Modulation of glycine transporters as a novel therapeutic strategy in neuropsychiatry

Abstract

Glycine plays a dual role in the central nervous system as a fast inhibitory neurotransmitter and a co-agonist at N-methyl-D-aspartate receptors (NMDARs). Its extracellular levels are tightly regulated by two glycine transporters (GlyTs): GlyT1, which modulates glycine near excitatory synapses to influence glutamatergic transmission, and GlyT2, which sustains presynaptic glycine for inhibitory signalling. Dysregulation of GlyT function has been linked to numerous neurological and psychiatric disorders, including schizophrenia, mood and anxiety disorders, neurodegeneration, epilepsy, stroke, addiction, and pain. This review examines recent preclinical and clinical progress in targeting GlyTs, with an emphasis on GlyT1 inhibition to enhance NMDAR function. Among GlyT1 inhibitors, sarcosine shows consistent promise, particularly for schizophrenia and depressive symptoms. However, the limited clinical success of other compounds underscores challenges in translating preclinical efficacy. Addressing issues such as selectivity, patient stratification, and novel regulatory mechanisms will be key. Future research leveraging imaging biomarkers and next-generation pharmacological tools may help unlock the full therapeutic value of GlyTs.

Introduction

Glycine is the structurally simplest amino acid with key roles in the central nervous system (CNS), functioning both as a primary fast-acting inhibitory neurotransmitter and as a modulator of excitatory neurotransmission. Its extracellular concentration within the synaptic milieu, is tightly regulated by high-affinity glycine transporters (GlyTs), including GlyT1 and GlyT2, which represent important therapeutic targets for modulating glycine’s function and related behavioral outcomes.

GlyT1 and GlyT2 both belong to the SLC6 family of sodium- and chloride-dependent neurotransmitter transporters. While they exhibit around 50% similarity in their nucleotide and amino acid sequences, they have different stoichiometries (Supplisson and Roux 2002), they are differentially distributed across the central nervous system (Guastella et al. 1992; López-Corcuera et al. 1998) and they fulfil distinct physiological roles (Liu et al. 1993; Ponce et al. 1998; Smith et al. 1992).

GlyTs are localized in neurons and astrocytes surrounding synaptic regions. At glycinergic inhibitory synapses, GlyT1 plays a key role in terminating synaptic transmission, while GlyT2 facilitates glycine reuptake into presynaptic terminals for reuse. Additionally, at excitatory synapses, GlyT1 helps regulate extracellular glycine levels in proximity to N-methyl-D-aspartate receptors receptors (NMDAR).

Current research suggests that GlyTs are involved in the pathophysiology of a wide range of neurological and psychiatric conditions, including among others, schizophrenia, neuropathic pain, substance use disorders, epilepsy, depression, stroke, and several neurodegenerative diseases (Cioffi and Guzzo 2016; Gallagher et al. 2022; Lemes Marques et al. 2020; Piniella and Zafra 2023; Söderpalm et al. 2017). GlyTs modulators, particularly targeting GlyT1, have emerged as valuable therapeutic options in these disorders, through pharmacological inhibition strategies aimed at enhancing glycine signalling, improving NMDAR activity.

Although GlyT1 inhibitors act in a similar manner, they can belong to diverse structural classes (Lechner 2006; Lindsley et al. 2006), may exert either competitive on noncompetitive inhibition, can be reversible or non-reversible, and may be selective for GlyT1 only or act also on GlyT2. For example, while sarcosine-based compounds (NFPS, (R)-NPTS and Org24589) appear to be irreversible and non-competitive selective inhibitors GlyT1, non-sarcosine-based GlyT1 inhibitors (PF 03463275, SSR504734 and N-methyl-SSR504734) appear to be potent, reversible, selective and orthosteric inhibitors of GlyT1 (Mezler et al. 2008).

However, the therapeutic potential of GlyT modulation remains inconclusive, with some studies reporting significant benefits and others failing to demonstrate clinical efficacy.

In the present review, we comprehensively examine the most recent preclinical and clinical findings on the therapeutic modulation of GlyTs in neuropsychiatric and neurodegenerative conditions, with a particular focus on schizophrenia, mood and anxiety disorders, and major neurodegenerative diseases such as Alzheimer’s and Parkinson’s disease.

Expression, characterization, and function of glycine, its receptors, and glycine transporters in the brain

Although glycine is widely distributed throughout the CNS, its highest concentrations are found in the spinal cord, brainstem, and retina (Aprison and Werman 1965; Werman et al. 1968). In these regions, glycine mediates fast inhibitory transmission by binding to the strychnine-sensitive glycine-A site on ionotropic glycine receptors (GlyRs), leading to chloride ion influx and hyperpolarization of the postsynaptic membrane (Lynch 2009).

Beyond its inhibitory role, glycine also exerts essential excitatory functions. It acts as a co-agonist at NMDARs, where it is required for receptor activation alongside glutamate (Johnson and Ascher 1987). Glycine also functions as an agonist at the excitatory GluN1/GluN3A receptor complex (Grand et al. 2018) and at the metabotropic G protein–coupled receptor GPR158 (Laboute et al. 2023).

The extracellular concentration of glycine within the synaptic cleft is accurately regulated by high-affinity GlyTs, specifically GlyT1 and GlyT2, which are respectively encoded by SLC6A9 and SLC6A5 genes (Jones et al. 1995; Morrow et al. 1998). GlyT1 is broadly expressed throughout the CNS (Adams et al. 1995; Zafra et al., 1995a), and it is predominantly found on glial cells, particularly astrocytes, in the spinal cord (Cubelos et al. 2005; Jursky and Nelson 1995), on both glial cells and subpopulations of glutamatergic neurons in forebrain regions (Cubelos et al. 2005) and in selected amacrine and ganglion neurons in the retina (Eulenburg et al. 2018; Pow and Hendrickson 1999). GlyT1 is present on both pre- and postsynaptic sites of glutamatergic synapses, where it co-localizes with NMDA receptors (Cubelos et al. 2005). In addition, GlyT1 has been detected also in synaptic vesicles of hippocampal areas (Cubelos et al. 2014). Positron emission tomography (PET) studies confirmed the expression of GlyT1 in glutamatergic pathways of both rodents and primates (Fuchigami et al. 2011; Herdon et al. 2010; Hoffmann et al. 2021; Passchier et al. 2010; Zeng et al. 2008).

