Physiological relevance and functional potential of central nervous system-derived cell lines.

Central nervous system (CNS)-derived neural cell lines have proven to be extremely useful for delineating mechanisms controlling such diverse phenomena as cell lineage choice and differentiation, synaptic maturation, neurotransmitter synthesis and release, and growth factor signalling. In addition, there has been hope that such lines might play pivotal roles in CNS gene therapy and repair. The ability of some neural cell lines to integrate normally into the CNS following transplantation and to express foreign, often corrective gene products in situ might offer potential therapeutic approaches to certain neurodegenerative diseases. Five general strategies have evolved to develop neural cell lines: isolation and cloning of spontaneous or mutagenically induced malignancies, targeted oncogenesis in transgenic mice, somatic cell fusion, growth factor mediated expansion of CNS progenitor or stem cells, and retroviral transduction of neuroepithelial precursors. in this article, we detail recent progress in these areas, focusing on those cell lines that have enabled novel insight into the mechanisms controlling neuronal cell lineage choice and differentiation, both in vitro and in vivo.

Lumbar transplants of immortalized serotonergic neurons alleviate chronic neuropathic pain.

The RN46A cell line was derived from embryonic day 13 rat medullary raphe cells by infection with a retrovirus encoding the temperature-sensitive mutant of SV40 large T antigen. This cell line is neuronally restricted and constitutively differentiates following a shift to non-permissive temperature. Brain-derived neurotrophic factor (BDNF) induced the serotonergic phenotype and increased the survival of RN46A cells in vitro. After transfection of the rat BDNF gene into RN46A cells, an autocrine BDNF-secreting cell line, 46A-B14, was isolated and transplanted into the rat CNS. Transplanted 46A-B14 cells had increased survival and enhanced serotonin (5HT) synthesis compared to 46A-V1 cells, RN46A cells transfected with vector-alone. When 46A-B14 cells were transplanted in the lumbar subarachnoid space of the spinal cord 1 week after a chronic constriction injury (CCI) of the sciatic nerve, they survived longer than 6 weeks on the pia mater. Furthermore, the tactile and cold allodynia and thermal hyperalgesia induced by CCI was significantly reduced during a 4-6- week period. The maximal effect occurred 1 week after transplantation. 46A-V1 cells, transplanted after CCI, did not survive beyond 2-3 weeks and had no effect on the allodynia and hyperalgesia induced by CCI. Acute intrathecal injection of the 5HT receptor antagonist methysergide decreased the antinociceptive effects of the 46A-B14 cells to pre-transplant levels. These data suggest that a chronically applied, low local dose of serotonin near the dorsal horn was able to reverse the development of chronic neuropathic pain following CCI. The use of neural cell lines that are able to deliver inhibitory neurotransmitters such as serotonin, in a model of chronic pain offers a novel approach to pain management.

Differentiation of an immortalized CNS neuronal cell line decreases their susceptibility to cytotoxic T cell lysis in vitro.

RN33B cells are a temperature-sensitive neuronal cell line derived from rat E12 medullary raphe nucleus (Whittemore and White (1993) Brain Research 615, 27-40). Undifferentiated RN33B cells express class I but not class II antigens of the major histocompatibility complex (MHC), and intercellular adhesion molecule-1 (ICAM-1), a ligand for lymphocyte function associated antigen-1 (LFA-1), expressed on cytotoxic T lymphocytes (CTLs). Treatment of undifferentiated RN33B cells with interferon-gamma (IFN-gamma) upregulated both class I MHC and ICAM-1. After neuronal differentiation, expression of class I MHC antigens or ICAM-1 was undetected, even after IFN-gamma treatment. The neuronally differentiated RN33B cells were also markedly less susceptible to lysis by alloantigen-specific CTLs. These data suggest that intrinsic to the differentiation of CNS neurons is a mechanism to escape CTL-mediated cell lysis.

Expression of the receptors for the C5a anaphylatoxin, interleukin-8 and FMLP by human astrocytes and microglia.

