Address correspondence and reprint requests to Dr. J. N. Barrett at Department of Physiology and Biophysics, R-430, University of Miami School of Medicine, P.O. Box 016430, Miami, FL 33101, U.S.A. E-mail: email@example.com
Abstract: We demonstrate that brief (30-min) exposure of cultured embryonic rat septal neurons to neurotrophins (NTs) increases choline acetyltransferase (ChAT) activity by 20-50% for all tested NTs (nerve growth factor, brain-derived neurotrophic factor, neurotrophin-3, and neurotrophin-4, each at 100 ng/ml). The increase in ChAT activity was first detected 12 h after NT exposure, persisted at least 48 h, and was not mediated by increased neuronal survival or action-potential activity. Under some conditions, the response to brief NT exposure was as great as that produced by continuous exposure. NT stimulation of ChAT activity was inhibited by inhibitors of p75- and Trk kinase-mediated signaling, by removal of extracellular Ca2+ during the period of NT exposure, and by buffering intracellular Ca2+ with BAPTA. Application of nerve growth factor and brain-derived neurotrophic factor transiently increased [Ca2+] within a subpopulation of neurons. NT stimulation of ChAT activity was not affected significantly by cyclic AMP agonists or antagonists. These findings suggest that brief exposure to NTs can have a long-lasting effect on cholinergic transmission, and that this effect requires Ca2+, but not cyclic AMP.
Embryonic basal forebrain cholinergic neurons express both Trk and p75NTR receptors (Dawbarn et al., 1988; Batchelor et al., 1989; Holtzman et al., 1992; Steininger et al., 1993) and respond to all identified neurotrophins (NTs). Exposure of cultured septal neurons to nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), neurotrophin-3 (NT-3), or neurotrophin-4 (NT-4) for 1-3 weeks increases the activity of acetyl-CoA:choline-O-acetyltransferase (ChAT; EC 220.127.116.11) by two- to threefold (for review, see Whittemore and Seiger, 1987; Alderson et al., 1990; Friedman et al., 1993; Li et al., 1995; Nonner et al., 1996). This study focuses on the effects of briefer NT exposures, for several reasons. First, brief exposure to NTs appears sufficient to produce both short- and long-term signaling events. For example, Knüsel et al. (1992) demonstrated that a 4-min exposure of embryonic rat basal forebrain cultures to NGF induces tyrosine phosphorylation of Trk family proteins, and Finkbeiner et al. (1997) showed that BDNF increases tyrosine phosphorylation of phospholipase Cγ in rat cortical neurons within 5 min. A longer-term effect is the demonstration of Toledo-Aral et al. (1995) that a 1-min exposure to NGF induces expression of the PN-1 type of Na+-channel protein in PC12 cells.
Hippocampal and cerebral cortical neurons, which are innervated by basal forebrain cholinergic neurons, rapidly increase NGF and/or BDNF release in response to depolarizing agents (Blöchl and Thoenen, 1995, 1996; Androutsellis-Theotokis et al., 1996; Goodman et al., 1996). Thus, knowledge of how basal forebrain cholinergic neurons respond to brief NT applications is potentially important for understanding the response of these neurons to phasic NT release from their target tissues. Comparison of the effects of brief and continuous NT application also has potential application for determining optimal patterns of in vivo NT administration.
Brief NT applications offer the possibility of probing the pathways by which NTs increase ChAT activity in central cholinergic neurons, because cultured neurons may tolerate brief exposures to inhibitors of second messenger pathways better than prolonged exposures. Intracellular second messengers/pathways tested here included Ca2+, Ca2+/calmodulin-dependent protein kinase (CaMK), and signaling cascades including mitogen-activated protein kinase kinase (MEK), all implicated in NT-mediated signaling in neurons (Finkbeiner and Greenberg, 1996; Courtney et al., 1997; Finkbeiner et al., 1997). We also tested the involvement of cyclic AMP (cAMP) and protein kinase A (PKA), because the promoter region of the ChAT gene includes a cAMP response element (Toussaint et al., 1992), and cAMP enhances some actions of NTs (e.g., Meyer-Franke et al., 1998; Boulanger and Poo, 1999).
We report that a 30-min exposure of embryonic basal forebrain neurons to any of four tested NTs produced a long-lasting increase in ChAT activity. For NGF and NT-3, the response to repeated brief NT exposures was as great as that produced by continuous exposure. The response to brief NT application was reduced by a blocking antibody against p75NTR receptors, by an inhibitor of Trk kinase activity, and by reduction of bath [Ca2+]. However, the response was not blocked by inhibitors of PKA, CaMKII, or MEK.
