Address corespondence and reprint requests to Andrzej Szutowicz, Chair of Clinical Biochemistry, Department of Laboratory Medicine, Medical University of Gdañsk, Dêbinki 7 Str., 80–211 Gdañsk, Poland. E-mail: firstname.lastname@example.org
Nerve growth factor (NGF) is a peptide displaying multiple cholinotropic activities. The aim of this work was to explain mechanisms of the positive and negative effects of NGF on phenotypic properties and viability of cholinergic cells. To discriminate these effects we used two p75NTR receptor-positive lines of cholinergic neuroblastoma cells, SN56 and T17 that are devoid of or express high affinity NGF (TrkA) receptors, respectively. cAMP and retinoic acid caused differentiation of both cell lines. In addition to the morphologic maturation, the increase of choline acetyltransferase activity, acetylcholine, Ca and cytoplasmic acetyl-CoA levels and decrease of mitochondrial acetyl-CoA and cell viability were observed. NGF caused similar effects in non-differentiated T17 cells but had no influence on non-differentiated SN56 cells. On the contrary, in both cAMP/all-trans-retinoic acid (RA) differentiated cell lines, NGF resulted in a similar suppression of cholinergic phenotype along with an increase of mitochondrial acetyl-CoA and cell susceptibility to nitric oxide and amyloid-β25–35. These effects of NGF were prevented by an antibody against the p75NTR receptor. Data indicate that: (i) positive cholinotrophic effects of NGF required activation of both TrkA and p75NTR receptors; (ii) cAMP/RA-evoked differentiation inhibited NGF effects mediated by TrkA receptors and activated its p75NTR-dependent suppressing influences and (iii) a differentiation-evoked decrease of mitochondrial acetyl-CoA and an elevation of mitochondrial Ca could augment impairment of cholinergic neurons by neurotoxic signals.
Nerve growth factor (NGF) is one of the neurotrophins with distinct cholinotropic activity in the brain. It was found to increase activities of choline acetyltransferase (choline-O-acetyltransferase, ChAT, EC 22.214.171.124) and vesicular acetylcholine transporter, the acetylcholine (ACh) level and rate of its release/synthesis as well as density of cholinergic M2 autoreceptors in septum, hippocampus and other cortical areas containing cholinergic neurons and their axonal terminals (Tonnaer et al. 1994). NGF resulted in regeneration of transsected cholinergic septo-hippocampal fibres (Horner and Gage 2000). Similar cholinotrophic effects of NGF were observed in primary and clonal cultures of cholinergic cells where it brought about increase of expression of mRNA, protein and activity of ChAT along with elevation of ACh content and release (Oosawa et al. 1999). These effects were mediated by activation of specific, high affinity NGF TrkA receptors autophosphorylated at tyrosine residues found on the cytoplasmic domain with subsequent activation of multiple signal transduction pathways. Cooperation of the p75NTR low affinity, non-specific receptor with the TrkA receptor plays an important role in modulation of NGF effects on cholinergic cells (Mamidipudi and Wooten 2002). Activation of p75NTR receptors resulted in suppression of cholinergic neurons in the brain and increased their death rate (Coulson et al. 2000). On the other hand, in p75NTR knock-out mice, the septohippocampal brain region displayed larger cholinergic neurons with higher ChAT activity than in control animals (Naumann et al. 2002). Accordingly, over-expression of p75NTR caused increased sensitivity of cholinergic neurons to excitotoxic neurodegeneration (Oh et al. 2000). The similar hypersensitivity to amyloid-β peptide (Aβ) was found in p75NTR-enriched 3T3 cells (Yaar et al. 1997). Hence, variable interactions between these two classes of neurotrophin receptors may cause either differentiation or suppression of cholinergic phenotype.
Loss of cholinergic neurons is a hallmark of several encephalopathies. Decreases of ChAT, vesicular ACh transporter and high affinity choline uptake activities, as well as ACh content were found in post mortem examination of brains of Alzheimer's disease (AD) patients (Tonnaer and Dekker 1994; Szutowicz et al. 1996). These alterations corresponded to the degree of mental retardation found shortly before the patient's death. The depletion of NGF and its TrkA receptors in affected regions of AD brains has been also claimed to contribute to cholinergic neuron loss (Dubus et al. 2000).
