Search

Ependymal cells have been considered one of prime targets for gene therapy in the central nervous system as they can secrete proteins directly into the cerebrospinal fluid. In this study, we have explored the probability of permanent exogenous gene expression using a combined adenovirus/transposon system. To this end, we created three adenoviruses; adenovirus #1 containing a CAG promoter-driven enhanced green fluorescent protein tagged with a palmitoylation site (palEGFP), whose DNA sequence was flanked by two different Tol2 ends, #2 containing a human FoxJ1 promoter-driven T2TP transposase, and #3 containing an EF-1 alpha promoter-driven T2TP transposase. We injected these adenoviruses into the lateral ventricles of adult rats to assess the duration of transgene expression, by which adenoviruses selectively infected to ependymal cells because they express the specific receptor. In animals injected with only adenovirus #1, we found palEGFP-expressing ependymal cells 1week after injection, but these cells had disappeared by 2weeks. In animals that received adenoviruses #1 and #2 in combination, despite detecting many palEGFP-expressing ependymal cells within the initial 2weeks, transgene expression in ependymal cells was almost disappeared 1month after injection. In contrast, many palEGFP-expressing astrocytes, oligodendrocytes, and neurons were found near the sites injected with adenoviruses #1 and #3, even 1month after injection. There was no prominent infiltration of immunological cells during the observation period. These findings indicate that an adenovirus-mediated transposon gene transfer system can lead to prolonged, but not permanent, expression of exogenous genes in ependymal cells of adult rats.

The filum terminale (FT) is a potential source of ependymal cells for transplantation. The present study was performed to clarify the characteristics of ependymal cells of the central canal (CC) of the FT in rats. The FT was a thin strand continuous with the conus medullaris (CM), a caudal end of the main spinal cord, situated at the L3-4 level in adult rats. The border between the CM and FT was not visible, but could be defined as the site where the strand was as thin as its more caudal segment. While the CM contained an appreciable amount of white and grey matter associated with the CC at its center, the FT had no or only a negligible amount of such spinal cord parenchymal tissue. The FT was tracked ca. 4 cm from the site defined above to the level of S4-5 in adult rats. The rostral part of the FT (FTI) included within the cauda equina is exposed to cerebrospinal fluid, whereas the more caudal part (FTE) was surrounded by a dense layer of connective tissue. Almost all ependymal cells were immunostained for Sox2, Sox9, FoxJ1, and CD133, generally recognized immunochemical markers for ependymal cells of the CC in the spinal cord. Ependymal cells of the CC of FT exhibited almost the same structural and immunohistochemical characteristics as those of the CC of the main spinal cord. Ependymal cells of FTI covered by a thin layer of connective tissue are considered appropriate for transplantation.

Nuclear factor one X (NFIX) has been shown to play a pivotal role during the development of many regions of the brain, including the neocortex, the hippocampus and the cerebellum. Mechanistically, NFIX has been shown to promote neural stem cell differentiation through the activation of astrocyte-specific genes and via the repression of genes central to progenitor cell self-renewal. Interestingly, mice lacking Nfix also exhibit other phenotypes with respect to development of the central nervous system, and whose underlying causes have yet to be determined. Here we examine one of the phenotypes displayed by Nfix(-/-) mice, namely hydrocephalus. Through the examination of embryonic and postnatal Nfix(-/-) mice we reveal that hydrocephalus is first seen at around postnatal day (P) 10 in mice lacking Nfix, and is fully penetrant by P20. Furthermore, we examined the subcommissural organ (SCO), the Sylvian aqueduct and the ependymal layer of the lateral ventricles, regions that when malformed and functionally perturbed have previously been implicated in the development of hydrocephalus. SOX3 is a factor known to regulate SCO development. Although we revealed that NFIX could repress Sox3-promoter-driven transcriptional activity in vitro, SOX3 expression within the SCO was normal within Nfix(-/-) mice, and Nfix mutant mice showed no abnormalities in the structure or function of the SCO. Moreover, these mutant mice exhibited no overt blockage of the Sylvian aqueduct. However, the ependymal layer of the lateral ventricles was frequently absent in Nfix(-/-) mice, suggesting that this phenotype may underlie the development of hydrocephalus within these knockout mice.

The subependymal zone at the lateral ventricular wall represents a major neurogenic niche of the adult mammalian brain and continuously provides new neurons for the olfactory bulb. A mosaic of stem and progenitor cells in this niche has the potential to respond to multiple signals including growth factors such as EGF. Recent studies using long-term ventricular infusion of EGF demonstrate intense cell proliferation around the ventricular wall, implicating the presence of EGF-reactive cells also outside the classical neurogenic lateral niche. Here we show that intraventricular injection of EGF induces within minutes CREB and ERK phosphorylation in astrocyte-like progenitor cells (type B cells) and EGF receptor-expressing transit-amplifying progenitor cells-both in the striatal and septal ventricular walls. EGF infusion for 6 days induced continued CREB and ERK activation in nestin+ cells paralleled by intense periventricular cell proliferation. In addition, the ependyma became EGF receptor-immunoreactive, revealed intense CREB phosphorylation and underwent partial de-differentiation. Our results demonstrate that intraventricular application of EGF induces CREB and ERK phosphorylation along the entire ventricular walls and thus permits a direct identification of EGF-responsive cell types. They further support the notion that not only the striatal ventricular wall where the SEZ is located but also the septal ventricular wall carries latent potential for the formation of neurons and glial cells.

The ependymocytes lining the central canal of the spinal cord in rats have prominent F-actin networks in the apical cytoplasm, exhibiting significant regional heterogeneity between differing spinal levels. It is not clear when and how this regional heterogeneity develops. We examined F-actin in the central canal during postnatal development using fluorescein isothiocyanate-conjugated phalloidin staining and confocal laser scanning microscopy. The apical F-actin network in the central canal increased gradually in width during the first 2 weeks after birth, peaked rapidly in the third postnatal week, then decreased gradually to adult levels. Regional differences in the F-actin network did not appear until 21 postnatal days. When the intervertebral spine movements of rats aged 14 days are continuously restrained for 1 week, no regional differences of F-actin arose. These findings suggest that regional differentiation of the organization of F-actin network is related to the differing kinematics of the spine at different levels.

