Basal Sodium-Dependent Vitamin C Transporter 2 polarization in choroid plexus explant cells in normal or scorbutic conditions

Vitamin C is incorporated into the cerebrospinal fluid (CSF) through choroid plexus cells. While the transfer of vitamin C from the blood to the brain has been studied functionally, the vitamin C transporter, SVCT2, has not been detected in the basolateral membrane of choroid plexus cells. Furthermore, it is unknown how its expression is induced in the developing brain and modulated in scurvy conditions. We concluded that SVCT2 is intensely expressed in the second half of embryonic brain development and postnatal stages. In postnatal and adult brain, SVCT2 is highly expressed in all choroidal plexus epithelial cells, shown by colocalization with GLUT1 in the basolateral membranes and without MCT1 colocalization, which is expressed in the apical membrane. We confirmed that choroid plexus explant cells (in vitro) form a sealed epithelial structure, which polarized basolaterally, endogenous or overexpressed SVCT2. These results are reproduced in vivo by injecting hSVCT2wt-EYFP lentivirus into the CSF. Overexpressed SVCT2 incorporates AA (intraperitoneally injected) from the blood to the CSF. Finally, we observed in Guinea pig brain under scorbutic condition, that normal distribution of SVCT2 in choroid plexus may be regulated by peripheral concentrations of vitamin C. Additionally, we observed that SVCT2 polarization also depends on the metabolic stage of the choroid plexus cells.

Choroid plexus epithelial cells are connected by tight junctions that restrict the paracellular movement of molecules between the blood and cerebrospinal fluid (CSF) [1][2][3] . Thus, the expression of zonula occludens 1 (ZO-1), claudins (1, 2 and 11) and occluding that generate a polarized epithelium (separating the apical membrane that contacts the CSF from the basolateral membrane, which is in contact with the blood) has been demonstrated in these cells 4,5 . Different studies have been carried out to examine choroid plexus function in primary cell cultures and immortalized cell lines 6-10 using bicameral systems that show different characteristics of polarized epithelial cells [11][12][13][14][15] and represent a blood-CSF barrier model for transcellular transport studies 13,14,16,17 . However, a limitation of these in vitro systems for studying the blood-CSF barrier is the exclusion of key structural components of the choroid plexuses, such as blood capillaries and stromal cells. In this way, choroid plexus explants represent an interesting study model of the blood-CSF barrier. Using this model, the subcellular localization of metal transporters, transferrin receptor (DMT1, MTP1 and TfR), and organic anion transporter (rROAT1-GFP) 18,19 has been studied. In explants of rat and shark choroidal plexus, transcellular transport and stroma fluorescein accumulation have also been studied 20,21 Vitamin C is an essential micronutrient for the normal metabolic functioning of the organism [22][23][24][25][26][27][28] . It is used as a cofactor in hydroxylation reactions and is a powerful water-soluble antioxidant; its participation in differentiation processes in different cell types has recently been determined [29][30][31][32][33][34][35][36][37][38][39] . In blood plasma, a concentration close to 50 μM has been detected, mainly in its reduced form, ascorbate (AA), with 5-10% in its oxidized form, dehydroascorbic acid (DHA). Independent of the ability to synthesize vitamin C, it must be efficiently incorporated into the different cells of the body. AA is actively incorporated by the cytoplasmic membrane through the sodium-ascorbate cotransporters (SVCTs) 40 and DHA is transported using the facilitated hexose transporters, GLUTs [41][42][43][44][45][46][47][48][49][50] In vivo studies injecting 14 C-labeled AA and subsequent autoradiography have shown that the radioactivity does not penetrate directly from the blood to the brain 51 . A high concentration of labeled AA was observed in the choroid plexus 2 min after the injection. Radioactivity spread from these areas throughout the brain by 24 h, with a high concentration of AA in the hippocampus and cerebellar cortex 51 . In studies carried out with rabbit choroid plexus cells in vitro, it was determined that it incorporated AA through a saturable type transport with an approximate Km of 44 μM 52,53 . Once the sodium-ascorbate cotransporters, SVCTs, were cloned, high SVCT2 mRNA expression was detected in the epithelial cells of the choroid plexus 30,46 . The expression of the mRNA added to the AA incorporation studies was suggestive of the presence of an active, saturable transporter, SVCT2, which would serve to transfer AA from the blood plasma to the CSF and subsequently the brain 54,55 Using porcine choroid plexus epithelial cells in bicameral culture systems, it was possible to detect active sodium-dependent AA transport in the basolateral membrane 16 . However, attempts to determine the basolateral localization of SVCT2 have failed, and only intracellular expression of SVCT2 has been observed in the epithelial cells of rat choroid plexus 56 . A similar location has been observed in human choroid plexus papilloma cells, even though kinetic studies to analyze AA uptake indicate that SVCT2 is expressed in these cells 57 .