Another common feature of cholinergic encephalopathies is a decrease in glucose and pyruvate utilisation along with suppression of pyruvate dehydrogenase (pyruvate : lipoate oxidoreductase acceptor acetylating, PDH, EC 126.96.36.199) activity proportional to the loss of cholinergic markers. Despite this, most of the various non-cholinergic neurons remain essentially preserved under these pathological conditions. It is claimed that this particular sensitivity of cholinergic neurons to neurodegeneration may result from the fact that they require acetyl-CoA both for energy production and for neurotransmitter synthesis (Szutowicz et al. 1996). Under pathological conditions, ACh release by neurons is markedly increased due to sustained membrane depolarization. This release triggers the re-synthesis of the transmitter maintaining the equilibrium of the ChAT reaction and restoring the intracellular ACh pool. However, such conditions suppress PDH activity, and facilitate acetyl-CoA transfer from mitochondria to the cytoplasmic compartment to be utilised for ACh synthesis. This could lead to depletion of acetyl-CoA in the mitochondrial compartment of extensively activated cholinergic neurons (Szutowicz et al. 2000).
Moreover, different groups of brain cholinergic neurons displayed variable sensitivity to the same pathogenic conditions. For instance, primary cholinergic neurons from medial septum lost their ChAT activity and became apoptotic after S-nitro-N-acetyl-penicillamine treatment, whereas those from brain stem were resistant to these conditions (Fass et al. 2000). Amyloid-β peptide (Aβ) brought about marked inhibition of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide reduction in ciliary neurotrophic factor-differentiated cholinergic RN-46 A neurons without an effect on brain-derived neurotrophic factor-differentiated serotoninergic cells (Olesen et al. 1998). This differential sensitivity to similar toxic inputs could be, at least in part, explained by variable proportions between rates of acetyl-CoA and ACh synthesis in different groups of brain cholinergic neurons (Szutowicz et al. 1996). Indeed, differentiated cholinergic SN56 neuroblastoma cells contained lower levels of acetyl-CoA than the non-differentiated ones (Jankowska et al. 2000). The same was the case for SN56 cells transfected with an additional copy of ChAT cDNA and expressing very high enzyme activity (Bielarczyk et al. 2003b). In addition, neurotoxic conditions such as Al accumulation or nitric oxide (NO) excess were found to cause more profound suppression of acetyl-CoA levels in differentiated rather than in non-differentiated cholinergic SN56 cells (Jankowska et al. 2000; Szutowicz et al. 2000).
It has been demonstrated, that NGF exerts divergent effects on differentiation of cholinergic neurons depending on their phenotype (Berse et al. 1999). It is, however, not known whether NGF is able to modify acetyl-CoA metabolism in cholinergic cells expressing different cholinergic phenotypes. Here, we present evidence that NGF and cAMP/all-trans-retinoic acid (RA) may significantly change the relationships between acetyl-CoA and ACh metabolism that may influence functional and survival abilities of cholinergic neurons under harmful conditions.
Materials and methods
Unless otherwise specified, biochemicals were obtained from Sigma-Aldrich (Poznañ, Poland), β-NGF was from Alomone Laboratories (Jerusalem, Israel), Aβ25–35 was from Bachem (Heidelberg, Germany), phospholine iodide was a generous gift from Wyeth (Manmouth Junction, NJ, USA), rabbit polyclonal antibody against fragment 43–161 of the mouse p75NTR receptor was from (Chemicon, Temecula, CA, USA) growth media and components were provided by Gibco Life Technologies (Warsaw, Poland), cell culture disposables derived from Sarstedt (Stare Babice, Poland), [1-14C-acetyl]acetyl-CoA (4 mCi/mmol) was from Perkin-Elmer (Boston, MA, USA).
