Interferon protects neurons in culture infected with vesicular stomatitis and herpes simplex viruses

The protective effect of crude rat interferon against infection by vesicular stomatitis virus (VSV) and herpes simplex virus (HSV) was assessed in two culture systems: the PC12 cell line and dissociated rat neurons derived from the superior cervical ganglion (SCG). Interferon was induced in rat embryo cells by inactivated Newcastle disease virus, and its effect was assessed by reduction of viral yields and prevention of viral cytopathology. Interferon protected PC12 cells, both in the presence and absence of nerve growth factor (NGF), against infection by both viruses, although at differing concentrations: protection against VSV was noted at approximately a 10-fold lower interferon concentration than that required to inhibit HSV infection. Dissociated SCG neurons were also protected, but higher interferon concentrations were required. These results demonstrate that the antiviral state can be established in neurons in response to interferon.     

The response of adrenergic neurones to axotomy and nerve growth factor.

Division of the axons of adrenergic neurones by crushing the postganglionic nerve trunks of rat superior cervical ganglia (SCG) at 6 days of age resulted in a permanent atrophy of the SCG reflected by a persistent decrease in the total protein content and in the activities of the enzymes tyrosine hydroxylase and DOPA decarboxylase. Administration of nerve growth factor (NGF) to rats with unilateral axotomy at a dose of 10 mug/g/day for the period 7-21 days of age resulted in hypertrophy of both normal and axotomised SCG. There was a progressive rise in the total protein content and in the activities of the two enzymes till the end of the treatment period in both SCG. After treatment ceased there was a progressive fall in the total protein content and activities of the two enzymes reaching a stable level after 4 weeks. The level reached for treated unoperated SCG remained elevated when compared to untreated control SCG. Axotomised treated SCG had approximately the same biochemical parameters as untreated control SCG and very much elevated over untreated axotomised SCG. These final levels persisted for at least 56 days after treatment had ceased. Animals showed a persistent ptosis after axotomy at 6 days of age but treatment with NGF resulted in a functional recovery by 11 weeks of age. It is suggested that there is normally a retrograde transfer of a factor durind development from the target cell to the perikarya of the neurone permitting survival if the appropriate connections are made. Failure to make such a contact results in cedd death. The cell death occurring normally, and the cell death resulting from axotomy, can both be prevented by NGF treatment leading to an hypertrophy of both SCG. This consistent with the hypothesis than NGF is the retrograde trophic agent for the sympathetic nervous system in the developing animal.     

Transynaptic regulation of low-affinity p75 nerve growth factor receptor mRNA precedes and accompanies lesion-induced collateral neuronal sprouting.

The bilateral sympathetic innervation of the rat pineal gland from the two superior cervical ganglia (SCG) is a useful model system to investigate the mechanisms by which intact neurons compensate for neuronal losses. Cutting of the internal carotid nerve (ICN) on one side has been shown to result in the removal of approximately one-half of the innervation to the pineal gland within 2 days. This denervation is followed by the development of collateral neuronal sprouting from the contralateral "intact" SCG, most of which takes place during the next 2 days. Using a solution hybridization protection assay, levels of low-affinity NGF receptor p75NGFR mRNA (pg/microgram total RNA) were found to be increased 25%, with no change in cyclophilin mRNA, in the SCG contralateral to the lesion performed 1 or 3 days earlier. In situ hybridization with a 35S riboprobe complementary to p75NGFR mRNA demonstrated a large increase in this mRNA in some cells of this intact SCG at both 1 and 3 days after a contralateral ICN cut lesion. The clustering of these cells toward the rostral portion of the SCG suggests that they may overlap with the population of sympathetic neurons which provides innervation to bilaterally innervated structures such as the pineal gland. The nature of the signals involved in the regulation of NGF receptor mRNA levels and their role in initiating and maintaining collateral sprouting remain to be fully established. Nevertheless, the time course of the changes in mRNA levels suggests that regulation of the low-affinity NGF receptor gene may be involved in the sequence of events associated with the collateral sprouting response by intact sympathetic nerve cells following partial denervation.     

