Quantitative analysis of cerebellar lobulation in normal and agranular rats

Cerebellar pattern formation was investigated in rats treated with DNA modifying agents. Animals were subjected to combinations of daily injections of methylazoxymethanol acetate (MAM) for the last 6 days gestation and/or localised X-irradiation of the hindbrain on postnatal days 1 and 5 (P1 and P5). Animals were analysed on embryonic day 18 (E18), P0, P3, P7, and P14. Five parameters of the cerebellum were recorded from midsagittal sections: the number of primary lobules; the thickness of the external germinal layer (EGL); the density of cells in the internal granule cell layer (IGL) region; and the midsagittal area and perimeter. In addition, the laterolateral cerebellar distance was calculated. The data demonstrate that pre- and postnatal reduction of the EGL results in reduced cerebellar growth and folding. Cessation of the treatment at birth results in a recovery and eventual overproduction of EGL, but cerebellar growth and the development of fissures lags behind that of normal rats. Pre- and postnatal destruction of the EGL severely limited cerebellar growth and fissuration, and the cerebella contained only five primary lobules at P14. Rats subjected to postnatal X-irradiation alone had a similar low density of granule cells relative to those treated with a combination of prenatal MAM injections and postnatal X-irradiation, and yet the cerebella contained deeper fissures and more lobules (nine at P14). The data indicate that there are two phases of cerebellar folding: the establishment of five lobules that arise independent of granule cell production, and the granule cell-dependent expansion and partitioning of these five principal lobules during postnatal development. We propose that the lack of correlation between the severity of the granule cell loss and degree of lobulation in agranular rats indicates that granule cells exert an inductive influence over lobulation that is in part independent of the forces generated by their production and differentiation.     

Developmental fates and migratory pathways of dividing progenitors in the postnatal rat cerebellum

To investigate the developmental fates and the migratory pathways of dividing progenitors in both the white matter (WM) and the external granule layer (EGL) in the early postnatal rat cerebellum, a replication-deficient retrovirus carrying the beta-galactosidase gene (BAG) was injected into the deep cerebellar tissue or the EGL of postnatal rats to label dividing progenitors. After 1-3 days post-injection (1-3 dpi) of BAG into the deep cerebellar tissue of postnatal day 4/5 (P4/5) rats, labeled immature, unipolar cells were found mainly in the WM. From 4 to 6 dpi, similar cells appeared in the internal granule (IGL), Purkinje cell, and molecular layers, although about half of the labeled cells still resided in the WM and appeared immature. The first morphologically definable Bergmann glia, astrocytes, and oligodendrocytes were also observed. From 14 to 20 dpi, most labeled cells had developed into Bergmann glia, astrocytes, oligodendrocytes, and interneurons in their appropriate layers. When BAG injections were performed at P14, unipolar cells were initially observed, but the majority of these differentiated into myelinating oligodendrocytes in the WM and IGL by 17 dpi. Few immature cells were labeled by injections administered at P20, and these did not develop into mature glia, but into cells with lacy, fine processes, possible representing immature oligodendrocytes. In contrast, BAG-labeled progenitors of EGL produced only granule neurons. Thus, within the first 2 postnatal weeks, dividing progenitors in the WM migrate as immature cells into the cortex before differentiating into a variety of glia and interneurons. The genesis of oligodendrocytes continues through the 2nd postnatal week and largely ceases by P20. EGL cells do not produce glia, but only granule cells.     