Native SN56.B5.G4 cholinergic murine neuroblastoma and derived cell line stably transfected with rat trkA cDNA (T17) were used in this study (both cell lines were gift from Dr J. K. Blusztajn, Boston MA, USA; Hammond et al. 1990; Berse et al. 1999). T17 cells expressed functional TrkA receptors as indicated by phosphorylation of MAP kinase upon treatment with NGF (Berse et al. 1999). Thus, T17 were either TrkA+/p75NTR+ and SN56 TrkA–/p75NTR+ cells. They were grown for 3 days to subconfluency in Dulbecco's modified Eagle's medium (DMEM) containing 1 mm l-glutamine, 2500 IU streptomycin per 1 mL and 10% fetal bovine serum at 37°C in atmosphere 5%CO2, 95% air. Differentiation was achieved by addition of 1 mm dibutyryl cAMP (cAMP) with 0.001 mm all-trans-retinoic acid (RA). Differentiated and non-differentiated cells were harvested and plated again in DMEM with 1% N2 supplement, containing no differentiating agents. NGF and/or neurotoxic agents were added as indicated in the Results section. For studies of ACh metabolism, 16 h before harvesting the cells, 0.005 mm phospholine iodide was added into the growth medium. Other changes in culture protocol are indicated in Results section. Cells were harvested into 10 mL of ice-cold HEPES buffered 0.9% NaCl, washed twice by centrifugation at 500 g for 5 min with same solution and suspended in 0.32 m sucrose containing 10 mm HEPES buffer (pH 7.4) and 0.1 mm EDTA to obtain protein concentration 10.0 mg/mL. Immediately after collection, the cells were used for Trypan blue exclusion assay and for metabolic studies. For enzyme assays, samples were kept frozen at −20°C for 2–7 days.
Trypan blue exclusion assay
Cell suspensions were mixed with an equal volume of 0.4% isotonic trypan blue solution. Total cell number and fraction of non-viable, dye accumulating cells were counted after 2 min in Fuchs–Rosenthal haemocytometer under light microscope (Wang et al. 2001).
Incubation medium contained in a final volume of 1.0 mL, 2.5 mm pyruvate, 2.5 mm l-malate, 90 mm NaCl, 30 mm KCl, 20 mm NaHEPES (pH 7.4), 1.5 mm Na-phosphate, 0.01 mm choline chloride, 0.015 mm eserine sulfate, 0.02 mm EDTA, 32 mm sucrose and 1–2 mg of cell protein. Incubation was started by the addition of cell suspension and continued for 30 min at 37°C with shaking at 100 cycles per min. For assay of acetyl-CoA content in cell mitochondria, 0.5 mL of incubation medium was mixed with equal volume of lysing solution containing 1.4 mg digitonin/mL in 125 mm KCl with 20 mm NaHEPES buffer (pH 7.4) and 3 mm EDTA. Lysate was transferred on 0.5 mL of silicone oil mixture (AR20 and AR200, 1 : 2). After 30 s, the mitochondrial fraction was separated from the soluble by centrifugation for 40 s at 12 000 g. After removal of the soluble fraction and silicon oils, the mitochondrial pellet was deproteinised by suspension in a small volume of 5 mm HCl and incubation in a boiling bath for 1 min. For determination of total acetyl-CoA content, 0.3 mL of incubation medium was centrifuged at 5 000 g for 2 min. Supernatant was removed and the cell pellet was deproteinised as described above.
Deproteinised extracts of whole cells and mitochondria were treated with maleic anhydride solution in ethyl ether for 2 h to remove CoA-SH. A cycling-reaction was carried for 60 min in 0.1 mL of medium containing 1.9 mm acetyl phosphate, 1.2 mm oxaloacetate, 0.72 IU phosphotransacetylase and 0.12 IU citrate synthase. This reaction was stopped by heating samples to 95°C for 6 min and the citrate formed was determined (Szutowicz and Bielarczyk 1987). Cytoplasmic acetyl-CoA levels were calculated by subtraction of mitochondrial acetyl-CoA from total acetyl-CoA content (Szutowicz and Bielarczyk 1987).
For the ACh assay, 0.2 mL incubation medium was centrifuged for 3 min at 10 000 g. The pellet and supernatant were used for assays of intracellular and released ACh, respectively. ACh was determined by an HPLC method with an enzymatic reactor containing acetylcholinesterase (EC 188.8.131.52) and choline oxidase (184.108.40.206) and electrochemical detection (commercial kit, Bioanalytical Systems, West Lafayette, IN, USA) (Pedersen et al. 1995; Jankowska et al. 2000).