Destruction of sympathetic and sensory neurons in the developing rat by a monoclonal antibody against the nerve growth factor (NGF) receptor

The ability of the monoclonal antibody, 192-IgG, directed against the rat nerve growth factor (NGF) receptor to mimic or inhibit the actions of NGF was examined in vitro and in vivo. 192-IgG had no effect on morphology, survival, or protein synthesis rates of sympathetic neuronal cultures. When injected into newborn rats, destruction of sympathetic, but not sensory, neurons was produced. Injection prenatally produced more dramatic destruction of sympathetic neurons and, in addition, destruction of neural crest-derived sensory neurons. Therefore, although 192-IgG had no discernible effects in vitro, it produced a pattern of neuronal destruction in vivo qualitatively similar to that produced by antibodies to NGF itself.     

Suppression of the in vitro and in vivo cytotoxic effects of guanethidine in sympathetic neurons by nerve growth factor

Repeated administration of guanethidine to newborn or adult rats produces destruction of sympathetic neurons. Addition of guanethidine to sealed plasma clot cultures (pH 7.8–8.0) of newborn rat or embryonic chick sympathetic ganglia produces cell destruction if the culture contains low concentrations (1 U) of nerve growth factor (NGF). Addition of higher concentrations of NGF suppresses the cytotoxic effect of guanethidine. The cytotoxic effects of guanethidine are not seen if the cultures are carried out at a lower pH (7.0–7.2). If NGF (2000 U/g/day) is given concurrently with guanethidine (100 μg/g/day) to newborn rats, the destruction of the sympathetic neurons normally caused by the guanethidine is prevented. Other guanidinium adrenergic neuron blocking agents (debrisoquine, guanoxan, guanadrel, and bethanidine) were testedin vitro andin vivo. All producedin vitro cytotoxic effects. None produced sympathetic neuron destructionin vivo. The results are consistent with a mechanism of guanethidine cytotoxicity involving inhibition of mitochondrial respiration. The inability of adrenergic neuron blocking agents, other than guanethidine, to destroy sympathetic neuronsin vivo may result from a lesser degree of accumulation in the sympathetic neuron cell body.     

Effects of gonadal steroids on nerve growth factor receptors in sympathetic and sensory ganglia of neonatal rats

The numbers of neurons in the rat superior cervical sympathetic ganglion (SCG) differ in males and females, with the males having 30% more SCG neurons than females at 60 days of age. This sex difference arises during the early postnatal period, when testosterone administration increases the numbers of neurons and alters the nerve growth factor (NGF) content of the rat SCG. In contrast, there is no gender difference in number of neurons in the L1 dorsal root ganglion. In both males and females, the amount of NGF bound per ganglion increased between postnatal days 5 and 15 (P5 and P15) in both dorsal root ganglia (DRGs) and the SCG. There is also a gender difference in NGF binding: SCGs and DRGs of female rats at both P5 and P15 bind more NGF per ganglion than do those of males. This effect was more marked in DRGs than in the SCG. Treatment of neonatal females with testosterone reduced NGF binding in both SCGs and DRGs to levels comparable to males at P5, and in DRGs at P15. In contrast, treatment of males with testosterone from birth resulted in a 2-3 fold increase of NGF binding in both SCGs and DRGs as compared to controls at P15. At P15, testosterone treatment of females increased NGF binding in the SCG. Males and females had opposing responses to neonatal exposure to estradiol. Treatment with estradiol from birth increased NGF binding in SCGs and DRGs of females, but had no effect on NGF binding of SCGs, and reduced NGF binding in DRGs of males.(ABSTRACT TRUNCATED AT 250 WORDS)     

The effect of nerve growth factor and its antibody on mast cells in vivo

Repeated subcutaneous administration of nerve growth factor (NGF) into newborn rats caused an increase in number and size of free and tissue-adherent mast cells (MCs). In vivo and in vitro autoradiographic studies show that MCs exhibit specific binding sites for NGF. Peritoneal MCs of adult rats exposed to high titers of endogenous (autoimmunization) and exogenous NGF antibodies showed severe cytological alterations and decreased in number. Light and electron microscopic studies revealed that MC destruction took place slowly (7-10 days) and required continual exposure to high titers of NGF antibodies. The role played by NGF in the hematopoietic cell lineage is at present under study in our laboratory.     