In this work, we have analyzed the polarization of SVCT2 in choroid plexus explants. SVCT2 is widely observed with basolateral polarization in all plexus cells isolated from different ventricles. Its expression is detected early in development, and its polarization can be observed on E15 in the rat brain. In the scorbutic Guinea pig choroid plexus, SVCT2 increases its basolateral polarization, detecting the fall in the plasma vitamin C concentration. Choroid plexus cultured in the absence of glucose showed SVCT2 depolarization; polarization is re-stablished in the presence of glucose or lactate.

SVCT2 and GLUT1 are widely expressed in choroid plexus cells. We first demonstrated that GLUT1
and SVCT2 are widely expressed in choroid plexus cells from the lateral ventricle of the mouse brain using RT-PCR and Western blot analysis (Fig. 1A,B,D,E). Using sections of brain tissue incubated in 4% paraformaldehyde for 2 days, GLUT1 was detected in the basolateral membrane of choroid plexus cells (Fig. 1C, arrows); however, SVCT2 was mainly internalized (Fig. 1F). Due to this unexpected result, we fixed brain tissues by perfusion with Bouin solution, and immunofluorescence analysis showed that although most SVCT2 is intracellular (Fig. 1G), a fraction was detected at the basolateral membrane, colocalizing with GLUT1 (Fig. 1G, arrow and merge). At the ependymal level of the third ventricle, GLUT1 and SVCT2 were preferentially detected at the apical membrane (Fig. 1H, arrow heads), suggesting differential polarization of SVCT2 in both cell types. Thus, methods of fixation (i.e., perfusion vs immersion) change SVCT2 polarization exclusively in choroid plexus cells, but not in ependymal cells or neurons (data not shown). Because vascular perfusion or immersion techniques induce hypoglycemia, we hypothesize that the polarization of SVCT2 in the choroid plexus is extremely sensitive to glucose levels; therefore, we added glucose to subsequent experiments in which brain tissues were fixed. Under these conditions, SVCT2 and GLUT1 are colocalized in the basolateral membrane of choroid plexus cells (Fig. 1I, arrows). However, the morphology of the epithelial cells of the choroid plexuses was altered, generating vacuoles in the basal region of the cells (Fig. 1I, arrows). Due to this artifact, we isolated the choroid plexus quickly and developed a method of fixation for choroidal plexus explants.

SVCT2 is mainly polarized to the basolateral membrane in choroid plexus explant cells.
Because standard procedures to fix brain tissue do not allow for the detection of SVCT2 basolateral polarization, we standardized the rapid dissection of the choroid plexus ( Fig. 2A, arrows) to generate tissue explants maintained for a short time in culture ( Fig. 2A, In vitro Choroid plexus). The lateral ventricle and fourth ventricle plexus (data not shown) were isolated and maintained as a compact structure in culture (Fig. 2B,C). Scanning electron microscopy showed that the cells remain polarized, forming a continuous epithelium, where the cells present small microvilli on their apical membrane (Fig. 2B, arrows). Using confocal microscopy, we confirmed intracellular transthyretin (TTR) distribution (Fig. 2C) and monocarboxylate transporter 1 (MCT1) apical localization (Fig. 2C, arrows and inset). Finally, ZO-1 was detected at the tight junctions (Figs 2D and 3D reconstruction and orthogonal image, arrows), which maintain the integrity of the epithelial layer (blue and red borders) (Fig. 1D, digital reconstruction). Thus, we conclude that choroid plexus cells in vitro maintain the normal polarization of different proteins in their membranes.
