Rescue of motoneurons from cell death by a purified skeletal muscle polypeptide: effects of the ChAT development factor, CDF

Rat skeletal muscle contains a 22 kd polypeptide that increases the level of choline acetyltransferase (ChAT) activity in cultures of embryonic rat spinal cord neurons and has been purified to homogeneity. The application of this factor, ChAT development factor or CDF, to developing chick embryos during the period of naturally occurring motoneuron cell death significantly increased the survival of motoneurons but did not affect the survival of dorsal root ganglion neurons or sympathetic preganglionic neurons (column of Terni). These results provide the first demonstration that an isolated, skeletal muscle-derived molecule can selectively enhance the survival of motoneurons in vivo and suggest that CDF may function in vivo to regulate the survival and development of motoneurons.     

Necdin protects embryonic motoneurons from programmed cell death

NECDIN belongs to the type II Melanoma Associated Antigen Gene Expression gene family and is located in the Prader-Willi Syndrome (PWS) critical region. Necdin-deficient mice develop symptoms of PWS, including a sensory and motor deficit. However, the mechanisms underlying the motor deficit remain elusive. Here, we show that the genetic ablation of Necdin, whose expression is restricted to post-mitotic neurons in the spinal cord during development, leads to a loss of 31% of specified motoneurons. The increased neuronal loss occurs during the period of naturally-occurring cell death and is not confined to specific pools of motoneurons. To better understand the role of Necdin during the period of programmed cell death of motoneurons we used embryonic spinal cord explants and primary motoneuron cultures from Necdin-deficient mice. Interestingly, while Necdin-deficient motoneurons present the same survival response to neurotrophic factors, we demonstrate that deletion of Necdin leads to an increased susceptibility of motoneurons to neurotrophic factor deprivation. We show that by neutralizing TNFα this increased susceptibility of Necdin-deficient motoneurons to trophic factor deprivation can be reduced to the normal level. We propose that Necdin is implicated through the TNF-receptor 1 pathway in the developmental death of motoneurons.     

Relevance of motoneuron specification and programmed cell death in embryos to therapy of ALS

The molecular cues that generate spinal motoneurons in early embryonic development are well defined. Motoneurons are generated in excess and consequently undergo a natural period of programmed cell death. Although it is not known exactly how motoneurons compete for survival in embryonic development, it is hypothesized that they rely on the ability to access limited amounts of trophic factors from peripheral tissues, a process that is tightly regulated by skeletal muscle activity. Attempts to elucidate the molecular mechanisms that underlie motoneuron generation and programmed cell death in embryos have led to various effective strategies for treating injury and disease in animal models. Such studies provide great hope for the amelioration of human amyotrophic lateral sclerosis (ALS), a devastating progressive motoneuron degenerative disease. Here we review the clinical relevance of studying motoneuron specification and death during embryonic development.     

Ethanol influences on the chick embryo spinal cord motor system: Analyses of motoneuron cell death,motility, and target trophic factor activity and in vitro analyses of neurotoxicity and trophic factor neuroprotection

A series of in vivo and in vitro experiments were conducted to determine the influence of prenatally administered ethanol on several aspects of the developing chick embryo spinal cord motor system. Specifically, we examined: (1) the effect of chronic ethanol administration during the natural cell death period on spinal cord motoneuron numbers; (2) the influence of ethanol on ongoing embryonic motility; (3) the effect of ethanol exposure on neurotrophic activity in motoneuron target tissue (limbbud); and (4) the responsiveness of cultured spinal cord neurons to ethanol, and the potential of target-derived neurotrophic factors to ameliorate ethanol neurotoxicity. These studies revealed the following: Chronic prenatal ethanol exposure reduces the number of motoneurons present in the lateral motor column after the cell death period [embryonic day 12 (E12)]. Ethanol tends to inhibit embryonic motility, particularly during the later stages viewed (E9-E11). Chronic ethanol exposure reduces the neurotrophic activity contained in target muscle tissue. Such diminished support could contribute to the observed motoneuron loss. Direct exposure of spinal cord neurons to ethanol decreases neuronal survival and process outgrowth in a dose-dependent manner, but the addition of target muscle extract to ethanol-containing cultures can ameliorate this ethanol neurotoxicity. These studies demonstrate ethanol toxicity in a population not previously viewed in this regard and suggest a mechanism that may be related to this cell loss (i.e., decreased neurotrophic support). © 1995 John Wiley & Sons, Inc.     