MATERIALS AND METHODS
Cell culture and NT exposure
The septal region was dissected from embryos (day 15 of gestation) obtained from anesthetized Sprague-Dawley rats (Charles River Laboratories, Wilmington, MA, U.S.A.) and dissociated by gentle trituration. The experimental protocols were approved by the University of Miami Animal Care and Use Committee and meet National Institutes of Health guidelines. Dissociated cells were plated at a density of 1,200-1,600 cells/mm2 into 96-well culture plates coated with poly-L-lysine in a defined culture medium [N5 (Kawamoto and Barrett, 1986)] supplemented with an acid-stable fraction of horse serum [5% (Kaufman and Barrett, 1983)] that contains seleno-protein-P (Yan and Barrett, 1998). This medium exhibits no NT bioactivity. Under these conditions, cultures remain neuronrich [>90% neurons as assessed by morphology and staining with anti-neurofilament antibody (Kaufman and Barrett, 1983; Tedeschi et al., 1986)], and neuronal survival (including survival of the cholinergic subpopulation) is independent of exogenous NTs in the absence of applied stresses (Nonner et al., 1996).
In most experiments, one of the four tested recombinant NTs was added to the culture wells after 7-10 days in vitro. rHu-NGF, rHu-BDNF, rHu-NT-3, and rHu-NT-4 were all generously supplied by Regeneron Pharmaceuticals (Tarrytown, NY, U.S.A.), courtesy of Drs. Ron Lindsay and George Yancopoulos. Unless otherwise noted, the NT concentration was 100 ng/ml (equivalent to ≤7.4 nM monomer, ≤3.7 nM dimer); previous work using longer NT exposures revealed no inhibitory effects on ChAT activity at this concentration, even when all four NTs were applied together (Nonner et al., 1996). After the indicated NT exposure time (usually 30 min), cells were washed three times with NT-free culture medium (∼5-min interval between washes) and maintained in this medium until assay. Control cultures never exposed to exogenous NTs received the same number and timing of washes as experimental cultures. Unless otherwise specified, tested drugs were added 15-30 min before NT addition and were removed by the same three washes used to remove the NTs.
In a control experiment using NGF covalently labeled with a fluorescent dye (Alexia 594, using protocols from Molecular Probes, Eugene, OR, U.S.A.), we found that three washes with control medium reduced the level of dye-labeled NGF to 0.03% of its original level in septal cultures. This result suggests that the triple wash used to terminate NT exposure reduced medium [NT] from the typical test concentration of 100 ng/ml to only 0.03 ng/ml. However, even after NT was washed out of the medium, cells might continue to be exposed to NT bound to receptors or internalized into the cell.
Assays for ChAT activity and cell viability
ChAT activity was assayed using the method of Fonnum (1969), usually 24 h after NT addition. Just before measurement of ChAT activity, the effects of tested drugs on cell viability were assayed by visual inspection and by measuring Alamar Blue reducing activity (protocols and reagents from AccuMed International Companies, Westlake, OH, U.S.A.). Drugs were applied at concentrations that had no effect on cell viability assayed 24 h later.
Measurements of intracellular Ca2+ concentration ([Ca2+]i) and cAMP
To test for NT-induced increases in [Ca2+]i, septal cultures were loaded with fura-2 acetoxymethyl ester (AM; 5 μM) for 30 min and then washed with normal medium for 60 min before imaging. Fura-2 fluorescence was measured using a cooled charge-coupled device camera (CH-250, Photometrics, Tucson, AZ, U.S.A.) with an imaging system described by David et al. (1997). Emissions excited by illumination at 340 and 380 nm were corrected for background and then ratioed. In some experiments, the 340/380 ratio (averaged over the soma) was calculated in single neurons as a function of time after NT addition (see Fig. 6). To determine the percentage of neurons whose [Ca2+]i increased in response to NT addition, we used a macro program written in PMIS (Photometrics) to calculate background-subtracted 340/380 ratios for all neurons in a microscope field. Neurons were identified as cells with processes and rounded somata. The criterion for a NT-responsive neuron was an increase in the 340/380 ratio of at least 30% over control beginning within 5 min after addition of NT (100 ng/ml) and lasting at least 5 min. Calibrations performed in ionomycin-permeabilized cells exposed to different Ca and BAPTA [1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid] mixtures suggested that a 30% increase in the 340/380 ratio corresponded to a [Ca2+]i increase of ∼50 nM.
Changes in [cAMP] were assayed using a TiterFluor Dual Range cAMP EIA kit with a fluorescence multiwell plate reader (both from PerSeptive Biosystems, Inc., Framingham, MA, U.S.A.). After a 30-min exposure to NTs or forskolin (5 μM), the medium was replaced with sodium acetate buffer, and the company's protocol for the acetylated enzyme immunoassay was followed. The fluorescence of standard and test wells was measured using excitation and emission wavelengths of 450 and 590 nm, respectively.