The expression of chemotactic receptors in the central nervous system is largely unexplored. In this study, we examined human astrocytes and microglia as well as the conditionally immortalized human astrocyte cell line HSC2 for expression of the C5a-anaphylatoxin receptor (C5aR), the interleukin-8 receptor (IL-8R) and the f-Met-Leu-Phe receptor (FMLPR). Using flow cytometry, indirect immunofluorescence and RT-PCR analysis, we demonstrated that astrocytes, microglia and HSC2 cells contain specific RNA and express surface protein for all three chemotactic receptors. These are the first studies to demonstrate definitively the expression of these chemotactic receptors astrocytes and microglia, thereby expanding the types of cells known to express chemotactic receptors. Moreover, these data suggest that these chemotactic receptors may play an important role in mediating the inflammatory response and perhaps other yet undescribed biological phenomena in the central nervous system.

Developmental regulation of tryptophan hydroxylase messenger RNA expression and enzyme activity in the raphe and its target fields.

Tryptophan hydroxylase is the rate-limiting enzyme in the synthesis of serotonin and during development, brain serotonin levels and tryptophan hydroxylase activities increase. Increased tryptophan hydroxylase activity could result from alterations in tryptophan hydroxylase messenger RNA levels, translation, and/or post-translational regulation. Tryptophan hydroxylase messenger RNA levels in the dorsal raphe nucleus increased 35-fold between embryonic day 18 and postnatal day 22, measured by quantitative in situ hybridization, then decreased by 40% between postnatal days 22 and 61. These changes correlated with tryptophan hydroxylase enzyme activities in the raphe nuclei as expected, but not in cortical or hippocampal targets. Tryptophan hydroxylase messenger RNA expression in the nucleus raphe obscuris increased 2.5-fold between postnatal days 8 and 22 but did not correlate with enzyme activity in the spinal cord. Using an in vitro model of serotonergic raphe neuron differentiation, serotonergic differentiation was associated with an increase in both tryptophan hydroxylase promoter activity and protein expression. In vivo, tryptophan hydroxylase messenger RNA levels per single cell and per brain section were correlated during development up to postnatal day 22, but not beyond for both the dorsal raphe nucleus and nucleus raphe obscuris. Between postnatal days 22 and 61 single cell levels of tryptophan hydroxylase messenger RNA in the dorsal raphe nucleus did not change yet the levels per brain section significantly decreased by 40%. During the same period in the nucleus raphe obscuris, tryptophan hydroxylase messenger RNA levels per single cell signifcantly increased by 30% yet levels per brain section did not change. Comparison of tryptophan hydroxylase messenger RNA levels per cell and per brain section indicated a serotonergic loss between postnatal days 22 and 61 in both the dorsal raphe nucleus and nucleus raphe obscuris and may reflect either a loss of neurotransmitter phenotype or cell death. This study is the first to characterize the expression of brain tryptophan hydroxylase messenger RNA during rat development. In addition, this study is the first to report the activity of tryptophan hydroxylase in the spinal cord and hippocampus in the embryonic and neonatal rat. Together, the data provide a better understanding of the intricate relationship between patterns of tryptophan hydroxylase messenger RNA expression and enzyme activity.

Functional analysis of a novel human serotonin transporter gene promoter in immortalized raphe cells.

To investigate the structural basis for genetic regulation of the human serotonin transporter gene, a 1.8 kb fragment upstream to the cap site was cloned and sequenced. The promoter possesses a polymorphic repeat region with 16 and 14 repeats, respectively. Both were cloned and characterized. The promoter sequence revealed an internal 379 bp fragment not reported in previous publications. This novel fragment contains consensus sequences for several transcription factors including SpI and GATA. DNA from 48 unrelated individuals was PCR amplified, in this region, to test for allelic variations. All were found to possess the additional 379 bp fragment. The integrity of the promoter was furthermore confirmed by genomic Southern blotting. The promoter activity was analyzed by reporter gene assays in neuronal and non-neuronal serotonergic cell lines. In immortalized serotonergic raphe neurons, RN46A, three cis-acting, cell specific, activating elements and a silencer were located. One of the activators and the silencer are located in the repeat region and one activator is positioned in the novel fragment. A fourth activating element was found to be active in both RN46A cells and in a non-neuronal serotonergic cell line, JAR. A 3.5 kb fragment from intron 1 was cloned and found to possess cell specific activity in JAR cells indicating the presence of an alternative promoter in intron 1.

