Trophic Dependence Research Articles

Apoptosis of spinal interneurons induced by sciatic nerve axotomy in the neonatal rat is counteracted by nerve growth factor and ciliary neurotrophic factor.

We have previously shown that not only motoneurons and dorsal root ganglion cells but also small neurons, presumably interneurons in the spinal cord, may undergo apoptotic cell death as a result of neonatal peripheral nerve transection in the rat. With the aid of electron microscopy, we have here demonstrated that apoptosis in the spinal cord is confined to neurons and does not involve glial cells at the survival time studied (24 hours). To define the relative importance of the loss of a potential target (motoneuron) and a potential afferent input (dorsal root ganglion cell) for the induction of apoptosis in interneurons in this situation, we have compared the distributions and time courses for TUNEL labeling, which detects apoptotic cell nuclei, in the L5 segment of the spinal cord and the L5 dorsal root ganglion after sciatic nerve transection in the neonatal (P2) rat. In additional experiments, we studied the effects on TUNEL labeling of interneurons after treatment of the cut sciatic nerve with either ciliary neurotrophic factor (CNTF) to rescue motoneurons or nerve growth factor (NGF) to rescue dorsal root ganglion cells. The time courses of the TUNEL labeling in motoneurons and interneurons induced by the lesion show great similarities (peak at 8-48 hours postoperatively), whereas the labeling in dorsal root ganglion cells occurs later (24-72 hours). Both CNTF and NGF decrease the number of TUNEL-labeled interneurons, but there is a regional difference, in that CNTF preferentially saves interneurons in deep dorsal and ventral parts of the spinal cord, whereas the rescuing effects of NGF are seen mainly in the superficial dorsal horn. The results are interpreted as signs of a trophic dependence on both the target and the afferent input for the survival of interneurons neonatally.

Augustus V. Waller and plastic surgery.

Augustus V. Waller's name is associated eponymously with the regressive changes in transected nerves. This article intends to present the many other aspects of Waller's scientific ventures. Apart from pioneering intravital microscopy, Waller made significant contributions to the understanding of the process of tissue inflammation. In addition to his observations of the degenerative changes in severed nerve fibers, he developed a concept of trophic dependence that is relevant for the management of the regeneration of peripheral nerves in plastic surgery.

Laminar variation in neuronal viability and trophic dependence in neocortical slices.

Organotypic slices are used frequently in studies of central nervous system development and function because they provide excellent experimental access with significant preservation of cellular context and relationships. Within a slice, however, a variety of factors may cause individual classes of neurons to respond differently to the culture environment. Differences in deafferentation, cellular maturation, trophic dependence and ongoing naturally occurring cell death may produce changes in the neuronal population that are transparent to the experimenter but that could affect experimental results significantly. In this study, we examined the distribution and prevalence of cell death among neurons in each cortical layer in organotypic slices. In addition, we assessed the ability of several neurotrophic factors to ameliorate neuronal death in each cortical layer. Within the first 24 hr in culture, there was striking laminar variation in the extent of neuronal death in culture, which could not be accounted for by the pattern of programmed cell death in vivo. In addition, neurons in the six layers of the neocortex differed in the degree to which they could be rescued by neurotrophic factors. These data suggest that differential neuronal death and rescue are important considerations in studies utilizing organotypic slices and may represent particularly confounding variables in studies of effects of trophic factors in such preparations.

Postmigratory Enteric and Sympathetic Neural Precursors Share Common, Developmentally Regulated, Responses to BMP2

The development of enteric and sympathetic neurons from neural crest precursor cells is regulated by signals produced by the embryonic environments to which the cells migrate. Bone morphogenetic proteins (BMPs) are present in the developing embryo and act to induce neuronal differentiation and noradrenergic properties of neural crest cells. We have investigated the role of BMP2 in regulating the appearance of distinct populations of autonomic neurons from postmigratory, HNK-1-positive neural crest precursor cells. BMP2 promotes neuronal differentiation of sympathetic and enteric precursor cells isolated from E14.5 rat. The effects of BMP2 change over time, resulting in a decrease in neuron number that can be attributed to apoptotic cell death. BMP2-dependent neuron death is rescued by gut-derived factors that provide trophic support to maturing neurons, indicating that BMP2 regulates the acquisition of trophic dependence of developing peripheral neurons. In addition to regulating neuron number, BMP2 promotes both panneuronal maturation and the acquisition of an enteric phenotype, as measured by lineage-specific changes in the expression of tyrosine hydroxylase and MASH-1. While BMP2 is sufficient to induce neuronal differentiation and panneuronal development, these results suggest that additional factors in the environment must collaborate with BMP2 to promote the final noradrenergic phenotype of sympathetic neurons.