Interactions between spinal cord stimulation and activity blockade in the regulation of synaptogenesis and motoneuron survival in the chick embryo

The present study investigated the effects of spinal cord stimulation, neuromuscular blockade, or a combination of the two on neuromuscular development both during and after the period of naturally occurring motoneuron death in the chick embryo. Electrical stimulation of the spinal cord was without effect on motoneuron survival, synaptogenesis, or muscle properties. By contrast, activity blockade rescued motoneurons from cell death and altered synaptogenesis. A combination of spinal cord stimulation and activity blockade resulted in a marked increase in motoneuron death, and also altered synaptogenesis similar to that seen with activity blockade alone. Perturbation of normal nerve–muscle interactions by activity blockade may increase the vulnerability of developing motoneurons to excessive excitatory afferent input (spinal cord stimulation) resulting in excitotoxic-induced cell death. © 1993 John Wiley & Sons, Inc.     

Development and survival of thoracic motoneurons and hindlimb musculature following transplantation of the thoracic neural tube to the lumbar region in the chick embryo: anatomical aspects

Thoracic spinal cord transplanted to the lumbar region at the time of neural tube closure in the chick embryo survives and initially differentiates normally similar to in situ thoracic cord. Normal numbers of motoneurons are produced that innervate the host hindlimb musculature. In control thoracic cord approximately 70% of the motoneurons are lost by normal cell death between embryonic day (E) 6 and E11-E12. By contrast, the transplanted thoracic cord loses only about 30% of the motoneurons during this period. Transplantation of one hindlimb to the thoracic region also reduces the normal loss of in situ thoracic motoneurons. We conclude that some factor(s) associated with the increased target size provided by the hindlimbs promotes the survival of thoracic motoneurons. In contrast, by E16-E18 motoneuron numbers in the thoracic transplants decrease to below control levels. Dorsal root ganglion cells in the transplant were also initially increased (on E8) but later decreased to below control values. Hindlimb muscles innervated by thoracic motoneurons in the transplant also differentiated normally up to E10 to E12. Myotube size and numbers, muscle size and myotube types (fast versus slow) all developed normally in several thoracically-innervated hindlimb muscles. However, beginning on E14 myotube numbers and muscle size were markedly decreased resulting in muscle atrophy. Injections of horseradish peroxidase (HRP) into the thoracic transplants labelled neurons in the host spinal cord and brainstem rostral to the transplant thereby indicating an anatomical continuity between host and transplant neural tube. Injections of HRP into specific thoracically innervated hindlimb muscles on E8 labelled distinct pools of motoneurons in the transplants.(ABSTRACT TRUNCATED AT 250 WORDS)     

Conditional gene ablation of Stat3 reveals differential signaling requirements for survival of motoneurons during development and after nerve injury in the adult.

Members of the ciliary neurotrophic factor (CNTF)/leukemia inhibitory factor (LIF)/cardiotrophin gene family are potent survival factors for embryonic and lesioned motoneurons. These factors act via receptor complexes involving gp130 and LIFR-beta and ligand binding leads to activation of various signaling pathways, including phosphorylation of Stat3. The role of Stat3 in neuronal survival was investigated in mice by Cre-mediated gene ablation in motoneurons. Cre is expressed under the neurofilament light chain (NF-L) promoter, starting around E12 when these neurons become dependent on neurotrophic support. Loss of motoneurons during the embryonic period of naturally occurring cell death is not enhanced in NF-L-Cre; Stat3(flox/KO) mice although motoneurons isolated from these mice need higher concentrations of CNTF for maximal survival in culture. In contrast, motoneuron survival is significantly reduced after facial nerve lesion in the adult. These neurons, however, can be rescued by the addition of neurotrophic factors, including CNTF. Stat3 is essential for upregulation of Reg-2 and Bcl-xl expression in lesioned motoneurons. Our data show that Stat3 activation plays an essential role for motoneuron survival after nerve lesion in postnatal life but not during embryonic development, indicating that signaling requirements for motoneuron survival change during maturation.     

