pontine cholinergic neurons depend on three neuroprotection systems to resist nitrosative stress
TRANSCRIPT
www.elsevier.com/locate/brainres
Brain Research 1002 (2004) 100–109
Research report
Pontine cholinergic neurons depend on three neuroprotection systems
to resist nitrosative stress
Michael McKinney*, Katrina Williams, David Personett, Caroline Kent, David Bryan,John Gonzalez, Karen Baskerville
Department of Pharmacology, Mayo Clinic Jacksonville, 4500 San Pablo Road, Jacksonville, FL 32224, USA
Accepted 23 December 2003
Abstract
Brainstem cholinergic populations survive in neurodegenerative disease, while basal forebrain cholinergic neurons degenerate. We have
postulated that variable resistance to oxidative stress may in part explain this. Rat primary cultures were used to study the effects of several
nitrosative/oxidative stressors on brainstem (upper pons, containing pedunculopontine and lateraldorsal tegmental nuclei; BS) cholinergic
neurons, comparing them with medial septal (MS), and striatal cholinergic neurons. BS cholinergic neurons were significantly more resistant
to S-nitro-N-acetyl-D,L-penicillamine (SNAP), sodium nitroprusside (SNP), and hydrogen peroxide than were MS cholinergic neurons, which
in turn were more resistant than striatal cholinergic neurons. Pharmacological analyses using specific inhibitors of neuroprotective systems
also revealed differences between these three cholinergic populations with respect to their vulnerability to SNAP. Toxicity of SNAP to BS
neurons was exacerbated by blocking NF-nB activation with SN50 or ERK1/2 activation by PD98059, or by inhibition of phosphoinositide-3
kinase (PI3K) activity by LY294002. In contrast, SNAP toxicity to MS neurons was augmented only by SN50, and SNAP toxicity to striatal
cholinergic neurons was not increased by any of these three pharmacological agents. In neuron-enriched primary cultures, BS cholinergic
neurons remained resistant to SNAP while MS cholinergic neurons remained vulnerable to this agent. Immunohistochemical experiments
demonstrated nitric oxide (NO)-induced increases in nuclear levels of phospho-epitopes for ERK1/2 and Akt, and of the p65 subunit of NF-
nB, within BS cholinergic neurons. These data indicate that the relative resistance of BS cholinergic neurons to toxic levels of nitric oxide
involves three intrinsic neuroprotective pathways that control transcriptional and anti-apoptotic cellular functions.
D 2004 Elsevier B.V. All rights reserved.
Theme: Development and regeneration
Topic: Neuronal death
Keywords: Neurodegeneration; Neuroprotection; Acetylcholine; Oxidative stress; Kinase
1. Introduction amounts of nitric oxide (NO) formed by calcium-dependent
Basal forebrain cholinergic neurons degenerate in Alz-
heimer’s disease and several other age-related disorders
(reviewed in Ref. [36]). Other populations of cholinergic
neurons are less affected. Notably, the cholinergic groups in
the pedunculopontine nucleus (PPN) and lateral dorsal teg-
mental nucleus (LDTN) located in the upper pons of the
brainstem are preserved [51]. These cholinergic neurons are
unusual in their relatively high expression of the neuronal
nitric oxide synthase (nNOS, Type I; [46]). NMDA receptor-
mediated toxicity is mediated largely by the excessive
0006-8993/$ - see front matter D 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.brainres.2003.12.021
* Corresponding author. Tel.: +1-904-953-8713; fax: +1-904-953-7117.
E-mail address: [email protected] (M. McKinney).
activation of nNOS [9,10]. The NO that is produced can
readily diffuse through membranes to endanger other cells.
Interestingly, the brainstem nNOS-positive cholinergic neu-
rons are relatively resistant to NMDA [18], and there are other
reported cases of resistance of nNOS-positive neurons to
NMDA toxicity [13,14,25,26,49]. Such findings suggest that
neuroprotective systems are coordinately expressed with
nNOS. Consistent with this idea, we have found that cultured
BS cholinergic neurons resist nitrosative stress while MS
neurons die by apoptosis when exposed to NO [12].
NO can activate both protective and destructive pathways
in cells [7,30,40]. The reaction of NO with superoxide anion
to form peroxynitrite can lead to oxidative modifications of
protein and nucleic acids [19] and the NO-mediated damage
M. McKinney et al. / Brain Research 1002 (2004) 100–109 101
to nuclear DNA can lead to activation of repair systems, e.g.,
poly-ADP polymerase (PARP), and apoptosis [33,43]. Cells
exposed to NO excess can also activate p53 [22,50,52], a key
initiator of apoptosis in response to DNA damage. Con-
versely, low levels of NO like those generated endogenously
can mediate survival from oxidative stressors by regulation
of mitochondrial function [41]. NO is known to mediate
antiapoptotic effects by inhibiting caspases by S-nitrosyla-
tion [7,43] or by stimulating Ras [27], which then activates
downstream survival mechanisms [15,53]. These down-
stream Ras-dependent cellular survival mechanisms may
involve activation of ERK1/2 [5,22] and/or of Akt, also
known as protein kinase B, an important protection element
with many anti-apoptotic functions [8]. These pathways can
cross-talk at a number of levels [1], as well as interact with
other survival systems like the atypical protein kinases C or
the transcription factor NF-nB [34]. Many of these activities
of NO are effected by nitrosylation or oxidation of cysteine
residues within signaling proteins [33]. Additionally, there is
evidence supporting direct effects of NO on certain tran-
scription factors [2,33].
Our previous investigations of the brain cholinergic pop-
ulations showed that they differ in their expression of
mRNAs for nNOS [43,46] and the superoxide dismutases
[20,21]. Since destruction of superoxide prevents peroxyni-
trite-mediated toxicity, our results suggested that manganese-
dependent superoxide dismutase could be one neuroprotec-
tive system for nNOS-expressing cholinergic neurons in
vivo. However, rigorous testing of hypotheses about protec-
tion systems in neuronal populations requires experimental
models of living cholinergic cells. To convincingly demon-
strate that regional differences in vulnerability actually exist,
the populations in question ideally would be evaluated in the
same experiment under the same conditions. The reticular
pontine cholinergic neurons in the LDTN and PPN are
generated about embryonic day 13 (E13), about 2 days after
the generation of the brainstem cholinergic motor nuclei [45].