For Ca accumulation studies, cells were incubated in depolarising medium containing 1 mm Ca, as described above, collected by centrifugation and washed with 1.5 mL Ca-free medium containing 1 mm EDTA to remove cation bound to the cell surface. EDTA was removed by washing the pellet with Ca/EDTA free incubation medium. The pellet was deproteinised with 5% (w/v) trichloroacetic acid (TCA). Supernatant was adjusted to pH 7.0 with NaOH. Samples (0.1 mL) were added to 1.0 mL of 0.009 mm arsenazo III solution in 20 mm Na-morpholinepropanesulfonic acid (MOPS) buffer (pH 7.0) and absorbance was read at 650 nm (Scarpa 1979).
Immediately before the assay, samples were thawed and diluted to desired protein concentration in 0.2% v/v Triton X-100. ChAT activity was assessed by the radiometric method using [1-14C]acetyl-CoA as a substrate (Fonnum 1975). PDH was assayed by trapping acetyl-CoA formed with citrate synthase in the presence of oxaloacetate with subsequent determination of citrate. (Szutowicz et al. 1981).
Protein was assayed by the method of Bradford (1976) with human immunoglobulin as a standard.
Statistical analyses were performed by one-way anova with Bonferroni multiple comparison test with p < 0.05 being considered statistically significant.
Concentration- and time-dependent effects of nerve growth factor on choline acetyltransferase activity
Three day culture of native non-differentiated SN56 (NC), expressing p75NTR but no TrkA receptors, with β-NGF concentrations from 0.1 to 500 ng/mL caused no significant change in ChAT activity and cell morphology (Figs 1a and 2). On the contrary, in T17TrkA+/p75NTR+ cells, increasing NGF concentrations resulted in gradual elevation of ChAT activity. Highest activation (≈ 160%) was observed at NGF concentrations from 1 to 100 ng/mL (Fig. 1a). This effect was accompanied by morphological maturation (Fig. 2). Higher concentrations of NGF caused decrease of enzyme activity. A 1 day lag period was observed for NGF-evoked ChAT activation in T17TrkA+ cells, with a maximal response attained after 3 days (Fig. 1B). On the basis of this observation, 100 ng of NGF per ml of medium and a 3-day culture period were used in subsequent experiments. It has been reported that cAMP/RA-evoked differentiation (DC) is maintained throughout subsequent passages without these agents (Bielarczyk et al. 2003a). On the contrary, NGF-induced differentiation of T17TrkA+ was not stable and disappeared during subsequent passage without this peptide (Fig. 3).
Differentiation-dependent effects of nerve growth factor on choline acetyltransferase activity
It is known that NGF interferes with ciliary neurotrophic factor-evoked increases of ChAT expression in SN56TrkA– cells by activation of the mitogen-activated protein kinase kinase/mitogen-activated protein kinase (MEK/MAPK) pathway (Berse et al. 1999). Therefore, one can expect that NGF effects on ChAT activity may depend on the actual activation of cAMP response element (CRE) and retinoic acid receptor in the cholinergic locus of these cells. In SN56TrkA– NC, NGF had no effect on ChAT activity whereas in T17TrkA+ NC it caused appromiately twofold increase of enzyme activity (Fig. 4). Addition of cAMP/RA resulted in several-fold increase of ChAT activity in both cell lines (Fig. 4). Under these conditions, NGF brought about a similar 40% suppression of ChAT activities in both cell lines (Fig. 4).
As the inhibitory effect of NGF was seen in both lines of DC, it was reasonable to think that p75NTR receptors, expressed in both of them, are involved in this process. Antibody to p75NTR receptor, when added alone (1 : 150 dilution), did not affect ChAT activity in SN56TrkA(–) nor in T17TrkA(+) DC and NC (Fig. 4). On the other hand, this antibody, added 6 h before NGF, totally reversed its suppressing effect on ChAT activity in DC of both lines (Fig. 4). It also abolished the activating influence of NGF on ChAT activity in T17TrkA+ NC (Fig. 4).