Elucidating the molecular mechanisms that underlie the target control of motoneuron death

Approximately half of the motoneurons generated during normal embryonic development undergo programmed cell death. Most of this death occurs during the time when synaptic connections are being formed between motoneurons and their target, skeletal muscle. Subsequent muscle activity stemming from this connection helps determine the final number of surviving motoneurons. These observations have given rise to the idea that motoneuron survival is dependent upon access to muscle derived trophic factors, presumably through intact neuromuscular synapses. However, it is not yet understood how the muscle regulates the supply of such trophic factors, or if there are additional mechanisms operating to control the fate of the innervating motoneuron. Recent observations have highlighted target independent mechanisms that also operate to support the survival of motoneurons, such as early trophic-independent periods of motoneuron death, trophic factors derived from Schwann cells and selection of motoneurons during pathfinding. Here we review recent investigations into motoneuron cell death when the molecular signalling between motoneurons and muscle has been genetically disrupted. From these studies, we suggest that in addition to trophic factors from muscle and/or Schwann cells, specific adhesive interactions between motoneurons and muscle are needed to regulate motoneuron survival. Such interactions, along with intact synaptic basal lamina, may help to regulate the supply and presentation of trophic factors to motoneurons.     

Developmental expression of voltage-dependent calcium currents in identified mouse motoneurons.

Previous work has shown that during chick embryonic development, large changes occur in the density of specific, motoneuronal calcium currents just prior to the period of naturally occurring motoneuron cell death. Here we report on calcium currents in mouse motoneurons isolated from embryos at the time of peak cell death and also during a subsequent developmental stage when supernumerary synapses are being eliminated. In mouse motoneurons, the density of high-voltage-activated calcium current increases significantly after the phase of cell death, during the period of synapse elimination.     

Brain-derived proteins that rescue spinal motoneurons from cell death in the chick embryo: Comparisons with target-derived and recombinant factors

Spinal motoneurons that normally die during early development can be rescued by a variety of purified growth or neurotrophic factors and target tissue extracts. There is also indirect evidence that brain or supraspinal afferent input may influence lumbar motoneuron survival during development and that this effect may be mediated by central nervous system–derived trophic agents. This report examines the biological and biochemical properties of motoneuron survival activity obtained from extracts of the embryonic chick brain. Treatment with an ammonium sulfate (25% to 75%) fraction of embryonic day 16 (E16) or E10 brain extracts rescued many spinal motoneurons that otherwise die during the normal period of cell death in vivo (E6 to E10). The same fractions also enhanced lumbar motoneuron survival following deafferentation. There were both similarities and differences between the active fractions derived from brain extracts (BEX) when compared with extracts derived from target muscles (MEX) or with purified neurotrophic factors. Survival activity from E10 BEX was as effective in promoting motoneuron survival as E10 MEX and more effective than astrocyte-conditioned media. Unlike MEX, the active fractions from BEX also rescued placode-derived nodose ganglion cells. In addition, unlike nerve growth factor and brain-derived neurotrophic factor, active BEX fractions did not rescue neural crest-derived dorsal root ganglion cells or sympathetic ganglion neurons. Interestingly, among many cranial motor and other brainstem nuclei examined, only the survival of motoneurons from the abducens nucleus was enhanced by BEX. Active proteins obtained from BEX were further separated by gel filtration chromatography and by preparative isolelectric focusing techniques. Activity was recovered in a basic (pI8) and an acidic (pI5) small molecular weight protein fraction (20 kD or less). The specific activity of the basic fraction was increased ×66 when compared with the specific activity of crude BEX, and the basic fraction had a slightly higher specific activity than the acidic fraction. The biological and biochemical properties of these fractions are discussed in the context of known neurotrophic factors and their effects on normal and lesion-induced motoneuron death during development. © 1995 John Wiley & Sons, Inc.     

Cholinergic differentiation factor (CDF/LIF) promotes survival of isolated rat embryonic motoneurons in vitro.