On the other hand, GlyT2 expression is limited to glycinergic neurons in regions such as the dorsal and ventral horns from the spinal cord, brainstem, cerebellum, nuclei of cranial nerves and in the auditory system (Jursky and Nelson 1995, 1996; Lemes Marques et al. 2020; Zafra et al., 1995a). Multiple studies have shown that GlyT2 is commonly located near strychnine-sensitive GlyRs (Araki et al. 1988; Kolston et al. 1992; Ottersen et al. 1990; Van den Pol and Gorcs 1988). Glycine transport by GlyT proteins is driven by a sodium gradient, which is tightly regulated by the activity of the Na⁺/K⁺-ATPase pump. The coordinated binding of glycine, sodium, and chloride ions triggers a conformational shift in the transporter, transitioning it from an outward-facing to an inward-facing state. This structural rearrangement allows the glycine-binding pocket to face the intracellular space, enabling the release of glycine along with sodium and chloride ions into the cytosol (Martínez-Maza et al. 2001). Functional distinctions between GlyT1 and GlyT2 arise from differences in their voltage-dependent conformational behaviors and ion flux characteristics: GlyT1 transport kinetics are primarily influenced by glycine binding to the outward-facing configuration, whereas GlyT2 activity is largely governed by sodium ion interactions at the binding site (Erdem et al. 2019).

GlyT1 and GlyT2 serve distinct yet complementary roles in regulating glycine availability across synaptic and extrasynaptic compartments of the CNS (Eulenburg et al. 2005). While GlyT2 primarily mediates presynaptic glycine reuptake to sustain inhibitory transmission, GlyT1 modulates glycine levels in glial cells and near excitatory synapses, thereby influencing NMDAR activation and overall neurotransmitter homeostasis (Zafra et al. 2017; Harvey and Yee 2013).

Evidence from knockout mouse studies indicates that the complete loss of either GlyT1 or GlyT2 function leads to lethality (Gomeza et al., 2003 A; Gomeza et al., 2003B), predominantly due to excessive glycinergic signalling in the case of GlyT1 deficiency, or insufficient glycinergic inhibition in GlyT2-null mice.

One of the critical functions of GlyT1 consists in indirectly modulating NMDAR activity and glutamatergic neurotransmission (Singer and Yee et al., 2024). Indeed, GlyT1 regulates the availability of glycine at the co-agonist site of NMDAR (Supplisson 2024), which is necessary for receptor activation (Kleckner and Dingledine 1988). Moreover, GlyT1 also acts to prevent glycine spillover beyond the synaptic cleft. By tightly regulating extrasynaptic glycine concentrations, GlyT1 limits the activation of extrasynaptic NMDA receptors, and avoids potential excitotoxicity or interference with glutamatergic signalling. Such process appears in fact to be implicated in excitotoxicity and synaptic dysregulation. By maintaining spatially confined glycine signaling, GlyT1 contributes to the fidelity of excitatory neurotransmission and supports the functional integrity of glutamatergic networks (Eulenburg and Gomeza 2010; Harvey and Yee 2013; Papouin et al. 2012).

In contrast, GlyT2 plays a fundamental role in inhibitory neurotransmission by maintaining adequate glycine levels for synaptic vesicle refilling in glycinergic neurons, thereby ensuring effective inhibitory signalling (Zafra et al. 2016, 2017; Zafra and Giménez 2008) as demonstrated by the significant deficits in neuronal function in GlyT2 KO mice, due to a reduction in the amount of Gly available for release in glycinergic neurons (Gomeza et al., 2003B). GlyT2 allows the recycling of previously released glycine in the extracellular space, ensuring appropriate levels of glycine for future releases from glycinergic neurons are maintained (Brasnjo and Otis 2003).

Moreover, both GlyT1 and GlyT2 exhibit dynamic expression during neurodevelopment, with distinct and tight expression patterns that reflect their specialized roles in neurogenesis and circuit maturation. GlyT1 is prominently expressed during early embryonic stages, where it is implicated in neural tube formation and the establishment of early neuronal architecture. In contrast, GlyT2 expression emerges later in development, coinciding with the functional maturation of glycinergic inhibitory circuits, particularly within the brainstem and spinal cord (Salceda 2022). Collectively, GlyT1 and GlyT2 serve not only as transporters but also as modulators of synaptic physiology, neurodevelopment, and potential therapeutic targets for neuropsychiatric disorders and pain.

Therapeutic modulation of GlyTs in neuropsychiatric disease

Schizophrenia

Schizophrenia is a multifaceted neuropsychiatric disorder characterized by core clusters of positive symptoms, negative symptoms, disorganized behavior, cognitive impairments, and deficits in social cognition (DSM-5 2013); McCutcheon et al. 2023). In addition to the classical, well-established theories explaining the pathophysiology of this complex disorder, a more recent hypothesis highlights NMDAR hypofunction, particularly in relation to negative (Hu et al. 2015; Javitt et al. 2004; Olney et al. 1999) and cognitive symptoms (Pei et al. 2021; Tsapakis et al. 2023; Yang and Svensson 2008), suggesting that enhancing NMDAR activity could help alleviate these symptoms of schizophrenia. This dysfunctional NMDAR hypothesis is reinforced by post-mortem studies, showing lower NMDAR subunit expression in patients with schizophrenia (Catts et al. 2016) and by multi-stage genome-wide association studies, which identify variants in NMDA signaling genes as risk factors for schizophrenia (Schizophrenia Working Group of the Psychiatric Genomics Consortium 2014). Specific NMDAR genetic variants also associated with cognitive impairments in schizophrenic patients (Ohi et al., 2015).