An axotomy model for the induction of death of rat and mouse corticospinal neurons in vivo

To study trophic dependencies of rat and mouse corticospinal neurons (CSN), we established a lesion model for the induction of death of analogous populations of CSN in these rodent species. Before lesion, CSN were retrogradely labeled with Fast Blue (FB). A stereotaxic cut lesion through the entire internal capsule (ICL) was used to axotomize CSN. The extent of axotomy was determined by application of a control tracer. In both species, FB-labeled CSN were localized in three major areas: (1) the sensory motor cortex; (2) the supplementary motor and medial prefrontal cortex; and (3) the somatosensory cortex. ICL does not lead to complete axotomy of CSN of the rat and mouse somatosensory cortex. In rats, ICL results in complete axotomy of CSN of the sensory motor cortex and incomplete axotomy of the caudal portion of the supplementary motor and medial prefrontal cortex. In mice, the area of axotomized CSN extends significantly further frontally. In both species, axotomy-induced death of CSN is observed in the center of the sensory motor cortex. This lesion model is useful for investigations on the response of CSN of the sensory motor cortex to lesion and therapeutic drugs.

A helminth community in breeding Yellow-legged Gulls (Larus cachinnans): pattern of association and its effect on host fitness

We studied the community of helminths parasitizing adult Yellow-legged Gulls (Larus cachinnans) sampled during three consecutive breeding seasons in a western Mediterranean colony. The helminth community was depauperate, probably because of the high trophic dependence of the gullery on waste food from refuse dumps. The prevalences of two helminth species varied both among years and between the sexes of gulls, while the intensity of parasitism did not vary significantly for any helminth species. No clear pattern of association between helminth species was detected, showing an unstructured helminth community in which the occurrence of any species was not restricted or favoured by the occurrence of any other species. The lack of associations might be influenced by the fact that only one core species occurred in the helminth community. Negative relationships were found between occurrence of certain helminths and body condition of hosts. Gulls infected by Tetrabothrius (Tetrabothrius) erostris were leaner than uninfected gulls; interestingly this negative relationship varied with the presence/absence of Cosmocephalus obvelatus. Moreover, body condition of gulls infected by C. obvelatus varied according to the year of sampling, while that of uninfected individuals showed no variability. High intensities of the parasites Eucoleus contortus and Brachylaima sp. were associated with poorer body condition of the host. In the case of E. contortus, this negative relationship with gulls' body condition was more pronounced in male than in female gulls. The results show that helminth pathogenicity can be affected by helminth community structure, environmental factors, and traits of the host.

A dynamic regulation of GDNF-family receptors correlates with a specific trophic dependency of cranial motor neuron subpopulations during development.

Glial cell line-derived neurotrophic factor (GDNF) family ligands promote the survival of developing motor neurons in vivo and in vitro. However, not all neurons survive with any single ligand in culture and GDNF null mutant mice display only a partial motor neuron loss. An interesting possibility is that subpopulations of motor neurons based on their function and/or their myotopic organization require distinct members of GDNF family ligands. Because responsiveness to the different ligands depends on the expression of their cognate ligand-binding receptor we have herein addressed this issue by examining the expression of GDNF-family receptors (gfr) during development and in the adult in cranial motor nuclei subpopulations. We have furthermore examined the in vivo role of GDNF for cranial motor neuron subpopulations. The shared ret receptor was expressed in all somatic, branchial and visceral cranial embryonic motor nuclei examined, showing that they are all competent to respond to GDNF family ligands during development. At early stages of development both the GDNF receptor, gfralpha1, and the neurturin (NTN) receptor, gfralpha2, were expressed in the oculomotor, facial and spinal accessory, and only gfralpha1 in the trochlear, superior salivatory, trigeminal, hypoglossal and weakly in the dorsal motor nucleus of the vagus and the ambiguous nucleus. The abducens nucleus was negative for both gfralpha1 and gfralpha2. The artemin (ART) receptor, gfralpha3, was expressed only in the superior salivatory nucleus. A motor neuron subnuclei-specific expression of gfralpha1 and gfralpha2 was seen in the facial and trigeminal nuclei which corresponded to their dependence on GDNF in null mutant mice. We found that the expression was dynamic in these nuclei, which may reflect developmental changes in their trophic factor dependency. Analysis of GDNF null mutant mice revealed that the dynamic receptor expression is regulated by the ligand in vivo, indicating that the attainment of changes in dependency could be ligand induced. Our results indicate that specific GDNF family ligands support selective muscle-motor neuron circuits during development.

