| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Department of Neuroendocrinology, Max Planck Institute of Psychiatry, 80804 Munich, Germany;
Laboratory of General and Comparative Physiology, UMR CNRS 8572, Muséum National d’Histoire Naturelle, 75231 Paris, France; and
Department of Neurobiochemistry, Max Planck Institute of Neurobiology, 82152 Martinsried, Germany
1Correspondence: Max Planck Institute of Psychiatry, Kraepelinstrasse 2–10, D-80804 Munich, Germany. E-mail: osa{at}mpipsykl.mpg.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Key Words: apoptosis • Bcl-2 • Bax • bax knockout mouse • corticosterone • dexamethasone • mineralocorticoid receptor • glucocorticoid receptor
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
. Two types of CS
receptor are found in the brain: mineralocorticoid (MR) and
glucocorticoid (GR) receptors. In the rat, these receptors share a
common endogenous ligand corticosterone (CORT); selective MR or GR
activation is apparently conferred by the ~10-fold greater affinity
of MR for CORT (1)
. Differential activation of this dual
receptor system may account for the opposing actions of corticosteroids
on neuronal proliferation, survival, and death in one hippocampal
subfield, the dentate gyrus (2
3
4)
. The dentate gyrus
displays both MR and GR (5)
, and there is evidence showing
that GR occupation as occurs during stress, for example, leads to an
arrest of neurogenesis in the dentate gyrus (6)
; at the
same time, GR occupation stimulates apoptosis within the granular and
hilar cell populations of the dentate gyrus (3)
. On the
other hand, MR activation is essential for maintenance of the granule
cell layer (7
8
9)
, and in a previous study we demonstrated
that MR activation can protect dentate gyrus neurons against acute GR
ligand-mediated apoptosis (3)
. In the latter study, we
also observed that aged rats are more vulnerable to GR-induced cell
death, a finding attributed to age-associated declines in basal CORT
secretion and changes in the availability of MR and GR
(5)
.
The regulation of apoptosis is complex, but it is now established that
proteins encoded by the bcl-2 gene family are major
regulatory components of the apoptotic pathway (10
, 11)
.
The Bcl-2 family comprises death inducer (e.g., Bax,
Bcl-xS) and death repressor (e.g., Bcl-2,
Bcl-xL) proteins, several of which can form homo-
and heterodimers (10
11
12)
. These proteins are activated by
physiological or injurious stimuli and appear to operate upstream of
events leading to the final execution phase of the apoptotic process,
which involves the activation of cysteine proteases—the caspases
(13
14
15)
. Whereas the susceptibility of an individual cell
to apoptosis-triggering stimuli is largely determined by its genetic
content, metabolic state, and developmental/proliferative state, as
well as its distinct receptor and signal transduction pathways
(15
16
17)
, the ratio of death inducer to death repressor
proteins seems to exert a rheostat-like control over whether an
individual cell will ultimately survive or die (18)
.
bcl-2 expression in the brain declines rapidly after birth,
but areas that display postnatal neurogenesis (e.g., the dentate gyrus;
see refs 19
20
21
) continue to express bcl-2 for
an extended period of time (11
, 22)
; in addition,
bcl-2 expression can be induced in the adult brain,
including the hippocampus (23
24
25)
. Unlike
bcl-2, the expression of another member of this family,
bcl-xL, occurs in the brain throughout
embryonic and postnatal life (25
26
27
28
29)
. The large
antiapoptotic splice variant of bcl-x
(bcl-xL) appears to be located exclusively
in neurons (11)
; bcl-xL has
been localized in both pyramidal and granular neurons of the
hippocampus (27
, 28)
. In contrast, levels of the smaller
proapoptotic bcl-x splice variant,
bcl-xS, are only barely detectable in
mature brain (26
, 27)
. Given that bcl-x
expression does not display the spatial and temporal constraints
associated with the expression of bcl-2, it has been
suggested that Bcl-xL may gradually substitute
for Bcl-2 during development (27)
. This molecule has also
been suggested to be critical for the survival of postnatal neurons
(26
, 28
, 30)
. The proapoptotic protein Bax
(31
32)
is widely distributed in the brain, its spatial
distribution in the hippocampus (dentate gyrus and CA2 areas >
CA3 and CA1 areas) corresponding closely to that of MR (see ref
5
). A reciprocal distribution pattern of Bax and Bcl-2 is
frequently seen among cell populations that are the targets of
apoptotic stimuli (24
, 29
, 33
, 34)
.