Basal forebrain cholinergic neurons are generated throughout
a 6-day period in caudal-to-rostral fashion covering E12–
E17 [6]. The cholinergic interneurons in the rat caudate-
putamen are born during the period E13–E17 [44]. Thus, to
obtain all three populations for culture, the brain must be
dissected on E17 or later.
After plating these three embryonic populations, the cells
must be cultured long enough to allow for maturation. An
example of an early step in differentiation in the LDTN/PPN
is the expression of nNOS, which occurs in vivo 3–4 days
after the birth of the cholinergic neurons [47,48]. However,
later steps in cholinergic differentiation include dendritic
development, axonal growth, and establishment of function-
al connections within target regions. For the subdivisions of
the basal forebrain, these events occur during the first and
second postnatal weeks [11,16,35,38,39]. To study these
three populations in parallel cultures, we made use of the
protocols developed by Knusel and Hefti [23] and Knusel et
al. [24], who compared trophic responses of pontine and
medial septal (MS) cholinergic neurons to various growth
factors. Using this culture method, we have observed that an
indicator of the cholinergic maturation, choline acetyltrans-
ferase activity, increases after plating on E18 to reach a
maximum after about 2 weeks in culture (unpublished data).
In our view, the glia are probably also necessary for each
cholinergic cell type to properly differentiate in its regionally
specific manner.
However, the presence of glia and many non-cholinergic
neurons in these cultures places severe limits on the power
of experiments to identify neuroprotection events in the
cholinergic minority. Only the activity or immunohisto-
chemical properties of choline acetyltransferase (or the
vesicular acetylcholine transporter) can be used to establish
the location, development, or survival of the cholinergic
neurons in such experiments. Nevertheless, considerable
insight can be gained by use of pharmacological agents that
specifically target putative neuroprotection systems. Certain
manipulations of culture conditions can address the ques-
tion of glial contributions. Dual immunocytochemistry for
phenotypic markers and activation-dependent signaling
antigens provides a way to localize signaling events within
cholinergic neurons. Combination of these methods provid-
ed good evidence showing that pontine cholinergic neurons
activate three important neuroprotection pathways to resist
nitrosative stress, while the two other cholinergic popula-
tions do not. The findings may have important implications
for understanding cholinergic pathophysiology in human
disease.
2. Materials and methods
2.1. Cell culture
All work with animals was performed under an approved
Institutional Animal Care and Use Committee protocol.
Embryos were obtained from Sprague–Dawley mother rats
at days E17–E18 and placed in ice-cold media (L15 with
modifications; [23,24]). The brains were removed and kept
in media on ice while dissections were conducted. The
medial septal area (along with the contiguous basal fore-
brain), the upper pons, and the striatum were dissected
under ice-cold media and collected into culture tubes
containing 2 ml ice-cold media. The cells were dissociated
by trituration as described [23,24], washed, counted, and
plated at 5� 105 cells per mm2 of six-well culture plates
coated with poly-L-lysine. The cells were grown for 2–3
weeks in modified L15 media and fed every 3 days after the
fifth day.
2.2. Choline acetyltransferase (ChAT) enzyme assay
The activity of ChAT in individual culture wells was
assayed as previously described [12]. Cells were washed
with cold isotonic phosphate-buffered saline (PBS; pH 7.4),
M. McKinney et al. / Brain Rese102
then harvested by addition of homogenization buffer (50
mM Tris–Hcl; 0.02% Triton X-100) and scraping to sus-
pend the cells. The cells were sonicated at 4 jC and debris
was removed by centrifugation. About 50 Ag protein in the
supernatant was assayed in duplicate in 100 Al reaction
mixture containing 10 nmol [14C]acetyl-Coenzyme A, 200
mM NaCl, 6 mM choline chloride, 50 mM sodium phos-
phate (pH 7.0), 150 mM physostigmine sulfate, and 0.05%
(wt/vol) bovine serum albumin. Reaction mixtures were
incubated at 37 jC for 60 min in a shaking water bath
(65 rpm). Reactions were stopped by adding 400 Al ice-coldwater. [14C]Acetylcholine was purified on Dowex 1� 8
columns and radioactivity was assessed by liquid scintilla-
tion. Protein was determined in the sample by the bicin-
choninic acid method (Micro BCA; Pierce, Rockford, IL,
USA). ChAT activity was expressed as nmol [14C]acetyl-
choline per mg protein per h.
2.3. Dual immunohistochemical procedures
For these studies, primary cells were plated at 5� 105
cells/mm2 on to round (18 mm dia), poly-L-lysine-coated
coverslips and cultured in 12-well plates. After treatments
(e.g., administration of S-nitro-N-acetyl-D,L-penicillamine
(SNAP)), the cells were fixed in 4% buffered paraformalde-
hyde and visualized by the avidin–biotin method (using
reagents mainly from Sigma, Jackson Laboratories, and
Vector Laboratories) were used for histochemistry. In our
dual-labeling protocol, the order of immunostaining was not
important; i.e., ChAT/VAChT staining could be performed
either before or after the second immunolabeling stage with
equivalent results. This is because we made use of an
intervening peroxide treatment and the Vector avidin block-
ing reagent to completely inactivate residual HRP and Extra-
vidin activities after the first stage of immunostaining.
Cholinergic neurons in the cultures were visualized with
primary antibodies to ChAT (goat polyclonal, Cat. No.
AF144P; Chemicon; 1:200) or to the vesicular acetylcholine
transporter (VAChT; goat polyclonal, Cat. No. AB1578;
Chemicon; 1:6000), incubated with an appropriate biotiny-
lated secondary antibody, then with horseradish peroxidase
(HRP)-coupled avidin (‘‘Extravidin’’, Sigma); the immuno-
complexes were then colorized, typically with diaminoben-
zidine (DAB) but alternatively with nickel-enhanced DAB or
Vector SG. Primary antibodies for phospho-Akt (Ser-473;
Promega Cat. No. G7441 or Cell Signaling Technology Cat.
No. 9277), phospho-ERK (anti human phospho-Tyr 204; Cat.