Tested drugs and reagents were purchased from the following companies: anti-NGF antibody, Boehringer Mannheim (Indianapolis, IN, U.S.A.); BAPTA-AM and fura-2-AM, Molecular Probes; Sp-adenosine-3′,5′-cyclic monophosphothioate, Rp-adenosine-3′,5′-cyclic monophosphothioate, K252a, and K252b, Calbiochem (La Jolla, CA, U.S.A.); ω-conotoxins GVIA and MVIIC, Alomone Labs (Jerusalem, Israel); nimodipine, 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), DL(-)-2-amino-5-phosphonovaleric acid (APV), KN-93, and KN-62, Research Biochemicals International (Natick, MA, U.S.A.); PD098059, New England Biolabs (Beverly, MA, U.S.A.). Ab9651, an anti-p75NTR antiserum directed against the extracellular domain of the receptor, was the generous gift of Dr. Moses Chao (New York University Medical Center, New York, NY, U.S.A.). This antiserum blocks NGF binding to p75NTR (Huber and Chao, 1995). All other chemicals came from Sigma (St. Louis, MO, U.S.A.).
NT effects on ChAT activity were measured by comparing ChAT activities in cultures exposed to NT with those of sister cultures that received the same number of washes but were maintained throughout in control medium (NT/control in Table 3). Drug effects on basal ChAT activity were measured similarly (Drug/control in Table 3). Drug effects on NT stimulation of ChAT activity were measured by comparing ChAT activities obtained from sister cultures exposed to drug plus NT or to drug alone [(NT + drug)/drug in Table 3]. This ratio was compared with the NT/control ratio measured in sister cultures not exposed to the drug. For the drugs in Table 3, the results with all four NTs (tested individually at 100 ng/ml) were not statistically different from each other, and thus were pooled together.
Table 1. Drug effects on basal and NT-stimulated ChAT activity
(NT + drug)/control
(NT + drug)/drug
Drugs were applied 15-30 min before NT application. NTs were then added for 30 min and washed out together with the tested drug. ChAT activities were assayed 24 h after the onset of NT exposure, except as noted below. Columns labeled “Drug/control,”“(NT + drug)/control,” and “NT/control” give, respectively, the ratio of ChAT activities measured following exposure to the indicated drug, to drug + NT, or to NT alone, divided by ChAT activities measured in sister cultures washed the same number of times but exposed throughout to control medium only. The column labeled “(NT + drug)/drug” indicates the extent to which NTs were able to stimulate ChAT activity in the presence of the drug; this ratio was calculated from ChAT activities measured following NT + drug and drug treatments (not as the ratio of the first two data columns). In most cases, NGF, BDNF, NT-3, and NT-4 (each at 100 ng/ml) were tested individually with each drug. Results were similar for all NTs, so values for all NTs tested with a given drug were averaged together. Numbers indicate means ± SEM of ratios for the number of culture wells indicated in parentheses. Values in a given row came from sister cultures from one or more platings (n values exceeding 16 in the drug/control column indicate multiple platings). Data from a given plating were included in the average only if NT addition to control medium produced an increase in ChAT activity that was ≥20% (i.e., ratio ≥1.2) and/or reached statistical significance at p < 0.05 or better. Tested drug concentrations fall within the range reported to be effective in other systems. None of the tested drugs produced detectable changes in neuronal viability at the indicated concentrations following 45-60-min exposure times; viability was assessed by morphological inspection and/or by assaying Alamar Blue reducing ability just prior to assaying ChAT activity. Sp-cyclic AMPS, Sp-adenosine-3′,5′-cyclic monophosphothioate; Rp-cyclic AMPS, Rp-adenosine-3′,5′-cyclic monophosphothioate.
aSignificant difference from control. This statistic was not determined for the (NT + drug)/control column, because the appropriate control for NT + drug is Drug only.
bSignificant difference from Drug only.
cIn these experiments, which shared a common control group, NT was a combination of NGF and BDNF, and ChAT activity was assayed 48 h after the brief exposure.
Pairwise comparisons used Student's t test. Comparison of multiple experimental treatments with a common control used Dunnett's test. p≤ 0.05 was considered significant. Unless otherwise noted, the n values reported in text and tables indicate the number of culture wells assayed to produce a given measured value.
Increases in ChAT activity produced by brief NT exposures
Figure 1A shows the ChAT activities of sister septal cultures exposed to NGF, BDNF, NT-3, or NT-4 for 15-120 min and assayed 24 h later. For exposure times ≥30 min, all four NTs significantly increased ChAT activity. At this NT concentration (100 ng/ml), the increase was similar for all tested NTs, 25-50% greater than control. ChAT activities following 30-, 60-, and 120-min NT exposures were not significantly different from each other. Most experiments used 30-min NT exposures on cultures maintained 7 days in vitro. Brief NT exposures produced the same percentage increase in ChAT activity in cultures ranging from 3 to 14 days in vitro (data not shown). Thirty-minute exposures to pair-wise combinations of the various NTs, as well as the combination of all four tested NTs, produced no greater enhancement than a single NT (each NT at 100 ng/ml; data not shown).