To date, none of the Bcl-2 family of proteins have been shown to be
directly regulated by CS. Evidence from other studies, however,
suggests that the link between CS and Bcl-2-related molecules may be
provided by the tumor suppressor protein p53, which shows a ubiquitous
distribution and whose trans-activation potential can be
modulated by GR (35)
. Furthermore, there is evidence that
p53 can, respectively, induce and repress bax and
bcl-2 transcription (36
, 37)
. This work
represents the first attempt to define some of the molecular signaling
pathways underlying the differential effects of corticosteroids on
neuronal survival (MR activation) and apoptosis (GR activation)
in vivo. Studies were performed in both young and aged rats
in light of previous work showing the greater vulnerability of
hippocampal neurons to glucocorticoid treatment in aged animals
(3
, 4)
.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
; see also ref
3
). Saline, CORT, and DEX were administered i.p. in two
divided daily doses. In some studies, wild-type (wt) and
bax knockout (bax -/-) mice (see below), aged
20–24 months, were used for testing the apoptotic effects of DEX (100
µg/kg). All the in vivo procedures conformed to local
animal welfare legislation. Animals were killed between 13:00 and
14:00, at which time their brains were immediately removed, frozen in
cooled isopentane, and stored at -80°C before cryosectioning (10
µm). Tissue sections (mounted on coded slides) were processed using
in situ nick end labeling (TUNEL) histochemistry, in
situ hybridization histochemistry, or reverse
transcriptase-polymerase chain reaction (RT-PCR) as described below.
For the histochemical analyses, six anatomically matched cryosections,
each separated by 50 µm (to circumvent the need for correction of
counting error during quantification), were taken from the dorsal
hippocampus (between bregma -3.3 and -4.3; ref 39
).
bax knockout (bax -/-) mice
The mouse bax gene consists of six exons assembled
into multiple, alternatively spliced transcripts encoding a 21 kDa
membrane (
) and two forms of cytosolic protein (ß and 
)
(31)
. To generate mice lacking all isoforms, we targeted
the first four exons, which encompass most of the protein coding
sequence. The bacterial lacZ gene, under the transcriptional
regulatory sequences of the bax gene, was inserted into the
construct to serve as a genetic marker of the mutation (Fig. 1
).
|
Construction of the targeting replacement vector and generation of
bax knockout mice
Genomic clones spanning the entire mouse bax gene
were isolated by screening of a 
FIX mouse genomic library (strain
129/SV DNA; Stratagene, Heidelberg, Germany) with a 311 bp mouse
bax cDNA corresponding to nucleotides 49 to 360 of the rat
bax cDNA sequence (31)
. The clones were
characterized in detail by standard molecular biology techniques.
Initially, a 5.2 kb HindIII-HindIII DNA fragment
containing the last two exons as well as sufficient flanking sequences
to achieve a good targeting efficiency was subcloned into pGNA plasmid
(ref 40
; pGNA plasmid kindly provided by Dr. P.
Brûlet). The targeting vector was generated by the subsequent
cloning of an AseI x BglI DNA fragment
comprising ~2 kb of 5'-flanking sequences of the gene
(BglI cuts 40 bp upstream of the AUG translation initiation
codon). The replacement vector finally contained 7.5 kilobases of
homologous bax genomic DNA, in which the first four exons of
the gene (contained in the BglI-HindIII DNA
region) were removed and replaced by the lacZ. To achieve an
enrichment of the homologous recombination events, we used a modified
version (41)
of the positive-negative selection procedure
(42)
to introduce the negative selection cassette (the HSV
thymidine kinase gene under the promoter of the phosphoglycerate kinase
gene) at the ends of the targeting vector after its linearization with
NotI.
