Introduction

Nerve growth factor (NGF) was the first identified member of the neurotrophin family. Proteins in this family are essential for growth and survival in the developing nervous system.1,2,3 Neurotrophins are also vital for the maintenance of the mature nervous system, and increasing evidence suggests these growth factors act as mediators of synaptic plasticity.4,5 However, recent data point to a paradoxical role for neurotrophins. Under specific conditions NGF can have a neurotoxic effect and mediate the elimination of certain cell populations.6,7,8

The diverse biological activities of neurotrophins may be explained by two classes of receptors with which these proteins associate. Each neurotrophin binds to a specific member of the tyrosine kinase receptor (Trk) family, and all interact with the 75-kDa common low affinity neurotrophin receptor p75NTR.9,10 The latter receptor belongs to the tumor necrosis factor (TNF) family. NGF binds specifically to TrkA, and this interaction has been well documented to support cell survival and differentiation.10 In contrast, the role of p75NTR is poorly understood, and may depend on the cellular context in which it is expressed.11,12 When co-expressed with TrkA, p75NTR can augment the trophic effects of NGF by enhancing TrkA autophosphorylation, and affinity for NGF binding.13,14 On the other hand, in cells which express only p75NTR, NGF may mediate a signaling cascade which leads to apoptotic cell death,6,7,8 although this does not always occur.15,16 When expressed in the absence of TrkA, p75NTR binds NGF with a higher affinity than when expressed alone, however this parameter has not been tested in cells which undergo p75NTR mediated apoptosis.17

The signal transduction pathways of p75NTR-mediated apoptosis have yet to be fully characterized, but are thought to involve the production of ceramide,18 activation of c-jun kinase,7 the accumulation of p53,19 and activation of various caspases.20 NGF interaction with p75NTR also leads to the translocation of nuclear factor κB (NF-κB), but this effect is thought to be anti-apoptotic.21,22 Ligand-mediated apoptosis is a trait shared by many members of the TNF receptor family.23

The transition metal cation Zn2+ is found in abundance in the environment, and has been shown to influence many functions in the nervous system, including the regulation of apoptosis and inflammatory states.24,25 This metal is packaged in presynaptic vesicles and released with chemical or electrical stimulation. Upon release, Zn2+ concentrations in the perisynaptic area can reach 100–300 μM.26 At this concentration, it has been suggested that Zn2+ contributes to neuronal death in cerebral ischemia,27 epilepsy,28 Alzheimer's disease,29 Parkinson's disease and amyotrophic lateral sclerosis.24

Zinc has previously been shown to block NGF-mediated neurite outgrowth in chick dorsal root ganglia (DRG), and rescue pheochromocytoma (PC12) cells from oxidative stress.30,31 Both effects are mediated via the TrkA receptor. Theoretical studies have predicted that the ability of NGF to bind to p75NTR can be attenuated in the presence of Zn2+ and other transition metal cations,32 and Zn2+ blocks binding to ‘low affinity’ p75NTR.30 These results led to the hypothesis that Zn2+ may block the pro-apoptotic effects of NGF. In this study we characterize the effects of Zn2+ on the binding of NGF to p75NTR and p75NTR-mediated apoptosis in embryonic day 5.5 (ED5.5) chick retinal cells.

Results

Binding affinity of NGF to p75NTR

Saturation binding was performed using dissociated ED5.5 chick neural retina and 125I-NGF concentrations between 0.005 and 1 nM. Linear transformation of the steady state saturation binding data of 125I-NGF to retinal cells demonstrated homogenous binding affinity parameters consistent with one binding site. Previously, it has been shown that chick retinal cells express p75NTR at ED4 and do not express TrkA until after ED6,6 although one report shows that specific populations of precursor amacrine cells may express TrkA as early as ED6.37 TrkA could not be detected in these cells with chemical cross-linking and immunoprecipitation (data not shown) thus, p75NTR is the only receptor available for interacting with NGF. The measured dissociation equilibrium constant (Kd) was 0.42 nM±0.12 (Figure 1a). To confirm this observation we also determined the IC50 value using various concentrations of unlabeled NGF (0.04–40 nM) to inhibit 125I-NGF (0.15 nM) from chemically cross-linking p75NTR in retinal and in PC12 cells which express both p75NTR and TrkA.17 Results were evaluated with SDS–PAGE and autoradiography (Figure 1b,c), followed by excision of bands and counting of radioactivity incorporated into both monomers and dimers of p75NTR (Figure 1d). In both cell types, the displacement of 125I-NGF from monomers and dimers did not differ statistically. The IC50 values for the monomer of p75NTR in retinal cells and PC12 cells were 0.23±0.03 and 0.55±0.11 nM respectively, suggesting that p75NTR has a higher affinity in retinal cells than in PC12 cells.