SVCT2 was widely detected in the cellular membrane of choroid plexus explant cells (Fig. 3A,B arrows). When performing a comparative analysis with MCT1, we defined that SVCT2 is fundamentally polarized to the basolateral membrane (Fig. 3C, Bas arrows) whereas MCT1 is found at the apical membrane (Fig. 3C, Ap arrows). Using confocal three-dimensional analysis (Fig. 3D, arrow 3D and 3D images), ZO-1 is extensively localized in the apical region of the cells; however, SVCT2 is observed in the basolateral region (Fig. 3D, 3D-images). When analyzing the immunofluorescence in three different optical sections and orthogonal projections (Fig. 3D, images 1, 2 and 3), SVCT2 does not colocalize with ZO-1 (Fig. 3D, images 2 and 3, insets). However, SVCT2 and GLUT1 showed colocalization in the lateral and fourth ventricle choroid plexus cells (Fig. 3E,F Merge-yellow and arrows, and H and Merge-magenta). Finally, we confirmed that MCT1 and SVCT2 are polarized to different cellular membranes ( Fig. 3G and H-Merge). Following triple labeling for SVCT2 (red), MCT1 (green) and GLUT1 (blue) (Fig. 3I,J) (high-power view analysis), GLUT1 is mainly concentrated in the lateral and basal membrane (arrows) with MCT1 in the apical membrane (arrow head) (Fig. 3I). Similarly, SVCT2 is only detected in the basal membrane of the epithelial cells (arrows) with MCT1 in the apical membrane (arrow head) (Fig. 3J). Merge high-power view analysis confirmed GLUT1 and SVCT2 colocalization (arrows-magenta) (Fig. 3J). Thus, differential sorting of transporters is detected in choroid plexus cell explants.  58 to over-express pEYFP-hSVCT2wt and detect its polarization at 17 days. Surprisingly, SVCT2 is clearly polarized in some cells of the plexus that are positively electroporated (Fig. 4B, arrow and inset). By immunofluorescence analysis, GLUT1 is widely expressed in the cellular membranes of the plexus, but without clear polarization (Fig. 4B, arrows). At this stage, SVCT2 and GLUT1 did not colocalize (Fig. 4B, arrows in lower panels and inset). www.nature.com/scientificreports www.nature.com/scientificreports/ Because cells of the choroid plexus can polarize SVCT2 during embryonic development, we next defined SVCT2 polarization in postnatal stages P5, P10, P15 and P20 (Fig. 4C,D-E). Using low-magnification analysis, SVCT2 and ZO-1 are widely expressed in these stages (Fig. 4C, upper panels), although without colocalization (Fig. 4C, lower panels, merge with low-and high-magnification analysis and insets). Greater SVCT2 polarization is clearly detected at P20; a thickened positive band is observed in the basal region of the choroid plexus cells (Fig. 4C, arrows). In the same stages, SVCT2 colocalizes with GLUT1 ( Fig. 4D, arrows in insets), but not with MCT1 (Fig. 4E, insets and arrow heads). We concluded that SVCT2 is intensely expressed in the second half of embryonic development, showing basolateral polarization that increases strongly during the first 20 days of postnatal development.

SVCT2 is functionally over-expressed with basolateral polarization in choroid plexus cells using lentiviral vectors in vitro and in vivo.