Compartmental organization of Purkinje cells in the mature and developing mouse cerebellum as revealed by an olfactory marker protein-lacZ transgene

In a line of transgenic mice (HpY-1), the pattern of expression of an olfactory marker protein (OMP)-lacZ fusion gene was analyzed in the cerebellum, where, in adult mice, OMP-lacZ was expressed primarily in Purkinje cells (PCs) of the posterior lobe. The transgene-expressing PCs were organized in parasagittal bands, with a boundary of expression roughly corresponding to the primary fissure that separates the cerebellum into anterior and posterior compartments. The regional expression of the lacZ gene was also analyzed during embryonic and postnatal development of the cerebellum. Within the cerebellum-isthmus region, transgene expression first was detected at embryonic day 13.5 (E13.5) in a cluster of postmitotic cells. By E14.5, lacZ was also expressed by a subpopulation of migrating PCs in the postisthmal and lateral cerebellar primordium, and, by E16.5, transgene-positive PCs formed caudally four sagittal bands symmetric to the medial embryonic fissure. The caudal pattern was retained in postnatal cerebella, where, by postnatal day 0 (P0), transgene-positive PCs in vermal lobules VIII and IX appeared to be organized in two prominent parasagittal compartments on either side of a negative midline band. In early postnatal animals, the transgene was expressed transiently in the anterior lobe vermis. Hence, from P5 onward, transgene expression appeared mostly restricted to the posterior lobe, where it followed a caudal-to-rostral gradient. In the paraflocculus, transgene-expressing PCs were confined to the rostrodorsal portion. The results indicate that the anterior and posterior cerebellar lobes are regulated by distinct ontogenetic programs, and PCs of functionally distinct cerebellar regions express the transgene differentially. Furthermore, the data suggest that ectopic expression of OMP-lacZ in the cerebellum is under the control of regulatory elements that provide positional information for the regional specification of PC subsets.     

Developmental changes of inhibitory synaptic currents in cerebellar granule neurons: role of GABA(A) receptor alpha 6 subunit

Eye opening and increased motor activity after the second postnatal week in rats imply an extensive development of motor control and coordination. We show a parallel development change in spontaneous IPSC (sIPSC) kinetics in cerebellar granule neurons. sIPSCs were studied by whole-cell recordings in cerebellar slices, prepared from 7-30 postnatal day old rats. Early in development, sIPSCs had slow decay kinetics whereas in older rats faster decaying sIPSCs were found in larger proportion. Currents elicited by 1 mM GABA pulses (GABACs) in nucleated patches excised from cerebellar granule neurons revealed that GABACs kinetics better approximate sIPSC decay in young but not in more developed rats. The expression of alpha 6 subunit of GABAA receptors, unique in cerebellar granule neurons, has been shown to increase during development. Therefore, we took advantage of the recently reported selective inhibition of GABAA receptors by furosemide to characterize the relative contribution of alpha 6 subunits to native receptors in inhibitory synapses of cerebellar granule neurons. Although furosemide inhibition of sIPSCs amplitude was highly variable among distinct granule cells, it increased during development. At the same time, furosemide failed to inhibit sIPSCs recorded from Purkinje neurons. From the comparison of furosemide inhibition and kinetics of sIPSCs with GABACs recorded from mammalian HEK293 cells transfected with combinations of alpha 1 and alpha 6 GABAA receptor subunits together with beta 2 gamma 2 subunits, we propose that an increased alpha 6 subunit contribution in the molecular assembly of postsynaptic receptors in cerebellar glomeruli is responsible for the developmental changes observed.     

Tissue plasminogen activator mRNA expression in granule neurons coincides with their migration in the developing cerebellum

Tissue plasminogen activator activity in the developing cerebellum, as quantified by zymography of cerebellar homogenates from embryonic day (E) 17 to adult mice, shows a peak of activity at postnatal day (P) 7, followed by a steady 75% decrease into adulthood. Northern blot analysis reveals a similar pattern for tissue plasminogen activator mRNA levels, which are low at E17 but increase dramatically, reaching their highest levels of specific mRNA/micrograms RNA in P1-P7 mice and declining about threefold in the adult mouse. In situ hybridization of whole mouse brain sections with a tissue plasminogen activator antisense cRNA probe shows pronounce reactivity in the cerebellum. Although some binding is associated with the cerebellar meninges, the external granule layer is devoid of tissue plasminogen activator mRNA at all ages. However, highly labeled elongated cells, which also bind antibody to neuronal nuclear antigen and are adjacent to Bergmann glial fibers (i.e., migrating granule neurons), are readily visible throughout the molecular and Purkinje layers at P7 and P14. In the adult mouse cerebellum, tissue plasminogen activator mRNA labeling is restricted to cells in the Purkinje/internal granule layers. Thus, tissue plasminogen activator gene expression is induced as granule neurons leave the external granule layer and begin their inward migration.     