Consequently, several pharmacological strategies now aim to inhibit GlyT1, which is proposed to elevate synaptic glycine levels by reducing its reuptake, thereby enhancing NMDAR activity indirectly (Cioffi et al., 2016; Chaki et al. 2015; Chang et al. 2020; Fadgyas-Stanculete and Capatina 2025; Hashimoto 2014), as represented in Fig. 1.

Fig. 1.

Structural and Functional Insights into GlyT1 and Sarcosine Modulation. GlyT1 consists of 12 transmembrane (TM) helices, with both the N- and C-termini facing the cytoplasmic side of the membrane. The protein exhibits pseudo two-fold symmetry, where TM helices 1–5 mirror TMs 6–10, aligned along an axis parallel to the membrane plane. Notably, TM1 and TM6 show partial unwinding at their midpoints, intersecting near the symmetry axis. These structural breaks are believed to play a critical role in shaping the ligand-binding site. The binding cavity, which houses both the glycine and associated ion sites, is located midway through the membrane bilayer and is sealed off from both extracellular and intracellular access points. This pocket is primarily formed by residues from TM1, TM3, TM6, and TM8. Sarcosine, an N-methylated analogue of glycine, preferentially targets GlyT1 over GlyT2. It binds at the S1 substrate-binding site, the same location utilized by glycine. Within this pocket, sarcosine is stabilized through multiple polar interactions involving TM1, TM3, TM6, and TM8, and is flanked by sodium and chloride ions that contribute to the binding stability and transporter conformation (Wei et al. 2024). By competing with glycine for this binding site, sarcosine inhibits glycine reuptake, leading to elevated levels of extracellular glycine. This, in turn, enhances NMDAR activity, since glycine functions as a co-agonist at the glycine modulatory site (GMS) located on the NR1 subunit of the NMDAR. Furthermore, evidence suggests that sarcosine may not only act indirectly but could also directly activate NMDARs by interacting with the GMS on NR1, providing a dual mechanism of action (Pei et al. 2019)

This approach may promote receptor function without triggering adverse effects typically associated with direct NMDAR activation, such as seizures or neurotoxicity (Chen et al. 2022; Kinney et al. 2003; Kemp and Leeson 1993; Perry et al., 2008). Chemical structures and a summary of main pharmacological properties of the GlyT1 inhibitors most extensively investigated in the context of schizophrenia are shown in Fig. 2 and Table 1 respectively.

Fig. 2.

Chemical structures of GlyT1 inhibitors extensively studied in the treatment of schizophrenia. This figure illustrates the chemical structures of five key GlyT1 inhibitors that are structurally and mechanistically distinct. Bitopertin (C₂₁H₂₀F₇N₃O₄S) is a sulfonamide-based GlyT1 inhibitor characterized by a trifluoromethyl- and fluorine-substituted pyridine ring, a methylsulfonyl-phenyl group, and a piperazine linker (see Bitopertin | C21H20F7N3O4S | CID 24946690 - PubChem for more molecular information of this compound). Iclepertin (C₂₀H₁₈F₆N₂O₅S) is a fluorinated GlyT1 inhibitor with a trifluoromethyl-substituted oxazole, a methylsulfonyl-trifluoropropyl ether phenyl ring, and a rigid azabicyclo[3.1.0]hexane scaffold (see Iclepertin | C20H18F6N2O5S | CID 155259577 - PubChem for more molecular information of this compound). Sarcosine (C₃H₇NO₂) is an endogenous methylated aminoacetic acid (see Sarcosine | C3H7NO2 | CID 1088 - PubChem for more molecular information of this compound). Org24598 (C₁₉H₂₀F₃NO₃) is an N-methylglycine derivative GlyT1 inhibitor featuring a terminal acetic acid Group, a methylated nitrogen, and a chiral hydrophobic side chain incorporating a phenyl ring and a 4-(trifluoromethyl)phenoxy moiety linked via an ether bond (see Org 24598 | C19H20F3NO3 | CID 5311285 - PubChem for more molecular information of this compound). PF-03463275 (C₁₉H₂₂ClFN₄O) is a GlyT1 inhibitor characterized by an imidazole-4-carboxamide core, methylated at N1, and linked to a chiral azabicyclo[3.1.0]hexane ring via a methylene bridge. An additional substitution occurs at the amide nitrogen with a 3-chloro-4-fluorobenzyl group (see N-[(3-chloro-4-fluorophenyl)methyl]−1-methyl-N-[[(1R,5 S)−3-methyl-3-azabicyclo[3.1.0]hexan-6-yl]methyl]imidazole-4-carboxamide | C19H22ClFN4O | CID 44156901 - PubChem for more molecular information of this compound)

Table 1.

Main molecular features and pharmaceutical properties of five GlyT1 inhibitors relevant in the treatment of schizophrenia. Bitopertin is part of the benzoylpiperazines, and it acts as a potent, noncompetitive, reversible and selective GlyT1 inhibitor, with an IC₅₀ 25 nM (Porter and Dawson 2014). Iclepertin acts as a potent, noncompetitive, reversible and selective GlyT1 inhibitor with an IC₅₀ of 5 nM (Rosenbrock et al. 2018). Sarcosine acts as a weak, competitive, reversible but highly selective substrate/inhibitor of GlyT1, with an IC₅₀ of 91 µM (Porter and Dawson 2014). Org24598 acts as a noncompetitive, irreversible, selective GlyT1 inhibitor with an IC₅₀ of 6.9 nM (Mezler et al. 2008). PF-03463275 acts as a competitive, reversible and selective GlyT1 inhibitor (Wei et al. 2024) and with an IC₅₀ of 32.7 nM (D’Souza et al., 2018)

Inhibitor	Molecular formula	Mol. weight (g/mol)	Binding type	Reversibility	Selectivity	IC₅₀
Bitopertin	C₂₁H₂₀F₇N₃O₄S	543.5	Noncompetitive	Reversible	Selective	25 nM
Iclepertin	C₂₀H₁₈F₆N₂O₅S	512.4	Noncompetitive	Reversible	Selective	5 nM
Sarcosine	C₃H₇NO₂	89.09	Competitive	Reversible	Selective	91 µM
Org24598	C₁₉H₂₀F₃NO₃	367.4	Noncompetitive	Irreversible	Selective	6.9 nM
PF-03463275	C₁₉H₂₂ClFN₄O	376.9	Competitive	Reversible	Selective	32.7 nM

By selectively targeting this mechanism, three compounds in particular have caught the interest of researchers and pharmaceutical companies: bitopertin, iclepertin and sarcosine.

