ASC Transporters Mediate D-Serine Transport into Astrocytes Adjacent to Synapses in the Mouse Brain

D-serine is an important signalling molecule, which activates N-methyl D-aspartate receptors (NMDARs) in conjunction with its fellow co-agonist, the neurotransmitter glutamate. Despite its involvement in plasticity and memory related to excitatory synapses, its cellular source and sink remain a question. We hypothesise that astrocytes, a type of glial cell that surrounds synapses, are likely candidates to control the extracellular concentration of D-Serine by removing it from the synaptic space. Using in situ patch clamp recordings and pharmacological manipulation of astrocytes in the CA1 region of the mouse hippocampal brain slices, we investigated the transport of D-serine across the plasma membrane. We observed the D-serine-induced transport-associated currents upon puff-application of 10 mM D-serine on astrocytes. Further, O-benzyl-L-serine and trans-4-hydroxy-proline, known substrate inhibitors of the alanine serine cysteine transporters (ASCT), reduced D-serine uptake. These results indicate that ASCT is a central mediator of astrocytic D-serine transport and plays a role in regulating its synaptic concentration by sequestration into astrocytes. Similar results were observed in astrocytes of the somatosensory cortex and Bergmann glia in the cerebellum, indicative of a general mechanism expressed across a range of brain areas. This removal of synaptic D-serine and its subsequent metabolic degradation are expected to reduce its extracellular availability, influencing NMDAR activation and NMDAR-dependent synaptic plasticity.


Introduction
Serine is one of the few bioactive molecules that is present as a dextro isomer (D-serine) in the mammalian body [1]. In the brain, it is synthesised from the amino acid L-serine by the enzyme serine racemase [2], with approximately 25-30% of serine being present as the D-isomer in the forebrain [1]. While D-serine is not used as a building block for protein synthesis, it does play an important role as a neuromodulator via its action on the N-methyl-D-aspartate (NMDA) class of glutamate receptors [3].
Electrophysiological patch-clamp recording: All experiments were performed at room temperature,~25 • C. Astrocytes were visually identified using infra-red DIC optics and were voltage-clamped using a Multiclamp 700B amplifier, controlled by Clampex 10.7 software (Molecular devices, San Jose, CA, USA). Data were acquired via a Digidata 1440A (Molecular devices), filtered at 10 kHz, sampled at 1 kHz, and analysed with Clampfit 10.7 (Molecular devices). Holding potentials were −80 mV and not corrected for liquid junction potentials. Series resistance was monitored throughout the experiment, and cells were excluded from recording if the series resistance increased by >30% of the initial values or if the holding current increased by >20% of the initial value.
Recording D-serine transporter currents: To investigate D-serine transport in stratum radiatum astrocytes, 10 mM D-serine was puff applied from a puff-pipette positioned approximately 40 µm from the astrocytic soma with a pressure of <20 psi for 5 s, using a Picosprizer II (General Valve, Fairfield, NJ, USA). Puffs were repeated every 2 min, and 5-8 stable recordings were averaged for measurement. For the current vs. voltage relationship, a voltage ramp protocol from −80 mV to +40 mV (100 mV per sec) was performed during the D-serine puff and the current from control ramps without D-serine was subtracted.
Statistics: Data are represented as mean ± SEM. Statistical significance was determined by means of a linear mixed-effects model using RStudio (R version 3.6.3), using function lme in the nlme package v3.1-162 [43]. For paired data (all data in Figures 3,4 and 6), the individual cell was treated as the random factor. Data were regarded as significant if p ≤ 0.05, and the significance is shown by *** for p ≤ 0.001.