A helminth community in breeding Yellow-legged Gulls (<i>Larus cachinnans</i>): pattern of association and its effect on host fitness

We studied the community of helminths parasitizing adult Yellow-legged Gulls (Larus cachinnans) sampled during three consecutive breeding seasons in a western Mediterranean colony. The helminth community was depauperate, probably because of the high trophic dependence of the gullery on waste food from refuse dumps. The prevalences of two helminth species varied both among years and between the sexes of gulls, while the intensity of parasitism did not vary significantly for any helminth species. No clear pattern of association between helminth species was detected, showing an unstructured helminth community in which the occurrence of any species was not restricted or favoured by the occurrence of any other species. The lack of associations might be influenced by the fact that only one core species occurred in the helminth community. Negative relationships were found between occurrence of certain helminths and body condition of hosts. Gulls infected by Tetrabothrius (Tetrabothrius) erostris were leaner than...

The Spatial–Temporal Gradient of Naturally Occurring Motoneuron Death Reflects the Time of Prior Exit from the Cell Cycle and Position within the Lateral Motor Column

Embryonic lumbar spinal motoneurons (MNs) are characterized by a period of programmed cell death (PCD) that spans several days and occurs in a rostrocaudal gradient. The generation of these MNs also takes place in a temporal–spatial gradient, such that MNs within rostral lumbar segments exit the cell cycle earlier and MNs within progressively caudal regions are born later. In vitro studies have shown that the latest born spinal MNs, presumably through the possession of endogenous “survival properties,” are also the last to acquire their trophic dependence. If the birth date and therefore spinal cord location of lumbar spinal MNs influence the spatial–temporal pattern of PCD, then earlier born MNs should die sooner and be located more rostrally than those generated later. Alternatively, if the time at which MNs die during development is unrelated to their prior exit from the cell cycle, those born at various phases should die throughout the period of PCD. We report here that lumbar MNs generated during the earliest part (embryonic day 2–3) of the proliferative period in the developing chick spinal cord tend to die during the earliest stages of the PCD period and that MNs born in successive 12-h intervals die at correspondingly later periods during PCD. Furthermore, the spatial progression of PCD of these subpopulations of MNs occurs in a rostrocaudal gradient. Finally, while MNs do appear to die in a mediolateral gradient during the period of MN PCD, this pattern is only partly accounted for by MNs born in consecutive intervals. These data support the notion that the timing and rostrocaudal location of MNs undergoing PCD reflect their time of exit from the cell cycle.

Development of Survival Responsiveness to Brain-Derived Neurotrophic Factor, Neurotrophin 3 and Neurotrophin 4/5, But Not to Nerve Growth Factor, in Cultured Motoneurons from Chick Embryo Spinal Cord

During embryonic development, most neuronal populations undergo a process usually referred to as naturally occurring neuronal death. For motoneurons (MTNs) of the lumbar spinal cord of chick embryos, this process takes place in a well defined period of time, between embryonic days 6 and 10 (E6-E10). Neurotrophins (NTs) are the best characterized family of neurotrophic factors and exert their effects through activation of their specific Trk receptors. In vitro and in vivo studies have demonstrated that rodent motoneurons survive in response to BDNF, NT3, and NT4/5. In contrast, the trophic dependencies of chicken motoneurons have been difficult to elucidate, and various apparently conflicting reports have been published. In the present study, we describe how freshly isolated motoneurons from E5.5 chick embryos did not respond to any neurotrophin in vitro. Yet, because motoneurons were maintained alive in culture in the presence of muscle extract, they developed a delayed specific survival response to BDNF, NT3, and NT4/5 that is clearly dose-dependent, reaching saturation at doses of 100 pg/ml. This trophic response correlated with increasing expression of the corresponding functional receptors TrkB and TrkC. Moreover, TrkB receptor is able to become autophosphorylated and to activate classical intracellular signaling pathways such as the extracellular signal-regulated protein kinase when it is stimulated with its cognate ligand BDNF. Therefore, our results reconcile the reported differences between in vivo and in vitro studies on the ability of chicken MTNs to respond to some members of the neurotrophin family of trophic factors.