In other experiments, ChAT activity was assayed at times ranging from 3 to 48 h after a 30-min NT exposure. Significant increases were first detected 12 h after NT addition and persisted undiminished for at least 48 h (Fig. 1B). Unless otherwise specified, a 24-h interval between NT exposure and bioassay (as in Fig. 1A) was used in subsequent experiments.
Was the increase in ChAT activity due mainly to brief exposure to a high NT concentration or to prolonged exposure to residual NT remaining in the medium after NT washout (∼0.03 ng/ml; see Materials and Methods) or bound to surfaces in the culture? Observations suggesting the importance of the brief exposure to high [NT] include the finding that NT exposures of <30 min did not significantly elevate ChAT activity (Fig. 1A), and that a 24-h exposure to 0.05 ng/ml NGF or BDNF did not increase ChAT activity significantly (data not shown). Figure 2 shows results of experiments that addressed this question by comparing the effects of simultaneous versus delayed addition of a blocking anti-NGF antibody. In Fig. 2A, antibody was either washed out simultaneously with a 30-min NT application or applied 4 h after the NT washout, remaining on the cultures until assay at 24 h. In Fig. 2B, antibody was added either simultaneously with the NGF or 1 or 4 h afterwards, and both NGF and antibody remained on the cultures until assay at 24 h. The results show that when applied simultaneously with NT, this antibody consistently inhibited the increase in ChAT activity produced by NGF (Fig. 2A and B), but not that produced by other NTs (Fig. 2A), consistent with previous studies demonstrating this antibody's specificity for NGF (Rodríguez-Tébar et al., 1990). The antibody usually had little or no effect on ChAT activity when applied 1-4 h after NGF washout (Fig. 2A and B). These results suggest that a brief exposure to 100 ng/ml NGF was sufficient to increase ChAT activity, with no need for continued signaling from residual NGF accessible to the medium (see Discussion).
Figure 3 plots the concentration dependence of ChAT activity in sister cultures exposed to NTs for 30 min (A) or 24 h (B) and assayed 24 h after the onset of NT exposure. For the 30-min exposure, significant increases in ChAT activity were measured for BDNF and NT-4 concentrations as low as 1 ng/ml; for all tested NTs, ChAT activity increased as NT concentration increased from 1 to 100 ng/ml. With the longer 24-h exposure, maximal responses were achieved at 1 or 10 ng/ml. A striking feature of Fig. 3 is that the percent increase in ChAT activity produced by the highest NT concentration (100 ng/ml) was similar for the 30-min and 24-h exposures.
Figure 4 shows results of an experiment testing whether the similar effects of brief and continuous exposures to high [NT] noted in Fig. 3 might also apply to treatments extended over several days. Sister cultures were exposed to the indicated NT (100 ng/ml) for either 30 min daily for 5 days (open columns) or continuously for 7 days (hatched columns), with all cultures assayed after 7 days. Continuous-exposure cultures were washed as frequently as brief-exposure cultures; this is important, because washing often results in some neuronal loss. Repeated brief NT exposures produced two- to fourfold increases in ChAT activity. For NGF and NT-3 brief and continuous exposures were equally effective, but for BDNF and NT-4 continuous exposure was more effective. The greater increase in ChAT activity produced by continuous exposure to BDNF and NT-4 may have been due to a stress-protective effect of these NTs. Microscopic examination showed that these frequently washed cultures contained fewer cells, and ChAT activities were lower than those measured in other experiments. Nonner et al. (1996) found that BDNF and NT-4 (but not NGF or NT-3) significantly increased both total neuronal survival and ChAT activity in low-density cultures.
Evidence for involvement of both Trk and p75 receptors
K252a and K252b inhibit the tyrosine kinase activity of Trk receptors (for review, see Knüsel and Hefti, 1992; Berg et al., 1992; Muroya et al., 1992; Nye et al., 1992; Tapley et al., 1992). Table 1 shows that when K252a was applied just before and during NGF exposure at the frequently used concentration of 200 nM, the NGF-induced increase in ChAT activity was not reduced. However, K252a did inhibit this response when allowed to remain in contact with cells until the assay at 24 h, even when the initial application of K252a was delayed until 1-4 h after NGF washout. Table 3 shows that exposure to a higher concentration of K252a (2 μM), or to a high concentration of K252b (5 μM), before and during NT exposure did block the increase in ChAT activity, but at these high concentrations the K252 drugs might have additional nonspecific inhibitory effects. Results with all four tested NTs (each at 100 ng/ml) were similar and are averaged together in Table 3. Assays for cell viability performed 24 h after drug exposure indicated that the inhibitory effects of the K252 compounds were not due to cell death.