Identification of Stratum Radiatum Astrocytes
Astrocytes located in the stratum radiatum layer of the CA1 hippocampal region of mice brain slices were identified visually based on their size and morphology (Figure 1a,b). The astrocytes were whole-cell voltage-clamped, displaying linear and passive currents in response to voltage steps, with low membrane resistance and a highly negative resting membrane potential (Figure 1c). The inclusion of 5% biocytin hydrochloride in the patch pipette and post hoc staining with streptavidin-alexa fluor-594 allowed for a visualisation of the astrocytic morphology and a confirmation of the cell type ( Figure 1d). 1a,b). The astrocytes were whole-cell voltage-clamped, displaying linear and passive rents in response to voltage steps, with low membrane resistance and a highly neg resting membrane potential (Figure 1c). The inclusion of 5% biocytin hydrochloride in patch pipette and post hoc staining with streptavidin-alexa fluor-594 allowed for a v alisation of the astrocytic morphology and a confirmation of the cell type ( Figure 1d

D-Serine Uptake into Stratum Radiatum Astrocytes
To investigate whether D-serine is taken-up by astrocytes, 10 mM D-serine was applied from a glass pipette onto voltage-clamped stratum radiatum astrocytes and resulting membrane currents were measured. To isolate the D-serine-induced memb currents, all recordings were performed in the presence of a cocktail of compound block ion channels and receptors, which could potentially generate a membrane cur (see methods). A D-serine-induced membrane current (IDS) of −18.8 ± 0.7 pA (n = 68) observed (Figure 2a). To characterise the nature of IDS, the current-voltage relation was investigated by comparing the current response to a ramp of membrane voltag

D-Serine Uptake into Stratum Radiatum Astrocytes
To investigate whether D-serine is taken-up by astrocytes, 10 mM D-serine was puff applied from a glass pipette onto voltage-clamped stratum radiatum astrocytes and the resulting membrane currents were measured. To isolate the D-serine-induced membrane currents, all recordings were performed in the presence of a cocktail of compounds to block ion channels and receptors, which could potentially generate a membrane current (see methods). A D-serine-induced membrane current (I DS ) of −18.8 ± 0.7 pA (n = 68) was observed (Figure 2a). To characterise the nature of I DS , the current-voltage relationship was investigated by comparing the current response to a ramp of membrane voltage before and during the D-serine application. I DS was found to be minimally affected by membrane voltage and did not exhibit a reversal of current at the reversal potential of any of the ions in the recording solution ( Figure 2b). This result is consistent with I DS being mediated by a membrane transporter current and not by activation of an ion channel or ionotropic receptor.
Biomolecules 2023, 13, x FOR PEER REVIEW 5 o fore and during the D-serine application. IDS was found to be minimally affected by m brane voltage and did not exhibit a reversal of current at the reversal potential of an the ions in the recording solution ( Figure 2b). This result is consistent with IDS being diated by a membrane transporter current and not by activation of an ion channel or otropic receptor. To ensure that puff application does not evoke a current due to mechanical stim tion, a puffer solution containing an external solution without D-serine was puff app onto the astrocytes. A negligible current (−0.25 ± 0.2 pA; n = 6) was observed, which in cates that the astrocytic membrane current does not arise due to an artefact of puff ap cation (Figure 2a).
To further support that the membrane currents originate from ASC transporter a vation, we used L-serine as an alternative ASC transport substrate. 10 mM L-serine puff applied onto voltage-clamped stratum radiatum astrocytes and a membrane curr (ILS) was observed, consistent with the substrate profile of ASC transport. Overall, 10 m O-benzyl-L-serine significantly reduced ILS, by 85.6% (ILS = −33.6 ± 2.0 pA in control −4.8 ± 0.8 pA in the presence of 10 mM benzyl serine; n = 4; p < 0.001; Figure 3c,d), indi ing that ILS is also primarily mediated by ASC transporters.  To ensure that puff application does not evoke a current due to mechanical stimulation, a puffer solution containing an external solution without D-serine was puff applied onto the astrocytes. A negligible current (−0.25 ± 0.2 pA; n = 6) was observed, which indicates that the astrocytic membrane current does not arise due to an artefact of puff application ( Figure 2a).