Selective lesion of the cholinergic basal forebrain causes a loss of cortical neuropeptide Y and somatostatin neurons

Degeneration of the cholinergic basal forebrain (CBF) and changes in cortical neuropeptide levels have been reported in Alzheimer's disease. In the present study, we sought to determine if a selective cholinergic lesion of nucleus basalis magnocellularis (Nbm) could affect the number and distribution of neuropeptide Y (NPY) and somatostatin (SS) immunoreactive neurons in the frontoparietal and occipital cortices of rats. Brain sections were evaluated at survival times of 1, 2, 4, 8, 12, 24, 48, 78 and 100 weeks after intraventricular injection of 192-saporin, an immunotoxin directed at the low affinity neurotrophin receptor (p75 NGFr), that selectively destroys the CBF. Following the immunotoxin lesion of the Nbm, the number of NPY-labeled neurons decreased 33% in the frontoparietal cortex and 60% in the occipital cortex compared to age-matched normal controls at most survival time points. A significant loss of SS-labeled neurons in both cortical regions was seen 12 weeks after 192-saporin injection with no further change up to 100-week survival time. The effect of age on neuropeptidergic populations was evaluated in normal control rats. The number of NPY and SS immunoreactive neurons in aged rats (21–26 months) decreased by 42% in the frontoparietal cortex and 27% in the occipital cortex when compared with young (3–6 months) and middle-age (9–14 months) rats. When both non-lesioned and lesioned animals with different ages were pooled for linear regression, a significant correlation was found between the number of cortical NPY- and SS-labeled neurons and cortical acetylcholinesterase (AChE) histochemical staining intensity. These findings indicate that: (1) cholinergic denervation of the Nbm is associated with an irreversible loss of neocortical NPY and SS immunoreactive neurons analogous to that observed in Alzheimer's disease and aging; (2) the degree of the loss of cortical NPY and SS immunoreactive neurons seems to be related to the extent of the reduction of cortical AChE intensity in both toxin-injected and normal aged rats. These findings may reflect a trophic dependence of NPY and SS neurons on cortical cholinergic input.

Terminal dendritic sprouting and reactive synaptogenesis in the postnatal organ of Corti in culture.

Synaptogenesis in the organ of Corti between the primary receptors, the inner hair cells, and the peripheral processes of their afferent spiral ganglion neurons in the mouse lasts for 5 days postnatally (Sobkowicz et al. [1986] J. Neurocytol. 15:693-714). The transplantation of the organ into culture at the fifth postnatal day induces a reactive sprouting of dendritic terminals and an extensive formation of new ribbon synapses within 24 hours. This reactive synaptogenesis differs strikingly from the primary synaptogenesis and has been seen thus far only in the inner hair cells. The synaptically engaged neuronal endings sprout a multitude of filopodia that intussuscept the inner hair cells. The filopodial tips contain a heavy electron-dense matter that appears to attract the synaptic ribbons, which form new synaptic contacts with the growing processes. The intensity of the filopodial growth and synaptogenesis subsides in about 3 days; the filopodia undergo resorption, leaving behind fibrous cytoplasmic plaques mostly stored in the supranuclear part of the hair cells. However, occasional filopodial growth and formation of new synaptic connections continued. The data demonstrate that any disruption or disturbance of the initial synaptic contacts between the inner hair cells and their afferent neurons caused by transplantation results in prompt synaptic reacquisition. Furthermore, we suggest that the transitory phase of terminal sprouting and multiribbon synapse formation manifests a trophic dependence that develops postnatally between the synaptic cells.

Degeneration of Cajal-Retzius cells in the developing cerebral cortex of the mouse after ablation of meningeal cells by 6-hydroxydopamine

In the central nervous system, the neurotoxic drug 6-hydroxydopamine (6-OHDA) selectively deletes central catecholaminergic neurons and meningeal cells. Meningeal cells are known to contribute to brain development and their specific degeneration leads to disorganized neuronal positioning. We have analyzed whether a particular population of cortical pioneer neurons, the Cajal-Retzius (CR) cells, which lie just below meningeal cells, is also affected by 6-OHDA treatment. We show that application of 6-OHDA to the cortical surface leads to a rapid degeneration of CR cells, without affecting other cortical neurons. The ablation of CR cells was prevented by normetanephrine, which blocks the 6-OHDA uptake into meningeal cells. These results indicate that the disappearance of CR cells after 6-OHDA treatment may be a result of the ablation of the meningeal cells and suggest a trophic dependence of CR cells upon meningeal cells.