Using a lentivirus that over-expresses hSVCT2wt-EYFP, which efficiently transduces Neuro2a cells (Fig. 5A), SVCT2 is polarized in the plasma membrane (Fig. 5A, arrows); a positive reaction for anti-GFP is also shown. Epithelial cells of the choroid plexus explants are also transduced in vitro by hSVCT2wt-EYFP lentivirus (Fig. 5C,D). Thus, hSVCT2wt-EYFP lentivirus was injected into the lateral ventricle at 1 day to analyze the expression and polarization of SVCT2 at days 20 and 70 (Fig. 5E). A significant number of cells were positive for hSVCT2wt-EYFP at days 20 (Fig. 5F,G) and 70 (Fig. 5H,I), where SVCT2 was www.nature.com/scientificreports www.nature.com/scientificreports/ detected in the cell membrane of the transduced epithelial cells (H, arrows). 3D-reconstruction of the positive cells (Fig. 5G, xy) shows SVCT2 in the silhouette of the cells, which is confirmed in the orthogonal planes xz and yz, showing SVCT2 basolateral polarization (Fig. 5G). Using 3D-reconstruction, we analyzed positive cells in 9 optical confocal planes (planes 1-9), confirming that SVCT2 is found in the cellular membranes of transduced cells (Fig. 5I, 1-9 images). Additionally, SVCT2-overexpressing cells also express GLUT1 in the basolateral membrane and MCT1 in the apical membrane (Fig. 5J,K). The highest SVCT2 colocalization is observed with GLUT1 (Fig. 5K, arrow), which is confirmed in 9 optical planes with yz and xz orthogonal projection (Fig. 5K, optical planes 1-9 with positive cell-arrow).
Finally, using a fixed cannula with a stereotaxic frame in the lateral ventricle, three injections of hSVCT2wt-EYFP lentivirus (see experimental scheme) were made in the lateral ventricle, and CSF samples were obtained from the cisterna magna at 48, 96 and 168 h. Thirty minutes before the second and third CSF extraction (96 and 168 h), AA was injected intraperitoneally for 30 min, and its concentration was analyzed in the CSF using the FRACS method. We did not observe significant differences in the concentration of AA in the CSF at 48 or 96 h of control animals or those intraperitoneally injected with AA. However, at 168 h after the first injection of the hSVCT2wt-EYFP lentivirus, a significant increase in AA transfers to the CSF in animals overexpressing SVCT2 (Fig. 5L). Thus, SVCT2, which can be over-expressed functionally in choroid plexus cells in vivo, is localized to the basolateral membrane by incorporating AA from the blood to the CSF. www.nature.com/scientificreports www.nature.com/scientificreports/

The choroid plexus explants from scorbutic Guinea pigs increase SVCT2 basolateral polarization.
We induced vitamin C deficiency in Guinea pig brains to define whether the polarization of SVCT2 can be modulated in vivo by the scorbutic condition. First, we defined that the choroid plexus of scorbutic animals maintains the barrier structure (Fig. 6A) at the ultrastructural level, in addition to its polarization. Microvilli and cilia were observed on the apical side of the cells. We also detected that ZO-1 expression was maintained in the explant cells generating a bee-like panel image (Fig. 6B). Under normal feeding conditions with vitamin C in the diet, SVCT2 is www.nature.com/scientificreports www.nature.com/scientificreports/ polarized to the basolateral membrane, colocalizing with GLUT1 ( Fig. 6C and inset). An intense reaction for GLUT1 is also observed in red blood cells. However, in the epithelial cells of the Guinea pig plexus, the reaction is heterogeneous in different cells (Fig. 6C, high-magnification images). Under vitamin C deficiency, SVCT2 is observed with great intensity, forming a well-defined, continuous band in the basal region of the cells (Fig. 6C, arrows), showing colocalization with GLUT1 (Fig. 6C, arrows). These results suggest that vitamin C deficiency increases SVCT2 polarization in the choroid plexus; thus, SVCT2 expression can be regulated by in vivo vitamin C concentration.
Finally, we used choroid plexus explants to define whether glucose deprivation changes SVCT2 polarization (Fig. 6D). Under these conditions, SVCT2 loses its polarization rapidly (Fig. 6D, arrows), a condition that is recovered by adding glucose as well as lactate to the medium (Fig. 6D, +Glucose or +Lactate, arrows). Interestingly, the absence of glucose does not affect GLUT1 polarization. These data suggest that the normal distribution of SVCT2 in the choroid plexus depends on the metabolic stage of the cells, which is heavily altered during brain tissue immersion or perfusion.