Mechanisms for ganciclovir resistance in gastrointestinal tumor cells transduced with a retroviral vector containing the herpes simplex virus thymidine kinase gene

Transfer of the herpes simplex thymidine kinase (HSV-TK) gene into tumor cells confers sensitivity to the cells to the viral drug ganciclovir (GCV). Although the efficacy of the HSV-TK/GCV approach is well studied, the mechanisms for the resistance of HSV-TK-transduced tumor cells to GCV are poorly understood. Here, we examined the mechanisms for GCV resistance in HSV-TK-transduced gastrointestinal (GI) cell lines. Our results show that GCV sensitivities vary in vitro and in vivo among the different HSV-TK-transduced GI tumor cell lines. GCV-resistant colonies were isolated from several different HSV-TK-transduced GI tumor cell lines after 14 days of GCV treatment. Characterization of GCV-resistant colonies demonstrated that the HSV-TK gene was either partially or completely deleted from the resistant HSV-TK-transduced cells. In the HT-29 RM and MIAPACA-2 RM cells, a 220-bp deletion of the gene was found, whereas in the HT-29 R1-R5-resistant cells, the whole TK gene was found to be absent. Immunocytochemical studies using a polyclonal antibody to the TK protein demonstrated that the HSV-TK protein was absent in the GCV-resistant, HSV-TK-transduced cells. Transfection of the resistant cells with an adenoviral vector containing a HSV-TK gene restored sensitivity to GCV. The presence of GCV-resistant cells was only demonstrable in GI tumor cell lines that also demonstrated a poor bystander effect. Our results suggest that GCV resistance found in tumor cells transduced with a retroviral HSV-TK gene is due to the lack of a functional TK protein in the tumor cells rather than any intrinsic resistance of the cells to GCV. In tumor cells with a good bystander effect, the small percentage of TK-transduced cells that do not express the TK protein are probably killed by the bystander effect because GCV-resistant tumor cells were not found in these cell lines. GCV-resistant tumor cells were found only in tumor cell lines with a poor bystander effect, by which, presumably, the transduced tumor cells lacking a functional TK gene were not killed by the bystander killing effect.     

The development of the cerebellar cortex of the Syrian hamster, Mesocricetus auratus. Foliation, cytoarchitectonic, Golgi, and electron microscopic studies

Although a number of investigations of abnormalities of cerebellar development have been carried out in the hamster, no detailed Golgi or ultrastructural studies of cerebellar development in this species have been reported. This report describes the development of the hamster cerebellar cortex from birth (day 0) through postnatal day 78, as studied by light, Golgi, and electron microscopic methods. Foliation patterns correlate with the expansion of the cerebellar layers and of total cerebellar area. Cytologic and morphologic development of the major cerebellar cell types--Purkinje, Golgi, basket, stellate, granule, and Bergmann glial cells--correlate with those of other species, such as the rat and mouse. Electron microscopic observations at selected developmental ages allow identification and classification of synapses in the early postnatal hamster. Parallel fiber and climbing fiber synapses are already present at birth. Although synaptogenesis probably continues through the first two postnatal months, all major cell types have developed initial synapses by postnatal day 6, at a time when little cellular maturation has occurred. By using gestational rather than natal age, close developmental correlations between hamsters and rat and mouse are possible. Since the gestational period of the hamster is only 16 days, the hamster cerebellum is less mature at birth than that of either the rat or mouse. Thus, the hamster is a convenient animal in which to investigate the effects of various procedures on early cerebellar development.     