Cortical target depletion and the developing lateral geniculate nucleus: implications for trophic dependence.

The dependence of the developing dorsal lateral geniculate nucleus (LGd) on visual cortex for survival has been well documented. Complete removal of visual cortex during early postnatal development results in degeneration of the LGd. To further explore the nature of this trophic relationship, we depleted variable proportions of the principal targets of geniculocortical axons, layer IV neurons, and also variable proportions of the supragranular neurons by intraperitoneal injections of different dosages of a mitotic inhibitor MAM (methylazoxymethanol acetate) into pregnant hamsters at the time when these neurons were being generated in the ventricular zone. We demonstrate that after more than 75% loss of layer IV there is no reduction in cell number in the LGd. HRP (horseradish peroxidase) injections into the LGd in adult animals reveal an essentially normal pattern of termination without evidence of rerouting of geniculocortical axons to other cortical areas, nor compensatory increase in arborization in layer VI and VIb (subplate). Geniculocortical axons terminate principally in the middle stratum of the depleted cortex above layer V, with obvious reduction in both the extent and density of arborization. After higher dosages of MAM treatment resulting in more severe cell loss in layers II-IV with the apparent loss of layer IV, the extent and density of geniculocortical arborization are further reduced. Reduction in size as well as total number of geniculate neurons become detectable. Above depletions of 75% of layer IV neurons, the number of surviving LGd neurons is linearly related to the total number of remaining layer II-IV neurons in the cortex. These findings are discussed in light of the possible trophic mechanisms that match cell populations in number during development.

The course of chloroplast DNA replication and its relationship to other reproductive processes in the chloroplast and nucleocytoplasmic compartment during the cell cycle of the algaScenedesmus quadricauda

DNA containing structures (cellular, chloroplast and mitochondrial nuclei) were stained with the fluorochrome DAPI. Fluorescence intensity, as a measure of DNA content, was estimated during the mitotic cycle in synchronized populations of the chlorococcal alga,Scenedesmus quadricauda. In cells yielding eight daughter cells, three consecutive steps in chloroplast DNA increase occurred over one mitotic cycle. The first step was performed shortly after releasing the daughter cells, the second and third steps occurred consecutively during the first half of the mitotic cycle. Commitment to chloroplast DNA replication was chronologically separated from commitment to division of chloroplast nuclei, revealing that these two chloroplast reproductive steps were under different control mechanisms. The replication of chloroplast DNA occurred at a different time to that of cell-nuclear DNA. The coordination of chloroplast reproductive processes and those in the nucleocytoplasmic compartment were governed by the mutual trophic and metabolic dependency of these compartments rather than by any direct or feedback control controlled by either of them.

Survival of newly postmitotic motoneurons is transiently independent of exogenous trophic support

We compared the survival requirements of early- and late-born motoneurons from E5 chicken spinal cord. Density gradient centrifugation followed by immunopanning using SC1 antibody allowed us to purify two size classes of motoneuron. Large motoneurons retained by 6.8% metrizamide were shown by BrdU labeling in ovo to be born on average 1.5 d earlier than the small motoneurons recovered from the metrizamide pellet. Large motoneurons were both biochemically and functionally more mature: they expressed higher levels of choline acetyltransferase and low-affinity neurotrophin receptor, and had an acute requirement for trophic support from muscle-derived factors. After 24 hr in culture in basal medium, all early-born motoneurons died, whereas 60% of late-born motoneurons survived. Small motoneurons can develop into large motoneurons in ovo, suggesting that they represent a general transitional stage in motoneuron development. Our results suggest that a defined period elapses between birth of a motoneuron and its acquisition of trophic dependence, possibly corresponding to the time required for target innervation. This property may have important consequences for the timing and regulation of developmental motoneuron death.