Figure 1
figure 1

(a) Scatchard analysis of steady state binding of 125I-NGF to ED5.5 chick retinal cells displays homogenous binding consistent with a single receptor binding of NGF with an equilibrium dissociation constant of 0.42±0.12 nM (S.E.M. n=7). To compare the affinity of the p75NTR monomer and homodimer in retinal and PC12 cells, the ability of various concentrations of unlabeled NGF to inhibit chemical cross-linking of 0.15 nM 125I-NGF to p75NTR was evaluated. In both cell types, the displacement of 125I-NGF in the monomer and homodimer did not statistically differ. The loss of cross-linking of 125I-NGF with increasing concentration of unlabeled NGF was determined by autoradiography followed by excising bands and counting incorporated radioactivity. Displacement analysis revealed that the p75NTR monomers had an IC50 value of 0.23±0.03 nM in retinal cells (b), and 0.55±11 nM in PC12 cells (c). The shift in monomer displacement profile between both cell types is shown in (d). (Results shown ±S.E.M. n=4)

Zn2+ inhibits binding of NGF to p75NTR

Previously, we proposed that Zn2+ alters the conformation of NGF,30,32 and inhibits binding to p75NTR, in a low affinity state.30 Here we demonstrate that Zn2+ blocks the binding of 125I-NGF to p75NTR in dissociated ED5.5 retinal cells. As p75NTR is the only neurotrophin receptor which interacts with NGF in these cells, we could use both total binding and chemical cross-linking analysis to evaluate the dose response profile of Zn2+. At concentrations of 100 μM, Zn2+ (and copper, data not shown) effectively blocks 125I-NGF binding (Table 1). This finding was selective, as equal concentrations of Mg2+ had no effect on binding. Dose response analysis revealed that the addition of 50 and 25 μM Zn2+ partially blocked binding, while concentrations of 10 μM had no effect. The inhibitory effects of Zn2+ could be prevented by the addition of equimolar CaEDTA into the incubation buffer. Inhibition of EDC/SNHS cross-linking of 125I-NGF to p75NTR with Zn2+ was seen with a similar dose response profile (Figure 2).

Table 1 Effect of Zn2+ on binding to retinal cells. Binding of 125I-NGF to p75NTR in chick retinal cells is blocked by 100 μM Zn2+. The binding efficiency was restored with the adition of equimolar CaEDTA. Partial inhibition was achieved with the addition of 50 and 25 μM Zn2+, while 10 μM had no effect. With the addition of 100 μM Mg2+, no change in binding was observed
Figure 2
figure 2

Zinc inhibits EDC/SNHS receptor cross-linking of 125I-NGF to p75NTR in retinal cells at concentrations of 100 μM, partially at 50 and 25 μM and has no effect at 10 μM