During the course of murine neurocysticercosis (NCC), disruption of the unique protective barriers in the central nervous system (CNS) is evidenced by extravasation of leukocytes. This process varies according to the anatomical sites and diverse vascular beds analyzed. To examine mechanisms involved in the observed differences, the expression and activity of eight matrix metalloproteinases (MMPs) were analyzed in a murine model of NCC. The mRNA expression of the MMPs studied was upregulated as a result of infection, and active MMPs were mainly detected in leukocytes migrating into the brain. Polarized expression and gelatinolytic activity of several MMPs were identified in immune cells extravasating pial vessels as early as 1 day post infection. In contrast, leukocytes expressing active MMPs and extravasating parenchymal vessels were not observed until 5 weeks post infection. In ventricular areas, most of the MMP activity was detected in leukocytes traversing the ependyma from leptomeningeal infiltrates. In addition, immune cells continued to express active MMPs after exiting vessels suggesting that enzymatic activity of MMPs is not just required for diapedesis. These results correlate with our previous studies showing differential kinetics in the disruption of the CNS barriers upon infection and help document the important role of MMPs during leukocyte infiltration and inflammation.

In peripheral organs, gonadal and adrenal steroids regulate diazepam-binding inhibitor (DBI) mRNA expression. In order to further investigate the involvement of peripheral steroid hormones in the modulation of brain DBI mRNA expression, we studied by semiquantitative in situ hybridization the effect of adrenalectomy (ADX) and castration (CX) and short-term replacement therapy on DBI mRNA levels in the male mouse hypothalamus. Cells expressing DBI mRNA were mostly observed in the arcuate nucleus, the median eminence and the ependyma bordering the third ventricle. In the median eminence and the ependyma bordering the third ventricule, the DBI gene expression was decreased in ADX rats and a single injection of corticosterone to ADX rats induced a significant increase in DBI gene expression at 3 and 12 h time intervals without completely restoring the basal DBI mRNA expression observed in intact mice. In the arcuate nucleus, ADX and corticosterone administration did not modify DBI mRNA expression. CX down-regulated DBI gene expression in the ependyma bordering the third ventricle. The administration of dihydrotestosterone (3-24 h) completely reversed the inhibitory effect of CX. In the median eminence and arcuate nucleus, neither CX or dihydrotestosterone administration modified DBI mRNA levels. These results suggest that the effects of glucocorticoids on the hypothalamo-pituitary-adrenocortical axis and androgens on the hypothalamo-pituitary-gonadal axis are mediated by DBI.

The CD14/toll-like receptor 4 (TLR4) complex plays a vital role in initiating lipopolysaccharide (LPS) signaling during inflammation. In this study, we assessed innate immune responses and inflammatory transmission in the rat brain following intracerebroventricular (i.c.v.) administration of LPS. I.c.v. LPS induced the widespread increase in CD14 mRNA but did not change levels of TLR4 transcription in the brain. An increase in TLR4 immunoreactivity, coincident with cell death, leukocyte infiltration and neural tissue damage, was found in the meninges, choroid plexus and ventricular ependyma. In addition to CD14, rapid increases in gene expression of IkappaBalpha, IL-1beta, and TNF-alpha occurred along the meninges and ventricular ependyma. The response was most intense along the borders of the brain and declined in intensity in the adjacent periventricular areas and cerebral cortex. In the brain parenchyma, increased TLR4 immunoreactivity was confined to the vasculature and neighboring tissues along with strong vascular expression of IkappaBalpha and mPGES-1. These results suggest involvement of TLR4 in both brain inflammation and neural tissue injury and support the hypothesis that local diffusion and vascular transmission of inflammatory molecules are two major routes for developing inflammation in the brain.

Ependymal primary cultures are a model for studying ependymal energy metabolism. Intracellular glycogen is built up in the cultures dependent on culture age and the presence of glucose and glutamate. This energy store is mobilized upon glucose withdrawal, stimulation with isoproterenol, forskolin or serotonin and after uncoupling of oxidative phosphorylation from ATP production. Serotonin regulates ependymal glycogen metabolism predominantly via 5-HT receptor (5-HTR) 7, which elicits an increase in the level of ependymal cyclic AMP. Although the most abundant mRNAs for serotonin receptors are those of 5-HTR 2B and 5-HTR 3A, ependymal cells in primary culture do not respond to serotonin with an increase in their concentration of cytosolic calcium ions. The mRNAs of 5-HTRs 1A, 6, 1B, 5B, 7, 1/2C and 5A are also detectable in order of decreasing abundance. The mRNAs for 5-HTRs 1D, 1F, 3B and 4 are absent from the cultured cells. The ability of serotonin to mobilize ependymal glycogen depends on the culture age and the time allowed for glycogen buildup. During glycogen buildup time, glutamate is consumed by the cells. An increased ability of 5-HT to mobilize ependymal glycogen stores is noticed after the depletion of glutamate from the glycogen buildup medium. In ependymal primary cultures, cilia are colocalized with glycogen phosphorylase isozyme BB, while the MM isoform is not expressed. It is known from the literature that an increase in the concentration of cytosolic cAMP in ependymal cells leads to a decrease in ciliary beat frequency. Therefore, the present data point towards a function for ependymal glycogen other than supplying energy for the movement of cilia.

The detailed distribution of neural nitric oxide synthase (nNOS)-positive cerebrospinal fluid-contacting neurons (CSF-CN) was studied in the wall of the third ventricle of rats by anti-nNOS immunohistochemistry. The coexistence of nNOS and 8-arginine vasopressin (AVP) or oxytocin (OT) was also investigated in the CSF-CN using double labeling immunohistochemistry. The results demonstrated a widespread occurrence of nNOS-CSF-CN throughout the wall of the hypothalamic third ventricle. The vast majority of nNOS-CSF-CN cell bodies were of magnocellular type, commonly classified as oval, fusiform, multipolar, and inverted pear shape. These cell bodies were located in the ependyma, the subependyma, or the parenchyma, and their processes inserted in the ependymal layer or directly contacted with the CSF space. Electron microscopy demonstrated many nNOS-immunoreactive somas, dendrites, and/or axons that were situated at the subependyma, the ependyma, or the supraependyma. Generally, the distribution of OT-CSF-CN in the third ventricular wall was similar to the nNOS-CSF-CN and the ratio of NOS/OT co-expression was approximately 88%. In comparison, the distribution of AVP-CSF-CN was mainly restricted to the rostral part of the third ventricle and the ratio of nNOS/AVP co-expression was only about 6%. The widespread presence of nNOS-CSF-CN-expressing OT in the third ventricular region suggests that NO is an important messenger in the CSF-hypothalamo-hypophyseal neuroendocrine regulation that may in part act in concert with OT.