Neuronal replacement strategies for spinal cord injury.

Functional recovery following central nervous system (CNS) trauma or neurodegenerative disease is likely to require the transplantation of exogenous neurons. Given the logistical constraints of the potential widespread use of fetal human CNS tissues for therapeutic treatment, alternative sources of exogenous neurons for grafting will likely be necessary. Described here are studies examining the use of an immortalized, CNS-derived, neuronal precursor cell line, RN33B, as such a source. Results demonstrate that RN33B cells show remarkable plasticity to respond to local microenvironmental cues to differentiate in a direction that is morphologically indistinguishable from endogenous neurons at the site of transplantation. Concomitantly, the adult CNS retains the capacity to direct very specific differentiation of those engrafted precursor cells. However, the type and extent of site-specific appropriate differentiation is influenced by the type and extent of the lesion at the graft site. Attempts at immortalizing human CNS cells wtih similar approaches led to significant chromosomal aberrations, obviating such a strategy for therapeutic treatment. Thus, we utilized mitogen expansion of fetal human spinal cord cells as a means to generate populations of undifferentiated human neural precursor cells. Cells were expanded in the presence of epidermal growth factor and fibroblast growth factor 2, were readily passaged, and retained a pluriopotential to differentiate into neurons and astrocytes through at least 4 passages, after which the proliferating precursors became restricted to the astrocytic lineage. Further delineation of the variable needed to maintain these cells as undifferentiated, pluripotent precursor cells should ultimately enable examination of the ability of these cells to restore function in the damaged CNS.

The temporal and spatial activation of microglia in fiber tracts undergoing anterograde and retrograde degeneration following spinal cord lesion.

The role of microglia in the response to CNS injury is not fully understood. We characterized the temporal activation of microglia in the adult spinal cord following a lesion that severed the axons of the dorsal columns and corticospinal tract at T8. Two days after lesion, microglia in the severed T4-T5 fasciculus (f.) gracilis were ameboid and expressed intense OX42 and increased class I major histocompatibility complex (MHC) antigen (OX18) immunoreactivities. No activated microglia were seen in the intact f. cuneatus or the corticospinal tract. Five days postlesion, OX42 immunoreactivity was slightly decreased in the f. gracilis, and OX18 expression was slightly enhanced. By 12 days postlesion, OX42 and OX18 immunoreactivities were near control levels. At L1-L2, activated microglia with increased OX18 expression were restricted to the corticospinal tract and were maximal 5 days postlesion, returning to near control levels by 12 days postlesion. In the medulla, enhanced OX42 and OX18 immunoreactivities were seen in the nucleus (n.) gracilis, but not the n. cuneatus, at 2 days postlesion. At 5 days postlesion, OX42 immunoreactivity was markedly decreased, but class I MHC antigen expression was still enhanced. GFAP immunoreactivity increased only in the n. gracilis and remained elevated 2-12 days postlesion. Microglial activation is an early lesion-induced event in the CNS, and activated microglia may play a role in mediating the regenerative capacity of injured CNS axons.

Embryonic brain precursors transplanted into kainate lesioned rat spinal cord.

Embryonic day 14 rat cerebral cortex-derived precursors were expanded with FGF2 and labeled with BrdU prior to being transplanted into the kainic acid-lesioned adult rat spinal cord. While these precursors give rise to cells with neuronal, astrocytic and oligodendroglial phenotypes vitro, they remained largely undifferentiated up to 12 weeks in vivo. Numerous BrdU-labeled cells were found in injured gray matter, and also lining the dilated central canal that sometimes accompanies these lesions. BrdU-labeled cells never co-expressed Map2ab, rarely co-expressed GFAP but often co-expressed nestin, even after 12 weeks in vivo. These observations suggest that the environment of the kainic acid-injured spinal cord is not hostile to transplanted embryonic cerebral cortex-derived precursors, but also is not conducive to their neuronal differentation.