Bitopertin

Bitopertin (RG1678) was the first GlyT1 inhibitor to undergo extensive clinical evaluation for the treatment of schizophrenia, and it initially showed promising results. An 8-week randomized, placebo-controlled, double-blind Phase II study, revealed a greater clinical improvement in schizophrenic patients when bitopertin (10 mg) was added to their current antipsychotic therapy (Umbricht et al. 2014). Similar positive results were found after 52-week adjunctive bitopertin in patients with persistent negative or suboptimally controlled symptoms (Hirayasu et al. 2016), although without establishing a placebo arm. However, subsequent studies failed to support the therapeutic potential of bitopertin. A 4-week, double-blind Phase 2–3 CandleLyte trial evaluated bitopertin monotherapy against placebo and olanzapine in patients with acute schizophrenia. Although a 30 mg dose of bitopertin demonstrated modest improvements in positive symptoms and increased rates of hospital discharge readiness, neither the 10 mg nor 30 mg dose produced statistically significant benefits on the primary outcome measure (Bugarski-Kirola et al. 2014). Similar results were found in the SearchLyte program, comprising three large Phase 3 trials, which evaluated adjunctive bitopertin in patients with schizophrenia and persistent positive symptoms. While one study (NightLyte) showed a modest improvement with the 10 mg dose on positive symptoms, the other trials failed to demonstrate a significant benefit over placebo, suggesting that bitopertin offers at best limited efficacy as an add-on treatment (Bugarski-Kirola et al. 2016). Following the SearchLyte program, three additional Phase 3 trials, SunLyte, DayLyte, and FlashLyte, investigated adjunctive bitopertin for persistent negative symptoms in stable schizophrenia patients. Despite the extended 24-week treatment period, none of the studies showed a significant advantage over placebo, reinforcing the limited clinical efficacy of bitopertin, even in symptom domains previously linked to glutamatergic modulation (Bugarski-Kirola et al. 2017). Following these evidence, the pharmaceutical company Roche discontinued its development for schizophrenia (Keshavan et al. 2017).

Nonetheless, preclinical research continues to highlight bitopertin’s mechanistic promise. In a schizophrenia-relevant mouse model with parvalbumin (PV) interneuron dysfunction, neurons that play a key role in maintaining excitatory/inhibitory (E/I) balance, a network feature disrupted in multiple neuropsychiatric conditions including schizophrenia and bipolar disorder (Pinna and Colasanti 2021), bitopertin was shown to restore E/I balance and improve cortical network function (Chen-Engerer et al. 2022). Additionally, bitopertin improved recognition memory, and attenuated NMDAR antagonist-induced working memory deficits in rodents, further supporting its potential to target cognitive impairments associated with NMDAR hypofunction (Deiana et al. 2022).

Iclepertin

Iclepertin (BI 425809) is another highly potent and selective GlyT1 inhibitor (Rosenbrock et al. 2018). In rodent models, it reduced MK-801-induced deficits in sensory processing and neural circuit function, and improved both social recognition and working memory (Rosenbrock et al. 2022). Notably, a Phase 2 trial (NCT02832037) demonstrated cognitive enhancement following 12 weeks of treatment (Fleischhacker et al. 2021). These promising results led the FDA to Grant iclepertin breakthrough therapy designation for schizophrenia in 2021. However, although data from another phase 2 trial (NCT03859973) confirmed its efficacy and safety in 2024 (Harvey et al. 2024), in 6-month Phase III trials (the CONNEX program), iclepertin did not meet the primary and key secondary endpoints, showing no statistically significant improvement in cognition or functional outcomes compared to placebo (NCT04846868; NCT04846881; NCT04860830), deciding to discontinue the long-term extension trial, CONNEX-X, effective immediately (NCT05211947; Update Phase III CONNEX clinical program schizophrenia | Boehringer Ingelheim).

Sarcosine

Sarcosine is another proven selective GlyT1 inhibitor (Guastella et al. 1992; Zhang et al. 2009). The positive effects of sarcosine administration are well described in pre-clinical studies, for example, by effectively reversing ketamine-induced behavioral, oxidative, and neuroinflammatory alterations in a rat model of schizophrenia, potentially by modulating oxidative stress, mitochondrial dysfunction, and neuroinflammation (Kumar et al. 2023). Similarly, pre-clinical in vivo imaging studies suggests that injections of sarcosine (500 mg or 1000 mg/kg) can reverse hippocampal neuronal dysfunction and impaired spatial encoding, under conditions of NMDA receptor hypofunction induced by MK801, indicating its potential relevance for treating cognitive and disorganization symptoms in schizophrenia (Hsiao et al. 2024), confirming previous studies (Pei et al. 2019). Positive effects have been reported also in restoring impaired long-term potentiation (LTP) in mice models (Varbanov et al. 2023).