Astrocytic D-Serine Transport Is Mediated by ASC Transporters
Having established the presence of a D-serine transport current in astrocytes, we sought to determine which membrane transporter was responsible for D-serine uptake. The principal D-serine transporters in the central nervous system are ASCT1 (Slc1a4), ASCT2 (Slc1a5), SNAT1 (Slc38a1), SNAT2 (Slc38a2), asc-1 (Slc7a10), and LAT1 (Slc7a5). In order to demonstrate which transporter mediates I DS , O-benzyl-L-serine was added to the external solution to inhibit ASCT1/2 [44]. Overall, 10 mM O-benzyl-L-serine significantly reduced I DS , by 84.4% (I DS = −22.4 ± 1.3 pA in control and −3.5 ± 0.4 pA in the presence of 10 mM benzyl serine; n = 5; p < 0.001; Figure 3a,b), indicating that I DS is primarily mediated by ASC transporters.

Control
In T-Pro

I DS Is Not Mediated by an Uncoupled Anion Channel Current
ASC transporters can generate trans-membrane currents through the coupled transport of charged substrates such as Na + [49,50] and via the opening of an integral anion channel, resulting in uncoupled anion currents [51]. In the previous experiments, I DS was recorded with an intracellular recording solution containing sodium and alanine to provide substrates for ASCT-mediated amino acid exchange, and SCN − to provide a permeant ion for any potential uncoupled anion channel (Figure 5a). To determine the relative contributions of coupled charge movement and uncoupled anion channel activation in I DS , intracellular SCN − was replaced by Cl − , which shows significantly reduced permeability though ASCT's integral anion channel [41]. However, the replacement of SCN − by Cl − did not alter I DS (I DS with intracellular SCN − = −18.8 ± 0.7 pA; n = 68, and with intracellular Cl − = −15.7 ± 0.9 pA; n = 9; p = 0.07; Figure 5b). Along with the lack of significant voltage dependence (Figure 2b), this result indicates that I DS is not mediated by the gating of an uncoupled anion channel and is therefore generated by the coupled transport of D-serine into astrocytes.

IDS Is Not Mediated by an Uncoupled Anion Channel Current
ASC transporters can generate trans-membrane currents through the coupled transport of charged substrates such as Na + [49,50] and via the opening of an integral anion channel, resulting in uncoupled anion currents [51]. In the previous experiments, IDS was recorded with an intracellular recording solution containing sodium and alanine to provide substrates for ASCT-mediated amino acid exchange, and SCN − to provide a permeant ion for any potential uncoupled anion channel (Figure 5a). To determine the relative contributions of coupled charge movement and uncoupled anion channel activation in IDS, intracellular SCN − was replaced by Cl − , which shows significantly reduced permeability though ASCT's integral anion channel [41]. However, the replacement of SCN − by Cl − did not alter IDS (IDS with intracellular SCN − = −18.8 ± 0.7 pA; n = 68, and with intracellular Cl − = −15.7 ± 0.9 pA; n = 9; p = 0.07; Figure 5b). Along with the lack of significant voltage dependence (Figure 2b), this result indicates that IDS is not mediated by the gating of an uncoupled anion channel and is therefore generated by the coupled transport of Dserine into astrocytes.
(a) (b) Figure 5. D-serine transport currents are not mediated by uncoupled anion movement through an integrated ion channel. (a) Functional structure of mouse ASCT1 highlighting its sodium dependence (2 or 3 Na + ), obligatory antiport mechanism (one intracellular amino acid exchanged for one extracellular amino acid), and the presence of a metabolically uncoupled anionic channel. The 3D alpha fold structure was modelled from open-source deep mind database [52,53] using EzMol [54] (b) Astrocytic D-serine current (IDS) is not altered by replacing intracellular SCNwith Cl -, indicating that the transport current is coupled to D-serine uptake and not due to activation of an uncoupled anion channel in the transporter structure.