Zn2+ inhibits p75NTR-mediated apoptosis

Apoptosis mediated via p75NTR signaling following exposure to NGF has previously been demonstrated in the normal development of chick neural retina, and in explanted cultures.6,38 Thus, to determine the effects of Zn2+ on cell death, dissociated cells from ED5.5 chick embryos were plated on 16-well chamber slides. Each well was treated independently and incubated overnight at 37°C with 5% CO2 and 95% air. To determine the number of apoptotic nuclei, nick translation TUNEL and anti-ssDNA MAb staining were performed. All slides had two control wells with either no treatment or control beads with cytochrome c. There was no significant difference between these two control groups in both assays. The basal percentage of apoptotic cells was taken from the cytochrome c treated groups which ranged between 9.5–12.7% following TUNEL staining, and 11.1–13.9% in MAb stained cultures. The control cytochrome c groups were normalized to 100% (Figure 3). With the addition of NGF coated glass beads, the number of TUNEL and MAb positive stained nuclei was significantly increased to 208±7.5 and 258.9±2.5% of control level respectively. When co-administered with NGF, Zn2+ (100 μM), proved to be neuroprotective as the number of TUNEL and anti-ssDNA MAb positive nuclei was decreased significantly to levels similar to control cultures. This protective effect could be blocked with equimolar CaEDTA (186±14.2 and 209±3.2% positive apoptotic cells). As predicted through binding analysis, this effect was specific for Zn2+, as 100 μM Mg2+ did not alter binding and did not significantly alter NGF-mediated cell death (195±9.2% TUNEL positive cells). As a positive control, we explored the anti-apoptotic effects of a known p75NTR antagonist, PD90780.39 When added at 100 μM this compound had a similar effect to Zn2+, and inhibited cell death in the ssDNA assay (Figure 4). As previously observed,6 the apoptotic effect was specific for NGF, as brain derived neurotrophic factor (BDNF) and neurotrophin-3 (NT-3) showed no significant increase in positive anti-ssDNA MAb stained cells compared to cytochrome c cultures when all proteins were bound to glass beads (Table 2).

Figure 3
figure 3

Exogenous NGF on glass beads induces chick retinal cells to undergo apoptosis shown using both TUNEL (a) and anti-ssDNA MAb (b) staining. The addition of 100 μM Zn2+ attenuates this effect, and NGF activity can be restored with the addition of equimolar CaEDTA. This effect is independent of the glass beads as the number of positively stained cells is not significantly different between the group treated with glass beads loaded with cytochrome c and untreated cultures. Zinc has no effect when administered without NGF. Results were analyzed using a one-way ANOVA and Dunnett's multiple comparison post hoc test (results shown ±S.E.M.). In each culture treatment, eight fields were counted with approximately 100 cells each. (TUNEL n=5; anti-ssDNA MAb, n=4; asterisks signify P<0.01)

Figure 4
figure 4

NGF mediated apoptosis is attenuated with the addition of the p75NTR antagonist PD90780 as seen when assayed for anti-ssDNA positive cells. Results were analyzed using a one-way ANOVA and Dunnett's multiple comparison post hoc test (results shown ±S.E.M.). In each culture treatment, eight fields were counted with approximately 100 cells each

Table 2 Per cent on anti ss-DNA Mab positive cells. Apoptosis mediated by p75 is specific for NGF as BDNF and NT-3 treated cells do not show greater number of postive cells than control cultures treated with cytochrome c. In each culture eight fields were counted with approximately 100 cells each (asterisk signifies P<0.01)

Discussion

There are three principal observations emerging from these studies. First, the pro-apoptotic effects of NGF on embryonic chick retina demonstrated initially by Frade et al.6,38 have been extended to dissociated cultures. Second, in cells where NGF mediates apoptosis through p75NTR signaling, 125I-NGF binds to p75NTR with high affinity. Third, Zn2+, which has been shown to alter NGF conformation and influence TrkA-mediated survival functions,30,31 also attenuates the pro-apoptotic activities mediated by the high affinity interaction of p75NTR and NGF.

There is an increasing body of evidence demonstrating that NGF can promote apoptosis in some cell populations which express p75NTR and not TrkA.6,7,8 A feature of this cell death which was only reported in chick retinal cells is that exogenous NGF in solution does not promote apoptosis, while NGF bound to glass beads (solid phase) does.38 The reason for this is unknown but it has been suggested that the way NGF is presented to developing cells in vivo by microglia may be critical.38 However, whether there are other factors involved in presenting NGF to p75NTR expressing retinal cells in vivo has yet to be explored. In our cultures, solid phase NGF treated cells demonstrated a twofold increase in the number of TUNEL positive nuclei, and 2.5-fold increase in the number of anti-ssDNA MAb positive nuclei compared to control and non-treated cultures. The greater number of MAb positive cells may be attributed to the fact that this antibody is more sensitive to the early stages of apoptosis than TUNEL labeling.36 The increased number of apoptotic cells can be attributed to the bound NGF, as groups treated with beads covered with cytochrome c had no significant increase in apoptotic nuclei compared to untreated control cultures.