No. sc-7383; Santa Cruz Biotechnology; 1:100), or the p65
subunit of NF-nB (rabbit polyclonal anti-p65; Cat. No. sc-
372; Santa Cruz Biotechnology; 1:500) were used in a second
histochemical procedure using a biotinylated secondary anti-
IgG antibody, HRP-labeled Extravidin, and a contrasting
color in the peroxidase reaction. Special attention was re-
quired in selecting the biotinylated secondary antibodies, so
that cross-reaction with the ChAT/VAChT immunocomplex
did not occur.
2.4. Statistical testing
To score the dual-labeled immunohistochemical prepara-
tions for the presence of low or high levels of phospho-
epitopes or p65 within cholinergic profiles, the identities of
the slides were blocked and each coverslip was scanned at
40� oil-immersion eyepiece (400� final magnification) in
an Leica DMLB microscope. Each specific VAChT or ChAT
profile was scored (� , ‘‘minus’’) if its level of the stain for
the signaling epitope was at or below the general level in the
majority of other cells in the vicinity, or (+, ‘‘positive’’) if
the epitope was noticeably higher than that observed in
other cells containing general background levels. The cov-
erslip was scanned in its entirety and the percent of all
scored cells that were (+) was calculated to determined
‘‘percent activation’’. Multiple coverslips from several in-
dependent cultures and staining experiments were assessed.
Because cells were cultured in individual wells for extended
periods (5–21 days) after seeding, we considered each well
an independent experiment. Comparisons between groups
were made using Student’s t-test and the level of signifi-
cance was p < 0.05.
arch 1002 (2004) 100–109
3. Results
3.1. Sensitivities of several populations of central cholin-
ergic neurons to oxidative/nitrosative stress
We used a variety of stressors to examine the vulnera-
bility of three cholinergic populations. Hydrogen peroxide is
frequently used to generate oxidative stress in cell culture
systems; typically, it induces apoptotic responses. Sodium
nitroprusside (SNP) and SNAP are two commonly used
‘‘NO donors’’ with differing mechanisms of inducing cel-
lular stress. SNP can be regarded as a nitrosonium ion
(NO+) producer, but it also releases iron and cyanide ions.
SNAP is a S-nitroso compound that may release NO in
media containing low amounts of transition metals, but it
also can transfer NO to thiol residues within cells, like those
in the reduced form of glutathione.
Striatal, MS, and BS mixed primary cultures (14–21
DIV) prepared from E18 fetal rat brain were exposed for
16–24 h to a range of concentrations of hydrogen peroxide,
SNP, or SNAP. The activity of ChAT was used as a measure
of the cholinergic neurons remaining alive. An immunos-
taining experiment with MS cultures showed that most of
the cholinergic neurons had degenerated by 6 h after
addition of SNAP (data not shown). Composites of data
from 4–12 independent experiments (11–24 independent
culture wells) are shown in Fig. 1. BS cholinergic neurons
were more able to survive all three of the treatments,
relative to MS or striatal cholinergic neurons. Striatal
neurons were the most vulnerable, with MS cells interme-
diate. The maximum differences in vulnerability were about
10-fold. SNAP was significantly more toxic to MS cholin-
Fig. 1. Concentration– response relationships for oxidative stress vulner-
ability to central cholinergic populations. Freshly prepared agents were
added to the media for brainstem (BS; squares), basal forebrain (MS;
circles), and striatal (triangles) mixed rat E17–18 primary cultures (14–21
DIV), and incubated overnight. ChAT activity was assayed as an indicator
of cholinergic neuronal survival. (A) Sensitivities to hydrogen peroxide.
These are averagesF standard errors (S.E.) for 9–24 wells from 12
different experiments. (B) Sensitivities to SNP. These are averagesF S.E.
for 10–18 wells from seven independent experiments. (C) Sensitivities to
SNAP. AveragesF S.E. for 10–18 wells from six independent experiments.
Significant differences between BS and MS: *p < 0.05; **p < 0.01;
***p< 0.001. Significant differences between BS and striatum: @p< 0.05;@@p< 0.01; @@@p< 0.001.
M. McKinney et al. / Brain Research 1002 (2004) 100–109 103
ergic neurons than to BS neurons, a result in agreement
with our previous results [12]. However, in the present
experiments, the difference was less pronounced. Notably,
the pattern of vulnerability to hydrogen peroxide differed
from those of the NO donors, in that a substantial fraction
of the cells (about 40% of the BS) appeared to be totally
resistant to this oxidative agent.
We concluded from these data that clear and significant
differences exist in the reactions of cholinergic neurons in
mixed primary cultures to nitrosative or oxidative stress. BS
cholinergic neurons exhibit relative resistance to both nitro-
sative and oxidative stressors, as compared to forebrain
cholinergic populations.
3.2. Conditioned media experiments
Our immunohistochemical analyses with glial fibrillary
acidic protein (GFAP) confirmed that the majority of cells
(probably more than 95%) present in our cultures were
astrocytes, as expected (not shown). Astrocytes secrete many
factors that influence neuronal survival. We conducted
experiments with conditioned media to determine to what
degree such factors might be influencing survival differences
for MS and BS cholinergic neurons exposed to NO excess.
Half of the media over MS or BS mixed primary cultures was
removed every other day and used to replace half the media in
the alternate culture (media ‘‘switching’’), over the 2-week
culture period after plating of E17–18 cells. Other wells
containing MS and BS cultures from the same dissections
were maintained in parallel as controls. It was necessary to
‘‘switch’’ only half the media, because the remaining volume
was replaced with fresh media to maintain the cells for the
required 2-week period. After 2 weeks, vulnerability to
SNAP (0.5 mM) was assessed. Fig. 2 shows composites of
the results from several independent experiments. The MS
control (no conditioned media treatment) cultures (Fig. 2A)
were vulnerable to SNAP (>60% loss of ChAT activity), and
the BS control cultures (Fig. 2B) were unaffected by SNAP.
In MS cultures fed conditioned media from BS cultures
significantly more ChAT activity remained after 0.5 mM
SNAP treatment than in control MS wells (Fig. 2A), but the
percentage loss of ChAT activity was about the same as with
unconditioned MS cultures. In BS cultures fed MS condi-
tioned media (Fig. 2B), SNAP was able to kill a minority of
cholinergic neurons ( p < 0.05). These data indicate that
factors released into the media in either MS or BS cultures
were probably not primarily responsible for the their differ-
ences in vulnerability to SNAP.