Procedures for electroporation and selection of R1 ES cells (generously
provided by Dr. A. Nagy, Department of Genetics, University of Toronto,
Toronto, Canada; 43
) have been described
(44)
. For the identification of homologous recombinant
cell clones, genomic DNA from individual double-drug resistant ES
clones was double digested with the restriction enzymes
EcoRI and EcoRV and subjected to Southern blot
analysis. As an external probe, we used a 152 bp
SalI-AseI fragment, which distinguishes a
9.2 kb band originating from the wild-type allele, and a 5 kb band that
derives from the disrupted allele (Fig. 1)
. The recombinant clones were
further analyzed in Southern blots using 3' external and internal
probes. Four of the 280 clones were found to carry the desired allele.
Homologous recombinant ES cells were injected into 3.5-day-old C57BL/6
blastocysts that were subsequently implanted into pseudopregnant
recipients. The resulting male chimeras were mated to C57BL/6 females,
and germline transmission of the mutant allele was identified by
Southern blot analysis of tail DNA from F1 offspring. Western blot
analysis was used to confirm the absence of bax expression
in the bax-/- mice (see Fig. 7
).
|
Histochemistry
TUNEL was used to detect cells undergoing apoptosis
(45)
; Giemsa-counterstained sections were then
microscopically examined according to previously described
morphological criteria (3)
. TUNEL-stained cells and
unlabeled healthy cells in the pyramidal blades (granule cell layer) of
the dentate gyrus were counted. Five readings were obtained for each
parameter from anatomically matched sections from each animal
(n=4). An ‘apoptotic index’ was derived from the number of
apoptotic cells occurring per 10,000 healthy cells (3)
. To
complement the RT-PCR measurements of bcl-2, bcl-x, and
bax mRNA, in situ hybridization histochemistry
was also performed on a sample of sections from each treatment:
35S-dATP-labeled 48-mer antisense
oligonucleotide probes complementary to bases 27–66 of
bcl-2, 141–180 of bcl-xL, and
15–54 of bax were used (46)
.
Detection of bcl-2, bcl-xL, and
bax mRNA
Relative levels of bcl-2, bcl-xL,
and bax mRNA were measured in hippocampal tissue using a
semiquantitative RT-PCR method. Total RNA, free from chromosomal DNA
contamination, was isolated and reverse transcribed with SUPERSCRIPT II
RNase H-Reverse transcriptase (Life Technologies, Inc., Eggenstein,
Germany) using custom-synthesized (Life Technologies)
oligo-dT12–18 primers for
bcl-xL and bax PCR, and a
sequence-specific antisense primer corresponding to nucleotides
638–657 of the cDNA sequence for bcl-2 PCR
(Table 1
); RT-PCR primers for glyceraldehyde-3-phosphate dehydrogenase (G3PDH)
were purchased from Clontech Labs (Palo Alto, Calif.). Briefly, 4 µg
of total RNA was incubated for 1 h at 42°C in a 20 µl reaction
mixture containing 25 mM of the primer, 10 mM DTT, 0.5 mM of dATP,
dCTP, dGTP, and dTTP, 200 units of reverse transcriptase, and 1x First
Strand Buffer (Life Technologies, Inc.). A 0.5 µl aliquot of reverse
transcribed cDNA was subjected to PCR in a 50 µl reaction mixture
containing 3' and 5' primers (each at 2 µM), 200 µM of each dNTP,
two units of AmpliTaq DNA Polymerase (Perkin Elmer, Überlingen,
Germany), and 1x GeneAmp buffer (Perkin Elmer). The reaction mixture
was overlayered with mineral oil and PCR amplification was conducted in
a Perkin-Elmer thermocycler for 21 cycles
(bcl-xL), 22 cycles (bax), 26
cycles (bcl-2), or 17 cycles (G3PDH) using the following
conditions: denaturation 40 s 94°C; annealing 40 s 53°C
(bcl-xL), 55°C (bcl-2), or
56°C (bax and G3PDH); primer extension 2 min. 72°C. PCR
products were electrophoresed on agarose gels and visualized under UV
light using SYBR Green DNA stain (Biozym, Hessisch Oldendorf, Germany).