ASC Transporters Also Mediate D-Serine Uptake in Cortical and Cerebellar Glial Cells
To investigate whether glial cells in other brain areas also transport D-serine using ASC transporters, astrocytes in layer 2 of the somatosensory cortex and Bergmann glia in the cerebellum were also studied. Similar to observations in hippocampal CA1 astrocytes the puff application of 10 mM D-serine onto glia in the cortex and cerebellum elicited trans-porter currents that were sensitive to O-benzyl-L-serine. Cortical astrocytes exhibited an IDS of −20.1 ± 0.8 pA, which was reduced to −6.7 ± 1.0 pA in the presence of 10 mM O-benzyl-L-serine (n = 4; p = 0.001; Figure 6a). Cerebellar Bergmann glia exhibited an IDS of −17.0 ± 2.6 pA, which was reduced to −1.3 ± 1.3 pA in the presence of 10 mM O-benzyl-L-serine (n = 4; p < 0.001; Figure 6b). These data demonstrate that the ASCT-mediated transport of D-serine into glial cells is wide-spread across a number of different brain areas. . D-serine transport currents are not mediated by uncoupled anion movement through an integrated ion channel. (a) Functional structure of mouse ASCT1 highlighting its sodium dependence (2 or 3 Na + ), obligatory antiport mechanism (one intracellular amino acid exchanged for one extracellular amino acid), and the presence of a metabolically uncoupled anionic channel. The 3D alpha fold structure was modelled from open-source deep mind database [52,53] using EzMol [54]. (b) Astrocytic D-serine current (I DS ) is not altered by replacing intracellular SCNwith Cl -, indicating that the transport current is coupled to D-serine uptake and not due to activation of an uncoupled anion channel in the transporter structure.

ASC Transporters Also Mediate D-Serine Uptake in Cortical and Cerebellar Glial Cells
To investigate whether glial cells in other brain areas also transport D-serine using ASC transporters, astrocytes in layer 2 of the somatosensory cortex and Bergmann glia in the cerebellum were also studied. Similar to observations in hippocampal CA1 astrocytes, the puff application of 10 mM D-serine onto glia in the cortex and cerebellum elicited trans-porter currents that were sensitive to O-benzyl-L-serine. Cortical astrocytes exhibited an I DS of −20.1 ± 0.8 pA, which was reduced to −6.7 ± 1.0 pA in the presence of 10 mM O-benzyl-L-serine (n = 4; p = 0.001; Figure 6a). Cerebellar Bergmann glia exhibited an I DS of −17.0 ± 2.6 pA, which was reduced to −1.3 ± 1.3 pA in the presence of 10 mM O-benzyl-L-serine (n = 4; p < 0.001; Figure 6b). These data demonstrate that the ASCT-mediated transport of D-serine into glial cells is wide-spread across a number of different brain areas.

Discussion
Our results demonstrate that D-serine uptake occurs into astrocytes located imme ately adjacent to excitatory synapses,in acutely isolated brain tissue. Previously, D-ser transport has been studied in cultured synaptosomes, astrocytes, and transfected cell li [20,28,55]. However, synaptosomes contain both neuronal and glial membrane, and culture conditions alter the expression of membrane transporters [56]. While our findi are broadly consistent with those of previous studies, our direct electrical recordings fr identified individual astrocytes in situ allow us to characterise the D-serine transpor cells in their native environment under physiological conditions.

Astrocytic D-Serine Uptake Is Mediated by ASCT1 Transporters
We observe electrogenic D-serine transport in hippocampal astrocytes, located in stratum radiatum of the CA1 region, where astrocytes regulate the excitatory neurotra mission of CA3-CA1 Schaffer collateral synapses [57,58]. This transport is highly sensit to O-benzyl-L-serine, indicating the involvement of ASC transporters (Figure 3a,b). Si lar results were observed in astrocytes in the somatosensory cortex and Bergmann gli the cerebellum, suggesting that this D-serine uptake mechanism is ubiquitous across ferent brain regions ( Figure 6). Of the two ASC transporter isoforms, ASCT1 (Slc1a4) a ASCT2 (Slc1a5), the inhibition by trans-4-hydroxy-proline indicates that our data predo inantly arise from the activation of ASCT1 (Figure 4f). This is consistent with gene a protein expression studies, which show that ASCT2 is present in cultured astrocytes [ 39], but expression is minimal or absent in astrocytes in the intact brain [37,40]. In contr ASCT1 is observed on the astrocytic soma and on processes throughout the br [33,34,55,59], making it the most likely candidate for mediating astrocytic D-serine upta The fact that benzyl serine does not occlude all of the observed D-serine induced c rent could be due to its inability to out-compete all of the D-serine binding. However, involvement of another transporter in contributing a minor proportion of the D-ser transport current cannot be ruled out. One possible candidate could be the system A tra porters SNAT1 and SNAT2 which are able to transport D-serine, though not as efficien as other amino acids [26,27,29]. Synaptosome and neuronal culture experiments h clearly shown D-serine uptake via SNAT1/2, which is sensitive to the classical system