Discussion
Vitamin C enters the brain continuously through the choroid plexus, epithelial cells that form the blood-CSF barrier. Although the transfer of AA to the brain through the choroid plexus has been measured in vivo 52,53 , SVCT2 polarization in cells of the choroid plexus has not been defined histologically. Unexpectedly, we observed that the polarization of SVCT2 is sensitive to glucose levels in the vascular perfusion fluid. Using fixatives without glucose, www.nature.com/scientificreports www.nature.com/scientificreports/ SVCT2 is internalized and within cytoplasmic compartment. However, when glucose is incorporated into the fixer fluid, virtually 100% of SVCT2 has basolateral distribution and colocalizes with GLUT1. Unfortunately, a fixative solution with glucose alters the tissue structure, generating cells with osmotically altered membranes with a "bubble'' appearance. To overcome this experimental difficulty, we prepared choroid plexus explants performing accelerated brain dissection, which after a few moments in culture medium, were fixed. Under these conditions, excellent tissue preservation was maintained with perfect ZO-1 distribution, as well as MCT1 apical localization and GLUT1 basolateral sorting. In addition, the TTR transporter was preferably observed intracellularly. We conclude that the choroid plexus explants maintain a cellular and functional structure similar to that observed in vivo.
Epithelial cells of the choroid plexus explants maintain interaction with the extracellular matrix, which regulates cellular polarization and maintains the histological structure of the barrier. Thus, this organoid-like structure should better maintain the barrier properties of the choroid plexus cells compared to cell cultures in monolayers using Transwell systems, which have been previously used to analyze transcellular vitamin C transport in vitro 11,16 . We showed that epithelial tight junctions were perfectly formed and maintained. In these conditions, we precisely defined the basolateral localization of SVCT2 in most choroid plexus cells. Our SVCT2 detection method allows us to conclude that the transporter is massively expressed in most of the epithelial cells of the choroid plexus. SVCT2 has a basolateral distribution in the cells, colocalizing with GLUT1. SVCT2 detection at the intracellular level was weak, and colocalization with TTR was not observed. Colocalization with MCT1, an apical transporter, was also not observed.
The phenotype of a polarized epithelium and the large amount of mitochondria present in the epithelial cells of the choroidal plexuses are characteristic of secretory epithelia 59 , which need high energy to maintain cellular processes related to transport and secretion. Acute hypoglycemic conditions lead to SVCT2 internalization and probably cause transient vitamin C deficit within the CSF. Long periods of starvation or malnutrition may directly affect SVCT2 sorting, decreasing the entry of vitamin C into the brain and may increase oxidative damage. These dietary conditions may also affect SVCT2 sorting in other tissues, probably decreasing the level of vitamin C in many peripheral organs; however, this condition must be studied. SVCT2 seems to have different types of sorting in different cells, which may each be affected by the hypoglycemic condition. For example, SVCT2 is located in the apical membrane in ependymal cells (present work), in rat epithelial cells of the bronchiole, and in syncytiotrophoblasts of the human placenta 60 . Following over-expression, basolateral SVCT2 has been observed in renal and intestinal epithelial cell lines 61,62 , as well as epithelial cells of the proximal and distal tubules of the kidney during postnatal development 63 . The basolateral polarization of hSVCT2 over-expressed in MDCK and Caco-2 cells 62,64 depends on a basolateral amino acid sequence (LMAI) present in the amino terminal domain, between amino acids 56 and 59. Although some domains that regulate its sorting are known, the pathophysiological conditions can regulate SCVT2 basolateral polarization are unknown. In this study, we have defined that hypoglycemia may modify the sorting of SVCT2. Additionally, we have analyzed other two conditions, i) early postnatal brain development and ii) vitamin C deficiency.
Formation of the choroid plexus occurs early in development 2,65-68 . By in situ hybridization, we confirmed SVCT2 mRNA expression in the rat choroidal plexus at 11 days of brain development 58 , however, immunohistochemical detection without polarization was observed at 15 days of development 58 . By in utero electroporation, we observed the polarization of hSVCT2-EYFP overexpressed in the choroidal plexus of 17-day-old mice, prior to the polarization of GLUT1. Therefore, SVCT2 is polarized early during brain development, probably during the blood-CSF barrier formation, and its location does not vary during postnatal development. GLUT1 polarization is induced postnatally, and is likely associated with the closure of the blood-brain barrier (BBB) 69 , condition that does not seem to regulate SVCT2 polarization. In this period, the CNS preferentially uses ketone bodies to obtain energy. In addition, there is an increase in the expression of monocarboxylate transporters, such as MCT1 70 . GLUT1 polarization in the cells of the choroid plexus and in the endothelial cells of the BBB coincides with dietary changes, where the brain begins to use glucose to supply the energy requirements. However, the data obtained for SVCT2 in the choroid plexus indicate that this transporter acquires its polarization early in embryonic development, and it remains polarized during postnatal life, regulating the concentrations of brain vitamin C independent of BBB postnatal formation 71,72 .