Clinically, a meta-analysis of seven randomized controlled trials (n = 326) found that sarcosine significantly improved overall clinical symptoms in patients with schizophrenia, particularly in those with stable symptoms or lower baseline severity but showed no statistically significant benefit for cognitive function compared to controls (Chang et al. 2020). A subsequent meta-analysis further clarified that while sarcosine did not produce uniform benefits across all patients with schizophrenia, it significantly improved symptoms in individuals with chronic and non-refractory illness, particularly those not treated with clozapine (Marchi et al. 2021), suggesting its therapeutic efficacy may depend on specific patient subtypes. In line with these findings, a 6-month randomized controlled trial demonstrated that sarcosine significantly ameliorates negative symptoms, general psychopathology, and overall symptom severity when added to stable antipsychotic treatment in patients with chronic schizophrenia (Pawlak et al. 2023). Furthermore, a recent network meta-analysis of 50 randomized controlled trials (n = 2,384) further reinforced sarcosine’s clinical potential, identifying it as one of the most effective and promising nutraceutical augmentation strategies in stable schizophrenic patients, with demonstrated benefits for both total and negative symptom domains, and good tolerability (Fornaro et al. 2025). The authors noted that the overall quality of evidence remains low, underscoring the need for further high-quality studies.

Other glyT1 inhibitors

On the other hand, trials with other GlyT1 inhibitors, such as Org 25,935 and PF-03463275, did not produce promising clinical results. Although preclinical studies initially suggested that Org 25,935 possessed antipsychotic-like properties (Harsing et al. 2003) and potential benefits in improving cognitive deficits (Castner et al. 2014), these findings were not confirmed in clinical trials. In a 12-week randomized, placebo-controlled study, Org 25,935 was evaluated as an adjunct to second-generation antipsychotics in patients with persistent negative symptoms of schizophrenia. The compound failed to produce significant improvements in negative symptoms, cognitive performance, or overall functioning compared to placebo, indicating limited clinical efficacy in this population (Schoemaker et al. 2014).

PF-03463275 showed similar results. In fact, while in primates it reduced ketamine-induced spatial working memory deficits (Roberts et al. 2010), in clinical trials it failed to reduce ketamine-related effects in healthy participants, although it might enhance LTP in patients with schizophrenia (D’Souza et al., 2018). Moreover, a crossover study evaluating its combination with computerized cognitive training in schizophrenia did not demonstrate any additional cognitive improvements beyond the training itself, despite favorable tolerability and high GlyT1 occupancy (Surti et al. 2023).

A descriptive summary of the overall therapeutic effects in schizophrenia of the main GlyT1 inhibitors reported is shown in Table 2.

Table 2.

Descriptive summary of efficacy data in schizophrenia based on the studies included in the review

Inhibitor	Overall efficacy data in Schizophrenia
Bitopertin	Promising pre-clinical results; no significant advantage over placebo in clinical trials
Iclepertin	No significant advantage over placebo in clinical trials
Sarcosine	Promising pre-clinical results; significant advantage over placebo in clinical trials, particularly effective as add-on therapy to conventional psychiatric treatments, with benefits for negative symptoms, general psychopathology, and overall symptom severity
Org24598	Promising pre-clinical results; no significant advantage over placebo in clinical trials
PF-03463275	Promising pre-clinical results; no significant advantage over placebo in clinical trials

Mood and anxiety disorders

Targeting the glutamatergic system has emerged as a promising approach for the development of next-generation antidepressant therapies (Hashimoto 2011; Krystal et al. 2002; Shimizu-Sasamata et al. 1996; Skolnick 1999; Stewart and Reid 2002; Tokita et al. 2012), after the surprising finding that D-cycloserine, a partial NMDAR agonist, exhibited antidepressant effects (Crane 1959; EPSTEIN et al., 1956). In addition, reduced expression of NMDAR subunits 1 and 2 A has been observed in postmortem brain tissue of individuals with major depressive disorder (Beneyto and Meador-Woodruff 2008) and decreased NMDAR binding has also been reported in suicide victims (Nowak et al. 1995).

These findings suggest that NMDAR hypofunction may play a role in the underlying neurobiology of depression, and the potential beneficial effect of potentiating its activity. Following these data, numerous pre-clinical and clinical studies evaluated the role of modulating GlyT1 in depression and anxiety disorders.

The administration of a GlyT1 inhibitor, SSR504734, showed antidepressant-like effects by reducing despair-related behaviors in the forced swim test in rats and tonic immobility in gerbils (Boulay et al., 2008). These preliminary suggestions were later consolidated by evidence showing that sarcosine administration exerted antidepressant-like effects (Chen et al. 2015; Huang et al. 2013) and reversed behavioral deficits caused by chronic unpredictable stress test in rats (Huang et al. 2013). Mechanistically, sarcosine was found to activate the AMPAR–mTOR signaling pathway. Pretreatment with either the mTOR inhibitor rapamycin or the AMPAR antagonist NBQX blocked both the behavioral effects and molecular activation induced by sarcosine. Additionally, sarcosine increased phosphorylation of the AMPAR subunit GluR1 at the PKA site, suggesting enhanced AMPAR membrane trafficking (Chen et al. 2015). Similar antidepressive effects have been reported in clinical data, in which sarcosine led to greater improvements than citalopram in measures such as the Hamilton Depression Rating Scale, Clinical Global Impression, and Global Assessment of Function. Patients receiving sarcosine showed a higher and faster remission rate and had lower dropout rates (Huang et al. 2013).

Other positive insights on the role of GlyT1 inhibitors in affective disorders come from studies using Specificity Protein 4 (Sp4) hypomorphic mice, a model of genetic vulnerability linked to schizophrenia and bipolar disorder. GlyT1 inhibition by Org 24598 improved attention but not reward-related behaviors, suggesting that targeting GlyT1 may selectively benefit cognitive impairments associated with SP4 dysfunction in psychiatric disorders (Young et al. 2015).

The effects of GlyT1 inhibitors on anxiety-related behaviors appear to be complex and context-dependent. While some evidence suggests that enhanced NMDAR activity may produce anxiogenic effects in rodents (Miguel and Nunes-de-Souza 2008) and that some GlyT1 inhibitors, such as NFPS increase anxiety-like behaviors (Labrie et al. 2009), other studies report anxiolytic outcomes with GlyT1 inhibition. For instance, the GlyT1 inhibitor SSR504734 has been shown to reduce both the acquisition and expression of contextual fear conditioning in rats (Nishikawa et al., 2010a) and to suppress maternal separation-induced ultrasonic vocalizations (USVs) in rat pups (Depoortère et al., 2005). Building on these findings, further research has later confirmed these findings, and notably, reported that anxiolytic response was reversed by the GlyA (strychnine-sensitive) receptor antagonist, but not by a GlyB antagonist, indicating that activation of GlyA plays a central role in mediating the anxiety-reducing effects of GlyT1 inhibitors. In contrast, traditional anxiolytics like diazepam and escitalopram were unaffected by GlyA blockade, suggesting a distinct mechanism of action for GlyT1-mediated anxiolysis (Komatsu et al. 2015).