Discussion
Our results demonstrate that D-serine uptake occurs into astrocytes located immediately adjacent to excitatory synapses, in acutely isolated brain tissue. Previously, D-serine transport has been studied in cultured synaptosomes, astrocytes, and transfected cell lines [20,28,55]. However, synaptosomes contain both neuronal and glial membrane, and cell culture conditions alter the expression of membrane transporters [56]. While our findings are broadly consistent with those of previous studies, our direct electrical recordings from identified individual astrocytes in situ allow us to characterise the D-serine transport in cells in their native environment under physiological conditions.

Astrocytic D-Serine Uptake Is Mediated by ASCT1 Transporters
We observe electrogenic D-serine transport in hippocampal astrocytes, located in the stratum radiatum of the CA1 region, where astrocytes regulate the excitatory neurotransmission of CA3-CA1 Schaffer collateral synapses [57,58]. This transport is highly sensitive to O-benzyl-L-serine, indicating the involvement of ASC transporters (Figure 3a,b). Similar results were observed in astrocytes in the somatosensory cortex and Bergmann glia in the cerebellum, suggesting that this D-serine uptake mechanism is ubiquitous across different brain regions (Figure 6). Of the two ASC transporter isoforms, ASCT1 (Slc1a4) and ASCT2 (Slc1a5), the inhibition by trans-4-hydroxy-proline indicates that our data predominantly arise from the activation of ASCT1 (Figure 4f). This is consistent with gene and protein expression studies, which show that ASCT2 is present in cultured astrocytes [37][38][39], but expression is minimal or absent in astrocytes in the intact brain [37,40]. In contrast, ASCT1 is observed on the astrocytic soma and on processes throughout the brain [33,34,55,59], making it the most likely candidate for mediating astrocytic D-serine uptake.
The fact that benzyl serine does not occlude all of the observed D-serine induced current could be due to its inability to out-compete all of the D-serine binding. However, the involvement of another transporter in contributing a minor proportion of the D-serine transport current cannot be ruled out. One possible candidate could be the system A transporters SNAT1 and SNAT2 which are able to transport D-serine, though not as efficiently as other amino acids [26,27,29]. Synaptosome and neuronal culture experiments have clearly shown D-serine uptake via SNAT1/2, which is sensitive to the classical system A inhibitor MeAIB [29]. Although SNAT1/2 are thought to primarily be expressed on neurons [35,60], and therefore mediate neuronal D-serine transport, there are also reports of these transporters being expressed on astrocytes [36]. However, the insensitivity of the astrocytic D-serine current to MeAIB excludes a role for SNAT1/2 in mediating the astrocytic D-serine uptake that is observed in our experiments (Figure 4a,b).
Another possible astrocytic D-serine transporter is asc-1 (Slc7a10), which mediates the transport of D-serine in neurons and plays a major role in neuronal D-serine release [20,21]. The expression level of asc-1 in astrocytes is minimal [61,62], but nonetheless asc-1 has been detected [63]. While the pharmacological inhibition of asc-1 does not reduce the D-serine transporter currents we observe in astrocytes (Figure 4c,d), the electroneutral uptake of D-serine via asc-1 could still occur. This seems unlikely though, as asc-1 mediated D-serine transport measured by radiolabelling studies is undetectable in astrocyte cultures [20,64]. Taken together, these findings strongly suggest that ASCT1, rather than SNAT1/2 or asc-1, is the mediator of astrocytic D-serine uptake in situ.

