at 24 h post-occlusion

Why This Matters

A wide and flexible therapeutic window with simple bolus dosing enables experimental protocols that more closely model clinically realistic treatment delays; this is a key selection criterion for translational stroke researchers choosing between glycine-site antagonist tool compounds.

• [1] Bordi F, Pietra C, Ziviani L, et al. The glycine antagonist GV150526 protects somatosensory evoked potentials and reduces the infarct area in the MCAo model of focal ischemia in the rat. Exp Neurol 1997; 145(2 Pt 1): 425-33. DOI: 10.1006/exnr.1997.6442. View Source
• [2] Di Fabio R, Conti N, De Magistris E, et al. Substituted analogues of GV150526 as potent glycine binding site antagonists in animal models of cerebral ischemia. J Med Chem 1999; 42(18): 3486-93. [Confirms 6 h therapeutic window for GV150526A]. View Source
• [3] Tatlisumak T, Takano K, Meiler MR, Fisher M. A glycine site antagonist, ZD9379, reduces number of spreading depressions and infarct size in rats with permanent middle cerebral artery occlusion. Stroke 1998; 29(1): 190-5. DOI: 10.1161/01.str.29.1.190. View Source

Safety Profile: Gavestinel Lacks Neuronal Vacuolization and Cognitive Impairment — Direct Head-to-Head Comparison with Dizocilpine (MK-801)

In a direct within-study comparison, gavestinel at 200 mg/kg i.v. (a dose 67-fold above the neuroprotective dose of 3 mg/kg) did not induce neuronal vacuolization in the posterior cingulate/retrosplenial cortex of rats, whereas dizocilpine (MK-801) at 1 mg/kg s.c. produced clear vacuolization [1]. In the same study, gavestinel at 3 mg/kg i.v. did not block perforant path-dentate gyrus long-term potentiation (LTP), while MK-801 at 1 mg/kg i.p. abolished LTP of both EPSP slope and population spike amplitude [1]. In the Morris Water Maze, rats receiving gavestinel at 10 mg/kg or 60 mg/kg p.o. showed no impairment in spatial learning compared to controls, whereas MK-801 at 0.08 mg/kg i.p. significantly impaired platform location [1]. These findings demonstrate that gavestinel, at doses far exceeding those required for neuroprotection, is devoid of the psychotomimetic, amnestic, and neurotoxic effects that are hallmark liabilities of NMDA channel blockers such as MK-801.

CNS Tolerability Endpoints

Head-to-head

No vacuolization, LTP block, or spatial learning deficits

Gavestinel 200 mg/kg i.v.; MK-801 1 mg/kg induced all

Supports behavioral endpoint interpretation without channel blocker confounds

Safety margin >67-fold above effective dose; rat in vivo models

Evidence Dimension	Induction of neuronal vacuolization, LTP blockade, and spatial learning impairment in rats
Target Compound Data	Gavestinel: 200 mg/kg i.v. — no vacuolization; 3 mg/kg i.v. — no LTP blockade; 10-60 mg/kg p.o. — no Morris Water Maze impairment
Comparator Or Baseline	MK-801 (dizocilpine): 1 mg/kg s.c. — vacuolization present; 1 mg/kg i.p. — LTP blockade; 0.08 mg/kg i.p. — significant Morris Water Maze impairment
Quantified Difference	Gavestinel: safety margin >67-fold above neuroprotective dose for vacuolization, >200-fold for cognitive endpoints. MK-801: adverse effects at or below neuroprotective doses (therapeutic index ~1)
Conditions	In vivo rat studies: histopathology (posterior cingulate/retrosplenial cortex), in vivo electrophysiology (perforant path-dentate gyrus LTP), Morris Water Maze behavioral paradigm

Why This Matters

For laboratories conducting in vivo neuroprotection or behavioral pharmacology studies, gavestinel's wide safety margin eliminates the confounding behavioral toxicity that complicates interpretation of experiments using MK-801 or other NMDA channel blockers, enabling cleaner assessment of glycine-site-mediated neuroprotection.