Analysis of gene action in the meander tail mutant mouse: examination of cerebellar phenotype and mitotic activity of granule cell neuroblasts

The meander tail (mea) gene results in a stereotypic pattern of cerebellar abnormalities, most notably the virtual depletion of granule cells in the anterior lobe of the cerebellum. The causal basis of this mutation is unknown. In this paper we have taken a three-part approach to the analysis of mea gene action. First, we quantitatively determined the effect of the mea gene on granule cell and Purkinje cell number. We found, in addition to the marked depletion of anterior lobe granule cells ( > 90%), there were also significantly fewer granule cells in the posterior lobe (20-30%) without a concomitant loss of Purkinje cells. Second, we explored the relationship between granule cell depletion caused by the mea gene and by the mitotic poison, 5-fluoro-2'-deoxyuridine (FdU). Prenatal and postnatal ICR mice were treated with FdU to ascertain the regimen that best produces a meander tail-like cerebellar phenotype. The similarity of the effects of the mea gene and injections of FdU at E17 and PO suggests the hypothesis that the mea gene acts to disrupt the cell cycle of cerebellar granule cell precursors. Thus, the third part of this study was to test this hypothesis by using injections of either BrdU (5-bromo-2'-deoxyuridine) or 3H-thymidine into homozygous and heterozygous meander tail littermates at E17 or PO. After processing the tissue for BrdU immunocytochemistry or 3H-thymidine autoradiography, counts were made of the number of labeled and unlabeled external granule layer (EGL) cells to determine the percentage that had incorporated the mitotic label (labeling index). No difference in the labeling index was found between homozygous meander tail mice and normal, heterozygous littermate controls. Therefore, the mitotic activity of the EGL neuroblasts is not disrupted by the mea gene. Furthermore, while a mitotic poison can produce a phenotype similar to the action of the mea gene, mea is phenomenologically different from FdU treatment.     

Developmental cues modulate GABAA receptor subunit mRNA expression in cultured cerebellar granule neurons

The developmental expression of mRNAs encoding the GABAA receptor was analyzed in the rat cerebellar cortex and in cultured cerebellar granule neurons. Our studies in vivo reveal that the alpha 1-, beta 2-, and gamma 2-subunit mRNA levels in the cerebellar cortex rise dramatically during the second post-natal week, a period temporally correlated with extensive cerebellar maturation. To determine if these increases were preprogrammed or dependent on extrinsic factors, we examined subunit mRNA expression in granule cell cultures prepared at embryonic day 19 (E19) and postnatal day 10 (P10), immature and mature stages of cerebellar development, respectively. In E19 cultures, the alpha 1, beta 2, and gamma 2 GABAA receptor subunit mRNAs were present and their levels remained constant over the 21 d culture period. These results suggest that GABAA receptor gene expression is not intrinsic to the immature granule cells. A different pattern was found in P10 cultures where the three subunit mRNAs were initially present at levels approximately sixfold higher than those found at E19. The beta 2- and gamma 2-subunit mRNAs remained constant for 4 d and then increased sixfold between 4 and 7 d in culture. The magnitude and time course of these increases were similar to the developmental changes that occurred in vivo. Thus, our findings raise the possibility that signals encountered during development are necessary to induce GABAA receptor subunit mRNA expression. Moreover, these cues have been received by granule neurons prior to P10.(ABSTRACT TRUNCATED AT 250 WORDS)     

Protection of herpes simplex virus thymidine kinase-transduced cells from ganciclovir-mediated cytotoxicity by bystander cells: the Good Samaritan effect