Altered differentiation of CNS neural progenitor cells after transplantation into the injured adult rat spinal cord.

Denervation of CNS neurons and peripheral organs is a consequence of traumatic SCI. Intraspinal transplantation of embryonic CNS neurons is a potential strategy for reinnervating these targets. Neural progenitor cell lines are being investigated as alternates to embryonic CNS neurons. RN33B is an immortalized neural progenitor cell line derived from embryonic rat raphe nuclei following infection with a retrovirus encoding the temperature-sensitive mutant of SV40 large T-antigen. Transplantation studies have shown that local epigenetic signals in intact or partially neuron-depleted adult rat hippocampal formation or striatum direct RN33B cell differentiation to complex multipolar morphologies resembling endogenous neurons. After transplantation into neuron-depleted regions of the hippocampal formation or striatum, RN33B cells were relatively undifferentiated or differentiated with bipolar morphologies. The present study examines RN33B cell differentiation after transplantation into normal spinal cord and under different lesion conditions. Adult rats underwent either unilateral lesion of lumbar spinal neurons by intraspinal injection of kainic acid or complete transection at the T10 spinal segment. Neonatal rats underwent either unilateral lesion of lumbar motoneurons by sciatic nerve crush or complete transection at the T10 segment. At 2 or 6-7 wk postinjury, lacZ-labeled RN33B cells were transplanted into the lumbar enlargement of injured and age-matched normal rats. At 2 wk posttransplantation, bipolar and some multipolar RN33B cells were found throughout normal rat gray matter. In contrast, only bipolar RN33B cells were seen in gray matter of kainic acid lesioned, sciatic nerve crush, or transection rats. These observations suggest that RN33B cell multipolar morphological differentiation in normal adult spinal cord is mediated by direct cell-cell interaction through surface molecules on endogenous neurons and may be suppressed by molecules released after SCI. They also indicate that the fate of immortalized neural progenitor cell lines in injured CNS must be stringently characterized.

In vitro labeling strategies for identifying primary neural tissue and a neuronal cell line after transplantation in the CNS.

Potential labels for identifying embryonic raphe neurons and a clonal, neuronally differentiating, raphe-derived cell line, RN33B, in CNS transplantation studies were tested by first characterizing the labels in vitro. The labels that were tested included 4',6-diamidino-2-phenylindole hydrochloride, 1,1'-dioctadecyl-3,3,3'-tetramethylindocarbocyanine perchlorate, the Escherichia coli lacZ gene, Fast Blue, Fluoro-Gold, fluorescein-conjugated latex microspheres, fluorescein isothiocyanate-conjugated or nonconjugated Phaseolus vulgaris leucoagglutinin, methyl o-(6-amino-3'-imino-3H-xanthen-9-yl) benzoate monohydrochloride, or tetanus toxin C fragment. Subsequently, the optimal in vitro labels for embryonic raphe neurons and for RN33B cells were characterized in vivo after CNS transplantation. In vitro, 1,1'-dioctadecyl-3,3,3'-tetramethylindocarbocyanine perchlorate (DiI) optimally labeled embryonic neurons. The Escherichia coli lacZ gene optimally labeled RN33B cells. Most labels were rapidly diluted in cultures of embryonic astrocytes and proliferating RN33B cells. Some labels were toxic and were often retained in cellular debris. In vivo, DiI was visualized in transplanted, DiI-labeled raphe neurons, but not in astrocytes up to 1 mo posttransplant. DiI-labeled host cells were seen after transplantation of lysed, DiI-labeled cells. beta-Galactosidase was visualized in transplanted, Escherichia coli lacZ gene-labeled RN33B cells after 15 days in vivo. No beta-galactosidase was seen in host cells after transplantation of lysed, lacZ-labeled RN33B cells. The results demonstrate that labels for use in CNS transplantation studies should be optimized for the specific population of donor cells under study, with the initial step being characterization in vitro followed by in vivo analysis. Appropriate controls for false-positive labeling of host cells should always be assessed.