Previous work has shown that the binding characteristics of p75NTR for NGF are dependent on the presence of TrkA.17 However, binding analysis of 125I-NGF to p75NTR has not been reported in cells that undergo cell death following this binding. In PC12 cells that express p75NTR and TrkA, p75NTR adopts a lower affinity state (Kd≈4.0 nM) than when p75NTR is expressed alone (PC12nnr5 cells; Kd≈0.2 nM). When co-expressed with TrkA, p75NTR serves a secondary role of promoting the binding of NGF to TrkA by influencing the allosteric properties of this receptor, possibly through the formation of a heteroreceptor complex.17 This concept is supported by studies which demonstrated that both receptors can be co-immunoprecipitated with antibodies to both p75NTR and TrkA when PC12 cells are incubated with EDC/SNHS prior to the addition of 125I-NGF.17 In ED5.5 chick retinal cells, we found that p75NTR displayed sub-nanomolar binding affinity measured by dissociation equilibrium analysis and showed that p75NTR binding in retinal cells displayed higher affinity than in PC12 cells by displacement studies. Therefore, apoptosis mediated by the interaction of NGF and p75NTR occurs when this receptor is in its high affinity conformation.

Through competition binding and receptor cross-linking studies, we have shown that 100 μM Zn2+ is sufficient to inhibit binding of NGF to p75NTR in ED5.5 chick retinal cells. Molecular modeling and ab initio calculations have been used to predict the most stable geometry of the Zn2+:NGF coordination complex.32 When Zn2+ interacts with NGF, it is chelated by four residues of the NGF dimer forming a 5-coordinate state with donor atoms Nε of His4, Nε of His8, Nδ of His84 and both δ-oxygens of Asp105.30 Transition metal cations split into three groups by their intrinsic stereochemical preference in the penta-coordinated environment. The optimal geometry of the coordination site for transition metal cations within NGF, distorted square pyramidal, is consistent with intrinsic stereochemical preferences of Zn2+ and Cu2+.32 The conformational changes of NGF induced by Zn2+ results in the decreased accessibility of the p75NTR binding determinant within NGF, loops I and IV,40,41 thereby inhibiting this receptor-ligand interaction.30,42

In our apoptosis assays, the addition of 100 μM Zn2+ to the NGF treated cells provided protection against p75NTR-mediated apoptosis, as the percentage of TUNEL and anti-ssDNA MAb positive nuclei were attenuated to levels similar to those seen in control cultures. Previous studies have shown that chick retinal cells can be protected against NGF induced apoptosis with the addition of the peptide compound dc28–36.38 This compound mimics loop I of NGF, a part of the binding determinant of p75NTR, and competitively inhibits receptor binding.40,42 The likely mode of action of this compound is through direct competitive inhibition of p75NTR and not through altering ligand structure. With Zn2+, we propose that protection is mediated by the ability of this cation to alter the conformation of NGF, which leads to non-competitive inhibition of binding to p75NTR. Consistent with this hypothesis, Zn2+ has been shown to inhibit protomer cross-linking of the two NGF monomers suggesting that a conformational alteration within NGF has taken place.30,32

Zinc has been shown to affect neuronal function in many ways, including both inhibition and activation of specific receptors, by entering post-synaptic cells and interrupting cellular homeostasis.25 The findings that Zn2+ can directly modulate cell functions are consistent with much of the existing experimental data indicating that Zn2+ contributes to the pathology of Alzheimer's disease, ischemia, amyotrophic lateral sclerosis and epilepsy.27,28,29,43 Following insult, presynaptic Ca2+-mediated release of vesicular Zn2+ results in chelatable concentrations of this transition metal between 100–300 μM.26 Our results demonstrate that this concentration is sufficient to inhibit NGF-mediated apoptosis and suggest a novel biological function of Zn2+.