3.3. Cholinergic vulnerability to SNAP in neuron-enriched
cultures
To further evaluate the possible involvement of non-
neuronal cells in cholinergic survival, we conducted some
experiments with ‘‘neuron-enriched’’ cultures (employing
‘‘Neuro-Basal’’ media); GFAP immunostaining indicated
Fig. 2. Test for the involvement of soluble neuroprotective factors in BS
resistance to nitrosative stress. MS and BS mixed rat primary cultures were
plated using dissociated E17–E18 cells. Every 3 days, half the volume of
conditioned media was transferred from BS to MS cultures, or MS to BS
cultures, and a half volume of fresh medium was added. Control MS and
BS wells were maintained by discarding half the media and adding fresh
media every 3 days. After 2–3 weeks of this procedure, 500 AM SNAP was
added to wells (DMSO carrier to control wells) and ChAT activity was
assessed the next day to determine toxicity. These data are the
averagesF S.E. for 21 wells for each condition in six independent
experiments. In control (no conditioning) wells, MS cholinergic neurons
were killed by SNAP (@@@p< 0.001) and BS cholinergic neurons were
unaffected by SNAP ( p>0.05). The amount of ChAT activity after SNAP
treatment of the ‘‘MS conditioned’’ wells was significantly higher than in
‘‘MS Control’’ wells (**p< 0.01), but the proportion of surviving cells was
similar. BS cells conditioned by MS media (‘‘BS Conditioned’’) were
significantly affected by SNAP (*p< 0.05), suggesting that MS media
might contain a substance that could render BS cells somewhat vulnerable
to SNAP.
Fig. 3. Relative vulnerability of MS and BS cholinergic neurons to SNAP is
maintained in neuron-enriched cultures. E17–E18 primary cells were
plated and cultured in Neuro-Basal media for several days before
administering 500 AM SNAP overnight. ChAT activity was used as a
measure of cholinergic neuron survival. The treatment killed almost half the
MS cholinergic neurons (***p< 0.001) but had no significant effect on BS
cholinergic neurons ( p>0.05). These results (averagesF S.E.) are from 23–
31 wells at each condition from eight independent experiments. The
differential toxicity of SNAP in these neuron-enriched cultures is virtually
identical to toxicity in mixed primary cultures.
M. McKinney et al. / Brain Rese104
that fewer than 10% of the cultured cells were astrocytes
(not shown). In this medium, we found that MS and BS
cholinergic neurons began to decline in viability after
about 5 days. Fig. 3 shows composite data from SNAP
toxicity experiments with MS and BS neuron-enriched
cultures, in which the donor was added at day 3 in culture
and ChAT activity was assayed the next day. The MS
cholinergic neurons still exhibited sensitivity to nitric oxide
excess (0.5 mM SNAP; p < 0.001), while the BS cholin-
ergic neurons were completely resistant. These data sug-
gest that population differences in vulnerability to SNAP
might best be explained by factors intrinsic to the cholin-
ergic neuron, and probably does not involve glial cells in
any major way.
3.4. Pharmacological blockade of putative neuroprotective
pathways
Further clues as to population differences in vulnerability
to NO excess were obtained in studies using pharmacological
blockers of key steps in three neuroprotective pathways. The
cell-permeable compound SN50 contains nuclear localiza-
tion sites for migration of the p65 subunit of NF-nB and can
block the nuclear actions of this transcriptional system by
diverting p65/p50 dimers from transmigrating into the nu-
cleus. PD98059 inhibits the activation of ERK1/2 by its
effects on MEK, and LY294002 inhibits phosphoinositide-3
kinase (PI3K), which is required for activation of the survival
kinase Akt (protein kinase B). To inhibit these pathways as
completely as possible, it was necessary to use maximal drug
concentrations. Preliminary experiments were performed
with BS cultures to determine the maximum concentrations
of SN50, PD98059, and LY 294002 that could be used
without major effects on ChAT activity in these cultures.
We found, for example, that control ChAT activity in BS
primary cultures remained near maximal with SN50 concen-
trations as high as 27 AM, and this was the concentration we
then used for comparing how inhibition of NF-nB affected
the toxicity of SNAP on three different cholinergic popula-
tions (Fig. 4).
Our results with SN50 confirmed our previous conclu-
sions with caffeic acid phenylethyl ester [12] that NF-nB was
involved in protecting MS and BS cholinergic neurons from
SNAP (Fig. 4). Furthermore, PD98059 (100 AM) and
LY294002 (100 AM) also rendered BS cholinergic neurons
vulnerable to 0.5 mM SNAP (Fig. 4C), while the toxicity of
SNAP to MS and striatal cholinergic neuronal populations
was not augmented by either drug (Fig. 4A,B, respectively).
arch 1002 (2004) 100–109
M. McKinney et al. / Brain Research 1002 (2004) 100–109 105
There were significant effects of the three inhibitors on
control ChAT levels in one or more of the cultures; for
LY294002 and PD98059, these effects were at most minor.
The toxic effect of SNAP to striatal cholinergic neurons
(>60% killing) was not potentiated by blockade of any of
the three putative neuroprotective pathways (P>0.05). These
data clearly show that BS cholinergic neurons are unique in
their apparent requirement for ERK1/2, PI3K, and NF-nBfunctions for their resistance to nitrosative stress.
3.5. Immunohistochemical evidence for NO-mediated acti-
vation of neuroprotection systems in BS cholinergic neurons
The pharmacological data (Fig. 4) suggested that BS
cholinergic neurons are unique in their involvement of NF-
nB, PI3K/Akt, and ERK1/2 in resisting SNAP toxicity.
However, since these are mixed cultures, it is possible that
these systems are not activated within cholinergic neurons
per se. We used dual immunohistochemistry for cholinergic
markers and for markers of neuroprotective systems to
investigate these pathways in identified BS cholinergic
neurons in situ. Fig. 5 shows typical profiles for NF-nB(p65 immunoreactivity), phospho-Akt, and phospho-ERK1/
2 in control cultures (A,C,E) and cultures treated with 0.5
mM SNAP (B,D,F). Multiple coverslips for each condition
and each immunostain were scored for the level (weak/
absent vs. moderate/strong) of stain for the signaling
epitope in BS cholinergic profiles (Table 1).