Endogenous nerve growth factor is required for regulation of the low affinity neurotrophin receptor (p75) in sympathetic but not sensory ganglia

During development, many sympathetic and sensory neurons are dependent on nerve growth factor (NGF) for survival. The low affinity neurotrophin receptor (p75), expressed in these neurons, is regulated by exogenous NGF in vitro and in vivo. However, whether p75 expression in vivo is under the control of endogenous NGF has not been determined. The role of NGF in regulating the expression of p75 in sympathetic and sensory nerves was investigated in Sprague-Dawley rats treated with an antiserum specific for NGF. P75 was differentially regulated. P75 immunoreactivity (-ir) within sympathetic neurons in the superior cervical ganglia (SCG) was reduced after 2 days, and disappeared after 5 days, of treatment with the NGF antiserum. In contrast, a significant increase in p75-ir was detected in nerve bundles within and close to the SCG from 3 to 14 days after treatment. A similar pattern of p75 expression was observed in the stellate and coeliac ganglia. In contrast, p75 expression in nerve terminals of the mesenteric arteries and irides was reduced. However, in the same animals the expression of p75 was not significantly affected by the treatment in dorsal root, trigeminal or nodose ganglia, salivary gland or small intestine. In contrast to p75, the NGF high affinity receptor trkA was little affected in sympathetic neurons by depletion of endogenous NGF for 2 weeks. These results indicate that endogenous NGF is required in sympathetic ganglia for the expression of p75 but not trkA in neurons, but for the down-regulation of p75 in glia. In contrast, endogenous NGF is not essential for the regulation of p75 in neurons or glia within sensory ganglia. © Wiley-Liss, Inc.     

Remodeling of adult sensory axons in the superior cervical ganglion in response to exogenous nerve growth factor

In previous studies, we found that a 2-week in vivo intracerebroventricular infusion of nerve growth factor (NGF) elicited a sprouting response by sympathetic perivascular axons associated with the intradural segment of the internal carotid artery. We hypothesized that NGF infused into the ventricular system would be internalized by responsive sympathetic cerebrovascular axons, retrogradely transported to parent cell bodies in the superior cervical ganglion (SCG), and subsequently released into the local ganglionic environment. Because fibers exhibiting immunoreactivity for calcitonin gene related peptide (CGRP) have been localized in the SCG, we used immunohistochemical methods to investigate whether a response by CGRP-immunoreactive axons in the SCG occurred following the proposed transport to and release of exogenous NGF in the ganglion. In consecutive tissue sections of the SCG stained for either CGRP or NGF, we found CGRP pericellular 'baskets' surrounding identified NGF-immunoreactive cell bodies. Nerve growth factor infusion resulted in a significant increase both in the number of CGRP pericellular baskets and in NGF-immunoreactive cell bodies. A significant positive correlation (r=0.95, P<0.05) between the pericellular baskets and NGF-immunoreactive cell bodies was observed, suggesting that intracranial projection neurons in the SCG released infused NGF (or possibly a converted signal) into the local ganglionic environment to elicit remodeling of CGRP fibers to form pericellular baskets. These findings were confirmed in sections double labeled for NGF and CGRP immunoreactivity. This remodeling suggests that exogenous NGF may mediate retrograde transneuronal plasticity, allowing for future in vivo examinations of the mechanisms involved in neurotrophin transport and release.     

The effects of nerve growth factor (NGF) and antiserum to NGF on the development of embryonic sympathetic neurons in vivo

The role of nerve growth factor (NGF) in the development of embryonic sympathetic neurons was examined in vivo. Individual mouse embryos received transuterine injections of NGF or antiserum to NGF (anti-NGF), and the effects on the superior cervical ganglion (SCG) were studied. Treatment with NGF at any gestational stage, from the time of ganglion aggregation to birth, increased ganglion tyrosine hydroxylase (T-OH) activity. Both the number of catecholaminergic neurons and T-OH activity per neutron were increased. Choline acetyltransferase (ChAc) activity was increased by NGF at early gestational stages, but not at later stages. These observations suggest that perikarya containing ChAc are responsive to NGF, whereas preganglionic nerve terminals are not. Treatment with anti-NGF rapidly and permanently decreased ganglion T-OH activity. The effects of anti-NGF were more pronounced at later gestational stages, suggesting that ganglia become increasingly dependent on NGF during development. Alteration of maternal levels of NGF had no effect on development of the embryonic SCG, suggesting that local embryonic concentrations of NGF are responsible for modulating sympathetic ontogeny.     