Table 2. Effect of Trk kinase inhibitor on NT stimulation of ChAT activity
ChAT activity (ratio to control)
In the brief-exposure groups, K252a was added 15 min before NGF addition; after a 30-min exposure to NGF, both NGF and K252a were washed out together. In other groups, K252a was added 1 or 4 h after the NGF exposure and was not washed out (thus, these cultures were exposed to K252a for ≥20 h). All cultures received the same number of washes. ChAT activity was measured 24 h after NGF exposure and is normalized to that measured in sister control cultures not exposed to NGF or K252a (0.59 ± 0.01 pmol/min/culture, n = 16). Values are means ± SEM for the number of culture wells indicated in parentheses.
Table 2 shows that an antiserum against the p75NTR receptor also blocked the increase in ChAT activity produced by brief exposure to both moderate (10 ng/ml) and high (100 ng/ml) concentrations of NGF or BDNF.
Table 3. Antiserum against p75NTR inhibits NT stimulation of ChAT activity
ChAT activity (treatment/control)
NT + anti-p75NTR
Normalized ChAT activities of sister cultures exposed to the indicated NT concentrations in the presence or absence of Ab9651 (75 μg/ml) are presented. Antiserum was added 1 h before a 30-min NT exposure and was washed out with the NTs. Values are means ± SEM normalized to sister cultures not exposed to NTs, with the number of culture wells indicated in parentheses. Each BDNF + antiserum value was significantly different from BDNF alone, and averaged NGF + antiserum values were significantly different from averaged NGF alone. Antibody alone had no significant effect on the ChAT activity measured in control medium.
Evidence that NT stimulation of ChAT activity is Ca2+-dependent
Figure 5A shows that the stimulation of ChAT activity following brief NT exposure was reduced when extracellular [Ca2+] was lowered below the normal 2 mM during NT exposure. Each plotted point was normalized to the ChAT activity measured in sister cultures that were exposed to the same changes in medium [Ca2+], but without addition of NT. Elevating extracellular [Ca2+] to 4 mM during the period of NT exposure did not increase ChAT activity above that measured with the normal 2 mM Ca2+ (data not shown).
Figure 5B shows that stimulation of ChAT activity following brief NT exposure was also reduced when cells were loaded with BAPTA (a Ca2+ buffer) by a 30-min exposure to 25 μM BAPTA-AM before, during, or after NT exposure. Each column is normalized to the ChAT activity measured in sister cultures that were exposed to BAPTA-AM at the same time, but with no addition of NT. NT stimulation of ChAT activity failed to reach significance in any BAPTA-loaded cells. The greatest inhibition occurred in cultures loaded with BAPTA before NTs were added, with less inhibition in cultures exposed to BAPTA at later times (see legend). In Fig. 5, results with all four tested NTs were similar and thus were averaged together.
NT-induced increase in [Ca2+]i
Figure 6 shows examples of transient increases in [Ca2+]i produced in fura-2-loaded septal neurons by applying NGF followed by BDNF (A) or BDNF followed by NGF (B). In those neurons in which the NT-induced increase in [Ca2+]i reached the criteria described in Materials and Methods (≥50 nM increase sustained for ≥5 min), the responses to NGF and BDNF were similar in magnitude and time course. The peak increase in [Ca2+]i, averaged over the soma, was 142 ± 24 nM (mean ± SEM) for NGF and 157 ± 28 nM for BDNF (n = 10 for each), with 50% of the peak response attained by 4.8 min after NGF addition and by 5.3 min after BDNF application (n = 2 for each). (These measurements of average [Ca2+]i may underestimate the maximal increase in [Ca2+] achieved near sites of Ca2+ release/entry.) [Ca2+]i remained at ≥50% of its peak value for an average of 17 ± 2.7 min for NGF and 25 ± 3.8 min for BDNF (n = 6 for each). The percentages of neurons responding to NGF and BDNF were similar: 6.1 ± 0.8% (17 microscope fields) for NGF and 7.6 ± 2.1% (seven fields) for BDNF.
Most tested NGF-responsive neurons (13 of 18, or 72%) also increased [Ca2+]i in response to BDNF (as in Fig. 6), as well as in response to neurotensin (1 μM, 18 of 24, or 75%). The Discussion summarizes evidence that these NT- and neurotensin-responsive neurons were cholinergic. Some neurons also increased [Ca2+]i in response to NT-3, but responses to NT-3 and NT-4 were not studied systematically.