Carbon flux on coral reefs:effects of large shifts in community structure

The effect of replacement of live coral cover by epilithic algae on patterns and magnitudes of carbon flux is examined for the shallow front slope of a midshelf reef in the Great Barrier Reef (GBR) complex of Australia. A steady-state network of carbon exchange among 19 trophic compartments is constructed for the coral-dominated state. From this, 2 scenarios for patterns of carbon flux when algae dominate are derived, viz. (1) the increase in algal production is channeled to detrital pathways (grazers do not respond), and (2) grazers utilise the increase in production of algal carbon so that transfers to detritus and grazers are in the same proportion as occurs when coral cover is high. The 3 models summarise current knowledge of carbon flux on GBR reef fronts and are compared using network analysis. Because fluxes in the reef front zone are dominated by exogenous imports and exports as a result of the high volume of water passing around and over the reef, the analyses ignore advective fluxes across the zone that are not internalised.The shift in structure to an algae-dominated system realises lower rates of benthic primary production, and thus system slze and activity (i.e. total system throughput, internal throughput, development capacity and ascendancy) are reduced, suggest- ing a disturbed system. With loss of coral cover, the proportion of the total flow that is recycled and transferred to the detritus pool increases (although the structure of recycling is not affected), and the balance of pathways in the network is changed: average path length increases, while the average trophic level of most of the second order consumers, and trophic efiiciencies of most trophic categories, decreases. Also, there are marked changes in dependencies of particular trophic groups on others. The analysis shows that, in the coral-dominated state, carbon fixed by zooxanthellae is used indirectly by most organisms in the system, even those seemingly remotely connected. Differences between the coral- and algae-dominated systems were much greater than differences between the 2 scenarios for the algae-dominated state. However, the exact fate of additional algae-derived carbon In the system is an important consideration since the 2 scenarios for the algae-dominated state yielded dissimilar values for some parameters (e.g. flow diversity, trophic dependencies and effective trophic levels of some com- partments, relative importance of recycling, trophic efficiency of some trophic categories).

Retinoic acid induces BDNF responsiveness of sympathetic neurons by alteration of Trk neurotrophin receptor expression

The expression of high affinity neurotrophin receptors (TrkA, TrkB, and TrkC) determines the survival response of different populations of neurons to specific members of the neurotrophin family, including nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), and neurotrophin-3 (NT-3). However, the mechanism which controls the expression of neurotrophin receptors during neuronal development is largely unknown. Here we show that the treatment of the cultured sympathetic neurons from newborn rat superior cervical ganglia (SCG) with retinoic acid (RA), a derivative of vitamin A, suppressed the expression of trkA mRNA and induced the expression of trkB mRNA. Expression of the functional TrkB receptor was confirmed by the emergence of trophic dependence of these neurons on BDNF in the absence of NGF. Differential regulation of trk mRNAs by RA provides a possible model for the establishment of neurotrophin dependence of peripheral neurons.

Development ofGlossiphonia complanata(Annelida, Hirudinea) and morphogenesis of helical muscle fibers

Summary The development of Glossiphonia complanata takes place in about 40 days. The young are attached to the parent's ventral side by means of their posterior sucker. At the time of hatching muscle cells are not recognizable. Therefore, the various morphogenetic steps of these round circomyarian fibers can be analyzed. The contractile material starts to be formed in a continuous peripheral ring encircling a central cytoplasmic core. Thus the round circomyarian condition is achieved from the very beginning. A comparison is made with the morphogenesis of muscle fibers in oligochaetes where an initial platimyarian condition is present. The formation of the gut in glossiphonids is completed in the last phases of embryonic development. When the young individual detaches from the parent some yolk store is still present in the gut diverticula. This condition is related to the peculiar trophic dependence of the young on the parent.

Sympathetic neuroblasts undergo a developmental switch in trophic dependence

Sympathetic neurons require NGF for survival, but it is not known when these cells first become dependent on neurotrophic factors. We have examined in vitro mitotically active sympathetic neuroblasts immuno-isolated from different embryonic stages, and have correlated this functional data with the expression of neurotrophin receptor mRNAs in vivo. Cells from E14.5 ganglia are supported by neurotrophin-3 (NT-3) in a serum-free medium, but not by NGF; NT-3 acts as a bona fide survival factor for these cells and not simply as a mitogen. By birth, sympathetic neurons are well-supported by NGF, whereas NT-3 supports survival only weakly and at very high doses. This change in neurotrophin-responsiveness is correlated with a reciprocal switch in the expression of trkC and trkA mRNAs by sympathetic neuroblasts in vivo. These data suggest that neurotrophic factors may control neuronal number at earlier stages of development than previously anticipated. They also suggest that the acquisition of NGF-dependence may occur, at least in part, through the loss of receptors for these interim survival factors.