Effect of nerve growth factor on calcium content
NGF caused no change in Ca content in SN56TrkA– NC but increased it by 106% in T17TrkA+ NC. Differentiation with cAMP/RA elevated Ca accumulation in both cell lines (Fig. 5). Under these conditions, NGF brought about a further increase in Ca both in T17TrkA+ DC and SN56TrkA– DC, by 35 and 32%, respectively (Fig. 5).
Effect of nerve growth factor on acetylcholine metabolism
Differentiation with cAMP/RA brought about 142 and 134% increases of ACh content and release in native SN56TrkA– cells and 308 and 85% elevations of these parameters in T17, respectively (Fig. 6a and b). NGF had no effect on ACh release by SN56TrkA– cells, but elevated it by 272 and 147% in T17TrkA+ NC and DC, respectively (Fig. 6b). On the other hand, NGF caused no change of ACh content in SN56TrkA– NC, whereas in T17TrkA+ NC an 88% increase was observed (Fig. 6a). In DC of both lines, NGF exerted a 40–61% drop in ACh content (Fig. 6a) that corresponded well with decreased ChAT activity under same experimental conditions (Fig. 4).
Effect of nerve growth factor on mitochondrial and cytoplasmic acetyl-CoA level
It is known that the acetyl-CoA for ACh synthesis in cytoplasm is provided from mitochondria (Szutowicz et al. 1996). Therefore, we investigated whether changes in ACh metabolism, described above (Fig. 6a and b), had an impact on acetyl-CoA levels in these compartments. Differentiation with cAMP/RA caused > 70% decrease of mitochondrial acetyl-CoA content in both cell groups (Fig. 7a). NGF brought about no significant depression of mitochondrial acetyl-CoA in SN56TrkA– NC but decreased it by 39% in T17TrkA+ NC (Fig. 7a). On the other hand, cAMP/RA-evoked decrease of mitochondrial acetyl-CoA in both DC cell lines was partially reversed by NGF (Fig. 7a). This NGF-evoked elevation of mitochondrial acetyl-CoA remained as an inverse relation to NGF-induced decreases in ChAT activities and ACh content in DCs (compare Figs 4, 6a and 7a).
Changes in cytoplasmic acetyl-CoA were opposite to those observed in mitochondria. Differentiation caused a similar increase of acetyl-CoA levels in cytoplasm of both cell lines (Fig. 7b). It correlated with simultaneous elevation of the level and release of ACh and the decrease of acetyl-CoA content in mitochondria of these cells (Figs 6a and 7a). NGF caused a much higher elevation of cytoplasmic acetyl-CoA in T17TrkA+ NC than in SN56TrkA– NC (Fig. 7b). On the contrary, in SN56TrkA– DC and T17TrkA+ DCs, NGF decreased or had no effect on the levels of this metabolite in cytoplasm, respectively (Fig. 7b).
Effect of sodium nitroprusside and nerve growth factor on pyruvate dehydrogenase activity
In several cholinergic encephalopathies, loss of cholinergic neurons and ChAT activity was found to be accompanied by inhibition of PDH activity in affected brain areas (Szutowicz et al. 1996). Therefore, in this study, we investigated whether NGF modifies neurotoxic effects of NO depending on the presence of one or two classes of NGF cell surface receptors. Differentiation with cAMP/RA caused no change and a 30% decrease of PDH activity in SN56TrkA– and T17TrkA+ cells, respectively (Fig. 8). NGF decreased PDH activity in native SN56TrkA– NC and DC by 20 and 28%, respectively. Also a 20% suppression of enzyme ativity was observed in T17TrkA+ NC treated with NGF (Fig. 8). Sodium nitroprusside (SNP), a NO generator, had no effect on PDH activity in SN56TrkA– NC, but in the presence of NGF it exerted a 36% inhibition (Fig. 8). In SN56TrkA– DC, SNP decreased PDH activity by 28% and NGF did not augment this effect (Fig. 8). On the contrary, PDH activity in T17TrkA+ NC was changed neither by SNP alone nor in combination with NGF. However, in T17TrkA+ DC, a significant 37% inhibition of PDH was found in the presence of SNP and NGF (Fig. 8). Thus, PDH activity in native SN56TrkA– cells appeared to be more sensitive to NO excesses than in T17TrkA+ cells.