When extracellular [Ca2+] was reduced to 1 mM, the percentage of neurons exhibiting a criterion response to NT application was not reduced (5% responding to NGF, 8% responding to BDNF), but the peak increase in [Ca2+]i was reduced to about half that measured in normal [Ca2+] (58 ± 46 nM for NGF, 72 ± 19 nM for BDNF, n ≥ 20 for both; significantly different from the peak values recorded in 2 mM Ca2+, p < 0.01).
We attempted to determine whether the Ca2+ dependence of NT effects involved Ca2+ entry across the plasma membrane and/or release of intracellular stores. Cd2+ (100 μM), which inhibits many routes of Ca2+ entry across the plasma membrane, inhibited the NT-induced increase in ChAT activity (Table 3; Cd2+ also reduced basal ChAT activity). Also, when 1 mM Mn2+ was present in the medium, addition of NTs to fura-2-filled cells reduced emissions at both exciting wavelengths (340 and 380 nm), consistent with quenching of fura-2 fluorescence by Mn2+ entry via NT-activated Ca2+ channels (Merritt et al., 1989). Thapsigargin (1 mM), which inhibits endoplasmic reticulum Ca2+-ATPase and thus reduces the amount of Ca2+ available for release from intracellular stores, had no significant effect on basal or NT-stimulated ChAT activities (Table 3).
We also tested the effects of drugs targeting certain depolarization-activated routes of Ca2+ entry. Addition of a combination of 1 μM tetrodotoxin to block voltage-dependent Na+ channels and 10 μM nifedipine to block L-type Ca2+ channels did not reduce the percentage of neurons exhibiting an NGF-induced increase in [Ca2+]i (7.7% of neurons in two experiments). A cocktail of inhibitors of depolarization-activated Ca2+ channels (nimodipine, ω-conotoxin GVIA, and ω-conotoxin MVIIC, each at 5 μM, to block L-, N-, and P/Q-type channels, respectively) likewise failed to reduce the percentage of responding neurons or the magnitude of the increase in [Ca2+]i (data not shown). Consistent with these results, tetrodotoxin, nimodipine, ω-conotoxin GVIA, and ω-conotoxin MVIIC did not inhibit NT stimulation of ChAT activity [NT/control and (NT + drug)/drug ratios in Table 3 were 1.23 and 1.21 for tetrodotoxin, 1.33 and 1.30 for nimodipine, and 1.70 and 1.65 for the conotoxin combination, respectively]. Dihydropyridine receptor antagonists, such as nimodipine, bind best to inactivated L-type channels, but nimodipine also failed to block in cultures depolarized with 30-50 mM KCl (Table 3). The NT-induced increase in ChAT activity also persisted when Ca2+ entry via kainate/α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid-type glutamate receptors was blocked with CNQX (0.3 μM), and when Ca2+ entry via NMDA-type glutamate receptors was blocked with APV (50 μM; Table 3). The effects of the drugs listed in Table 3 did not depend on which NT was used, so the reported ChAT activity ratios were averaged over all tested NTs.
Treatments that simply elevated [Ca2+]i did not increase ChAT activity. Depolarizing cells with 30-50 mM extracellular K+ had no significant effect on basal or NT-stimulated ChAT activity (Table 3). Also, a 30-min exposure to the Ca2+ ionophores ionomycin (100-200 nM) or A23187 (50-100 nM) yielded ChAT activities averaging 80-91% of control when assayed 24 h later.
Stimulation of ChAT activity by brief NT exposure is not blocked by inhibitors of PKA, CaMK, or MEK
Table 3 shows that NT stimulation of ChAT activity was not altered significantly by either the cAMP antagonist Rp-adenosine3′,5′-cyclic monophosphothioate nor the PKA activator Sp-adenosine-3′,5′-cyclic monophosphothioate, [each at 100 μM; NT/control and (NT + drug)/drug ratios were 1.24 and 1.26 for Sp-adenosine-3′,5′-cyclic monophosphothioate and 1.27 and 1.26 for Rp-adenosine-3′,5′-cyclic monophosphothioate, respectively]. Another PKA activator, forskolin (5 μM), was similarly ineffective (data not shown). Enzyme immunoassays verified that forskolin increased [cAMP] in septal cultures, but NT application produced no detectable change in cAMP levels: cAMP measured immediately following a 30-min exposure to NT (pooled results of four tested NTs) or forskolin was 48 ± 4.1 and 1,100 ± 100 fmol/culture, respectively, compared with 54 ± 18 in sister controls (n = 8-31 culture wells).