Transplants of neuronal cells bioengineered to synthesize GABA alleviate chronic neuropathic pain.

The use of cell lines utilized as biologic "minipumps" to provide antinociceptive molecules, such as GABA, in animal models of pain is a newly developing area in transplantation biology. The neuronal cell line, RN33B, derived from E13 brain stem raphe and immortalized with the SV40 temperature-sensitive allele of large T antigen (tsTag), was transfected with rat GAD67 cDNA (glutamate decarboxylase, the synthetic enzyme for GABA), and the GABAergic cell line, 33G10.17, was isolated. The 33G10.17 cells transfected with the GAD67 gene expressed GAD67 protein and synthesized low levels of GABA at permissive temperature (33 degrees C), when the cells were proliferating, and increased GAD67 and GABA during differentiation at nonpermissive temperature (39 degrees C) in vitro, because GAD67 protein expression was upregulated with differentiation. A control cell line, 33V1, transfected with the vector alone, contained no GAD67 or GABA at either temperature. These cell lines were used as grafts in a model of chronic neuropathic pain induced by unilateral chronic constriction injury (CCI) of the sciatic nerve. Pain-related behaviors, including cold and tactile allodynia and thermal and tactile hyperalgesia, were evaluated after CCI in the affected hind paw. When 33G10.17 and 33V1 cells were transplanted in the lumbar subarachnoid space of the spinal cord 1 week after CCI, they survived greater than 7 weeks on the pia mater around the spinal cord. Furthermore, the tactile and cold allodynia and tactile and thermal hyperalgesia induced by CCI was significantly reduced during the 2-7-week period after grafts of 33G10.17 cells. The maximal effect on chronic pain behaviors with the GABAergic grafts occurred 2-3 weeks after transplantation. Transplants of 33V1 control cells had no effect on the allodynia and hyperalgesia induced by CCI. These data suggest that a chronically applied, low local dose of GABA presumably supplied by transplanted cells near the spinal dorsal horn was able to reverse the development of chronic neuropathic pain following CCI. The use of neural cell lines that are able to deliver inhibitory neurotransmitters, such as GABA, in a model of chronic pain offers a novel approach to pain management.

Target regulation of neuronal differentiation in a temperature-sensitive cell line derived from medullary raphe.

Following infection of dissociated embryonic day 13 rat medullary raphe cells with a retrovirus encoding the temperature-sensitive mutant of SV40 large T antigen, a clonal cell line, RN33B, was isolated by serial dilution. At 33 degrees C, RN33B cells divide with a doubling time of 48 h and show T antigen, vimentin, nestin, diffuse neuron-specific enolase, and low and medium molecular weight neurofilament immunoreactivities. RN33B cells are immortal, but not transformed, as they will not grow in soft agar. At non-permissive temperature (38.5 degrees C), T antigen expression is markedly decreased and RN33B cells cease mitotic activity and differentiate with phase bright cell bodies and 'neuritic-like' processes. Differentiated RN33B cells express enhanced neuronal-specific protein expression but do not synthesize astrocytic or oligodendrocytic-specific proteins. Moreover, differentiated RN33B cells returned to 33 degrees C re-express T antigen, but do not de-differentiate or begin dividing. Co-culture with embryonic hippocampus and cerebral cortex, but not medullary raphe or spinal cord, resulted in significantly greater survival, more complex neuronal morphology, and enhanced expression of neuronal-specific antigens. Immunohistochemical and Northern blot analysis revealed high levels of low affinity NGF receptor protein and mRNA in differentiated RN33B cells. PCR analysis demonstrated the presence of trkB, but not trkA or trkC, mRNA in both undifferentiated and differentiated RN33B cells. These data suggest that the observed target regulation of RN33B cell neuronal differentiation in co-culture may be mediated by neurotrophin(s).