Adult neural stem/progenitor cells proliferate throughout the life of the animal in the subependymal zone and the subgranular zone of the dentate gyrus (DG). Treatments such as enriched environment, dietary restriction, running and anti-depressants increase proliferation, however, stress and opiates have been shown to decrease proliferation. While models of binge ethanol drinking decreases proliferation, few studies have characterized the effect chronic ethanol usage has on progenitor cell proliferation. In this study, we have examined changes in the progenitor cell proliferation rate following chronic ethanol consumption. Animals were given a nutritionally balanced liquid diet containing 6.5% v/v ethanol or an isocalorically balanced liquid diet. Bromodeoxyuridine (BrdU) was administered (150 mg/kg×3) and the animals sacrificed 2 h after the last injection on days 3, 10 or 30 of the ethanol diet. Coronal brain blocks were paraffin embedded and 6 μm sections sliced and immunohistochemically stained for BrdU. Quantitation of the number of BrdU-labeled cells in the subgranular zone of the DG revealed a significant decrease only at the 3-day time-point, with recovery by the 10- and 30-day time-points. Thus, the progenitor cell proliferation rate is transiently decreased by chronic ethanol usage. This data suggests that chronic alcohol use results in a compensatory response that restores the progenitor cell proliferation rate.

Many studies have been made on the distributions of CSF contacting neurons (CSF-CNs) in the parenchyma of the brain with horseradish peroxidase (HRP) or autoradiographics. A significant amount of data has shown that both HRP and autoradiographical substances could pass freely through the spaces of ependyma into the parenchyma of the brain. It is therefore possible that the results were not exact. We found that CB-HRP was a dependable tracer to CSF-CNs and studied the distributions and the signaling directions of cerebrospinal fluid contacting neurons (CSF-CNs) in the parenchyma of the brain with the cholera toxin subunit B with horseradish peroxidase (CB-HRP) tracing combined with transmission electron microscopy. The results were as follows: (1) CSF contacting tanycytes existed not only in the wall of the third ventricle (3V), but also in the walls of the lateral ventricle (LV), the fourth ventricle (4V) and the central canal (CC) of the spinal cord. (2) Some CSF contacting glia cells were observed in the lateral septal nucleus (LS). (3)The distal CSF-CNs in the parenchyma were found in LS, the anterodorsal thalamic nucleus (AD), the supramammillary nucleus (SuM), the dorsal raphe nucleus (DR), the floor of 4V and the lateral superior olive (LSO), but they were mainly found in DR and divided into groups A and B. (4) Axon terminals labeled by CB-HRP were found in the cavity of the brain ventricle. (5) The synaptic relationships between the neurons were labeled by CB-HRP in DR and no-labeled by CB-HRP in the parenchyma. Both synapses Gray I and II were found. It was significant that the presynaptic elements were formed by the neurons no-labeled CB-HRP and the postsynaptic elements labeled CB-HRP. Our results suggested firstly that the signaling directions of CSF-CNs in DR were only from the parenchyma to CSF.

Cultivation of adult rat neural stem cells (RNSCs) from the ventricular subependyma has been reported to be more difficult than growth of mouse neural stem cells. This is unfortunate, because rats provide useful models of brain function and disease, and implantation of RNSCs in these models could provide critical information on allograft behavior. Growing the cells in an appropriate medium (NS-A+B27 supplement), plating at sufficient densities (5 cells per mm(2)), and minimizing opportunities for detachment from the substratum made it possible to isolate and cultivate these cells for over 6 months for50 passages with no apparent change in phenotype. Single clones could be expanded indefinitely and differentiated to form astrocytes, oligodendrocytes, and neurons, demonstrating that the cultures did indeed contain neural stem cells. The cells had a much shorter cell cycle time ( approximately 13 h) than doubling time ( approximately 35 h), suggesting that these cells produce post-mitotic cells in approximately two of three divisions, thus making expansion difficult. The optimization of methods to grow adult RNSCs and identification of characteristics that limit their growth should prove useful in increasing the use of RNSCs for studies of their potential role in brain health and disease.

Scanning the hypothalamus of rats for receptor binding sites of the octapeptide hormone angiotensin II (ANG II), we observed ANG II-sensitive fibres in the ventrolateral hypothalamus. The ANG II (AT1)-receptor-immunoreactive processes originate from cells—probably tanycytes—embedded in the base and the ventrolateral walls of the third ventricle and reach into the retrochiasmatic area, the ventrolateral hypothalamus and the median eminence.

The habenular-interpeduncular pathway is involved in the modulation of several functions including neuroendocrine and stress responses. Interleukin-18 (IL-18) is a pro-inflammatory cytokine predominantly studied as a modulator of immune functions and also produced in the adrenal cortex following activation of the hypothalamic-pituitary-adrenal axis. In the central nervous system, IL-18 was demonstrated to induce sleep and to influence long-term potentiation and was proposed to mediate local inflammatory reactions. The present study investigated the localization of IL-18 and its expression following either acute or chronic restraint stress in the brain of adult male Wistar rats. Using immunocytochemistry and in situ hybridization we report the unprecedented localization of IL-18 in the neurons of the superior part of the medial habenula (MHbS), their projections to the interpenducular nucleus and its expression in the ependymal cells surrounding the third and the lateral ventricles. In addition, acute (2 h) or chronic (6 h/day for 3 weeks) restraint stress induced a strong elevation of IL-18 immunostaining in the MHbS but not in ependymal cells. The present data suggest that IL-18 may participate in the modulation of stress responses in the MHbS. They also suggest that ependymal cells may be the source of IL-18 previously reported in the cerebrospinal fluid (CSF). The role of IL-18 in the ependyma and the CSF remains to be elucidated.