A study of the de- and regenerative changes in the sympathetic nervous system of the adult mouse after treatment with the antiserum to nerve growth factor.

In the adult mouse, the antiserum to nerve growth factor (NGF) induced marked atrophic changes of the ganglionic cell bodies in the superior cervical ganglion (SCG) and a disappearance of adrenergic nerve terminals in several peripheral tissues. By fluorescence histochemistry a lower-than-normal content of the noradrenaline (NA) transmitter was observed within the entire adrenergic neurone only 1 day after a single injection of NGF-antiserum (0.1 ml/g body weight). An atrophy of adrenergic nerve cell bodies and a disappearance of adrenergic nerve terminals were observed after 3 days, but the antiserum-induced effects did not appear maximally developed until 7 days after treatment. These fluorescence histochemical findings were paralleled by a gradual decrease of the endogenous NA levels in peripheral tissues and also of the weight of the SCG. A gradually proceeding restoration towards normal of the adrenergic innervation apparatus was observed fluorescence histochemically following a 5-day treatment with NGF-antiserum (0.1 ml/g body weight each dose), and after 6 weeks to 3 months a normal or close to normal fluorescence microscopical appearance was regained in the peripheral tissues and also in the SCG. These findings were parelleled by the results of the determinations of endogenous NA in peripheral tissues and by the results of weighing the SCG. We discuss some important differences between NGF-antiserum and 6-hydroxydopamine with respect to their mode of action on the mature sympathetic nervous system. Finally, we suggest that a decreased availability of NGF in a terminal area, due to an interference with endogenous NGF by NGF-antibodies, may temporarily result in an impaired function of the supplying adrenergic neurone, including a degeneration of nerve terminals.     

Different effects of nerve growth factor on cultured sympathetic and sensory neurons

Long-term cultures of dissociated nodose ganglion (NG) and superior cervical ganglion (SCG) neurons from newborn rabbits were used to compare their response to nerve growth factor (7S NGF). SCG neurons required added NGF for their survival and a concentration of 1 μg/ml was found to be optimal. NG neurons, on the other hand, survived well for a long term without addition of NGF, but its application (1 μg/ml) was found to be effective in accelerating the growth of fibers (neurites) and neuronal somata. It is concluded that unlike SCG, NG neurons do not depend on exogenous NGF but may require an intrinsic trophic-like factor which may be contained in the serum of the medium, emanating from glial cells or by metabolic cooperation between neurons.     

Comparative studies on the effect of the nerve growth factor on sympathetic ganglia and adrenal medulla in newborn rats

The effect of nerve factor (NGF) and nerve growth factor antiserum (NGF-AS) on sympathetic ganglia and adrenal chromaffin cells of newborn rats was investigated in a comparative study.     

Destruction of the sympathetic nervous system in neonatal rats and hamsters by vinblastine: prevention by concomitant administration of nerve growth factor

If the primary mechanism by which nerve growth factor (NGF) gains access to the cell body is a specific uptake and retrograde transport from the periphery, then an inhibitor of axoplasmic transport would be expected to produce cell death. Such an inhibitor was administered to neonatal and adult rats and hamsters. A single injection of vinblastine (0.25–0.5 mg/kg s.c.) to 2-day-old rats produced massive cell death in the superior cervical ganglia (SCG). The levels of tyrosine hydroxylase in the SCG were reduced 80% at doses of 0.4 mg/kg, a dose which was uniformly fatal. Doses of 0.25 mg/kg at two days of age resulted in a 50% mortality rate and the survivors showed a permanent, partial sympathectomy. The adrenal medulla and sensory neurons did not appear to be affected. Administration of a single injection of vinblastine at maximally tolerated doses did not destroy sympathetic neurons in animals 14 days of age or older. Chronic administration of vinblastine (0.5 mg/kg every other day for two weeks) to adult rats did not destroy sympathetic neurons. Similar results were obtained in hamsters.Concomitant administration of NGF appeared to completely prevent the cytotoxic effects of vinblastine assessed by both morphological and enzymatic criteria. Several explanations are discussed which might explain these results.The similarities in the effects of vinblastine, other drugs, and axonal crush on sympathetic neurons are noted. It is proposed that all drugs which destroy sympathetic neurons (guanethidine, 6-hydroxydopamine, vinblastine) as well as anti-NGF and axonal ligation destroy the neurons by depriving the cell body of NGF from the periphery.     