PD098059 blocks activation of MEK1 by Raf (Dudley et al., 1995) and reduces NGF-induced differentiation in PC12 cells (Alessi et al., 1995; Pang et al., 1995), as well as BDNF-induced phosphorylation of extracellular signal-regulated kinase 2 in cortical neurons (Finkbeiner et al., 1997). Table 3 shows that PD098059 (10-25 μM) increased basal ChAT activity, but did not block NT stimulation of ChAT activity [NT/control and (NT + drug)/drug ratios of 1.20 and 1.15, respectively]. Inhibitors of CaMKII and CaMKIV [KN-62 and KN-93, 10 μM (Tokumitsu et al., 1990; Enslen et al., 1994)] decreased basal ChAT activity, but also had no effect on NT stimulation of ChAT activity [NT/control and (NT + drug)/drug ratios of 1.28 and 1.26 for KN-62, 1.27 and 1.23 for KN-93, respectively]. Likewise, a combination of 10 μM KN-62 and 20 μM PD098059, which would be expected to inhibit most BDNF-induced activation of fos transcription in cortical neurons (Finkbeiner et al., 1997), failed to block NT stimulation of ChAT activity [NT/control and (NT + drug)/drug ratios of 1.18 and 1.17, respectively]. This drug combination was also ineffective when present during the entire 24-h interval between NT addition and ChAT assay (data not shown). These results suggest that the increase in ChAT activity produced by brief NT exposures is not mediated by cAMP and does not require activity of MEK1, CaMKII, or CaMKIV.
Brief NT exposures have long-lasting effects on ChAT activity
This study demonstrates that a brief (30-min) exposure to 100 ng/ml NGF, BDNF, NT-3, or NT-4 produces a 20-50% increase in ChAT activity assayed 24 h later. The magnitude of this increase is similar to that produced by a 24-h continuous exposure to these NTs. For BDNF and NT-4, brief exposure to concentrations as low as 1 ng/ml significantly increased ChAT activity. We did not determine whether the increase in ChAT activity was due to an increase in ChAT protein and/or to posttranslational modification of enzyme activity, but the prolonged onset and duration of the increase in ChAT activity provide sufficient time for an alteration in gene expression. Pongrac and Rylett (1998) found that a 2-day exposure to NGF increased by 50% the steady-state level of ChAT transcript in cultured rat basal forebrain, and that a 4-day exposure produced comparable increases in ChAT protein and ChAT activity (100 and 70%, respectively). An increase in ChAT gene expression, possibly combined with stimulation of the embedded vesicular acetylcholine (ACh) transporter gene (Bejanin et al., 1994; Erickson et al., 1994), would be expected to increase ACh synthesis and release and ACh-mediated synaptic responses (Moises et al., 1995).
High concentrations of the K252 inhibitors of Trk tyrosine kinase activity were required to block NT stimulation of ChAT activity in experiments where the inhibitor was washed out at the same time as the NT, but the standard 200 nM concentration was sufficient when the inhibitor was added 1-4 h after NT washout and allowed to remain on the cultures until assay at 24 h. This pattern of inhibition suggests a requirement for tyrosine kinase activity persisting following NT washout, mediated perhaps by persisting NT binding to Trk receptors on the plasma membrane (Ure and Campenot, 1997) and/or by internalized activated Trk receptors.
NT stimulation of ChAT activity is Ca2+-dependent, and NTs increase [Ca2+] within cholinergic neurons
Our data indicate that the increase in ChAT activity produced by brief NT application required elevation of [Ca2+]i, and that NGF and BDNF elevated [Ca2+]i in overlapping subsets of basal forebrain neurons. Several lines of evidence suggest that most of the neurons that showed NGF-induced [Ca2+]i elevations were cholinergic. First, the percentage of neurons that exhibited these increases in [Ca2+]i was similar to the percentage of cholinergic neurons identified by histochemical staining for acetylcholinesterase in these cultures [∼5-10% (Nonner et al., 1993)]. Second, in basal forebrain cultures, both types of NGF-responsive receptor (TrkA and p75NTR) are colocalized within, and largely confined to, the cholinergic subpopulation (Dawbarn et al., 1988; Batchelor et al., 1989; Holtzman et al., 1992; Steininger et al., 1993; Gibbs and Pfaff, 1994; Sobreviela et al., 1994). Third, most tested NGF-responsive neurons also increased [Ca2+]i in response to neurotensin, consistent with demonstrations that in the basal forebrain neurotensin receptors are localized mainly on cholinergic neurons (Szigethy et al., 1990), and that neurotensin application increases their Ca2+ conductance (Alonso et al., 1994).