The distribution of nestin immunoreactivity was studied in the whole normal adult human forebrains using new anti-human nestin mouse monoclonal and rabbit polyclonal antiserum. The nestin immunoreactive cells could be divided into three types according to their morphological characteristics. The first type contained neuron-like nestin immunoreactive cells, distributed in CA1-3 of hippocampus, septum, the nucleus of diagonal band, amygdala and basal nucleus of Meynert. The second type contained astrocyte-like cells, distributed in the subependymal zone and subgranular layer of dentate gyrus. The third type of cells had smaller cell bodies and fewer processes, also distributed in the subependymal zone and subgranular layer of dentate gyrus. Double immunohistochemical staining showed that the nestin positive, neuron-like cells in the nucleus of diagonal band and hippocampus also expressed NSE. However, the astrocyte-like nestin immunoreactive cells of the subependymal zone and subgranular layer of dentate gyrus were not double labeled with GFAP. Although some nestin immunoreactive fibers were distributed in the infundibulum, no nestin-immunoreactive cells were detected in the cortex. These data indicate that nestin exist in the adult human brain outside of the subependymal zone and dentate gyrus and also implies that nestin-immunoreactive cells may play a role in the modulation of basal forebrain function.

Ependymal cells form a single-layered, ciliated epithelium at the interface between the cerebrospinal fluid and the brain parenchyma. Although their morphology has been studied in detail, ependymal functions remain largely speculative. We have established and characterized a previously described cell culture model to investigate ependymal glycogen metabolism. During growth in minimal medium lacking many non-essential amino acids including L-glutamate, but containing glucose at physiological concentration, the cells contained negligible amounts of glycogen (7+/-3 nmol glucosyl residues/mg protein) despite the presence of insulin. However, during a period of 24 h, the cells accumulated glycogen to very high levels after transferal to a medium containing insulin, glucose at a 5-fold higher concentration, and all proteinogenic amino acids except L-asparagine and L-serine (990+/-112 nmol glucosyl residues/mg protein). Omission of insulin resulted in a 50% reduction in glycogen accumulation. Upon glucose deprivation, glycogen was degraded with a half-life of 21 min. The ependymal primary cultures contained 80+/-5 mU glycogen phosphorylase (Pho)/mg protein and stained positively with antibodies raised against this enzyme. Astroglial cultures built up less glycogen and had less Pho activity under identical conditions. Ependymal glycogen was mobilized by noradrenaline and serotonin. Our results indicate that ependymal cells maintain glycogen as a functional energy store, subject to rapid turnover dependent on the availability of energy substrates and the presence of appropriate signal molecules. Thus ependymocytes appear to be active players in the multitude of processes resulting in normal brain function, and ependymal primary cultures are suggested as a suitable model for studying the role of ependymal cells in these processes.

Human mutations in the transcription factor gene, SOX9, cause campomelic dysplasia (CD), a severe dwarfism associated with brain abnormalities including dilation of lateral ventricles, hypoplasia of the corpus callosum and cerebellum defects. To improve our understanding of how SOX9 contributes to the molecular genetic pathway of brain development we sought to investigate the distribution of SOX9 protein in rat and mouse brain. The regions of SOX9 expression identified in this study correlated with the sites of reported brain abnormalities in CD patients. SOX9 immunoreactivity was observed in nuclei of scattered cells throughout the brain, in the ependymal layer and cells of the choroid plexus. In the forebrain most SOX9-immunoreactive nuclei co-localised with the glial astrocyte marker S-100. In the cerebellum, SOX9 was observed mostly in cells surrounding Purkinje cells, which were identified, by electron microscopy, as Golgi epithelial cells, also known as Bergmann glia. Using SOX9 antibody as a marker for the precursors of Bergmann glia, we traced their origin during mouse development. At embryonic day (E)14.5 and E16.5, SOX9 immunoreactivity was present mainly in the primordial choroid plexus, and ventricular zone. By E18.5, SOX9 was observed in the granular cell and Purkinje cell layers but no labelling was detectable in the external granular layer. These results suggest that SOX9 immunoreactivity is a marker for Bergmann cells during development and favour the proposed origin of the secondary glial scaffold arising from Bergmann cells derived exclusively from the ventricular zone.

The neural adhesion molecule L1, a member of the immunoglobulin superfamily of cell recognition molecules, performs important functions in the developing and adult nervous system. This view is confirmed by the fact that mutations in the human L1 gene cause a severe neurological disease, termed CRASH (acronym for: corpus callosum hypoplasia, retardation, adducted thumbs, spastic paraplegia, and hydrocephalus). X-linked hydrocephalus is certainly the most prominent symptom of CRASH syndrome. Mouse mutants deficient in L1 also develop enlarged ventricles. Here, we report that ventricular dilation in L1-deficient mice is not correlated with stenosis of the aqueduct of Sylvius nor with ultrastructural abnormalities of ependymal cells lining the lateral ventricles or the aqueduct. However, a few L1 mutant mice displayed severe hydrocephalus, characterized by a significant enlargement of the skull and an almost complete atrophy of the cerebral cortex. The aqueduct of these severely affected animals was completely closed. Since mutant animals from two independently generated L1-deficient mouse lines displayed a similar phenotype, we consider severe hydrocephalus as a specific consequence of L1-deficiency. However, results of the present study also indicate that severe hydrocephalus represents a secondary rather than a primary defect of the L1 mutation; our combined data suggest that deformations of the brain as a result of massively enlarged ventricles secondarily cause stenosis of the aqueduct and subsequently high pressure hydrocephalus.

Transthyretin (TTR) is involved in the transport of thyroxine (T4) and retinol-binding protein (RBP) in cerebrospinal fluid (CSF) and serum. TTR is secreted in the CSF by the epithelial cells of choroid plexus. The binding of [(125)I]TTR to cultured ependymoma cells which form the brain cerebrospinal barrier, was studied to determine whether these cells carry receptor(s) for TTR. TTR was bound by ependymoma cells in a time-dependent manner reaching equilibrium within 2 h. Scatchard analysis was consistent with a single class of high-affinity binding sites with a K(d) of approximately 18 nM. Saturable high-affinity binding of human TTR has previously been described in rat primary hepatocytes and human renal adenocarcinoma, neuroblastoma, hepatoma and astrocytoma cells, and also transformed lung cells. Endocytosis of fluorescent or biotinylated TTR was observed in ependymoma cells in cytoplasmic vesicles but TTR did not colocalize with clathrin in endocytic coated vesicles. Endocytosis of TTR was inhibited by high sucrose concentration (0.45 M). Finally, ligand blotting and chemical-linking experiments revealed the presence of a approximately 100 kDa putative TTR receptor on the ependymoma cell membrane. Receptor binding of TTR provides a potential mechanism for the delivery of T4 within the central nervous system.