Changes in catecholamine neuronal uptake and receptor binding in the brains of spontaneously hypertensive rats (SHR).

In these studies we have characterized differences between spontaneously hypertensive rats (SHR) and normotensive Wistar/Kyoto (WKY) rats with respect to rates of neuronal uptake of norepinephrine (NE) and dopamine (DA), and beta-adrenergic receptor (dihydroalprenolol; [3H]DHA) binding in the central nervous system. We find that SHR have greater rates of NE uptake in the frontal cortex, cerebellum, hypothalamus and pons-medulla during early development, and that these changes are accounted for, at least in the cerebral cortex, by an increased Vmax of the NE uptake mechanism. In addition, we find a decrease in the Bmax for [3H]DHA binding, suggestive of down-regulation of beta-adrenergic receptors of this region. In contrast to the results for NE uptake, we have measured significant decreases in DA uptake in the frontal cortex of the SHR at several postnatal ages. Decreases in DA uptake were also observed in the striatum of SHR although these changes were found only in animals approximately 6 weeks of age. From these results we have suggested that NE neurons projecting to a number of brain regions have elevated functional activity, while more regionally selective decreases in dopaminergic functional activity are characteristic of the SHR. We have further proposed that these changes in catecholamine neurons of the central nervous system may play an important role in the development of both the hypertension and behavioral hyperactivity exhibited by these animals.

Regional distribution and ionic requirement of Cl-/Ca2+-activated and Cl-/Ca2+-independent glutamate receptors in rat brain.

Recent studies have shown that Cl- and Ca2+ ions increase [3H]glutamate binding to rat forebrain synaptic plasma membranes by expressing a new class of glutamate receptors. We examined the regional distribution of these two classes of glutamate binding sites and further characterized their ionic requirements. Significant differences in both Cl-/Ca2+-independent (basal) and Cl-/Ca2+-activated receptors, as well as the ratios of these two receptor classes were observed among different areas of the CNS. Cl- and Ca2+ appeared to act synergistically, with Cl-ion an absolute requirement for Ca2+ stimulation, in expressing these additional binding sites. Ca2+ alone did not affect glutamate binding.

Cations differentially affect subpopulations of L-glutamate receptors in rat synaptic plasma membranes.

Several cations were examined for their ability to specifically affect one of the 3 L-glutamate (L-Glu) binding sites in rat forebrain synaptic plasma membranes (i.e. Na+-dependent, Cl--dependent and Cl--independent). Na+-dependent binding was potently inhibited by K+ and NH4+ ions. Other monovalent cations tested (Cs+, Li+, triethylammonium) had no effect on this binding site. Polyvalent cations (Co2+, Ni2+, Cu2+, Zn2+, Cd2+ and Cr3+) also had little effect on the Na+-dependent L-Glu binding site. Cl--dependent L-Glu binding was potently inhibited by Na+ ions but was not affected by other monovalent ions. All of the divalent cations were potent inhibitors of both Cl--dependent and -independent binding. The results show that these binding sites of L-Glu can be distinguished by their response to cations and suggest possible novel modes of regulation in vivo.

Neuronotrophic activity for ciliary ganglion neurons. Induction following injury to the brain of neonatal, adult, and aged rats.

The induction of neuronotrophic activity following injury to the brain of neonatal, adult, and aged Sprague-Dawley rats was compared using an improved in vitro assay. The maximal levels of activity in tissue surrounding the wound were reached at 3, 10-15, and about 15 days postlesion in neonatal, adult, and aged animals, respectively. Tissue neuronotrophic levels were always much lower in neonatal animals relative to the older animals. Accumulation of neuronotrophic activity in the gelfoam placed into the wound cavities in neonatal and adult animals lagged behind the levels in tissue by 4-5 days, suggesting that either the neuronotrophic factor itself or the cells which produce it are transferred from the tissue into the gelfoam. Relatively little activity accumulated in the gelfoam taken from aged Sprague-Dawley rats, and this observation was confirmed in aged Fischer rats. Aged animals seem to be unable to produce or release one of a number of neuronotrophic factors in response to injury.