As there was extensive overlap between the neurons that exhibited NGF- and BDNF-induced increases in [Ca2+]i, and as basal forebrain cholinergic neurons also express BDNF-responsive TrkB receptors, it is likely that most neurons that showed BDNF-induced increases in [Ca2+]i were cholinergic. We were surprised that the percentage of BDNF-responsive neurons did not exceed the percentage of NGF-responsive neurons, because more basal forebrain neurons are immunoreactive for TrkB receptors than for TrkA receptors (Merlio et al., 1992; Cheng and Mattson, 1994). Also, TrkB-immunoreactive neurons cultured from other brain regions exhibit BDNF-induced increases in [Ca2+]i [hippocampal neurons (Berninger et al., 1993; Thoenen, 1995; Marsh and Palfrey, 1996); spinal motoneurons (Stoop and Poo, 1996); cortical neurons (Finkbeiner et al., 1997)]. Perhaps in the septum, NT-induced increases in [Ca2+]i as large and long-lasting as those measured here require some feature restricted to cholinergic neurons, such as p75NTR. Our finding that most of the neurons that showed NGF-induced elevations of [Ca2+]i also responded to BDNF, but that some neurons responded only to NGF or only to BDNF agrees with conclusions drawn from measurements of ChAT activity by Nonomura et al. (1995).
Consistent with the fact that basal forebrain cholinergic neurons express multiple NT receptors, several of our findings suggest that at least some of the effects of brief NT application were mediated by direct (rather than indirect) effects of NTs on cholinergic neurons. The increase in [Ca2+]i in NT-responsive neurons was rapid, reminiscent of rapid, Ca2+-dependent effects of NTs described in other neurons (Gundersen and Barrett, 1979, 1980; Lohof et al., 1993; Knipper et al., 1994a,b; Lessmann et al., 1994; Kang and Schuman, 1995; Levine et al., 1995; Berninger and Poo, 1996; Tanaka et al., 1997). Also, although some effects of NTs are activity-dependent (e.g., McAllister et al., 1996), neither the elevation in [Ca2+]i nor the increase in ChAT activity was inhibited by suppression of action-potential activity or by blockage of L-, N-, P-, or Q-type Ca2+ channels. Furthermore, the increase in ChAT activity was not blocked by drugs that would inhibit Ca2+ entry via ionotropic glutamate receptors. Our findings are thus consistent with the hypothesis that NTs are downstream mediators of the effects of neuronal activity on ChAT activity.
Elevation of [Ca2+]i by other means (high [K+], ionomycin, A23187) did not elevate ChAT activity, suggesting that the route and/or kinetics of the elevation of [Ca2+]i (and/or some additional NT-stimulated event) are important for the NT-induced increase in ChAT activity. The measurements of Marsh and Palfrey (1996) in hippocampal pyramidal neurons suggested that BDNF increased both Ca2+ influx across the plasma membrane and release from internal stores. Our data indicate that the NT-induced elevation of [Ca2+]i in basal forebrain cholinergic neurons probably does not occur via release of a thapsigargin-sensitive internal store. We have not yet identified route(s) of Ca2+ entry across the plasma membrane, but several possibilities have been eliminated (see above).
Elevations of [Ca2+]i and/or [cAMP] can result in activation of a cAMP-response element-binding protein (CREB) implicated in some NT-mediated effects (e.g., Finkbeiner et al., 1997), but several lines of evidence demonstrated that the stimulation of ChAT activity by brief NT exposures was independent of cAMP. These findings differ from those reported for a septal cell line, where cAMP increased expression of the ChAT gene (Berse and Blusztajn, 1995), but are consistent with the finding of Quirin-Stricker et al. (1997) in neuronal cell lines that some NGF-activated transcription factors that activate the ChAT gene promoter are cAMP-independent. A combination of drugs that inhibits most BDNF-induced phosphorylation of CREB in cortical neurons also failed to block NT stimulation of ChAT activity. Thus, further work will be needed to identify the Ca2+-dependent intracellular signaling pathways that mediate the increase in ChAT activity following brief NT exposures. There may also be differences in the pathways by which brief versus prolonged NT exposures increase ChAT activity. In PC12 cells, both the type of Na+ channel induced and the second messenger pathways involved are different for brief (minutes) versus prolonged NGF exposures (D'Arcangelo et al., 1993).
In summary, this work demonstrated that brief (30-min) exposure of embryonic rat septal neurons to exogenous NTs produces a persisting increase in ChAT activity whose magnitude is similar for all four tested NTs (NGF, BDNF, NT-3, and NT-4). We have presented evidence that this effect involves both p75NTR receptors and prolonged Trk kinase activity and is Ca2+-dependent. These findings suggest that depolarization-evoked bursts of NT release from hippocampal neurons may produce long-lasting increases in ChAT activity in septal cholinergic neurons.
We thank Regeneron Pharmaceuticals (Drs. Ron Lindsay and George Yancopoulos) for providing recombinant neurotrophins, and Dr. Moses Chao for providing the anti-p75 antiserum. This work was supported by the National Institutes of Health (NS 12207).