Type 2 iodothyronine deiodinase, an enzyme involved in the conversion of thyroxin to the biologically active 3,5,3′-triiodothyronine, is highly concentrated in a group of specialized ependymal cells, tanycytes, lining the wall and floor of the third ventricle. As this distribution is highly reminiscent of the distribution of cells containing the phosphatase inhibitor, DARPP-32, we raised the possibility that these two proteins may coexist in tanycytes and that DARPP-32 may modulate type 2 deiodinase activity by regulating the phosphorylation state of the cAMP regulatory factor, CREB. To address this question, double-labeling histochemical studies were performed for type 2 deiodinase mRNA and DARPP-32 immunoreactivity (IR), or DARPP-32- and CREB-IR in the same tissue sections. Type 2 deiodinase mRNA was found in the cell bodies of all DARPP-32-immunolabeled tanycytes. Both type 2 deiodinase mRNA and DARPP-32-IR also extended into tanycyte processes that ramified in the arcuate nucleus and median eminence, in close association with blood vessels and portal capillaries. In contrast, type 2 deiodinase mRNA was not present in the same cells that contained DARPP-32-IR in the pituitary gland. All tanycytes containing DARPP-32-IR also contained CREB-IR in their nucleus. Since type 2 deiodinase activity can be induced by substances that increase cAMP, we hypothesize that DARPP-32 may regulate the activity of type 2 deiodinase by prolonging the activation of CREB. Selectivity for the colocalization of these factors to tanycytes but not the pituitary gland, may explain the heterogeneous response of type 2 deiodinase activity in these two loci in response to specific stimuli such as fasting.

CRF and urocortin, administrated systemically, exert peripheral biological actions which may be mediated by brain pathways. We identified brain neuronal activation induced by intravenous (i.v.) injection of CRF and urocortin in conscious rats by monitoring Fos expression 60 min later. Both peptides (850 pmol/kg, i.v.) increased the number of Fos immunoreactive cells in the paraventricular nucleus of the hypothalamus, supraoptic nucleus, central amygdala, nucleus tractus solitarius and area postrema compared with vehicle injection. Urocortin induced a 4-fold increase in the number of Fos-positive cells in the supraoptic nucleus and a 3.4-fold increase in the lateral magnocellular part of the paraventricular nucleus compared with CRF. Urocortin also elicited Fos expression in the accessory hypothalamic neurosecretory nuclei, ependyma lining the ventricles and choroid plexus which was not observed after CRF. The intensity and pattern of the Fos response were dose-related (85, 255 and 850 pmol/kg, i.v.) and urocortin was more potent than CRF. Neither CRF nor urocortin induced Fos expression in the lateral septal nucleus, Edinger-Westphal nucleus, dorsal raphe nucleus, locus coeruleus, or hypoglossal nucleus. These results show that urocortin, and less potently CRF, injected into the circulation at picomolar doses activate selective brain nuclei involved in the modulation of autonomic/endocrine function; in addition, urocortin induces a distinct activation of hypothalamic neuroendocrine neurons.

The molecular composition and functional properties of cell-cell junctions of choroid plexus epithelial cells and the ependyma of the lateral ventricular wall were investigated in the rat brain. Expression studies of cadherin and alpha- and beta-catenins, as well as expression of occludin and ZO-1, indicated that cell adherens and tight junctions were present in both choroid plexus epithelial cells and in ependymal cells. We then tested the hypothesis that phorbolester in vivo can induce changes in the expression level of adherens and tight junction molecules at the blood-cerebrospinal fluid (CSF) barrier as well as in the ependyma. In addition, the functional properties of the ependymal junctions were tested by injection of dextran 3000 into the striatum after phorbolester application. Twenty-four hours after phorbolester-injection into the lateral ventricle of the rat brain, the expression patterns of tight and adherens junction molecules were markedly changed in the epithelial cells of the choroid plexus. The adherens junction proteins cadherin and beta-catenin were reduced in both the ependymal cells of the lateral ventricle and choroid plexus epithelial cells. In addition, the occludin-immunoreactivity of the choroid plexus epithelial cells was strongly reduced. However, the ZO-1 immunoreactivity was not affected by the phorbol ester-treatment and the alpha-catenin immunoreactivity was not changed. Furthermore, phorbol ester injection induced a reduction of the volume of intrastriatal injected biotinylated dextran (m.w. 3000), which is consistent with a modulatory influence of protein kinase C activation on the clearance capacity of the brain.

To investigate the function of the major non-muscle dystrophin isoform, Dp71, we substituted a beta-galactosidase (betagal) reporter gene for Dp71 by homologous recombination in embryonic stem cells. Staining for betagal activity in chimeric mice revealed Dp71 promoter activity in glial cells in the CNS, in neurons of the inner nuclear and inner plexiform layers of the retina, and in the kidney tubules and collecting ducts. Our observations demonstrate that Dp71 is widely expressed in the adult CNS (retina, cerebellum, cerebral cortex, ependyma, and choroid) as well as the adult kidney epithelium and suggest a broad function for Dp71 in differentiated tissues.

In scrapie infection, prion protein (PrP Sc ) is localized in areas where there is neurodegeneration and astrocytosis. It is thought that PrP Sc is toxic to neurons and trophic for astrocytes. In our study, paraffin sections from scrapie infected (263K and 139H) and control hamsters were examined with histological and immunocytochemical staining. We found that PrP Sc was present in the ependymal cells of both 263K- and 139H-infected hamsters. In 139H-infected hamsters, PrP Sc was found in the cytoplasm of neurons in cerebral cortex and in hypothalamic paraventricular (PVN) and supraoptic (SON) nuclei. In contrast, neuronal cytoplasm and nuclei, were positive for PrP Sc in most areas such as cortex, hippocampus, and thalamus in 263K-infected hamsters. Many aggregations of PrP Sc could be seen in the cortex, hippocampus, substantia nigra and around the Pia mater, corpus callosum, fimbria, ventricles, and blood vessels in sections from 139H- and/or 263K-positive animals. Furthermore, PrP Sc was also co-localized with glial fibrillary acidic protein (GFAP) in many reactive astrocytes (approximately 90%) in certain areas such as the hippocampus in 263K-infected hamsters, but not 139H-infected hamsters. The patterns of astrocytosis and PrP Sc formation were different between 139H- and 263K-infected hamsters, which may be used for a diagnosis purpose. Our results suggest a hypothesis that multiple cell-types are capable of PrP Sc production. Our results also confirm that reactive astrocytes can produce and/or accumulate PrP Sc during some scrapie strain infections. The findings suggest a `snowball effect', that is: astrocytosis might play an important role in amyloidosis, while amyloidosis may induce further astrocytosis at least in 263K-infected hamsters.