Effect of nerve groeth factor and sodium nitroprusside on choline acetyltransferase activity and cell viability
ChAT activity in SN56TrkA– NC was affected neither by SNP nor by NGF. In turn, exposure of SN56TrkA– DC to SNP caused a 30% decrease of ChAT activity, which was not modified significantly by NGF (Fig. 9) In T17TrkA+ NC, an activating effect of NGF was totally abolished by SNP which alone did not change basic ChAT activity. However, in T17TrkA+ DC, NGF augmented ChAT inhibition by SNP, which alone resulted in no significant alterations of enzyme activity (Fig. 9).
NGF alone had no effect on viability of SN56TrkA– NC and DC. However, it elevated the fraction of non-viable cells in T17TrkA+ DCs (Fig. 10). SNP increased the number of non-viable cells in overall populations of SN56TrkA– NC and T17TrkA+ NC from about 6% to 22 and to 11%, respectively. NGF did not increase the number of non-viable cells in NC treated with SNP (Fig. 10). DC of both lines appeared to be more sensitive to NO excesses than the NC. SNP increased the fraction of Trypan blue positive DCs of both lines to ≈ 30%. Addition of NGF in these conditions caused further increase of fraction of non-viable cells up to ≈ 37% (Fig. 10).
Role of p75NTR receptors in Aβ and SNP cytotoxicity
Aβ was found to evoke excessive NO synthesis in the brain (Eckert et al. 2003). Thus combined effect of both compounds may be involved in pathomechanisms of cholinergic encephalopathies. Also, NGF exerted divergent influences on differentiation of cholinergic neurons (Fig. 4). Therefore, we investigated whether NGF modifies the detrimental effects of SNP/Aβ in cholinergic cells. Addition of NGF to SN56TrkA– DC culture augmented inhibition of ChAT activity by SNP and Aβ25–35 from 50% to 74%(Table 1). These two toxins increased the fraction of Trypan blue positive cells from 8 to 26%. NGF itself did not change cell viablity but increased the detrimental effect of SNP and Aβ from 26 to 36% (Table 1) Anti-p75NTR antibody alone did not affect cell viability and ChAT activity. However, it totally overcame the sensitising effects of NGF on ChAT activity and cell viability in the presence of SNP and Aβ25–35 (Table 1).
Table 1. Effect of NGF on ChAT activity and viability of differentiated SN56 cells treated with amyloid- β(25–35) and SNP
ChAT activity (nmol/min/mg protein)
Trypan blue Positive cells (%)
Data are means ± SEM from four experiments. Significantly different from: acontrol; b1 mm SNP + 0.001 mmAβ(25–35); cSNP + Aβ(25–35) + NGF, at p < 0.01. Anti-p75NTR mouse receptor antibody was used in dilution 1: 150. SNP was added to the growth media as indicated, 24 h after plating for 10 min. All plates were washed with buffered saline. Appropriate media were restored as described and cultures continued for additional 48 h. Antibody was present in the medium during entire culture period. NGF (100 ng/mL) was added 6 h after plating.
Activation of ChAT and morphological differentiation in T17 by β-NGF remains in accord with the presence of TrkA receptor and its downstream signalling pathways in these cells. Lack of such response in SN56NC was due to the absence of TrkA receptors in this cell line (Figs 1a and 2) (Berse et al. 1999). Concentration of NGF required for half-maximal activating effect on ChAT activity was found to be about 0.4 nm (Fig. 1a). On the other hand, the inhibitory effect on ChAT activity seen at 8 nm NGF could result from activation of low affinity p75NTR receptors. These values corresponded to values of constants for high and low affinity NGF binding being in the high pm and low nm range, respectively (Fig. 1a)(Ross et al. 1998). They remain also in accord with findings that p75NTR–/– mice displayed cholinergic hypertrophy whereas those with over-expression of p75NTR presented a reduction of cholinergic innervation in basal forebrain, respectively (Naumann et al. 2002). It has been shown that cholinergic differentiation of SN56TrkA– through independent additive activation of cAMP response element binding protein (CREB) and retinoic acid nuclear receptors, persisted long time after cAMP/RA removal from the medium (Bielarczyk et al. 2003a). It was also claimed that NGF stimulates cell differentiation through CREB activation (Kaplan and Miller 2000). However, the fast disappearance of ChAT activation by NGF after its withdrawal from cultured T17TrkA+ NC indicates that this peptide causes cholinergic differentiation of these cells, by signalling pathways different from that stimulated by cAMP/RA (Fig. 3). This suggestion is supported by the fact that blockade of TrkA did not abolish CREB phosphorylation in sympathetic neurons (MacInnis et al. 2003). It is likely that NGF-evoked differentiation depends on MEK/MAPK activation, as inhibition of MEK with U0126 abolished the activating effect of this peptide in T17TrkA+ cells (A. Szutowicz and H. Bielarczyk, unpublished data) (Mellott et al. 2002).