• [1] Bordi F, Terron A, Reggiani A. The neuroprotective glycine receptor antagonist GV150526 does not produce neuronal vacuolization or cognitive deficits in rats. Eur J Pharmacol 1999; 378(2): 153-60. DOI: 10.1016/s0014-2999(99)00462-8. View Source

Oral Bioavailability: Gavestinel ED50 6 mg/kg p.o. In Vivo vs. GV196,771A <10% Oral Bioavailability

Gavestinel is orally active in vivo: in the NMDA-induced convulsion model in mice, it exhibits an ED50 of 6 mg/kg by the oral route, compared with an ED50 of 0.06 mg/kg by the intravenous route [1]. Pharmacokinetic studies in rat confirm oral bioavailability of up to 26%, with the compound predominantly excreted via the bile [3]. By contrast, GV196,771A — a structurally related indole-2-carboxylate glycine-site antagonist developed by the same group — shows oral bioavailability below 10% in both rat and mouse, with a plasma clearance of approximately 2 mL/min/kg [2]. This intra-series difference demonstrates that oral bioavailability among indole-2-carboxylate glycine antagonists is highly sensitive to structural modification and cannot be assumed across analogs. Other glycine-site antagonist scaffolds, such as certain quinoxalinediones and kynurenic acid derivatives, have negligible oral bioavailability or require prodrug strategies [2].

Oral Bioavailability

Reported

ED50 p.o. 6 mg/kg, F% up to 26%

Mouse NMDA convulsion assay; rat PK

Supports oral administration study designs

GV196,771A oral bioavailability

Evidence Dimension	Oral bioavailability and oral in vivo potency in rodent models
Target Compound Data	Gavestinel: ED50 = 6 mg/kg p.o. (NMDA convulsion model, mice); oral bioavailability up to 26% (rat). ED50 i.v. = 0.06 mg/kg
Comparator Or Baseline	GV196,771A: oral bioavailability <10% (rat and mouse); plasma clearance ~2 mL/min/kg. Quinoxalinediones generally have poor oral bioavailability
Quantified Difference	Gavestinel oral bioavailability ~2.6-fold higher than GV196,771A; gavestinel is orally active in behavioral models at 6 mg/kg, whereas GV196,771A requires higher doses or alternative routes to achieve comparable CNS exposure
Conditions	Mouse NMDA-induced convulsion assay (ED50); rat pharmacokinetic studies (oral vs. i.v. administration, [14C]-labeled compound)

Why This Matters

Oral bioavailability enables chronic dosing paradigms in preclinical models (e.g., neuropathic pain, epilepsy) that are impractical with compounds requiring continuous intravenous infusion, making gavestinel the preferred glycine-site antagonist tool for studies requiring repeated oral administration.

• [1] NCATS Inxight Drugs – GAVESTINEL. GV150526A inhibited convulsions induced by NMDA in mice by both IV and po routes (ED50 = 0.06 and 6 mg/kg, respectively). View Source
• [2] TargetMol – GV-196771A. GV196771A shows low oral bioavailability in rat and mouse (<10%), with plasma clearance of ~2 mL/min/kg. View Source
• [3] Kajbaf M, Barnaby RJ, Bottacini M, et al. Pharmacokinetics, metabolism and excretion of the glycine antagonist GV150526A in rat and dog. Xenobiotica 2003; 33(4): 415-28. DOI: 10.1080/0049825031000072469. View Source

Highest-Value Application Scenarios for Gavestinel Sodium Salt Based on Verified Differentiation Evidence

Selective Glycine-Site NMDA Receptor Pharmacology in Recombinant and Native Systems

Gavestinel's nanomolar glycine-site affinity (pKi 8.5) combined with >1,000-fold selectivity over glutamate recognition sites [1][4] makes it an optimal pharmacological tool for isolating glycine-dependent NMDA receptor signaling in electrophysiological and calcium-imaging experiments. Unlike ACEA 1021, which antagonizes AMPA and kainate receptors at concentrations above 10 μM [1], gavestinel can be used across a wide concentration range (0.03–0.35 μM for functional blockade of glutamate-induced Ca2+ increases) with minimal risk of non-NMDA receptor interference. This is particularly critical in mixed neuronal cultures or slice preparations where multiple glutamate receptor subtypes are co-expressed.