In our mixed primary cultures, NF-nB was constitutively
activated in a subpopulation (36%) of BS cholinergic neurons
(Table 1 and Fig. 5A/B). By 60min after administering SNAP
to BS primary cultures, increases in the fraction of profiles
containing strong nuclear p65 immunoreactivity were evi-
dent throughout the cultures, indicating activation of the NF-
nB system in response to nitrosative stress. The experiment
shown in Fig. 5A/B is one of four that were counted; in the
SNAP-treated cultures, most ChAT-positive BS neurons
(75%) exhibited moderate to very strong p65 nuclear reac-
tivity at this time point (ChAT immunoreactivity is colorized
brown with DAB, and p65 is the dark blue–black of nickel-
enhanced DAB). This was a highly significant increase
compared to control levels ( p < 0.001).
In the experiment in Fig. 5C,D, ChAT is immunostained
blue (with Vector SG) and pAkt (ser473) is colorized with
DAB (brown). In control BS cultures (n = 4), an average of
61% of the cholinergic neurons exhibited strong nuclear
pAkt reactivity (57% in the experiment in Fig. 5C/D),
indicating constitutive activation of the protein kinase.
Treatment with SNAP for 60 min caused the nuclei of the
large majority (83%) of the BS ChAT-positive cells (n = 4)
Fig. 4. Pharmacological profiles implicating involvement of neuroprotective
pathways in variable resistance of three cholinergic population to nitric oxide
excess. Rat E17–E18 mixed primary cultures were maintained for 2–3
weeks before conducting experiments. The inhibitor concentrations were:
LY294002, 100 AM; PD98059, 100 AM; SN50, 27 AM. Drugs were added 1
h prior to addition of 500 AM SNAP; surviving cholinergic neurons were
assessed the next day by quantitation of ChAT activity. (A) Striatal
cholinergic neurons. SNAP administration killed about half the cells
( p< 0.001). Inhibition of ERK1/2 activation with PD98059 and NF-nBwith SN50 suppressed ChAT activity in control wells (**p< 0.01 and
***p< 0.001, respectively), but none of the three inhibitors changed
significantly the degree of toxicity of nitric oxide excess ( p>0.05). These
data are averagesF S.E. for 11–14 wells in six independent experiments.
AveragesF S.E. for 9–10 wells in five independent experiments. (B) Medial
septal cholinergic neurons. SNAP killed about 60% of the cells ( p< 0.001).
Of the three inhibitors, only SN50 had any effect on control cells
(***p < 0.001) or exacerbated SNAP toxicity (@@p < 0.01). (C) BS
cholinergic neurons. SNAP (500 mM) did not significantly affect ChAT
activity in control wells ( p>0.05), but all three inhibitors significantly
potentiated SNAP toxicity (@@p < 0.01; @@@p < 0.001). SN50 and
LY294002 moderately suppressed ChAT activity in control wells
(*p< 0.05; **p< 0.01). Averages +/� for 9–10 wells in five independent
experiments.
Fig. 5. Immunohistochemical evidence for activation of neuroprotective systems in BS cholinergic neurons after exposure to nitric oxide. Red arrows indicate
BS cholinergic profiles. ChAT-NF-jB dual immunohistochemistry: (A) Control and (B) 500 AM SNAP, 60 min. ChAT is colorized with DAB (brown) and the
p65 subunit is visualized using Ni-DAB (blue–black). In the SNAP-treated well (B), the nucleus contains a heavy deposit of p65 immunoreactivity. ChAT-pAkt
dual immunohistochemistry: (C) Control and (D) 500 AM SNAP, 60 min. ChAT is visualized using Vector SG (blue) and pAkt (Ser 473) is colorized with DAB
(brown). The activation of Akt is suggested by the much stronger brown deposits in the nuclei in panel D. VAChT-pERK immunohistochemistry: (E) Control
and (F) 500 AM SNAP, 60 min. VAChT immunoreactivity is DAB (brown) and pERK (1/2) is colorized using Vector SG (blue).
M. McKinney et al. / Brain Research 1002 (2004) 100–109106
to acquire a strong pAkt signal, and this was a significant
increase ( p< 0.05).
In the experiment in Fig. 5E (Control) and Fig. 5F
(SNAP), VAChT is the cholinergic marker visualized with
DAB and pERK1/2 is visualized with Vector SG (blue). In
control wells, 41% of BS cholinergic neurons contained
strong cytosolic and/or nuclear pERK1/2 reactivity (n = 6),
Table 1
Immunostaining assessment of signaling activities in primary BS
cholinergic neurons
Second immunostain Control (carrier only) SNAP (0.5 mM)-treated
p65 (NF-nB) 36F 2% (4) 75F 4%*** (4)
Phospho-Akt (ser473) 61F 9% (4) 83F 4%* (5)
Phospho-ERK1/2 41F 4% (6) 61F 2%*** (6)
These numbers are the percent of ChAT- or VAChT-positive profiles that
also contained elevated levels of signaling epitope. Student’s t-test for the
different between Control and SNAP conditions.
*p< 0.05.
***p< 0.001.
suggesting constitutive activation. This MAP kinase system
was activated throughout the culture by 60 min after
addition of SNAP, and 65% of cholinergic neurons were
found to be strongly labeled with pERK1/2 (Fig. 5F), a
significant increase (n = 6, p < 0.001). Our results indicate
that nitric oxide excess leads to the activation of all three of
these putative neuroprotection pathways in BS cholinergic
neurons.
4. Discussion
Our experiments show that BS cholinergic neurons in
mixed primary culture are relatively resistant to oxidative
and nitrosative stressors, as compared to MS or striatal
counterparts. As many as 40% of the BS cholinergic
neurons survive up to 100 mM hydrogen peroxide, and
BS cholinergic cells are at least a log-order more resistant to
NO toxicity, as compared to MS or striatal cholinergic
neurons. Media-switching experiments indicated that hu-
M. McKinney et al. / Brain Research 1002 (2004) 100–109 107
moral factors, like those that might be released from
astrocytes, probably were not important in this resistance.
BS cholinergic neurons remained resistant, and MS neurons
remained vulnerable, to NO when astrocytes were largely
eliminated from the cultures, suggesting that reaction to
nitrosative stress is a property intrinsic to the cholinergic
neuron. Pharmacological data indicated that BS neurons
resisted nitrosative stress by mechanisms involving at least
three putative neuroprotective pathways, a feature very
different from either MS or striatal cholinergic populations.