The use of TrkA+ and TrkA– cells derived from a common line enabled us to discriminate TrkA and p75NTR receptor-mediated NGF effects on cholinergic cells. Thus, blockade of NGF-dependent increase of ChAT activity in T17TrkA+ NC by specific antibodies against p75NTR proved that co-activation TrkA by this low affinity neurotrophin receptor was indispensable for this process (Fig. 4). Lack of such interactions in SN56TrkA– NC confirms primary role of TrkA in NGF-dependent cholinergic differentiation (Fig. 4). These results remain in accord with data demonstrating that p75NTR acted as a helper receptor increasing affinity of NGF binding with TrkA (Canossa et al. 1996; Mamidipudi and Wooten 2002). On the other hand, they are in conflict with data showing that p75NTR receptors are not needed for cholinergic differentiation of PC12 cells by NGF (Niederhauser et al. 2000).
The presence of the suppressing effects of NGF on ChAT activity both in SN56TrkA– and T17TrkA+ DC indicates that it was mediated by p75NTR receptor activation in a manner independent of TrkA receptors (Fig. 4) (MacInnis et al. 2003). Similar efficacy of this NGF-evoked inhibition in both DC lines allow one to conclude that activation of the cAMP/RA pathway abolished TrkA-dependent signalling in T17TrkA+ DC (Fig. 4). In turn, reversal of NGF-evoked decrease of ChAT activity by anti-p75NTR antibody confirms significance of the low affinity neurotrophin receptor in suppression of cholinergic phenotype of cholinergic neurons differentiated by other mechanisms (Fig. 4). These findings allow one to conclude that activation of the CREB/RA pathway may block TrkA-mediated differentiation and facilitate the p75NTR suppressing signal in cholinergic neurons.
The increase of Ca content in SNTrkA– 56 and T17TrkA+ cells by cAMP/RA may be an important signal leading to cholinergic cell differentiation (Fig. 5). NGF elevated Ca content in T17TrkA+ NC but not in SN56TrkA– NC indicating that it is specific for differentiation of cells carrying TrkA receptors. Thus, different Ca-dependent mechanisms are involved in these two paradigms of cholinergic differentiation (Fig. 5). On the other hand, the suppressing effects of NGF on cholinergic phenotype in SN56TrkA– DC and T17TrkA+ DC were also accompanied by an increase of intracellular Ca (Figs 4 and 5). This excessive accumulation of Ca was not induced by cAMP/RA nor TrkA-dependent pathways as it took place in previously differentiated cells of both lines. Hence, this effect was probably mediated by activation of p75NTR present in both groups of cells. Excessive Ca accumulation in DCs caused by NGF may also explain cell sensitisation to NO and Aβ (Figs 5 and 10, Table 1) (Bielarczyk et al. 2003a; Hajnoczky et al. 2003).