Sympathetic ingrowth to the trigeminal ganglion following intracerebroventricular infusion of nerve growth factor

The objective of the present study was to examine the remodeling of uninjured sympathetic axons in the adult rat trigeminal ganglion following a 2-week in vivo intracerebroventricular infusion of NGF. The accumulation of infused NGF in the trigeminal was assessed using ELISA and sympathetic fibers were localized immunohistochemically with an antibody to tyrosine hydroxylase (TH). In addition, high performance liquid chromatography coupled with electrochemical detection (HPLC-ECD) allowed for biochemical measurements of the catecholamines norepinephrine (NE) and dopamine (DA). Increased NGF protein in the trigeminal ganglion was paralleled by a significant increase in sympathetic fibers and pericellular plexuses (i.e. baskets) in the cell body regions. Some ganglia showed elevated NE following NGF infusion, yet the 88% increase in mean NE did not reach significance. Following bilateral removal of the sympathetic superior cervical ganglia (SCG), a significant reduction was observed in overall NE levels and in TH-immunoreactive (-ir) fibers in the cell body regions and peripheral branches, suggesting the SCG as the origin of the sympathetic ingrowth. However, mean DA levels as well as TH-ir fibers within the trigeminal central branch were unaffected by NGF infusion or removal of the SCG and likely resulted from intrinsic dopaminergic cell bodies. In conclusion, our data provide evidence that the increased availability of NGF in the young adult rat trigeminal ganglion observed following in vivo NGF infusion enhanced sympathetic associations with the sensory neurons in the trigeminal, supporting a role for NGF in the regulation of sympathosensory interactions.     

Biological importance of retrograde axonal transport of nerve growth factor in adrenergic neurons.

Previous studies have shown that nerve growth factor (NGF) produces a selective induction of tyrosine hydroxylase (TH) in peripheral adrenergic neurons and that NGF is transported retrogradely with a high selectivity from the adrenergic nerve terminals to the perikaryon. In order to investigate the biological importance of retrograde NGF transport, the following experiments have been performed; (a) effect of NGF on TH activity in superior cervical ganglia (SCG) after unilateral injection into the anterior eye chamber and the submaxillary gland; and (b) effect of systemic injection of NGF on TH activity in SCG after blockage of retrograde axonal transport by axotomy. After unilateral injection of NGF into the anterior eye chamber and submaxillary gland of both 8-10-day-old rats and adult mice, the increase in TH activity in the SCG was considerably larger on the injected than on the non-injected side although the adrenergic neurons supplying the two organs do not account for more than 25% of the total number of adrenergic neurons in the SCG. A direct diffusion mechanism could be excluded by the fact that unilateral local injection of [125 I] produced no significant side difference in the accumulation of radioactivity in the SCG 2 after injection whereas after 14 h there was a several-fold difference between the injected and non-injected side. Moreover, the nodose ganglia which are located very close to the SCG exhibited no statistically significant difference in the accumulation of radioactivity at any time. Forty-eight hours after subcutaneous injections of 10 mg/kg of NGF the increase in TH activity of the SCG amounted to 154% on the intact side and to 92% on the axotomized side. However, these experiments do not permit decisions about the extent the axotomy, as such, impaired the response to NGF. It is concluded that the biological effect of NGF results to a considerable extent, from the moiety which reaches the cell body by retrograde transport from the nerve terminals.     

Protection from guanethidine-induced neuronal destruction by nerve growth factor: Effect of NGF on immune function

The chronic administration of guanethidine causes an immune-mediated destruction of sympathetic neurons in rats. Destruction can be prevented by various immunosuppressive agents, including gamma-irradiation and cyclophosphamide, as well as by administration with nerve growth factor (NGF). Experiments were conducted to determine whether: (1) NGF prevented accumulation of guanethidine within sympathetic neurons; and (2) NGF caused an inhibition of immune function by either blocking proliferation of immune-competent cells or by blocking effector function even in the presence of antigen and activated immune cells. NGF did not prevent accumulation of guanethidine within sympathetic ganglia in vivo, a necessary prerequisite for neuronal destruction, nor was it inhibitory on immune function using several assay systems. NGF, purified by either conventional methods or additionally by HPLC ("ultrapure'), did not inhibit either proliferation of cloned cytotoxic T lymphocytes (CTL) to antigen (class I major histocompatibility antigens) or lysis of target cells bearing the appropriate antigens. In addition, NGF did not exhibit growth stimulating effects in this assay system (i.e. it could not substitute for T cell growth factor). NGF also did not cause an inhibition of either murine or rat allogeneic mixed lymphocyte responses measured by lysis of appropriate target cells or proliferation, respectively. Finally, NGF did not inhibit, but rather appeared to stimulate the antibody response to sheep red blood cells generated in vivo in young rats. Thus NGF does not appear to prevent the immune-mediated neural destruction induced by guanethidine by acting as an immunosuppressive agent, but rather acts by some other mechanism such as preventing expression or recognition of antigen(s) on the sympathetic neuron.     