Vitamin C deficiency in vivo could be a determining factor in the basolateral polarization of SVCT2. Most mammals synthesize AA from glucose. However, species, such as human and Guinea pigs, must incorporate vitamin C into their diet as they lack L-gulono-γ-lactone oxidase, which is necessary for AA biosynthesis 73 . Generation of vitamin C deficiency in Guinea pigs by reducing their intake for 4 weeks produced greater SVCT2 polarization in epithelial cells of the choroidal plexus, probably as a way to optimize incorporation of the circulating vitamin C in these animals. We also analyzed if the metabolic condition affected SVCT2 and GLUT1 polarization. Choroidal plexus express MCT1 and MCT3, in addition to GLUT1, suggesting that these cells can capture lactate and glucose for energy. In the absence of these energy substrates, we observed a deregulation of the basolateral polarization of SVCT2 but not GLUT1, which also was detected in the basolateral membrane. This effect was reversed with both glucose and lactate. Modification of the localization of SVCT2 and GLUT1 suggests that the secretory cells need energy to maintain transport processes and polarization of the transporters.
Our laboratory has been the first to use lentiviral vectors to overexpress SVCT2 transporter in vivo and in vitro in the choroid plexus. We analyzed the overexpression of hSVCT2-EYFP in the choroid plexus in vivo using intracerebroventricular injections. The fluorescence pattern of hSVCT2-EYFP in most cells in vivo was similar to the immunofluorescence pattern observed in explants. Overexpression of the transporter in the choroid plexuses in vivo demonstrated the basolateral polarization of SVCT2 in a model of endogenous polarization of non-transformed cells. In vivo SVCT2 overexpression also allowed us to increase the concentration of vitamin C in the CSF. In this way, similar experimental tools can be designed in the future to compensate for decreased vitamin C in the CSF during aging or during a neurodegenerative disease. (2019) 9:14422 | https://doi.org/10.1038/s41598-019-50772-2 www.nature.com/scientificreports www.nature.com/scientificreports/ We concluded that SVCT2 is intensely expressed in the second half of embryonic brain development, showing basolateral polarization that increases strongly during the first 20 days of postnatal development. In adult choroid plexus cells, SVCT2 showed high colocalization with GLUT1 in the basolateral membranes of epithelial cells. Additionally, we confirmed that epithelial cells of the choroid plexus in vitro and in vivo are transduced by an hSVCT2wt-EYFP lentivirus, where SVCT2 was detected in the basolateral membrane of the epithelial cells, which incorporate AA (intraperitoneally injected) from the blood to the CSF in vivo. Finally, we induced vitamin C deficiency in Guinea pig brain to define whether the polarization of SVCT2 can be modulated by the scorbutic condition. These data suggest that the normal distribution of SVCT2 in the choroid plexus depends on the metabolic stage of the cells, which is heavily altered during brain tissue immersion or perfusion.
Increasing intracerebral concentrations of vitamin C is a complex process; however, by modifying the expression of SVCT2 in the choroid plexus, it is feasible to increase the concentrations of vitamin C in the CSF, a condition that could help in brain pathologies associated with aging and oxidative damage.

Materials and Methods
Animals. Animal  In situ hybridization. Analysis of SVCT2 mRNA in fetal rat brain was performed as previously described 58 .