GlyT1 and GlyT2 are not only pharmacologically relevant but also subject to intracellular regulatory mechanisms that may contribute to the pathophysiology of mood disorders. Notably, GSK3β, a kinase implicated in mood regulation and a known target of lithium, has been found to differentially modulate GlyT activity, suppressing GlyT1 while enhancing GlyT2 function (Jiménez et al. 2015). This points to a mechanistic link between mood disorder pathways and glycinergic neurotransmission, suggesting potential new avenues for therapeutic intervention.

However, despite these promising preclinical findings, clinical translation has proven challenging. In a randomized, placebo-controlled trial, the GlyT1 inhibitor Org 25935 failed to enhance the therapeutic effects of cognitive-behavioral therapy (CBT) in patients with panic disorder. Despite being well tolerated, Org 25935 showed no significant benefit over placebo in reducing anxiety symptoms, highlighting a potential gap between animal models and human outcomes in the context of NMDA receptor modulation in anxiety disorders (Nations et al. 2012).

Interestingly, sarcosine has shown preliminary promise in the treatment of obsessive-compulsive disorder (OCD), which, although no longer classified as an anxiety disorder, shares overlapping glutamatergic mechanisms with anxiety and related disorders (Cortese and Phan 2005; Karthik et al. 2020). In a 10-week open-label study, sarcosine led to a significant reduction in OCD symptom severity, with 32% of participants achieving a clinically meaningful response, particularly among drug-naive patients. These findings suggest that targeting GlyT1 may hold therapeutic potential beyond traditional anxiety disorders and warrant further investigation in larger, controlled OCD trials (Wu et al. 2011).

A descriptive summary of the overall therapeutic effects in mood and anxiety disorders of the main GlyT1 inhibitors reported is shown in Table 3.

Table 3.

Descriptive summary of efficacy data in mood and anxiety disorders based on the studies included in the review

Inhibitor	Overall efficacy data in mood and anxiety disorders
SSR504734	Promising antidepressant-like and anxiolytic effects in pre-clinical results
Sarcosine	Promising antidepressant-like effects in pre-clinical results; significant advantage over placebo in clinical trials with major depressive disorder patients and in obsessive-compulsive disorder patients
Org24598	Mixed pre-clinical results: improvement in attention but not in reward-related behaviors; no significant advantage over placebo in reducing anxiety symptoms in clinical trials

Neurodegenerative disorders

Alzheimer’s disease

Alzheimer’s disease (AD) is a progressive neurodegenerative condition marked by amyloid-β accumulation, which contributes to NMDAR-dependent synaptic weakening, loss of dendritic spines, and cognitive decline. Given its role in regulating glutamatergic signaling through NMDAR, GlyT1 has emerged as therapeutic target in AD for its potential neuroprotective effects (Cubelos et al. 2005; Imamura et al. 2008; Jiménez et al. 2011; Vergas-Medrano et al., 2011; Zafra et al., 1995a; Zafra et al., 1995b).

In an amyloid-β-induced mouse model of AD, pretreatment with the GlyT1 inhibitor NFPS prevented cognitive deficits and reduced neuroinflammation and neuronal damage. Mechanistically, NFPS enhanced neuroprotective signaling by upregulating BDNF–TRKB–mTOR and CaMKIV–CREB pathways, while promoting long-term potentiation and suppressing GluN2B expression, highlighting its potential as a therapeutic strategy for AD (Oliveira-Lima et al. 2024). These results are in accordance with previous findings with another novel and selective GlyT1 inhibitor, ASP2535, which demonstrated strong cognitive-enhancing effects in preclinical models of AD, improving working and spatial memory in aged and pharmacologically impaired rodents (Harada et al. 2012).

Further supporting this approach, SSR504734 significantly enhanced hippocampal network oscillations and LTP in rats, pointing to improved NMDAR-mediated synaptic plasticity relevant to cognitive deficits in AD (Ahnaou et al. 2020).

However, these preclinical promising findings have not yet translated into clinical success. In a 12-week, randomized, placebo-controlled Phase 2 trial, the GlyT1 inhibitor BI 425809 was evaluated in 610 patients with mild-to-moderate probable AD. Despite high treatment compliance and good tolerability, BI 425809 failed to show any significant improvement in cognitive outcomes on the ADAS-Cog scale across all tested doses, indicating no meaningful clinical benefit of BI 425809 in this population (Wunderlich et al. 2023). Further clinical studies with additional GlyT1 inhibitors are needed to determine whether targeting GlyT1 represents an effective therapeutic strategy in AD.

Parkinson’s disease

L-3,4-dihydroxyphenylalanine (L-DOPA) remains the cornerstone therapy for managing motor symptoms in Parkinson’s disease (PD), yet its long-term use is frequently accompanied by serious complications, including dyskinesia (Fox and Lang 2008) experienced by up to 95% of patients with advanced PD (Hely et al. 2005). Given the strong association between glutamatergic dysfunction and dyskinesia in advanced PD, modulation of this system has become a therapeutic focus. While traditional approaches have aimed to reduce glutamate transmission, recent studies aimed to explore the opposite strategy, enhancing NMDAR function by inhibiting glycine reuptake.