D-Serine Produces Coupled Transport Currents in Astrocytes
ASC transporters (Slc1a4 and Slc1a5) are members of the same transporter family as the high-affinity excitatory amino acid (glutamate) transporters EAATs: EAAT1-EAAT5; the products of genes Slc1a3, Slc1a2, Slc1a1, Slc1a6 and Slc1a7 [65]. The activation of substrate transport in EAATs results in a membrane current that is coupled to the translocation of substrates, along with an uncoupled anion channel current due to the opening of an intrinsic anion channel [66,67]. This anion channel has a significantly higher permeability to SCN − than to Cl − , as predicted by thermodynamic studies of anion selectivity through channels [68]. The relative amounts of coupled charge movement to uncoupled anion current are highly variable between family members, with the anion channel carrying a small proportion of current in EAAT1-3 and the majority of the current in EAAT4-5 [69]. Studies of ASCT1 expressed in Xenopus oocytes also indicate the gating of an uncoupled anion conductance when the transporter is active [41]. However, in our recordings, the lack of the effect of substitution of internal SCNfor Cland the absence of current reversal at positive membrane potentials are inconsistent with the activation of an uncoupled anion conductance (Figures 2b and 5b). In accordance with this, serine has been shown to be an inhibitor rather than an activator of the anion conductance in ASCT1 [51]. Electrogenic coupled amino acid transport has previously been reported for ASCT2 [49,50], and our data from ASCT1 are consistent with this type of mechanism.