Although considerable attention has been directed in the field of gene therapy toward elucidating the mechanism by which a transduced cell could kill a bystander cell, little is known about how bystander cells may affect transduced cells. We hypothesized that bystander cells, particularly if they were capable of gap junctional communication, could protect cells transduced with the herpes simplex virus thymidine kinase (HSV-TK) from ganciclovir (GCV)-induced cytotoxicity. To test this hypothesis, we used a rat hepatocyte cell line (WB) that can carry out efficient gap junctional communication, a WB clone transduced with HSV-TK (WB-TK), and a communication-incompetent subclone of WB cells (aB1). We cocultured WB-TK cells with either WB or aB1 cells, treated them with GCV, and then plated the cells into selective media that permitted us to quantify independently the surviving fraction of WB-TK cells or bystander cells. We found that WB bystander cells conferred up to a 1000-fold protection on WB-TK cells treated with GCV. aB1 cells conferred detectable, but significantly less, protection. These findings demonstrate that herpes simplex virus thymidine kinase-transduced cells can be significantly protected by bystander cells, particularly those that can carry out gap junctional communication. Whether this "Good Samaritan" effect improves the overall efficacy of gene therapy, by prolonging the survival of the source of toxic metabolites, or decreases effectiveness by increasing the survival of transduced cells will need to be determined in vivo.     

Cerebellar brain-derived neurotrophic factor-TrkB defect associated with impairment of eyeblink conditioning in Stargazer mutant mice

In the spontaneous ataxic mutant mouse stargazer, there is a selective reduction of brain-derived neurotrophic factor (BDNF) mRNA expression in the cerebellum. BDNF protein levels in the cerebellum are reduced by 70%. Despite normal levels of full-length and truncated TrkB receptor, constitutive and neurotrophin-4/5-induced tyrosine phosphorylation was significantly reduced in several signal transduction molecules, including phospholipase-Cgamma1, erk1, and erk2. Morphological examination revealed an increased number of external granule cells at postnatal day 15 and the presence of abnormal neurons resembling immature granule cells in the adult. These abnormalities are associated with a severe impairment in the acquisition of classical eyeblink conditioning, indicating cerebellar malfunction. Our data suggest that normal BDNF expression and TrkB signal transduction in the cerebellum are necessary for learning and plasticity in this model.     

The brain chondroitin sulfate proteoglycan brevican associates with astrocytes ensheathing cerebellar glomeruli and inhibits neurite outgrowth from granule neurons

Brevican is a nervous system-specific chondroitin sulfate proteoglycan that belongs to the aggrecan family and is one of the most abundant chondroitin sulfate proteoglycans in adult brain. To gain insights into the role of brevican in brain development, we investigated its spatiotemporal expression, cell surface binding, and effects on neurite outgrowth, using rat cerebellar cortex as a model system. Immunoreactivity of brevican occurs predominantly in the protoplasmic islet in the internal granular layer after the third postnatal week. Immunoelectron microscopy revealed that brevican is localized in close association with the surface of astrocytes that form neuroglial sheaths of cerebellar glomeruli where incoming mossy fibers interact with dendrites and axons from resident neurons. In situ hybridization showed that brevican is synthesized by these astrocytes themselves. In primary cultures of cerebellar astrocytes, brevican is detected on the surface of these cells. Binding assays with exogenously added brevican revealed that primary astrocytes and several immortalized neural cell lines have cell surface binding sites for brevican core protein. These cell surface brevican binding sites recognize the C-terminal portion of the core protein and are independent of cell surface hyaluronan. These results indicate that brevican is synthesized by astrocytes and retained on their surface by an interaction involving its core protein. Purified brevican inhibits neurite outgrowth from cerebellar granule neurons in vitro, an activity that requires chondroitin sulfate chains. We suggest that brevican presented on the surface of neuroglial sheaths may be controlling the infiltration of axons and dendrites into maturing glomeruli.     