In MPTP-lesioned marmosets with L-DOPA-induced dyskinesia and psychosis-like behaviors (PLBs), acute administration of NFPS significantly reduced both dyskinesia and PLBs by up to 51%, without diminishing the antiparkinsonian efficacy of L-DOPA (Frouni et al. 2021). These data were then confirmed with bitopertin, which significantly reduced both established and developing L-DOPA-induced dyskinesia and improved parkinsonian symptoms in rodent models (Frouni et al. 2022).

Following these lines of evidence supporting the anti-dyskinetic effects of GlyT1 inhibitors in animal models of PD, a recent study explored the anatomical basis for their efficacy. Using [³H]-NFPS binding, GlyT1 expression was mapped across key brain regions in L-DOPA-treated 6-OHDA-lesioned rats with varying severity of abnormal involuntary movements (AIMs). GlyT1 levels were found to correlate with dyskinesia severity, showing region-specific changes, most notably increased binding in the ipsilateral substantia nigra and decreased levels in the thalamus and subthalamic nucleus (Frouni et al. 2024). These findings furtherly suggest that GlyT1 inhibition modulates glutamatergic signaling in motor pathways implicated in dyskinesia, providing mechanistic support for its therapeutic potential in improving motor control and enhancing L-DOPA efficacy in PD.

Beyond their utility in managing L-DOPA-induced dyskinesia, GlyT1 inhibitors may also offer benefit for addressing non-motor complications of PD. In a small placebo-controlled clinical trial, sarcosine improved depression and behavioral symptoms in patients with Parkinson’s disease dementia, without worsening motor or cognitive function (Tsai et al. 2014).

Expanding its therapeutic scope, pretreatment with the GlyT1 inhibitor NFPS has demonstrated neuroprotective effects in mouse models of PD and Huntington’s disease (HD). Specifically, NFPS conferred significant protection against striatal damage induced by 6-hydroxydopamine (6-OHDA) and quinolinic acid. It markedly reduced neuronal degeneration, preserved dendritic spine density, and improved motor performance across multiple behavioral assays, including the rotarod, open field, and cylinder tests. Mechanistically, NFPS modulated NMDAR subunit expression by reducing GluN2A and GluN2B levels while preserving GluN1, thereby shifting the NMDAR subunit ratio toward a potentially less excitotoxic profile (Izidoro Ribeiro et al. 2024). These findings support GlyT1 inhibition as a promising approach to enhance striatal neuroprotection, regulate glutamatergic transmission, and improve motor function in neurodegenerative disease contexts.

In addition to their symptomatic and neuroprotective effects, GlyT1 inhibitors may also promote structural recovery. In a mouse model of dopaminergic degeneration, treatment with a 6-OHDA enhanced axonal sprouting and reinnervation in the dorsal striatum, leading to behavioral improvement; effects that were shown to depend on NMDAR signaling specifically in dopamine neurons (Schmitz et al. 2013). These findings point to a possible regenerative role for GlyT1 inhibition in PD.

 Other relevant therapeutic perspectives and novel GlyT interactions

Beyond their established roles in affective, psychotic, and neurodegenerative disorders, GlyTs have also been implicated in the pathophysiology and potential treatment of epilepsy. Epilepsy is characterized by disrupted neuronal excitability and neurotransmitter imbalances, including glycine and GABA. Glycinergic dysfunction, particularly involving GlyRs, has been associated with epileptiform activity (Chen et al. 2014; Straub et al. 1997). Several studies have demonstrated that GlyT1 inhibition may exert anticonvulsant effects. For example, sarcosine was shown to raise seizure thresholds in the maximal electroshock seizure (MES) test in mice, suggesting its potential as an adjunctive therapy (Socała et al. 2010). Furthermore, increased hippocampal GlyT1 expression was observed in rodent models of temporal lobe epilepsy, and both pharmacological inhibition and genetic deletion of GlyT1 led to seizure protection (Shen et al. 2015). Similarly, treatment with the GlyT1 inhibitor M22 elevated seizure thresholds without impairing motor performance (Zhao et al. 2016). Complementing these findings, SSR504734 was shown to raise seizure threshold in the maximal electroshock seizure threshold (MEST) test, though it had limited efficacy in other seizure paradigms and was associated with increased systemic inflammation, warranting further investigation into its therapeutic viability (Gapińska et al. 2025).

Glycine signaling and GlyT1 inhibition have also shown promise in the context of stroke and ischemic brain injury. Notably, GlyT1 inhibition has been shown to confer neuroprotection in both in vitro and in vivo models of cerebral ischemia. Pretreatment sarcosine and NFPS reduced neuronal injury, possibly by dampening GluN2B-mediated excitotoxicity and restoring balance in glycine and glutamate signaling (Pinto et al. 2012, 2014, 2015).

NMDAR signaling also plays a key role in addiction-related neuroplasticity and drug-seeking behaviors (Hopf 2017), and accumulating evidence suggests that modulating this receptor system may help reduce substance use (Donlon et al. 2024; Tomek et al. 2013).

In fact, also in this context, GlyT1 inhibition has shown promise. Pharmacological studies using Org24598 and genetic models have demonstrated that GlyT1 inhibition facilitates extinction of cocaine-conditioned behaviors and reduces drug-seeking in rodents (Achat-Mendes et al. 2012; Paolone et al. 2009). Notably, GlyT1 heterozygous knockout mice also display faster extinction of cocaine-associated behaviors (Nic Dhonnchadha et al. 2012; Puhl et al. 2015).

Beyond stimulants, GlyT1 inhibition has been investigated in alcohol dependence. Org 25935 reduced alcohol consumption and relapse-like behavior in animal models, likely through its effects on glycine and dopamine levels in reward-related brain regions such as the nucleus accumbens (Lidö et al. 2017; Molander et al. 2007). These data were recently confirmed with the use of Org24598, which when combined with varenicline and bupropion, it significantly reduced alcohol relapse behavior in rats, suggesting synergistic interactions between glycinergic and dopaminergic mechanisms. This triple combination not only abolished the alcohol deprivation effect but also modestly enhanced dopamine output in the nucleus accumbens, pointing to a potential strategy for improving relapse prevention while minimizing required dosages of individual agents (Olsson et al. 2024). This compound showed positive effects also in withdrawal symptoms; it ameliorated recognition memory loss due to ethanol withdrawal in rats, after binge-like ethanol exposure (Filarowska-Jurko et al. 2024). However, a phase 2 clinical trial evaluating Org 25935 for alcohol relapse prevention did not yield significant improvements (Szegedi et al. 2012), although this could be possibly due to suboptimal dosing and pharmacokinetic oversight.