Immunohistochemical experiments revealed significant
increases in nuclear intensities of phospho-epitopes for
ERK1/2 and Akt, and the p65 subunit of NF-nB, withinBS cholinergic neurons after exposure to NO.
Our previous work suggests that when MS cholinergic
neurons are exposed to nitrosative stress, they die by an
apoptotic mechanism [12]. Also in that study, the relative
resistance of the BS cholinergic neurons to SNAP was
shown to involve maintenance of normal levels of gluta-
thione. NO-induced apoptosis has been shown to involve
DNA damage [31] and activation of p53 [22,50]. NO
toxicity to neurons may [32] or may not [54] involve
activation of poly (ADP-ribose) polymerase, possibly
depending on whether peroxynitrite formation is involved.
Based on these reports, it seems that DNA damage and
activation of these systems may be involved in the loss of
MS or striatal cholinergic neurons in our cultures after
exposure to SNAP, although there are other possibilities.
The relative resistance of BS cholinergic neurons, which
seems to involve PI3K, ERK1/2, and NF-nB, thus eventu-ally may be explained by the regulatory influences of these
neuroprotective pathways on p53 activity, specifically on
its propensity to elicit apoptosis after DNA damage. By its
ability to regulate p53 accumulation, ERK1/2 prevents
nitric oxide-induced apoptosis of chondrocytes [22]. Akt,
which depends on PI3K for activation, prevents NO-in-
duced apoptosis of hippocampal neurons by blocking
transcription of p53 [52]. These neuroprotective functions
may be activated by NO in parallel with pro-apoptotic
events in the same cell. Thus, the survival of BS cholin-
ergic neurons may be the result of a peculiarity in this cell
type for stronger activation of anti-apoptotic pathways than
pro-apoptotic pathways, when exposed to NO. Some alter-
native hypotheses might relate to the ability of this cell type
to absorb or neurtralize excess NO or mobilize more robust
repair mechanisms for damaged DNA or protein.
Our original hypothesis was that the relative vulnerability
of BS and MS cholinergic neurons to oxidative stress is
related to the relative vulnerability of these cells in AD. This
hypothesis may appear now to be overly simplistic, because
in the present study, we found that striatal cholinergic neurons
are even more vulnerable in vitro than the MS population.
Striatal cholinergic neurons are generally thought to be
unaffected in AD. However, a 60% loss of the cholinergic
neurons in the ventral striatum has been reported [28]. There
are also substantial reductions in neurotrophin receptor ex-
pression on striatal cholinergic neurons, and this not limited
to the ventral region [3,4]. Thus, it appears that at least some
striatal cholinergic neurons are as vulnerable in AD as the
basal forebrain system.
An improved hypothesis for selective cholinergic degen-
eration in AD might combine differential oxidative stress
vulnerability, which we have clearly demonstrated, with
additional disease-related insults into a ‘‘multi-hit’’ hypoth-
esis. For example, it may be that basal forebrain cholinergic
neurons are more severely affected in AD because of their
susceptibility to oxidative or nitrosative stress, in combi-
nation with detrimental influences of cortical amyloid
deposition; the latter feature of the AD brain is presumably
less of a factor in the striatum. However, in our view,
amyloid deposition alone does not explain cholinergic loss.
We have found in transgenic mouse models that the basal
forebrain cholinergic population is preserved even with
extensive and nearly life-long cortical amyloid deposits
[17]. Moreover, in the Parkinson’s disease brain, in which
basal forebrain cholinergic degeneration can occur [20],
cortical pathology is absent.
It is apparent that much more study of these cholinergic
neuronal populations will be required before selective
vulnerability can be better explained. Recognizing this, we
have cloned cholinergic cell lines from the brainstem and
basal forebrain for more detailed molecular approaches
[37,42]. These cell lines exhibit a wide range of vulnerabil-
ity to nitrosative stress and appear to invoke the same three
neuroprotection systems we have inferred to exist in pontine
cholinergic neurons (manuscript in preparation). Clarifica-
tion of how the downstream signaling events in these three
cascades promote cholinergic survival of oxidative stressors
will likely help in understanding why some cholinergic
populations degenerate in human disease.
References
[1] B.A. Ballif, J. Blenis, Molecular mechanisms mediating mammalian
mitogen-activated protein kinase (MAPK) kinase (MEK)-MAPK cell
survival signals, Cell Growth Diff. 12 (2001) 397–408.
[2] C. Bogdan, Nitric oxide the regulation of gene expression, Trends
Cell Biol. 11 (2001) 66–75.
[3] F. Boissiere, S. Hunot, B. Faucheu, L.B. Hersh, Y. Agid, E.C. Hirsch,
Trk neurotrophin receptors in cholinergic neurons of patients with
Alzheimer’s disease, Dement. Geriatr. Cogn. Disord. 8 (1997) 1–8.
[4] F. Boissiere, B. Faucheux, M. Ruberg, Y. Agid, E.C. Hirsch, De-
creased TrkA gene expression in cholinergic neurons of the striatum
and basal forebrain of patients with Alzheimer’s disease, Exp. Neurol.
145 (1997) 245–252.
[5] A. Bonni, A. Brunet, A.E. West, S.R. Datta, M.A. Takasu, M.E.
Greenberg, Cell survival promoted by the Ras-MAPK signaling path-
way by transcription-dependent and -independent mechanisms, Sci-
ence 286 (1999) 1358–1362.
[6] D.R. Brady, P.E. Phelps, J.E. Vaughn, Neurogenesis of basal forebrain
cholinergic neurons in rat, Dev. Brain Res. 47 (1989) 81–92.
[7] H.-T. Chung, H.-O. Pae, B.-M. Choi, T.R. Billiar, Y.-M. Kim, Nitric
oxide as a bioregulator of apoptosis, Biochem. Biophy. Res. Com-
mun. 282 (2001) 1075–1079.
M. McKinney et al. / Brain Research 1002 (2004) 100–109108
[8] S.R. Datta, A. Brunet, M.E. Greenberg, Cellular survival: a play in
three Akts, Genes Dev. 13 (1999) 2905–2927.
[9] V.L. Dawson, T.M. Dawson, E.D. London, D.S. Bredt, S.H. Snyder,
Nitric oxide mediates glutamate neurotoxicity in primary cortical cul-
tures, Proc. Natl. Acad. Sci. U. S. A. 88 (1991) 6368–6371.