Enhanced tumor cell killing in the presence of ganciclovir by herpes simplex virus type 1 vector-directed coexpression of human tumor necrosis factor-alpha and herpes simplex virus thymidine kinase

Past studies have documented the promise of herpes simplex virus type 1 (HSV-1) thymidine kinase (TK) suicide gene therapy as a potential antitumor treatment. HSV-TK converts the pro-drug ganciclovir (GCV) into a toxic nucleotide analogue, the incorporation of which into cellular DNA blocks cell proliferation. In this report, we have examined the hypothesis that the effectiveness of HSV-TK suicide gene therapy can be enhanced by coexpression of the antitumor cytokine human tumor necrosis factor-alpha (TNF-alpha) from the same replication-defective HSV-1 vector. In vitro testing demonstrated that TNF-alpha expression from this vector potentiated the killing of both TNF-alpha-sensitive L929 tumor cells and TNF-alpha-resistant U-87 MG cells in the presence of GCV. Furthermore, treatment of established intradermal L929 tumors in vivo with the TNF-alpha/TK vector and GCV resulted in prolonged animal survival compared with treatment with parental HSV-TK vector in the presence or absence of GCV. Treatment of intracerebral U-87 MG tumors showed a clear benefit of TK therapy, but a significant further increase in survival using the TNF-alpha vector could not be demonstrated. We found that potentiation of cell killing in vitro required intracellular TNF-alpha because purified protein added to the culture medium of cells infected with HSV-TK vector failed to have the same effect. Accordingly, potentiation in vivo should depend on efficient infection, but immunohistochemical analysis indicated that virus administration by U-87 MG intratumoral injection was inadequate, resulting in an estimated <1% infection of all tumor cells. Moreover, the majority of infected tumor cells were localized at the tumor margin. Together, these results suggest that TNF-enhanced tk gene therapy should provide a useful treatment for TNF-alpha-sensitive tumors and perhaps also for TNT-alpha-resistant tumors if vector delivery can be improved to increase the percentage of transduced tumor cells.     

Mouse Zic1 is involved in cerebellar development

Zic genes encode zinc finger proteins, the expression of which is highly restricted to cerebellar granule cells and their precursors. These genes are homologs of the Drosophila pair-rule gene odd-paired. To clarify the role of the Zic1 gene, we have generated mice deficient in Zic1. Homozygous mice showed remarkable ataxia during postnatal development. Nearly all of the mice died within 1 month. Their cerebella were hypoplastic and missing a lobule in the anterior lobe. A bromodeoxyuridine labeling study indicated a reduction both in the proliferating cell fraction in the external germinal layer (EGL), from 14 d postcoitum, and in forward movement of the EGL. These findings suggest that Zic1 may determine the cerebellar folial pattern principally via regulation of cell proliferation in the EGL.     

The weaver mutation causes a loss of inward rectifier current regulation in premigratory granule cells of the mouse cerebellum

Considerable interest has recently focused on the weaver mutation, which causes inward rectifier channel alterations leading to profound impairment of neuronal differentiation and to severe motor dysfunction in mice (Hess, 1996). The principal targets of mutation are cerebellar granule cells, most of which fail to differentiate and degenerate in a premigratory position (Rakic and Sidman, 1973a,b). Two hypotheses have been put forward to explain the pathogenetic role of mutant inward rectifier channels: namely that inward rectifier channel activity is either lacking (Surmeier et al., 1996) or altered (Kofuji et al., 1996; Silverman et al., 1996; Slesinger et al., 1996). We have examined this question by recording inward rectifier currents from cerebellar granule cells in situ at different developmental stages in wild-type and weaver mutant mice. In wild-type mice, the inward rectifier current changed from a G-protein-dependent activation to a constitutive activation as granule cells developed from premigratory to postmigratory stages. In weaver mutant mice, G-protein-dependent inward rectifier currents were absent in premigratory granule cells. A population of putative granule cells in the postmigratory position expressed a constitutive inward rectifier current with properties compatible with mutated GIRK2 channels expressed in heterologous systems. Because granule cells degenerate at the premigratory stage (Smeyne and Goldowitz, 1989), the loss of inward rectifier current and its regulation of membrane potential are likely to play a key role in the pathogenesis of weaver neuronal degeneration.     