We present evidence that the cholinergic differentiation factor (CDF), originally purified from cardiac and skeletal muscle cell-conditioned medium and found to be identical to leukemia inhibitory factor (LIF), promotes survival of embryonic day 14 rat motoneurons in vitro. These neurons were retrogradely labeled with the fluorescent tracer Dil and enriched on a density gradient or purified to homogeneity by fluorescence-activated cell sorting. Subnanomolar concentrations of CDF/LIF supported the survival of 85% of the motoneurons that would have died between days 1 and 4 of culture. The enhanced survival was accompanied by a 4-fold increase in choline acetyltransferase (ChAT) activity per culture. CDF/LIF also increased ChAT activity in dorsal spinal cord cultures, but had no detectable effect on ChAT levels in septal or striatal neuronal cultures. For comparison, other neurotrophic molecules were tested on motoneuron cultures. Ciliary neurotrophic factor had effects on motoneuron survival similar to those of CDF/LIF, whereas basic fibroblast growth factor was somewhat less effective. Nerve growth factor had no effect on the survival of rat motoneurons.     

Early regional variations in motoneuron numbers arise by differential proliferation in the chick embryo spinal cord

Regional differences in the number of motoneurons in the spinal cord of the chick are thought to arise developmentally by region-specific cell death and cell migration. In this way, a numerically homogeneous motor column throughout the spinal cord is believed to be molded into the adult pattern. Region-specific differences in proliferation are not thought to play a significant role in this process. By counting motoneurons in serial sections throughout the rostral-caudal extent of the spinal cord on Embryonic Day 4 in the chick, we have found that the numerical variations in motoneurons in different spinal cord regions are already foreshadowed by this stage, which is before the onset of both cell death and the secondary migration of neurons out of the motor column. These results indicate that although nonproliferative events may contribute to the later regional variations in motoneuron numbers, the initial differences themselves are created early on by regionally specific proliferative events.     

Motoneuronal death during spinal cord development is mediated by oxidative stress

The involvement of reactive oxygen species (ROS) in neuronal death has been determined in culture, and in association with several neurodegenerative disorders. We examined whether ROS participate in the cell death observed during spinal cord development. We found that the general pattern of high ROS levels, gene expression for some antioxidant enzymes, and motoneuron death correlated positively along spinal cord development. ROS were reduced in spinal cords cultured in the presence of a synthetic superoxide dismutase and catalase mimetic, with a concomitant reduction in cell death and an increase in the number of motoneurons. The number of motoneurons was higher in spinal cords treated with the antioxidant than in those treated with caspase inhibitors. In general, the increase in motoneuron survival did not correlate with the reduction in cells undergoing DNA degradation in the motoneuronal region. These results suggest that ROS are signaling molecules controlling caspase-dependent and caspase-independent programmed motoneuron death, and support the hypothesis that this mechanism is abnormally turned on in some neurodegenerative disorders and aging.     

Motoneuron projection patterns in chick embryonic limbs with a double complement of dorsal thigh musculature

During the normal development of the chick, lateral motoneurons within the lumbosacral motor column of the spinal cord consistently project to muscles of dorsal origin within the limb while medial motoneurons project to muscles of ventral origin. To determine if specific cues arising from each type of target are the dominant guidance cues used by lateral and medial motoneurons to create this pattern, I examined motoneuron projections in embryonic chick limbs with a double complement of dorsal thigh musculature and no ventral musculature. Results indicate that cues associated with muscles of a specific developmental origin do not invariably dominate. Before and after the major period of motoneuron death, all muscles in dorsal limb regions (host) were innervated by lateral or dorsal pool neurons. Most ventrally positioned (donor) muscles were innervated by medial or ventral pool neurons. Only the donor iliofibularis, a muscle located very near to its original source of innervation, received projections from some lateral neurons. Within the limb proper, medial or ventral pool neurons projected to donor muscles in a patterned manner suggesting that they were following nonspecific regional cues and perhaps also responding to the availability of uninnervated target tissue. I conclude that axon sorting into distinct lateral and medial classes is independent of limb target complement and that subsequent pathway choice is a separate event governed by both specific target cues and other guidance mechanisms.     

Ethanol influences on the chick embryo spinal cord motor system. II. Effects of neuromuscular blockade and period of exposure