The presence and synthesis of transthyretin, a major carrier protein of thyroxine in rat cerebrospinal fluid, was investigated in choroid plexus epithelial cells and ependymal cells by immunocytochemistry, in situ hybridization, and analysis by Northern and Western blot using a specific oligonucleotide probe and a specific polyclonal antibody to transthyretin. Choroid plexus epithelial cells expressed transthyretin at high levels in developing rat cerebral hemispheres and in cultured cells. These cells secreted transthyretin into the cerebrospinal fluid. In the developing rat brain transthyretin was present in the cytoplasm of ependymal cells, in vesicles in contact with the apical membrane and in cilia. In ependymal cell cultures this protein was particularly abundant in the cilia of these cells. In contrast, ependymal cells did not synthesize transthyretin. It is postulated that transthyretin is transported to ependymal cells from the cerebrospinal fluid by endocytosis.

10-Formyltetrahydrofolate dehydrogenase (10-FTHFDH) is a folate-binding protein that is important in folate metabolism. In addition, 10-FTHFDH catalyzes the rate-limiting step in hepatic folate-dependent formate oxidation. We measured folate concentrations and examined cellular 10-FTHFDH expression in regions of the adult rat central nervous system (CNS), to study components of CNS oxidative formate metabolism. Folate was detected in every CNS region studied; the concentrations ranged from 3 to 14% of that detected in the liver. Immunohistochemical expression of 10-FTHFDH was identified in many CNS structures. 10-FTHFDH was mostly expressed by glia, especially astrocytes and ventricular ependyma. Neuronal expression was weak but evident in the cerebral cortex, basal ganglia, cerebellum, and spinal cord. Thus, CNS tissue has the chief components of folate-dependent formate oxidation and the chief site of this oxidation appears to be glia.

A ganglion-like system of intraventricular neuronal perikarya and processes known as the supraependymal neuronal complex (SENC) is located on the floor of the third ventricle of the golden hamster. Serotonergic neuronal elements were identified in the SENC by immunocytochemical techniques. Serotonin (5-hydroxytryptamine, 5-HT) was not detected within supraependymal neuronal perikarya. However, serotonergic nerve fibers were found to project via the median eminence both onto the surface and into the neuropil of the SENC. The SENC is thought to be involved in neuroendocrine regulation since it innervates the pituitary gland. The results of the present study suggest that serotonin plays a role in SENC/pituitary gland interaction.

Monoclonal antibodies (mAbs) D3G6 and C8A5, specific for amino acid residues 160-168 of S182 protein, immunolabeled neurons, ependymal and choroid plexus cells, and myocytes in brain sections from normal subjects and people with Alzheimer disease or Down syndrome and in rats and mice. Oligodendroglia, microglia, and the majority of astrocytes were negative. S182 protein or a fragment of the protein detected with these mAbs is not a constituent of amyloid-beta deposits or tangles.

We investigated the immunohistochemical localization of the macrophage migration inhibitory factor (MIF) in the rat brain. In addition to epithelial ependymal cells lining the ventricular wall, tanycytes in the basomedial hypothalamus were heavily immunostained. The immunoreactive processes of tanycytes made contacts to sinusoidal capillaries and reached the pial surface forming an immuno-positive structure at the floor of the hypothalamus. Other immunoreactive cells contained the subcommissural organ in the roof of the third ventricle and the epithelial lamina of the choroid plexus. The localization of MIF in cells which have contact with cerebrospinal fluid and blood vessels suggests that MIF might play a role as a humoral factor in the brain.

In the CNS of vertebrates, although the subcommissural organ (SCO) has been identified as an ependymal brain gland and Reissner's fiber (RF) as a condensed product of its secretion, the exact nature of the secretory substances has remained elusive. In the present study, immunocytochemical application of polyclonal antibodies against calcitonin revealed intense immunoreactivity in the cells, cell processes and cerebrospinal spinal fluid (CSF)-contacting apical terminals of the columnar ependymal cells of the SCO in the teleost, Clarias batrachus. Intense immunoreactivity was also seen throughout the length of the RF as it extended along the Sylvian aqueduct, fourth ventricle and central canal of the spinal cord. Control procedures were employed to confirm the specificity of the immunoreaction. The results for the first time suggest that calcitonin-like substance may be the synthetic and secretory product of the SCO that may be released into the CSF or stored in the RF. Presence of calcitonin-like immunoreactivity in the SCO-RF complex of the frog Rana tigrina and the lizard Calotes versicolor underscores wider significance of the phenomenon. In view of the potentials of these findings, it is felt that investigations aimed at establishing the precise nature of calcitonin-like immunoreative material in the SCO-RF complex may be rewarding. © 1997 Elsevier Science B.V. All rights reserved.

Polyclonal antibodies were generated against ADP-ribosylarginine hydrolase (AAH), using recombinant fusion protein of rat AAH and glutathione-S-transferase as a immunogen, and affinity-purified. Western blotting showed that the antibodies recognized in mouse brain homogenate a single protein with a molecular mass of 38 kDa, the expected size for mouse AAH. An analysis using the antibodies revealed that heavy labelings were apparent in various brain regions. In the cerebral cortex, pyramidal cells in layers III and V were the most heavily labeled. In the hippocampal formation, labeling was present on the pyramidal neurons and granule cells. The most heavily immunostained cell type was the pyramidal neuron of CA3. In the cerebellum, Purkinje cells were the most heavily labeled. Less intense staining was present over the granule cells. In the basal ganglia, neurons in the caudate nucleus and large multipolar cells in the amygdaloid complex were immunoreactive. Heavy labeling was seen in many midbrain and brainstem nuclei. Neurons in the habenula and ependymal cells were stained heavily. On Western blot analysis of rat cerebrospinal fluid (CSF), the anti-AAH antibodies recognized a protein with a molecular mass of 38 kDa. This is apparently the first evidence of a widespread but distinctive distribution of AAH in neurons of mouse brain and the presence of extracellular AAH in rat CSF.