Preclinical Stroke Research Requiring Clinically Relevant Therapeutic Window

Gavestinel's demonstrated ability to reduce infarct volume when administered as a single 3 mg/kg i.v. bolus up to 6 hours after MCAO onset [2] addresses a pervasive challenge in translational stroke research: most neuroprotective candidates show efficacy only when given before or within 1–2 hours of occlusion. Researchers modeling treatment delays that mimic clinical stroke care timelines (where median time-to-treatment in the GAIN trials was 5.2 hours) will find gavestinel uniquely suited among glycine-site antagonists for protocols with extended post-ischemia intervention windows.

In Vivo Neuroprotection Studies Requiring Absence of Confounding Behavioral Toxicity

For in vivo studies where behavioral endpoints (motor function, cognitive performance) are primary readouts, gavestinel's lack of neuronal vacuolization, LTP impairment, and spatial learning deficits — even at doses 67-fold above the neuroprotective range [3] — directly addresses a core limitation of NMDA channel blockers such as MK-801. This safety profile enables experimental designs where neuroprotection can be assessed without the interpretive confound of drug-induced sedation, ataxia, or amnesia. The Phase III GAIN clinical data further confirm that gavestinel's safety profile translates to humans, with no serious safety issues reported in over 1,300 patients [4].

Oral Dosing Paradigms for Chronic NMDA Receptor Modulation in Rodent Models

Gavestinel's oral bioavailability (ED50 = 6 mg/kg p.o. in the NMDA convulsion assay) [4] and favorable pharmacokinetic profile (low clearance, biliary excretion) [2] support repeated oral dosing in chronic disease models — a practical advantage over infusion-dependent glycine antagonists like ZD9379 and over poorly bioavailable indole-2-carboxylate analogs such as GV196,771A (<10% oral bioavailability). This makes gavestinel the compound of choice for multi-day or multi-week rodent studies in neuropathic pain, epilepsy, or neurodegeneration where sustained NMDA receptor glycine-site blockade is desired.

Application

Selection Property

Validation Focus

Glycine-site NMDA receptor pharmacology studies

Reported glycine-site affinity and selectivity profile

Off-target AMPA/kainate receptor interference profiling

Cerebral ischemia model studies

Reported delayed-dosing window in MCAO model

Infarct volume and electrophysiological endpoint reproducibility

In vivo CNS pharmacology without behavioral confounds

Reported CNS tolerability endpoint context (vacuolization, LTP, spatial learning)

Behavioral endpoint interpretation independent of NMDA channel blocker effects

Oral administration studies for sustained NMDA modulation

Reported oral exposure and CNS bioavailability

Chronic dosing feasibility and PK/PD model validation

• [1] Bertin Bioreagent – Gavestinel (sodium salt) CAT N°: 32845. Selectivity data and functional Ca2+ inhibition in rat cortical neurons (0.03-0.35 μM). View Source
• [2] Bordi F, Pietra C, Ziviani L, et al. The glycine antagonist GV150526 protects somatosensory evoked potentials and reduces the infarct area in the MCAo model of focal ischemia in the rat. Exp Neurol 1997; 145(2 Pt 1): 425-33. DOI: 10.1006/exnr.1997.6442. View Source
• [3] Bordi F, Terron A, Reggiani A. The neuroprotective glycine receptor antagonist GV150526 does not produce neuronal vacuolization or cognitive deficits in rats. Eur J Pharmacol 1999; 378(2): 153-60. DOI: 10.1016/s0014-2999(99)00462-8. View Source
• [4] NCATS Inxight Drugs – GAVESTINEL. ED50 values, selectivity, GAIN trial summary, oral bioavailability. View Source
• [5] Kajbaf M, Barnaby RJ, Bottacini M, et al. Pharmacokinetics, metabolism and excretion of the glycine antagonist GV150526A in rat and dog. Xenobiotica 2003; 33(4): 415-28. DOI: 10.1080/0049825031000072469. View Source