The study described below was performed as a continuation of a previous study in which we found reduced motoneuron number in lumbar spinal cord of the chick embryo following chronic ethanol administration from embryonic day 4 (E4) to E11. We sought to determine whether this reduction was due to primary ethanol toxicity or to enhancement of naturally occurring cell death (NOCD) and to determine whether administration of ethanol at a later period of development could also reduce motoneuron number. Earlier studies have shown that curare suspends NOCD in the chick embryo. By administering both ethanol and curare to these embryos from E4 to E11 and examining the lumbar spinal cord on E12, we determined that ethanol was directly toxic to motoneurons and reduced motoneuron number in the absence of NOCD. By administering ethanol from E10 to E15 and examining the lumbar spinal cord on E16, we determined that ethanol can reduce motoneuron number without altering spinal cord length during more than one stage of chick embryo development, and that ethanol toxicity is not dependent on NOCD. In addition, we demonstrated that ethanol does not affect the neurotrophic content of chick muscle when it is administered from E10 to E15. © 1997 John Wiley & Sons, Inc. J Neurobiol 32 : 684–694, 1997     

Vascular endothelial growth factor protects spinal cord motoneurons against glutamate-induced excitotoxicity via phosphatidylinositol 3-kinase

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease characterized by the selective death of motoneurons. Recently, vascular endothelial growth factor (VEGF) has been identified as a neurotrophic factor and has been implicated in the mechanisms of pathogenesis of ALS and other neurological diseases. The potential neuroprotective effects of VEGF in a rat spinal cord organotypic culture were studied in a model of chronic glutamate excitotoxicity in which glutamate transporters are inhibited by threohydroxyaspartate (THA). Particularly, we focused on the effects of VEGF in the survival and vulnerability to excitotoxicity of spinal cord motoneurons. VEGF receptor-2 was present on spinal cord neurons, including motoneurons. Chronic (3 weeks) treatment with THA induced a significant loss of motoneurons that was inhibited by co-exposure to VEGF (50 ng/mL). VEGF activated the phosphatidylinositol 3-kinase/Akt (PI3-K/Akt) signal transduction pathway in the spinal cord cultures, and the effect on motoneuron survival was fully reversed by the specific PI3-K inhibitor, LY294002. VEGF also prevented the down-regulation of Bcl-2 and survivin, two proteins implicated in anti-apoptotic and/or anti-excitotoxic effects, after THA exposure. Together, these findings indicate that VEGF has neuroprotective effects in rat spinal cord against chronic glutamate excitotoxicity by activating the PI3-K/Akt signal transduction pathway and also reinforce the hypothesis of the potential therapeutic effects of VEGF in the prevention of motoneuron degeneration in human ALS.     

The development of motoneurons in the embryonic spinal cord of the mouse mutant, muscular dysgenesis (mdg/mdg): survival, morphology, and biochemical differentiation

Motoneuron development was studied in the spinal cord of the mouse mutant, muscular dysgenesis, between embryonic days (E) 13 and 18. Dysgenic embryos are characterized by the absence of neuromuscular activity (motility) and exhibit a number of other striking changes in neuromuscular development. Many of these changes have also been observed in chick embryos chronically treated with neuromuscular blocking agents that suppress motility. Motoneuron survival, as well as several other aspects of neuronal development, was examined in the thoracic and lumbar spinal cords of mutant and control embryos. There was a significant decrease in motoneuron numbers in control embryos indicating the presence of naturally occurring cell death in the mouse spinal cord. At all ages examined, the dysgenic embryos had significantly more healthy and significantly fewer degenerating motoneurons than controls. There were no differences in the number of dorsal root ganglion neurons or in any of the other morphometric parameters examined between mutant and control embryos. Creatine kinase activity, a marker for myofiber maturation, was significantly reduced in the limb musculature of mutant embryos. Choline acetyltransferase activity was significantly increased in the spinal cord of mutant embryos. No significant differences were observed in spinal cord levels of acetylcholinesterase activity between control and mutant embryos. The absence of muscle contractions in the dysgenic mouse is associated with a number of changes in neuromuscular development, including a substantial reduction of naturally occurring motoneuron death.     

Interaction between hindbrain and spinal networks during the development of locomotion in zebrafish