[10] V.L. Dawson, T.M. Dawson, D.A. Bartley, G.R. Uhl, S.H. Snyder,
Mechanisms of nitric oxide-mediated neurotoxicity in primary brain
cultures, J. Neurosci. 13 (1993) 2651–2661.
[11] A. Dinopoulos, H.B.M. Uylings, J.G. Parnavelas, The development of
neurons in the nuclei of the horizontal and vertical limb of the diag-
onal band of Broca of the rat: a qualitative and quantitative analysis of
Golgi preparations, Dev. Brain Res. 65 (1992) 65–74.
[12] U. Fass, K. Panickar, D. Personett, D. Bryan, K. Williams, J.
Gonzales, K. Sugaya, M. McKinney, Differential vulnerability of pri-
mary cultured cholinergic neurons to nitric oxide excess, NeuroReport
11 (2000) 931–936.
[13] R.J. Ferrante, N.W. Kowall, M.F. Beal, E.P. Richardson, E.D. Bird,
J.B. Martin, Selective sparing of a class of striatal neurons in Hun-
tington’s disease, Science 230 (1985) 561–563.
[14] D.M. Ferriero, L.J. Arcavi, S.M. Sagar, T.K. McIntosh, R.P. Simon,
Selective sparing of NADPH-diaphorase neurons in neonatal hypo-
xia– ischemia, Ann. Neurol. 24 (1988) 670–676.
[15] M. Gonzalez-Zulueta, A.B. Feldman, L.J. Klesse, R.G. Kalb, J.F.
Dillman, L.F. Parada, T.M. Dawson, V.L. Dawson, Requirement for
nitric oxide activation of p21(ras)/extracellular regulated kinase in
neuronal ischemic preconditioning, Proc. Natl. Acad. Sci. U. S. A.
97 (2000) 436–441.
[16] E. Gould, T.W. Farris, L.L. Butcher, Basal forebrain neurons undergo
somatal and dendritic remodeling during postnatal development: a
single-section Golgi and choline acetyltransferase analysis, Dev.
Brain Res. 46 (1989) 297–302.
[17] D. Hernandez, K. Sugaya, T. Qu, E. McGowan, K. duff, M. McKin-
ney, Survival and plasticity of basal forebrain cholinergic systems in
mice transgenic for presenilin-1 and amyloid precursor protein mutant
genes, NeuroReport 12 (2001) 1377–1384.
[18] W.L. Inglis, K. Semba, Discriminable excitotoxic effects of ibotenic
acid, AMPA, NMDA and quinolinic acid in the rat laterodorsal teg-
mental nucleus, Brain Res. 755 (1993) 17–27.
[19] H. Ischiropoulos, Biological tyrosine nitration: a pathophysiological
function of nitric oxide and reactive oxygen species, Arch. Biochem.
Biophys. 356 (1998) 1–11.
[20] K.A. Jellinger, Pathology of Parkinson’s disease. Changes other
than the nitrostriatal pathway, Mol. Chem. Neuropathol. 14
(1991) 153–197.
[21] C. Kent, K. Sugaya, D. Bryan, D. Personett, M. McKinney, Expres-
sion of superoxide dismutase messenger RNA in adult rat brain cho-
linergic neurons, J. Mol. Neurosci. 12 (1999) 1–10.
[22] S.-J. Kim, J.-W. Ju, C.-D. Oh, Y.M. Yoon, W.K. Song, J.-H. Kim, Y.J.
Yoo, O.-S. Bang, S.-S. Kang, J.-S. Chun, ERK-1/2 and p38 kinase
oppositely regulate nitric oxide-induced apoptosis of chondrocytes in
association with p53, caspase-3, and differentiation status, J. Biol.
Chem. 277 (2002) 1332–1339.
[23] B. Knusel, F. Hefti, Development of cholinergic pedunculopontine
neurons in vitro: comparison with cholinergic septal cells and re-
sponse to nerve growth factor, ciliary neuronotrophic factor, and ret-
inoic acid, J. Neurosci. Res. 21 (1988) 365–375.
[24] B. Knusel, P.P. Michel, J.S. Schwaber, F. Hefti, Selective and nonse-
lective stimulation of central cholinergic and dopaminergic develop-
ment in vitro by nerve growth factor, basic fibroblast growth factor,
epidermal growth factor, insulin and the insulin-like growth factors I
and II, J. Neurosci. 10 (1990) 558–570.
[25] J.Y. Koh, D.W. Choi, Vulnerability of cultured cortical neurons to
damage by excitotoxins: differential susceptibility of neurons contain-
ing NADPH-diaphorase, J. Neurosci. 8 (1988) 2153–2163.
[26] J.Y. Koh, D.W. Choi, Cultured striatal neurons containing NADPH-
diaphorase or acetylcholinesterase are selectively resistant to injury
NMDA receptor agonists, Brain Res. 446 (1988) 374–378.
[27] H.M. Lander, J.S. Ogiste, S.F.A. Paerce, R. Levi, A. Novogrodsky,
Nitric-oxide-stimulated guanine nucleotide exchange on p21ras,
J. Biol. Chem. 270 (1995) 7017–7020.
[28] S. Lehericy, E.C. Hirsch, P. Cerbera, L.B. Hersh, J.J. Hauw, M.
Ruberg, Y. Agid, Selective loss of cholinergic neurons in the ven-
tral striatum of patients with Alzheimer disease, Proc. Natl. Acad.
Sci. U. S. A. 86 (1989) 8580–8584.
[30] S.A. Lipton, Neuronal protection and destruction by NO, Cell Death
Diff. 6 (1999) 943–951.
[31] Z. Liu, L.J. Martin, Motor neurons rapidly accumulate DNA single-
strand breaks after in vitro exposure to nitric oxide and peroxynitrite
and in vivo axotomy, J. Comp. Neurol. 432 (2001) 35–60.
[32] A.S. Mandir, M.F. Poitras, A.R. Berliner, W.J. Herring, D.B. Guastella,
A. Feldman, G.G. Poirier, Z.Q. Wang, T.M. Dawson, V.L. Dawson,
NMDA but not non-NMDA exctitotoxicity is mediated by poly (ADP-
ribose) polymerase, J. Neurosci. 20 (2000) 8005–8011.