Immunohistochemical localization of nerve growth factor in the rat pineal gland

Sympathetic nerve fibers arising from the superior cervical ganglia are the main innervation of the rat pineal gland. Since most organs innervated by these ganglia contain nerve growth factor (NGF), the hypothetical existence of NGF in the pineal gland was investigated. The peroxidase anti-peroxidase technique was applied for the immunohistochemical demonstration of NGF using a polyclonal antiserum on Bouin-fixed, paraffin-embedded pineal glands from adult, young and 6-hydroxydopamine (6-OHDA)-treated rats. Few immunopositive cells were observed in the adult pineal gland. A more conspicuous population of immunoreactive cells was noted in young animals (20-45 days old), especially in those chemically denervated with 6-OHDA. NGF immunoreactive cells displayed a stellate shape resembling the interstitial or glial cells previously described in the rat pineal gland. Since NGF plays a trophic effect on sympathetic neurons during development and adulthood, we postulate that its presence in the pineal gland may exert a trophic role on its sympathetic innervation.     

Activation of adenosine A2A receptor protects sympathetic neurons against nerve growth factor withdrawal

Adenosine mediates a range of effects in the central nervous system (CNS), including the promotion of neuronal survival, but its actions on sympathetic neurons are less well characterized. We therefore sought to understand the role of endogenous adenosine in contributing to the survival of neurotrophin-dependent sympathetic neurons. Rat superior cervical ganglion (SCG) cultures were maintained in the continuous presence of nerve growth factor (NGF) and then exposed to adenosine deaminase (ADA), to deplete endogenous adenosine. This resulted in a marked increase in cellular apoptosis, to a level that approximated the effect of NGF withdrawal. Furthermore, the addition of exogenous adenosine to NGF-deprived SCG neurons resulted in enhanced cell survival. Analysis of adenosine receptor (AR) subtypes on these neurons, using real-time RT-PCR and receptor binding analyses, revealed that the A2A receptor was the major subtype present. Accordingly, the A2A receptor agonist CGS21680 significantly reduced both ADA-induced and NGF-withdrawal-induced neuronal apoptosis, whereas the A1 receptor agonist R-PIA had no such effect. The survival-promoting effect of CGS21680 was eliminated when cells were coincubated with a molar excess of an A2A receptor antagonist. Finally, follow-up experiments revealed that CGS21680 prevented the induction of early apoptotic events, such as changes in mitochondrial integrity and caspase activation, and that it also triggered an increase in ERK activation, which was essential for neurotrophin-independent cell survival. Taken together, these findings provide evidence that endogenous adenosine may be important in mediating protection of sympathetic neurons and that it may act via the A2A receptor subtype.     

Local control of neurite sprouting in cultured sympathetic neurons by nerve growth factor

Sprouting of neurites by sympathetic neurons from newborn rats was studied in compartmentalized cultures. The neuronal cell bodies resided in proximal compartments, and neurites penetrated silicone grease barriers and elongated within distal compartments. Nerve growth factor (NGF) was initially supplied at 1 microgram/ml in all compartments, but was subsequently withdrawn from proximal compartments and for a time was only supplied to distal neurites. Little or no neurite growth was observed in proximal compartments after NGF withdrawal, but reintroduction of NGF resulted in substantial neurite growth over the next few days which was shown to have originated from local sprouting within the proximal compartments. This result is distinct from previous work on NGF-enhanced nerve fiber elongation in demonstrating that quiescent, NGF-deprived regions of sympathetic neurons sprout neurites in response to local reexposure to NGF.