Briefly, a cDNA of 0.62 kb encoding the rat SVCT2 sequence obtained by RT-PCR from cultured embryonic neurons was subcloned into pCR-4-Blunt-Topo (Clontech, Palo Alto, CA, USA) and used to generate sense and antisense digoxigenin (DIG)-labeled riboprobes by in vitro transcription with T3 or T7 RNA polymerase (Boehringer Mannheim, Mannheim, Germany). In situ hybridization was performed on fetal brain sections mounted on poly-L-lysine-coated glass slides. The sections were deparaffinized, rehydrated and treated with proteinase K (1 mg/mL in PBS; 3 min at 37 °C). Tissue samples were then fixed with 4% PFA, acetylated for 10 min at 20 °C, incubated in pre-hybridization solution for 15 min at 37 °C and 25 mL of hybridization mix with 1:20 to 1:100 dilutions of riboprobe at 42 °C overnight 40 . Samples were rinsed in SSC (0.15 M NaCl/0.015 M sodium citrate) and washed twice for 15 min at 42 °C. Finally, the slides were washed at 37 °C for 15 min each in 2x SSC, 1x SCC, 0.3x, and 0.03x SSC. Visualization of DIG was performed with a monoclonal antibody coupled to alkaline phosphatase (1:500, anti-DIG-alkaline phosphatase Fab fragments; Boehringer Mannheim) for 2 h at 20 °C. NBT/ BCIP (Boehringer Mannheim) was used for colorimetric detection of alkaline phosphatase activity. Controls included use of the sense riboprobe and omission of the probe as previously specified 75,76 . RT-PCR analysis. For RT-PCR analysis, we used samples from choroid plexus explants. Total RNA was purified using Trizol reagent (Invitrogen, Carlsbad, CA, USA) and quantified in a SmartSpec300 spectrophotometer (Bio-Rad, Hercules, CA, USA). For RT-PCR, 1 μg of RNA was pre-treated with DNase I (Fermentas, ON, Canada) and processed. The basic thermocycling conditions included the following: one cycle at 95 °C for 5 min; 35 cycles at 95 °C for 30 s, 60 °C for 30 s, and 72 °C for 30 s; and one cycle at 72 °C for 7 min. The following primers were used to analyze the expression of SVCT2: forward 5′-ACGTTTGGATGCAGGTTACCC-3′ and reverse 5′-TGAAGCAGAGCAGCCAGGATAC-3′ [expected product 517 base pairs (bp)] and GLUT1: forward 5′-CATGTATGTGGGGGAGGTGT-3′ and reverse 5′-GACGAACAGCGACACCACAG-3′ [expected product 559 base pairs (bp)].
In utero electroporation. Plasmids were prepared according to the manufacturer's protocol (HiSpeed plasmid midi Kit, Qiagen). After pregnant, rats were anesthetized with isoflurane (2-4%), a midline abdominal incision was performed, and the uterus was taken out. Microinjections of 1-3 uLof plasmid DNA solution (0.5-1 mg/mL) in PBS plus Fast Green (0.05% w/v) into the lateral cerebral ventricles of E14-E15 embryos were made through the uterine wall via a Hamilton syringe. Between 0.5 and 1 µg of EYFP-N1 or EYFP-N1-SCT2wt (n = 5) were used per brain embryo. After injection of the plasmids, five pulses of 50 V each were delivered across the head of the embryo within the uterus 68 . Finally, embryos were reincorporated into the abdominal cavity and allowed to develop for an additional 2-3 days. Electroporated animals were then sacrificed; the brains were fixed, and thick sagittal sections (40-60 mm) were obtained. The position and morphology of transfected cells were analyzed by fluorescent microscopy.
AA measurement. The animals were anesthetized with 5% isoflurane, and kept at 2% during the procedure.
The temperature was kept under control, and the heart rate was monitored in real time using the small animal physiological monitoring system (Hardvard apparatus # 75-1500). In total, 30 µL of CSF were extracted from the cisterna magna by stereotaxic technique, and 20 µL was used to determine AA levels by FRASC colorimetric assay (bioassay system # EASC-100) according to the manufacturer's instructions.
Statistical analysis. The statistical analysis was performed using Prism version 4.0 software (GraphPad Software, San Diego, CA, USA). The data represent the mean ± SD of three independent experiments with each determination, which were conducted in triplicate. Statistical comparison between two or more groups of data was carried out using analysis of t-student (two-tailed). p < 0.001 was considered statistically significant.

Data Availability
All data generated or analysed during this study are included in this published article (and its Supplementary  Information files).