Physiological Implications of ASC Transporters and Enzymes in D-Serine Regulation
The extracellular concentration of D-serine is an important factor when determining the activation of NMDA glutamate receptors and governing synaptic plasticity [12][13][14]70]. As this concentration is governed not only by the rate of D-serine release but also the rate of its removal back into cells, the astrocytic uptake pathway that we have identified is predicted to play a role in controlling the synaptic actions of glutamate. The constitutive knockout of ASCT1 [55] reveals some unexpected effects though, as not only is D-serine transport disrupted, but its precursor L-serine is also significantly affected. A global reduction in both D-and L-serine occurs and a compensatory increase in glycine production is observed. This glycine serves as an alternative NMDAR co-agonist and obscures any effects of D-serine changes [55]. Further studies of ASCT1 using a conditional knockout strategy would be revealing in determining the importance of astrocytic ASCT1 in D-serine transport, while minimising the possible effects of upregulating alternative transporters or metabolic pathways.
The fate of D-serine following its transport into astrocytes is uncertain. Potentially, it is released from the astrocytic compartment in a stimulus dependent manner, and, as such, it acts as a gliotransmitter to further modulate neuronal activity [14,71,72]. Alternatively, D-serine can be metabolised in astrocytes by one of two potential metabolic pathways. One pathway is via D-amino acid oxidase (DAAO), which catalyses the oxidative deamination of D-serine to β-hydroxy pyruvate [73]. DAAO is functionally expressed in astrocytes and Bergmann glia in the cerebellum and other hindbrain regions but appears absent from the forebrain [74][75][76]. As there is strong DAAO expression and activity in Bergmann glia, it is likely to catalyse the oxidation of the D-serine that we observe being transported into these cells (Figure 6b). The alternative D-serine metabolic pathway is via serine racemase [2], which, in addition to catalysing the formation of D-serine from L-serine, can also catalyse an α, β-elimination pathway for both D-and L-serine to pyruvate [77,78]. Unlike DAAO, the highest expression of serine racemase is observed in the forebrain, with significant amounts observed in astrocytes [2,15,56,[79][80][81][82]. This enzyme is therefore most likely to mediate the metabolism of the D-serine that we observe being sequestered into astrocytes in the hippocampus and somatosensory cortex. While serine racemase is able to convert D-serine to L-serine, this is not likely in astrocytes at physiological D-serine concentrations, with the α, β-elimination reaction predominating under normal conditions [78]. The production of astrocytic L-serine occurs via the alternative 3-phosphoglycerate dehydrogenase pathway [17,18], which supplies the majority of L-serine for the "serine shuttle" transfer to neurons (Figure 7).
Understanding how astrocytes control extracellular D-serine concentrations via ASCT1 could also be significant in identifying treatment strategies for schizophrenia. Models of schizophrenia involving the reduced action of glutamate at NMDA receptors have gained prominence in recent years [97]. The reduced activity of D-serine at NMDA receptors may be central to the aetiology of schizophrenia, and, in agreement with this theory, the reduction of serine racemase activity produces schizophrenia-like phenotypes in animal models [98][99][100][101]. In humans, genetic variation of the serine racemase gene is associated with schizophrenia [99,102]. The treatment of schizophrenia by DAAO inhibition, or by direct administration of D-serine has had some success [103]. However, the observation of reduced ASCT1 expression in schizophrenia [104] suggests that increasing serine racemase activity or ASCT1 expression may be beneficial in this condition.
In summary, our data show the direct electrical recording of ASCT1 transporters in identified cells in situ and demonstrate D-serine uptake into glia of the hippocampus, cortex, and cerebellum. This indicates that ASCT1 is a central mediator of D-serine transport and that astrocytes/Bergmann glia may modulate extracellular D-serine levels and play a role in the metabolic degradation of D-serine. Our findings imply that during normal physiological conditions, ASCT1 may influence the synaptic actions of glutamate, and, during pathological conditions, it may be involved in the aetiology of a range of neurological disorders including traumatic brain injury, Alzheimer's disease, neuropathic pain, and schizophrenia. Figure 7. Proposed physiological role of Alanine-Serine-Cysteine Transporters (ASCTs). L-Serine is produced exclusively in astrocytes from glucose via glycolysis (G) and phosphorylated pathways (PP). ASC transporters (ASCT) expressed on the astrocytic membranes can aid in the shuttle of Lserine into the neuronal compartment (pre-or postsynaptic). The neuronal L-serine transporter is yet to be characterised. Neuronal L-serine is converted into D-serine by serine racemase (SR) and released via asc-1 transporters. Released D-serine can act on NMDARs along with glutamate (Glu) to enhance neuronal communication. Following its action at NMDARs, D-serine is cleared from the synaptic cleft in a pathway involving astrocytic ASC transporters. Once inside the astrocyte, D-serine can be metabolised by D-amino acid oxidase (DAAO), most prominently expressed in the cerebellum. Alternatively, astrocytic D-serine can be metabolised by α, β-elimination to form pyruvate and ammonia, catalysed by SR. Under pathological conditions, D-serine concentrations are increased in the astrocyte, possibly via production from L-serine, and may be released by ASCT to cause over-activation of NMDA receptors. Figure 7. Proposed physiological role of Alanine-Serine-Cysteine Transporters (ASCTs). L-Serine is produced exclusively in astrocytes from glucose via glycolysis (G) and phosphorylated pathways (PP). ASC transporters (ASCT) expressed on the astrocytic membranes can aid in the shuttle of L-serine into the neuronal compartment (pre-or postsynaptic). The neuronal L-serine