Little is known about the role of the hindbrain during development of spinal network activity. We set out to identify the activity patterns of reticulospinal (RS) neurons of the hindbrain in fictively swimming (paralyzed) zebrafish larvae. Simultaneous recordings of RS neurons and spinal motoneurons revealed that these were coactive during spontaneous fictive swim episodes. We characterized four types of RS activity patterns during fictive swimming: (i) a spontaneous pattern of discharges resembling evoked high-frequency spiking during startle responses to touch stimuli, (ii) a rhythmic pattern of excitatory postsynaptic potentials (EPSPs) whose frequency was similar to the motoneuron EPSP frequency during swim episodes, (iii) an arrhythmic pattern consisting of tonic firing throughout swim episodes, and (iv) RS cell activity uncorrelated with motoneuron activity. Despite lesions to the rostral spinal cord that prevented ascending spinal axons from entering the hindbrain (normally starting at approximately 20 h), RS neurons continued to display the aforementioned activity patterns at day 3. However, removal of the caudal portion of the hindbrain prior to the descent of RS axons left the spinal cord network unable to generate the rhythmic oscillations normally elicited by application of N-methyl-d-aspartate (NMDA), but in approximately 40% of cases chronic incubation in NMDA maintained rhythmic activity. We conclude that there is an autonomous embryonic hindbrain network that is necessary for proper development of the spinal central pattern generator, and that the hindbrain network can partially develop independently of ascending input.     

Developmental characteristics of AMPA receptors in chick lumbar motoneurons

Ca2+ fluxes through ionotropic glutamate receptors regulate a variety of developmental processes, including neurite outgrowth and naturally occurring cell death. In the CNS, NMDA receptors were originally thought to be the sole source of Ca2+ influx through glutamate receptors; however, AMPA receptors also allow a significant influx of Ca2+ ions. The Ca2+ permeability of AMPA receptors is regulated by the insertion of one or more edited GluR2 subunits. In this study, we tested the possibility that changes in GluR2 expression regulate the Ca2+ permeability of AMPA receptors during a critical period of neuronal development in chick lumbar motoneurons. GluR2 expression is absent between embryonic day (E) 5 and E7, but increases significantly by E8 in the chick ventral spinal cord. Increased GluR2 protein expression is correlated with parallel changes in GluR2 mRNA in the motoneuron pool. Electrophysiological recordings of kainate-evoked currents indicate a significant reduction in the Ca2(+)-permeability of AMPA receptors between E6 and E11. Kainate-evoked currents were sensitive to the AMPA receptor blocker GYKI 52466. Application of AMPA or kainate generates a significant increase in the intracellular Ca2+ concentration in E6 spinal motoneurons, but generates a small response in older neurons. Changes in the Ca(2+)-permeability of AMPA receptors are not mediated by age-dependent changes in the editing pattern of GluR2 subunits. These findings raise the possibility that Ca2+ influx through Ca(2+)-permeable AMPA receptors plays an important role during early embryonic development in chick spinal motoneurons.     

Neurotrophic agents prevent motoneuron death following sciatic nerve section in the neonatal mouse

We have examined the ability of different neurotrophic and growth factors to prevent axotomy-induced motoneuron cell death in the developing mouse spinal cord. After postnatal unilateral section of the mouse sciatic nerve, most motoneuron (MN) loss occurs in the lateral motor column of the fourth lumbar segment (L4). Significant axotomy-induced cell death occurred after surgery performed on or before postnatal day (PN) 5. In contrast, no significant cell loss was found when axotomy was performed after PN10. Axotomy on PN2 or PN5 resulted in a 44% loss of L4 motoneurons by 7 days, and a 66% loss of motoneurons by 10 days postsurgery. Implantation of gelfoam presoaked in various neurotrophic factors at the lesion site rescued axotomized motoneurons. Nerve growth factor (NGF), nedurotrophin-4/5 (NT-4/5) and ciliary neurotrophic factor (CNTF) rescued 20%–30% of motoneurons, whereas brain-derived neurotrophic factor (BDNF), neurotrophin-3 (NT-3), and insulin-like growth factor 1 (IGF-1) rescued virtually all motoneurons from axotomy-induced death. By contrast, platelet-derived growth factor (PDGF)-AA, PDGF-AB, basic fibroblast growth factor (bFGF), and interleukin (IL-6) were ineffective on motoneuron survival following axotomy. NGF, BDNF, NT-3, IGF-1, and CNTF also prevented axotomy-induced atrophy of surviving motoneurons. These data show that mouse lumbar motoneurons continue to be vulnerable to axotomy up to about 1 week after birth and that a number of trophic agents, including the neurotrophins, CNTF, and IGF-1, can prevent the death of these neurons following axotomy. Our studies confirm and extend previous reports on the time course of axotomy-induced mouse motoneuron death and the survival promoting effects of neurotrophic factors. 1994 John Wiley & Sons, Inc.