In previous studies with animal models of spinal cord injury (SCI) pharmacological activation of peroxisome proliferator activated receptors (PPAR) and liver X receptors (LXR) were used to reduce tissue damage and promote behavioral recovery in animal models. We have studied the endogenous expression of the transcription factors PPARα and LXRβ in the chronic stage after SCI in rats. The immunohistochemical investigation revealed a long lasting increase in the level of PPARα in white matter in the vicinity of the lesion site. The source of this signal was identified in a subpopulation of astrocytes outside of the glial scar area. Intrathecal injections of oleic acid/albumin reduced the lesion-induced PPARα immunoreactivity. In addition, ependymal cells displayed a prominent PPARα signal in the non-injured spinal cord, and continued to express the receptor as they proliferated and migrated within the damaged tissue. The nuclear receptor LXRβ was detected at similar levels after SCI as in sham operated animals. We found high levels of immunoreactivity in the gray matter, while in the white matter it was present in subpopulations of astrocytes and oligodendrocytes. Macrophages that had accumulated within the center of the lesion contained LXRβ in their cell nuclei. Possible endogenous functions of PPARα and LXRβ after SCI are discussed, specifically the control of fatty acid and cholesterol metabolism and the regulation of inflammatory reactions.

This study shows that in the choroid plexus of Rana esculenta particulate guanylate cyclase (GC) is appreciably stimulated by porcine brain natriuretic peptide (BNP). Ultracytochemical tests for GC show that BNP notably increases the enzymatic reaction product along the apical surfaces of the epithelial cells. It can therefore be hypothesized that the apical zone of the epithelial cells possess receptors which have a particular affinity for BNP produced in the central nervous system and dumped into the cerebrospinal fluid. These results, together with those of a previous study [32], confirm that the choroid plexus is an organ which has receptors for the natriuretic peptides which are involved in the processes of osmoregulation and the control of cerebrospinal fluid production.

Our previous studies demonstrated that tooth pulp stimulation increases the pain threshold, whereas stimulation of the periaqueductal central gray (PAG) has the opposite effect. The aim of this study was to investigate the effect exerted by electrical stimulation of the nociceptive afferent terminals in the tooth pulp and analgesic electrical stimulation of the PAG on the release of immunoreactive β-endorphin (β-EP-IR) into the cerebrospinal fluid (CSF) and artificial cerebrospinal fluid (aCSF) perfusing the cerebral ventricles in rat, and to establish whether during such stimulation the μ-opioid receptor (MOR) was activated. The tongue jerk reflex was induced by dental pulp stimulation in rats under chloralose anesthesia. CSF was collected from the cerebello-medullary cistern, and then 30-minute perfusions of the lateral ventricles with aCSF were conducted with collection of perfusate portions from the cerebello-medullary cistern at rest (control), during electrical tooth pulp stimulation evoking the nociceptive tongue jerk reflex and during inhibition of that reflex with simultaneous stimulation of tooth pulp and the PAG. In the second series of experiments, a MOR-selective antagonist, β-funaltrexamine (β-FNA) was perfused through the cerebral ventricles 10min before tooth pulp stimulation. β-EP-IR was determined in the collected CSF and aCSF perfusates by radioimmunoassay (RIA). Stimulation of tooth pulp with electrical impulses resulted in a significant increase of β-EP-IR. On the other hand, simultaneous tooth pulp and PAG stimulation inhibited β-EP-IR release into the fluid perfusing the cerebral ventricles. β-FNA blocked evoked tongue jerks (ETJ) induced by PAG stimulation, but no enhancing effect of β-FNA on β-EP-IR release into the perfusate was observed. The obtained results indicate that stimulation of the tooth pulp increases β-EP release significantly, whereas PAG stimulation significantly inhibits β-EP-IR release into the CSF, and that these effects are mediated by MOR. The results of the experiments allow to conclude that endogenous β-EP, released as a result of electrical tooth pulp stimulation in orofacial pain, diffuses through the cerebroventricular ependyma into the CSF and exerts a modulatory effect, mediated by MOR, alterating the properties of neurons in the trigeminal sensory nuclei, interneurons, and motoneurons of the hypoglossal nerve.

Distribution of adult T cell leukemia derived factor (ADF/TRX), a human thioredoxin homologue, was studied by immunohistochemistry in gerbil brain during reperfusion after transient cerebral ischemia. In control brains, immunoreactivity was observed widely in the central nervous system, including the ependyma, tanycytes, endothelial cells as well as subcommisural organs, and weakly in the neuronal cell bodies. During reperfusion, ADF/TRX was expressed in glial cells in the CA1 and dentate hilus of the hippocampus, and in a few cases in the lateral portion of the caudateputamen. ADF/TRX positive glias first appeared after reperfusion for 24 h, and the intensity of staining peaked at 72 h and diminished after 7 days of reperfusion. They were concentrated around microvessels and identified to be astroglia by double labeling immunohistochemistry with glial fibrillary acidic protein as marker. Immunoblotting analysis demonstrated an increase of ADF/TRX in the postischemic hippocampus. Astroglial expression of ADF/TRX in postischemic injuries suggests a role in neuroprotection by hydrogen peroxide reducing and protein-refolding activities or in modulation of local immune responses.

Neurons containing immunoreactivity to the endogenous lactose-binding lectin, RL-29, were examined in the L4 segment of the spinal cord of rats with an experimental neuropathy. The cells appeared by 5 days and were also present at 14 and 28 days postoperatively. All neurons were found in the dorsal horn ipsilateral to the injury. The neurons were multipolar and the reaction product revealed the morphology of the primary and secondary dendrites and some axons. Most of the neurons were located in the reticulated region of the dorsal horn, corresponding to Rexed's lamina V. In 14-day neuropathy animals treated with the NMDA-receptor antagonist MK-801, the number of RL-29 cell profiles observed was significantly reduced. Double labeling experiments revealed that spinothalamic tract neurons did not contain RL-29. The results suggest that the neuropathic injury produces a long term, transynaptic change in a subpopulation of dorsal horn neurons, that is mediated by excitatory amino acid transmitters acting at NMDA receptors.