transporter is yet to be characterised. Neuronal L-serine is converted into D-serine by serine racemase (SR) and released via asc-1 transporters. Released D-serine can act on NMDARs along with glutamate (Glu) to enhance neuronal communication. Following its action at NMDARs, D-serine is cleared from the synaptic cleft in a pathway involving astrocytic ASC transporters. Once inside the astrocyte, D-serine can be metabolised by D-amino acid oxidase (DAAO), most prominently expressed in the cerebellum. Alternatively, astrocytic D-serine can be metabolised by α, β-elimination to form pyruvate and ammonia, catalysed by SR. Under pathological conditions, D-serine concentrations are increased in the astrocyte, possibly via production from L-serine, and may be released by ASCT to cause over-activation of NMDA receptors. ASCT1 is an obligate amino acid exchanger [41], with influx of D-serine causing antiport of another amino acid. ASCT1 activation using T-Pro has been shown to mediate the release of astrocytic L-serine [55], leading to the suggestion that D-serine transport into astrocytes via ASCT1 could stimulate the release of L-serine. As the neuronal release and subsequent astrocytic uptake of D-serine would cause a depletion of the neuronal D-serine supply, it will be advantageous to stimulate L-serine shuttling from astrocytes to neurons. The exchange of D-serine for L-serine by astrocytic ASCT1 will thus help to promote the resupply of D-serine in neurons (Figure 7). Conversely, there is little evidence to suggest that under normal physiological conditions astrocytic ASCT1 could mediate the release of D-serine in exchange for another amino acid [20,55]. The absence of D-serine release could be due to its metabolism, which would maintain a low cytoplasmic concentration [16,82]. Alternatively, the ability of ASCT1 to release D-serine may be reduced if the exchange of amino acids is asymmetric, allowing different selectivity for import and export, as observed for a range of antiporters including ASCTs [83,84].
D-serine release from astrocytes, while unlikely in normal conditions, has been shown to occur during a range of pathological conditions. As excess activation of NMDA receptors causes calcium influx and neuronal death [85], this elevation of extracellular D-serine concentration could enhance the neurotoxic effects of glutamate. For example, in traumatic brain injury, astrocytes become reactive and upregulate serine racemase, causing an increase in the cytoplasmic D-serine concentration [86]. ASCT1 is an exchanger that can mediate the uptake or release of amino acids, and the increase in cytoplasmic D-serine concentration promotes its release from astrocytes via ASCT1 [87]. Hence, D-serine release switches from being mainly neuronal to being astrocytic, and this underlies synaptic damage [86]. Under these conditions, ASCT1 inhibition by glial specific ASCT1 knockout or pharmacological inhibition is protective for synapses [87]. Similarly, in Alzheimer's disease, astrocytes become reactive and neurotoxic, upregulate serine racemase, and release D-serine [88,89]. This is in contrast to normal aging, where D-serine levels are known to fall [90], leading to the possibility that D-serine could be a biomarker for Alzheimer's disease [91,92]. The inhibition of astrocytic D-serine production and release in a mouse model of Alzheimer's disease improves synaptic function [93], raising the possibility that targeting the astrocytic production of D-serine, or ASCT1 mediated D-serine release would represent a novel Alzheimer's disease therapy, which is divergent from the current anti-amyloid strategies [94]. Similarly, for neuropathic pain, the upregulation of serine racemase in spinal cord astrocytes causes the release of D-serine and sensitisation of pain pathways [95,96]. This leads to the possibility that the suppression of ASCT1-mediated D-serine release may be beneficial in the treatment of neuropathic pain.
Understanding how astrocytes control extracellular D-serine concentrations via ASCT1 could also be significant in identifying treatment strategies for schizophrenia. Models of schizophrenia involving the reduced action of glutamate at NMDA receptors have gained prominence in recent years [97]. The reduced activity of D-serine at NMDA receptors may be central to the aetiology of schizophrenia, and, in agreement with this theory, the reduction of serine racemase activity produces schizophrenia-like phenotypes in animal models [98][99][100][101]. In humans, genetic variation of the serine racemase gene is associated with schizophrenia [99,102]. The treatment of schizophrenia by DAAO inhibition, or by direct administration of D-serine has had some success [103]. However, the observation of reduced ASCT1 expression in schizophrenia [104] suggests that increasing serine racemase activity or ASCT1 expression may be beneficial in this condition.
In summary, our data show the direct electrical recording of ASCT1 transporters in identified cells in situ and demonstrate D-serine uptake into glia of the hippocampus, cortex, and cerebellum. This indicates that ASCT1 is a central mediator of D-serine transport and that astrocytes/Bergmann glia may modulate extracellular D-serine levels and play a role in the metabolic degradation of D-serine. Our findings imply that during normal physiological conditions, ASCT1 may influence the synaptic actions of glutamate, and, during pathological conditions, it may be involved in the aetiology of a range of neurological disorders including traumatic brain injury, Alzheimer's disease, neuropathic pain, and schizophrenia.
Author Contributions: K.S.K. performed the experiments, analysed data, prepared data for publication, and wrote the initial manuscript. K.S.K. and B.B. contributed to the overall study design, interpretation, and editing of the manuscript. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement: The animal study protocol was approved by the Animal Experimentation Ethics Committee of The Australian National University (protocol code A2020/38; approved 27 October 2020).

Data Availability Statement:
The raw data from this study are available via Mendeley data at DOI: 10.17632/n8d2ymd85h.1.