[33] H.E. Marshall, K. Merchant, J.S. Stamler, Nitrosation and oxidation in
the regulation of gene expression, FASEB J. 14 (2000) 1889–1900.
[34] J.L. Martindale, N.J. Holbrook, Cellular response to oxidative
stress: signaling for suicide and survival, J. Cell. Physiol. 192
(2002) 1–15.
[35] D.A. Matthews, J.V. Nadler, G.S. Lynch, C.W. Cotman, Development
of cholinergic innervation in the hippocampal formation of the rat: I.
Histochemical demonstration of acetylcholinesterase activity, Dev.
Biol. 36 (1974) 130–141.
[36] M. McKinney, J.T. Coyle, The potential for muscarinic receptor sub-
type-specific pharmacotherapy for Alzheimer’s disease, Mayo Clin.
Proc. 66 (1991) 1225–1237.
[37] M. McKinney, K. Baskerville, D. Personett, K. Williams, J. Gonzales,
Cholinergic plasticity and the meaning of death, in: K. Maiese (Ed.),
Neuronal and Vascular Plasticity: Elucidating Basic Cellular Mecha-
nisms for Future Therapeutic Discovery, Kluwer Academic Publish-
ing, Norwell, MA, 2003, pp. 27–74.
[38] N. Mechawar, L. Descarries, The cholinergic innervation develops
early and rapidly in the rat cerebral cortex: a quantitative immunocy-
tochemical study, Neuroscience 108 (2001) 555–567.
[39] J.V. Nadler, D.A. Matthews, C.W. Cotman, G.S. Lynch, Development
of cholinergic innervation in the hippocampal formation in the rat. II.
Quantitative changes in choline acetyltransferase and acetylcholines-
terase activities, Dev. Biol. 36 (1974) 142–154.
[40] K. Nandagopal, T.M. Dawson, V.L. Dawson, Critical role for nitric
oxide signaling in cardiac and neuronal ischemic preconditioning and
tolerance, J. Pharmacol. Exp. Ther. 297 (2001) 474–478.
[41] E. Paxinou, M. Weisse, Q. Chen, J.M. Souza, C. Hertkorn, M. Selak,
E. Daikhin, M. Yudkoff, G. Sowa, W.C. Sessa, H. Ischiropoulos,
Dynamic regulation of metabolism and respiration by endogenously
produced nitric oxide protects against oxidative stress, Proc. Natl.
Acad. Sci. U. S. A. 98 (2001) 11575–11580.
[42] D. Personett, K. Sugaya, D. Hammond, M. Robbins, M. McKinney,
Use of capillary electroporesis with laser-induced fluorescence detec-
tion to assess messenger ribonucleic acid molecules amplified by the
polymerase chain reaction: applications in the cloning of cells, Elec-
trophoresis 18 (1997) 1750–1759.
[43] D. Personett, U. Fass, K. Panickar, M. McKinney, Retinoic acid-
mediated enhancement of the cholinergic/neuronal nitric oxide syn-
thase phenotype of the medial septal SN56 clone: establishment of a
nitric oxide-sensitive proapoptotic state, J. Neurochem. 74 (2000)
2412–2424.
[44] P.E. Phelps, D.R. Brady, J.E. Vaughn, The generation and differenti-
ation of cholinergic neurons in rat caudate-putamen, Dev. Brain Res.
46 (1989) 47–60.
[45] P.E. Phelps, L.A. Brennan, J.E. Vaughn, Generation patterns of im-
munocytochemically identified cholinergic neurons in rat brainstem,
Dev. Brain Res. 56 (1990) 63–74.
[46] K. Sugaya, M. McKinney, Differential expression of mRNA for the
calmodulin-dependent nitric oxide synthase within cholinergic neuro-
nal populations, Mol. Brain Res. 23 (1994) 111–125.
M. McKinney et al. / Brain Research 1002 (2004) 100–109 109
[47] M. Takemura, S. Wakisaka, K. Iwase, N.H. Yabuta, S. Nakagawa, K.
Chen, Y.C. Bae, A. Yoshida, Y. Shigenaga, NADPH-diaphorase in the
developing rat: lower brainstem and cervical spinal cord, with special
reference to the trigemino-solitary complex, J. Comp. Neurol. 365
(1996) 511–525.
[48] Y. Uehara-Kunugi, K. Terai, T. Taniguchi, I. Tooyama, H. Kimura,
Time course of in vitro expression of NADPH-diaphorase in cultured
rat brain neurons: comparison with in vivo expression, Dev. Brain
Res. 59 (1991) 157–162.
[49] Y. Uemura, N.W. Kowal, M.F. Beal, Selective sparing of NADPH–
diaphorase–somatostatin–neuropeptide Y neurons in ischemic gerbil
striatum, Ann. Neurol. 27 (1990) 620–625.
[50] X. Wang, D. Michael, G. de Murcia, M. Oren, p53 activation by nitric
oxide involves down-regulation of Mdm2, J. Biol. Chem. 277 (2002)
15697–15702.
[51] N.L. Woolf, R.W. Jacobs, L.L. Butcher, The pontomesencephaloteg-
mental cholinergic system does not degenerate in Alzheimer’s dis-
ease, Neurosci. Lett. 96 (1989) 277–282.
[52] A. Yamaguichi,M. Tamatani, H.Matsuzaki, K. Namikawa, H. Kiyama,
M.P. Vitek, N. Mitsuda, M. Tohyama, Akt activation protects hippo-
campal neurons from apoptosis by inhibiting transcriptional activity of
p53, J. Biol. Chem. 276 (2001) 5256–5264.
[53] H.Y. Yun, M. Gonzalez-Zulueta, V.L. Dawson, T.M. Dawson, Nitric
oxide mediates N-methyl-D-aspartate receptor-induced activation of
p21 ras, Proc. Natl. Acad. Sci. U. S. A. 95 (1998) 5773–5778.
[54] K. Yuyama, H. Yamamoto, I. Nishizaki, T. Kato, I. Sora, T. Yama-
moto, Caspase-independent cell death by low concentrations of nitric
oxide in PC12 cells: involvement of cytochrome C oxidase inhibition
and the production of reactive oxygen species, J. Neurosci. Res. 73
(2003) 351–363.