Increase of Ca in mitochondria of DC may be responsible for decreased PDH activity, due to activation of PDH-kinase (Figs 5 and 8) (Lai et al. 1988; Bielarczyk et al. 2003a). Inhibition of pyruvate decarboxylation led to decreased concentration of acetyl-CoA in mitochondrial compartment of DCs (Fig. 7a and b) (Szutowicz et al. 1999). In addition, excessive accumulation of Ca in mitochondria of differentiated SN56TrkA– and T17TrkA+ cells might activate permeability transition state of the mitochondrial membrane and direct release of acetyl-CoA into cytoplasmic compartment (Fig. 7a and b) (Green and Reed 1998). Combination of these two factors could augment depletion of acetyl-CoA in mitochondria, which in this experiment declined to 30% of control values (Figs 7a and 8) (Szutowicz et al. 1999). In addition, the increase of ACh accumulation and release in DC could stimulate acetyl-CoA utilisation for the re-synthesis of the transmitter pool. Activation of ACh metabolism was facilitated by simultaneous release of acetyl-CoA from mitochondria to the cytoplasmic compartment (Figs 6a and b, 7a and b). Under such conditions, any neurotoxic factor that inhibits/inactivates PDH, impairs mitochondrial membrane integrity and/or activates excessive transmitter release could bring about further limitation of availability of this metabolite for energy-producing pathways in mitochondria (Figs 6a and b, 7 and 8) (Jankowska et al. 2000). Accordingly, differentiated cholinergic neurons could display greater vulnerability to neurotoxic conditions than the non-differentiated ones (Fig. 10).
This report provides the first direct evidence that NGF and other differentiating factors inversely affect acetyl-CoA content in mitochondria of cholinergic cells through modulation of ACh metabolism. Cholinergic differentiation by any signalling pathway would increase permeability of the mitochondrial membrane to acetyl-CoA leading to decrease of its content in mitochondria and increase in the cytoplasmic compartment (Fig. 7a and b). Elevation of cytoplasmic acetyl-CoA shifted the equilibrium point of ChAT reaction towards an increase in the intracellular pool of ACh (Fig. 6a). Increase of ChAT activity and cytoplasmic acetyl-CoA could speed up the rate of ACh re-synthesis after its facilitated release (Figs 4, 6b and 7b). A similar mechanism explains the co-existence of low levels of acetyl-CoA and an increased vulnerability of SN56 cells over-expressing ChAT (Bielarczyk et al. 2003b).
Accordingly, the increase of mitochondrial acetyl-CoA in DC treated with NGF could result from its decreased utilisation for transmitter synthesis in cell cytoplasm (Figs 6a and 7a). Despite of that vulnerability NGF-treated DCs to NO and Aβ was greater than the not treated ones as demonstrated by decrease of ChAT activity and cell viability (Figs 9 and 10, Table 1). This inconsistency may be explained by either the stimulation by NGFp75NTR receptor-activated suppression pathways and/or by accumulation of additional amounts of Ca in mitochondria (Figs 5 and 7a) (Dawson and Dawson 1995; Eckert et al. 2003). It is known that moderate accumulation of Ca in mitochondria, after a single addition of cAMP/RA or NGF, may propagate signals increasing energy production and thereby stimulating cell differentiation (Figs 2, 4 and 5). In contrast, excessive Ca load, demonstrated here after joint application of cAMP/RA and NGF could induce inhibition of pyruvate utilization, energy production and release of proapoptotic factors from mitochondria (Figs 5 and 8) (Lai et al. 1988; Hajnoczky et al. 2003). Therefore, under these conditions, stimulation of ACh release, inhibition of aconitase and oxidative phosphorylation by NO, could markedly facilitate the damage to DCs (Figs 6, 7 and 10) (Bielarczyk et al. 2003a; Hajnoczky et al. 2003). The increase of trypan blue positive cells by NGF and the decrease of ChAT activity in T17TrkA+ DC remain in accord with such mechanism (Figs 9 and 10).
Data presented herein indicate that expression of the cholinergic phenotype in brain septal neuroblastoma cells may be activated through independent cAMP/RA and TrkA/p75NTR signalling pathways. Differentiation by cAMP/RA inhibited TrkA-mediated up-regulation and activated p75NTR-dependent suppressing effects of NGF on cholinergic cells. Increased Ca accumulation and decreased mitochondrial levels of acetyl-CoA in DC play an important role in phenotype-dependent variable susceptibility of septal cholinergic neurons to neurotoxic stimuli. Thus, variable interactions between these three signal transduction pathways may form the mechanism of differential expression of cholinergic phenotype and susceptibility of septal neurons to neurodegenerative inputs.
Work was supported by Ministry of Science and Informatization Republic of Poland, 6P05A 010 20 and Medical University of Gdañsk project St-57. The authors are indebted to Dr Krzysztof J. Blusztajn for SN56 and T17 cells. Phospholine iodide was a generous gift from Wyeth (Momouth Junction, NJ, USA).