GlyTs have also emerged as promising targets in managing pain and opioid tolerance. Neuropathic pain, caused by lesions in the somatosensory system, involves disrupted balance between excitatory and inhibitory neurotransmission.

Org 25935, sarcosine, and NFPS have shown analgesic effects in preclinical models (Barthel et al. 2014; Dohi et al. 2009; Morita et al. 2008). Lidocaine’s metabolite N-ethylglycine also acts as a GlyT1 inhibitor, reducing hyperalgesia by elevating glycine levels and enhancing GlyR activation (Werdehausen et al. 2015). Moreover, GlyT1 inhibition may delay opioid tolerance. In a recent study, co-treatment with NFPS preserved morphine’s analgesic effect in rats. The mechanism appears tied to NMDAR modulation via elevated glycine levels, rather than changes in opioid receptors, pointing to a novel adjunctive strategy for chronic pain (Galambos et al. 2024).

Targeting GlyT2 has yielded even more robust results. Compounds like ORG25543 and ALX1393 demonstrated potent antinociceptive effects in various mouse models, including herpetic, bone cancer, and bladder pain, particularly where GlyT1 is downregulated (Motoyama et al. 2014; Nishikawa et al. 2010a, b).

Clinically, the dual GlyT2/5-HT2A inhibitor VVZ-149 has shown promise. In a Phase 3 trial for postoperative pain, VVZ-149 significantly reduced pain scores and opioid use, offering rapid relief with good tolerability (Lee et al. 2025).

Conclusion and future directions

Overall, GlyTs remain a promising and innovative treatment approach for neuropsychiatric disorders, particularly through the modulation of GlyT1. Among GlyT1 inhibitors, sarcosine continues to demonstrate potential as a viable therapeutic option, especially in schizophrenia and in treating depressive symptoms across both mood disorders and PD. Despite encouraging results in preclinical models, most other GlyT-targeting compounds have not yet translated their preclinical efficacy successfully into clinical outcomes, highlighting a persistent gap between experimental promise and therapeutic application. Noteworthy, recent late-phase clinical trials of iclepertin did not yield significant improvements in cognition in schizophrenia, suggesting that a uniform treatment modality across heterogeneous patient populations may obscure therapeutic benefits. These outcomes highlight the need for precision strategies to identify subgroups of patients most likely to respond. Predictive biomarkers, spanning metabolomic profiles, genetic polymorphisms related to glycine and glutamatergic signaling, neuroimaging markers of cortical glycine availability, and electrophysiological signatures of NMDA receptor function, could play a central role in guiding patient selection. For example, baseline levels of sarcosine and related metabolites, or imaging readouts of cortical glutamatergic tone, may help identify individuals with glycine hypofunction who are more likely to benefit from GlyT1 modulation. Incorporating such stratification into trial design may improve the signal-to-noise ratio of efficacy outcomes, reducing the risk of false-negative findings and accelerating clinical development of this drug class. Such patient stratification is already being explored; for example, recent metabolomic profiling in drug-naïve schizophrenia patients identified baseline levels of sarcosine, along with specific phospholipid species such as phosphatidylcholine and sphingomyelin, as part of a predictive biomarker panel capable of distinguishing future treatment responders from non-responders (Wang et al. 2025). In addition, further clinical investigations using diverse GlyT1 inhibitors are still needed to determine with confidence whether GlyT1 modulation represents a viable therapeutic approach.

Although the underlying neurobiological rationale for targeting GlyTs remains strong, several challenges must still be addressed. These include the lack of transporter selectivity, particularly between GlyT1 and GlyT2; the risk of homeostatic compensation, which may blunt long-term efficacy; and potential interindividual variability in glycine system activity or GlyT expression, which may influence treatment response but remains poorly understood.

Overall, several strategic avenues may help realize the therapeutic promise of GlyT modulators. First, personalized medicine approaches, informed by genetic, neurochemical, or imaging biomarkers. Second, the development of next-generation compounds with improved pharmacodynamics, region-specific activity, reduced side effects, and minimal off-target effects could substantially improve therapeutic profiles. Third, the discovery of novel modulatory sites and mechanisms for GlyTs may offer alternative strategies beyond direct inhibition. In fact, recent findings have shown that membrane cholesterol can regulate GlyT1 activity by altering its conformational dynamics and transporter turnover through specific cholesterol-recognition motifs (Li et al. 2025). This unveils a promising direction for therapeutic development, where modulating membrane–protein interactions could allow more precise regulation of transporter activity with potentially greater specificity and fewer side effects. Lastly, the integration of advanced imaging techniques, including PET tracers targeting GlyTs, such as the recently developed [¹⁸F]ALX5406 which appeared to have good GlyT1 selectivity in pre-clinical imaging (Hoffmann et al. 2021), may offer critical tools for measuring target engagement and optimizing dosing in clinical trials.

In conclusion, while the clinical translation of GlyT inhibitors remains a work in progress, current evidence supports their mechanistic validity and therapeutic potential. Continued refinement in compound design, new modulatory options and better patient stratification will be crucial for unlocking their full utility in the treatment of complex neuropsychiatric and neurological disorders.

Author contributions

A.P. and A.P: Conceptualization, Resources, Writing-original draft preparation, Figure drawing. Both Authors contributed equally to this review.

Funding

The research was funded by the Medical University of Silesia grant No; BNW-1-005/N/4/I.

Data availability

The datasets analysed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethical approval

Not applicable.

Consent to participate

Not applicable.

Consent to publish

Not applicable.

Competing interests

Authors declare no conflict of interest.