Neuronotrophic factors for mammalian brain neurons: injury induction in neonatal, adult and aged rat brain.

Tissue from neonatal, adult and aged Sprague-Dawley rat brain contained low levels of survival promoting activity for embryonic neurons from various rat brain regions. This basal neuronotrophic activity was 2-fold higher in adult and 4-fold higher in aged rat brain, with respect to that in neonatal brain. Tissue extracts also contained 4-fold higher levels of neuronotoxic activity in adult and aged than in neonatal brain. Ablation of the occipital/entorhinal cortex caused a 3-fold increase in neuronotrophic activity in the tissue surrounding the wound in neonates and 4-fold in adult and aged brain, with respect to the basal levels. Maximal activities occurred at 3 days, 6 days and 15 days postlesion in neonatal, adult and aged tissue respectively, and subsequently returned to basal levels. Neuronotoxic activity was not induced following injury. The slower neuronotrophic response to injury in older animals may be one of the determinant factors of the slower recovery from brain damage observed in aging.

Decreased level of nerve growth factor (NGF) and its messenger RNA in the aged rat brain.

Trophic factors such as nerve growth factor (NGF) are thought to support survival, differentiation and maintenance of neurons. Recent results indicate that NGF produced in cortical and hippocampal areas is required for the function of cholinergic neurons in the basal forebrain. With the use of enzyme immunoassay and RNA blot hybridization we studied the NGF protein and NGF mRNA, respectively, in regions of the brain innervated by basal forebrain cholinergic neurons in adult and aged rats. Levels of NGF protein were decreased by 40% in hippocampus of aged (28 months) Fischer 344 rats compared with adults (6 months), whereas no alterations were observed in cerebral cortex. Moreover, a reduction by 50% in the NGF mRNA was found in samples of the aged forebrain (cerebral cortex, hippocampus, basal forebrain and hypothalamus) compared to the adult. NGF deficiencies may thus account for the loss of cholinergic neurons in the basal forebrain generally found to accompany normal aging and resulting in altered cognitive functions.

Acidic and basic fibroblast growth factor levels in spinal cord cultures are not regulated by alterations in heparan sulfate proteoglycan expression.

The present study was undertaken to assess both the levels of acidic and basic fibroblast growth factors in spinal cord cultures and to determine how they were presented to responsive cells. Western blots detected a single acidic fibroblast growth factor-like protein (17 kDa) and two (18 kDa, 24 kDa) basic fibroblast growth factor-immunoreactive proteins, the levels of which varied with the antibody used. Levels of all three proteins were unaltered in cultures grown in the presence of a mitotic inhibitor, which greatly reduced the number of astrocytes. Cell blots showed increased survival of spinal cord neurons at Mr that corresponded with the three proteins detected immunologically. Solubilized cultures separated on a P100 column showed mitogenic activity for NIH3T3 cells from 17-18 and 24 kDa fractions. Treatment of the cultures with heparitinase did not decrease the levels of acidic and basic fibroblast growth factors detected by Western blots, suggesting that these proteins were not associated with extracellular membrane heparan sulfate proteoglycans. The major fraction of both proteins appeared to be intracellular with a minor amount complexed with extracellular matrix proteins. An inhibitor of xylose-linked proteoglycan synthesis significantly altered heparan sulfate proteoglycan deposition into extracellular matrix, but did not alter the levels of acidic or basic fibroblast growth factors detected by Western blots, or the levels of choline acetyltransferase, glutamic acid decarboxylase, or aspartate aminotransferase activities. These results indicate that both acidic and basic fibroblast growth factors are stored predominantly intracellularly, with only a minor fraction complexed with extracellular proteins. We suggest that these intracellular proteins may be released following injury in the CNS and mediate a cascade of neuroprotective events.