The ability of p75NTR to contribute to neurodegeneration has been demonstrated under several conditions where Zn2+ may also be involved. Increased p75NTR expression has been observed following middle cerebral artery occlusion induced ischemia.44 Recent experimental data also suggest that β-amyloid protein binds to p75NTR, and this binding can lead to apoptosis.45 However, the most compelling evidence for the direct involvement of p75NTR in neuronal degeneration under pathologic conditions may be the observations on cell death following both pilocarpine induced seizure activity and unilateral kainic acid administration. The effects of both insults correlate strongly with the expression of p75NTR in hippocampus, cortex and basal forebrain regions.46,47 Further, the chronic administration of antibodies to rat p75NTR substantially decreases kainic acid induced cell death of cholinergic cells.47 The finding that Zn2+ inhibits apoptosis mediated by p75NTR and NGF is significant, considering that vesicular Zn2+ release following insult occurs in many areas of the nervous system where p75NTR has been implicated in cell death. Therefore, this mechanism for Zn2+-mediated protection offers novel insight into the potential interaction of Zn2+, NGF and p75NTR under conditions associated with cell death in the nervous system. The characterization of p75NTR-mediated cell death prior to the appearance of trkA (when NGF becomes trophic) will enhance further investigation of the co-modulation of these receptors.

Materials and Methods

Chick embryos

Fertilized white horn hen eggs were incubated at 38°C in 70% humidity, and staged as previously described.33 The dissection was performed by excising both eyes from the optic cups and removing the retinal layer. Single cell suspensions of retinal cells were prepared by digesting explants in 0.05 mg/ml DNase, 0.01 mg/ml trypsin in calcium magnesium free Gey's balanced salt solution (CMF; pH 8.1) for 15 min at 37°C. Cells were then suspended in RPMI 1640 with 10% heat inactivated horse serum (Gibco), spun through fetal calf serum (Gibco) and washed three times in RPMI 1640.

125I-NGF preparation

Full-length (1–118) mouse NGF was obtained from Cedarlane Laboratories. The radioination of NGF was performed as previously described,34 with modification.30

Receptor binding and cross-linking

Both binding and cross-linking were used to determine the affinity of p75NTR. PC12 cells tissue culture preparations and binding and cross-linking procedures were performed as previously described.17

Retinal cell culture

To culture cells for cell death determination, they were washed three times in RPMI and suspended in RPMI+2% B27 supplement (Gibco) and plated on 16-well chamber slides coated with poly-d-lysine (0.1 mg/ml overnight). Wells were individually treated and incubated at 37°C with 5% CO2 and 95% air.

NGF adsorbed beads

Full-length mouse NGF was adsorbed on controlled pore glass beads as previously described.35 Briefly, glass powder was incubated overnight in trifluracetic acid at room temperature, then beads were washed with distilled water and PBS six times each. Neurotrophins or cytochrome c were added to the activated beads at concentrations of 0.01 mg/ml and placed on ice for 1 h. Beads were washed with PBS six times and applied immediately on top of the plated cells.

Apoptosis determination

To determine the number of apoptotic cells, TUNEL and anti-ssDNA Mab staining were performed. For TUNEL staining, cells were fixed in PBS with 4% formaldehyde and 4% sucrose, then solubilized in 0.25% Triton X-100 for 5 min, then suspended in 3% H2O2 for 5 min. PBS with 2% BSA was used as a blocker (30 min). DNA nick translations were marked with biotin-16-dUTP as described in manufacturers instructions (Boehringer Mannheim). Adding 300 mM sodium chloride and 30 mM sodium citrate for 45 min stopped the reaction. Biotin labeled DNA was marked using ABC kit as per manufacturers instruction (Vectastain). Anti-ssDNA MAb staining was performed as described.36 Briefly, cells were fixed with methanol/PBS (6 : 1) at −20°C, then resuspended with 4.5 mM MgCl in PBS and heated to 100°C for 5 min. Following a wash with PBS, plated cells were incubated with 3% H2O2 for 5 min and 0.1% BSA for 30 min. Then washed and exposed to MAb F-27 (Alexis) for 30 min, biotin conjugated goat anti mouse IgM (Sigma) for 15 min, and to ExtraAvidin peroxidase (Sigma) for 15 min with washings between each step. In TUNEL and MAb assays, cells were stained with DAB (Vectastain) solution as described by manufacturer. Slides were then washed with water and dehydrated in ethanol (75–100%) and xylens for 5 min each. Permount was used to fix the coverslips for microscopy.