Our previous proteomic studies disclosed upregulation of alphaB-crystallin, a small heat shock protein, in the brain tissue of Ts65Dn mice, a mouse model for Down syndrome (DS). To validate data obtained in model animals, we studied at present the levels and distribution of total alphaB-crystallin and its forms phosphorylated at Ser-45 and Ser-59 in the brain tissues of DS subjects and age-matched controls at 4 months to 23 years of age. On immunoblots from frontal cortex and white matter, alphaB-crystallin and its form phosphorylated at Ser-59 were detectable already in infants, whereas alphaB-crystallin phosphorylated at Ser-45 appeared in small amounts in older children. Although the levels of total alphaB-crystallin were modestly increased in DS subjects, the amounts of both phosphorylated forms were much higher (up to approximately 550%) in the group of older children and young adults with DS than in age-matched controls. Immunoreactivity to alphaB-crystallin occurred not only in a subset of oligodendrocytes and some subpial and perivascular astrocytes, which was reported earlier, but also in GFAP-positive astrocytes accumulating at the sites of ependymal injury as well as some GFAP/platelet-derived growth factor receptor alpha-positive cells in both DS and control brains, which is a novel observation. Given that the chaperone and anti-apoptotic activities of alphaB-crystallin are phosphorylation-dependent, we propose that enhanced phosphorylation of alphaB-crystallin in the brains of young DS subjects might reflect a cytoprotective mechanism mobilized in response to stress conditions induced or augmented by the effect of genes encoded by the triplicated chromosome 21.

The organization and chemical development of the area postrema (AP) in the macaque monkey was studied by immunohistochemistry imaged with conventional and confocal microscopy from day 40 of gestation to adulthood. The thin ependyma of the adult was found to develop from a thick continuous structure beginning in the second trimester. It was later invaded by tyrosine hydroxylase immunoreactive (TH+) and dopamine beta-hydroxylase immunoreactive (DBH+) cells and fibers, suggesting a possible route for release of neurotransmitter directly into ventricular cerebrospinal fluid. Other TH+ and/or DBH+ fibers were found in close approximation to blood vessels. Prominent vascularity of the parenchyma of AP was present late in the first trimester (fetal day (Fd)57 in the macaque) and increased further until birth. By contrast, the underlying solitary nucleus was hypervascular at Fd57, but its vascularity rapidly declined by late in the second trimester. TH+ neurons first appeared late in the first trimester, and DBH+ neurons appeared in the second trimester; these findings are consistent with the view that catecholaminergic cells in AP are the earliest members of the A2 noradrenergic group. Catecholaminergic cells or fibers in AP contained little labeling for synaptic vesicular proteins, suggesting that the release of neurotransmitter there may not involve a synaptic mechanism. Synapses were first observed in mid-second trimester, and most were associated with GABA+ fibers.

Focal brain injury or perforant-path transections respectively led to an increase in the number of glial-fibrillary acidic protein (GFAP)-immunopositive astrocytes around the focal wound or in the terminal fields of the perforant path in the dentate molecular layer. This GFAP accumulation occured 48–72 h after focal brain injury or perforant-path transection (wallerian degeneration). Focal brain injury also led to an accumulation of c-fos protein (Fos) in glial cells, ependyma and cells in the pia mater of the brain within 6 h of injury and this effect dissipated within 72 h. However, perforant-path lesions were not associated with accumulation of Fos in glial cells in the dentate molecular layer suggesting that c-fos induction in glial cells after injury is not necessary for GFAP accumulation. Induction of Fos in glia, ependyma and pia after focal brain injury may be associated with proliferation of these cells after injury.

Neurons that utilize melanin-concentrating hormone (MCH) as a neuromodulator are localized within the postero-lateral hypothalamus and zona incerta. These neurons project diffusely throughout the central nervous system and have been implicated in critical physiological processes such as energy homeostasis and sleep. In the present report, we examined the distribution of MCH immunoreactivity in the brainstem of the cat. In addition to MCH+ axons, we found MCH-immunoreactive cells that have not been previously described either in the midbrain raphe nuclei or in the periaqueductal and periventricular areas. These MCH+ cells constituted: 1. ependymal cells that lined the fourth ventricle and aqueduct, 2. ependymal cells with long basal processes that projected deeply into the subventricular (subaqueductal) parenchyma, and, 3. cells in subventricular regions and the midbrain raphe nuclei. The MCH+ cells in the midbrain raphe nuclei were closely related to neuronal processes of serotonergic neurons. Utilizing Neu-N and GFAP immunohistochemistry we determined that the preceding MCH+ cells were neither neurons nor astrocytes. However, we found that vimentin, an intermediate-filament protein that is used as a marker for tanycytes, was specifically co-localized with MCH in these cells. We conclude that MCH is present in tanycytes whose processes innervate the midbrain raphe nuclei and adjacent subependymal regions. Because tanycytes are specialized cells that transport substances from the cerebrospinal fluid (CSF) to neural parenchyma, we suggest that MCH is absorbed from the CSF by tanycytes and subsequently liberate to act upon neurons of brainstem nuclei.

This study investigated the transfer of T4 from cerebrospinal fluid (CSF) into the choroid plexuses (CP) and ventricular brain regions, and the role of P-glycoprotein (P-gp), multidrug resistance protein 1 (mrp1) and organic anion transporting polypeptides (oatps). During in vivo ventriculo-cisternal (V-C) perfusion in the anesthetized rabbit (meditomidine hydrochloride 0.5 mg kg(-1), pentobarbitone 10 mg kg(-1) i.v.), 125I-T4 was perfused continuously into ventricular CSF with reference molecules 14C-mannitol and blue dextran. Over 2 h, 36.9+/-4.6% 125I-T4 was recovered in cisternal CSF. Addition of P-gp substrate verapamil increased CSF 125I-T4 recovery to 51.4+/-2.8%, although mrp1 and oatp substrates had no significant effect. In brain, 125I-T4 showed greatest accumulation in the CP (1.52+/-0.31 ml g(-1)), followed by ventricular regions (caudate putamen, ependyma, hippocampus, 0.05-0.14 ml g(-1)). At the CP, verapamil and probenecid (but not indomethacin) significantly increased 125I-T4 accumulation, implicating a role for P-gp and oatps. Of other brain regions, all three drugs increased accumulation in caudate putamen 3-5 times, and indomethacin and probenecid increased accumulation in ependyma 4-5 times. The role of P-gp was investigated further in isolated incubated CPs using 5 microg/ml C219 anti-P-gp antibody. Both 125I-T4 and 3H-cyclosporin accumulation increased by 80%, suggesting that P-gp is functional in the CP and has a role in removal of T4. Combined with the in vivo results, these studies suggest that P-gp provides a local homeostatic mechanism, maintaining CSF T4 levels. We conclude that P-gp and oatps contribute to the transfer of 125I-T4 between the CSF, CP and brain, hence regulating 125I-T4 availability in CSF.