In situ labeling of granule cells for apoptosis-associated DNA fragmentation reveals different mechanisms of cell loss in developing cerebellum

We have examined the role apoptosis plays during postnatal development of the mouse cerebellum by a new method utilizing T7 DNA polymerase for the in situ detection of DNA fragmentation associated with cell death. Granule cell loss between the third and fifth postnatal weeks, hypothesized to affect the granule cell to Purkinje cell stoichiometry, is not associated with DNA fragmentation. However, cerebellar granule cells undergo extensive nuclear DNA fragmentation between postnatal days 5 and 9. Cell death prior to synaptogenesis may help regulate granule cell number. Our results suggest that different mechanisms of cell death within the same neuronal cell population occur during development.     

Combination of antiviral immunotoxin and ganciclovir or cidofovir for the treatment of murine cytomegalovirus infections

The effects of two anti-murine cytomegalovirus (MCMV) immunotoxins used in combination with ganciclovir (GCV) or cidofovir (HPMPC) against MCMV were determined in vitro and in mice. The inhibitors were added to cell cultures 24 or 48 h after MCMV adsorption so as to not affect the initial infection rate. The immunotoxins (0.63, 1.25 and 2.5 micrograms/ml) combined with GCV (1.25, 2.5 and 5 microM) or HPMPC (0.03, 0.06 and 0.12 microM) caused synergistic inhibition of virus yield in C127I cells at most of the combinations tested. No toxic effect on cell growth in culture was observed at these immunotoxin/drug combinations. The effects of immunotoxin and GCV treatment were studied further in MCMV-infected severe combined immunodeficient (SCID) mice. Immunotoxin (1 mg/kg per day) given by intraperitoneal (i.p.) injection on days 1, 4 and 7 of the infection did not extend the mean day to death compared with the placebo group. Once daily i.p. treatment with GCV (50 mg/kg per day) for days starting at 24 h after virus inoculation extended survival time almost 11 days. The combination of immunotoxin plus GCV was better than GCV alone, extending the mean day to death an additional 2 to 3 days, which is suggestive of a synergistic effect.     

Embryonic and postnatal expression of four gamma-aminobutyric acid transporter mRNAs in the mouse brain and leptomeninges

The distribution of gamma-aminobutyric acid (GABA) transporter mRNAs (mGATs) was studied in mouse brain during embryonic and postnatal development using in situ hybridization with radiolabeled oligonucleotide probes. Mouse GATs 1 and 4 were present in the ventricular and subventricular zones of the lateral ventricle from gestational day 13. During postnatal development, mGAT1 mRNA was distributed diffusely throughout the brain and spinal cord, with the highest expression present in the olfactory bulbs, hippocampus, and cerebellar cortex. The mGAT4 message was densely distributed throughout the central nervous system during postnatal week 1; however, the hybridization signal in the cerebral cortex and hippocampus decreased during postnatal weeks 2 and 3, and in adults, mGAT4 labeling was restricted largely to the olfactory bulbs, midbrain, deep cerebellar nuclei, medulla, and spinal cord. Mouse GAT2 mRNA was expressed only in proliferating and migrating cerebellar granule cells, whereas mGAT3 mRNA was absent from the brain and spinal cord throughout development. Each of the four mGATs was present to some degree in the leptomeninges. The expression of mGATs 2 and 3 was almost entirely restricted to the pia-arachnoid, whereas mGATs 1 and 4 were present only in specific regions of the membrane. Although mGATs 1 and 4 may subserve the classical purpose of terminating inhibitory GABAergic transmission through neuronal and glial uptake mechanisms, GABA transporters in the pia-arachnoid may help to regulate the amount of GABA available to proliferating and migrating neurons at the sub-pial surface during perinatal development.