Relative differences in the amount of amplified PCR products between
treatment groups were assessed semiquantitatively by radioactive
(32P) dot blot hybridizations with gene-specific
antisense cDNA probes. Blot intensities (arbitrary optical density
units) were measured using a Bio-Imaging Analyzer system (Fuji)
combined with densitometry. Intensities for amplified bax,
bcl-xL, and bcl-2 were
normalized against those obtained for G3PDH in the same sample. The
linearity of PCR amplification was verified by amplifying serial
dilutions of reverse transcribed cDNA.
|
Western blot assays
Frozen hippocampi were processed according to the previously
described procedures (47)
, and equal amounts of protein
were subjected to gel electrophoresis and electrophoretically
transferred onto PVDF membranes (Amersham, Braunschweig, Germany).
Blots were washed with Tris-buffered saline, treated with 2% (v/v)
H2O2 in 20% methanol/0.1 M
Tris (pH 7.5) for 30 min, and then preblocked (3 h) in TBST buffer (50
mM Tris, pH 7.5; 190 mM NaCl; 0.1% Tween 20) containing 8% (w/v)
dried nonfat milk (Bio-Rad, Munich, Germany), and 1% normal donkey
serum. The preblock solution was removed and replaced with fresh
solution containing 0.025–0.1% (v/v) of either an anti-Bax,
anti-Bcl-xL, or anti-Bcl-2 antiserum (PharMingen,
Hamburg, Germany) or a pan-p53 antibody (Oncogene Research, Cambridge,
Mass.). After overnight incubation at 4°C, blots were washed in TBST
buffer before incubation (1 h, room temperature) with 0.1 µg/ml of
horseradish peroxidase-labeled donkey anti-rabbit IgG (Amersham) in
TBST containing 5% milk and 0.5% normal rat serum. After washing with
TBST, specific protein bands were detected using a chemiluminescent
detection system (ECL-Plus, Amersham).
Data analysis
Statistical analysis was carried out using the computed values
for each individual animal. Data were scrutinized by one-way analysis
of variance and appropriate post hoc tests; comparisons were
made with respect to age-matched controls and treatment paradigm. The
level of significance was preset at P 
0.05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
); no effects were observed in any other hippocampal subfield.
Specifically, exposure to CORT led to a reduction of TUNEL-positive
(apoptotic) cells within the dentate gyrus compared to saline-treated
animals (P<0.05), whereas treatment with the GR specific
ligand DEX significantly increased the apoptotic index
(P<0.05). As also shown in Fig. 2
, older animals showed
higher basal levels of apoptosis (P<0.05); the older
animals showed greater vulnerability to the apoptotic effects of GR
occupation. In both age groups, DEX-induced apoptosis occurred
predominantly within the subgranular zone.
|
Expression of genes regulating cell death as a function of age and
differential CS receptor stimulation
Initially, in situ hybridization histochemistry was
performed on selected brain sections from saline-treated control
animals in order to monitor the expression patterns of bax,
bcl-xL, and bcl-2 mRNA in the adult
dentate gyrus. All three genes were found to be expressed in the
granular layer of the dentate gyrus. These findings were used as a
basis for the remainder of the studies in which changes in the
steady-state levels of bax, bcl-xL, and
bcl-2 mRNA were determined by semiquantitative RT-PCR
analysis.
As shown in Fig. 3
, expression of the cell death promoter gene bax was
maintained at a similar level in both young and old saline-treated
control rats. In contrast, there was a striking age-related decrease
(P<0.05) in the expression of the cell death inhibitor gene
bcl-2, with older animals showing ~50% fewer
bcl-2 transcripts than younger animals. The older animals
showed a significantly lower number of
bcl-xL transcripts than their younger
counterparts; the splice-variant bcl-xS proved
to be undetectable.
|
Treatment of young and old rats with a dose of CORT sufficient to
preferentially activate MR resulted in a significant
(P<0.05) lowering of bax gene expression
(Fig. 4A
). In contrast, treatment with the GR agonist DEX resulted
in a marked increase in the levels of bax mRNA in both young
and aged rats (Fig. 4A
). The altered patterns of
bax expression after the different hormonal manipulations
were matched by changes in Bax protein concentrations: Western blot
analysis demonstrated that DEX stimulated, whereas CORT inhibited, Bax
levels (Fig. 5
).
|
|
Although the relative number of bcl-xL mRNA
transcripts were increased (P<0.05) in old rats after
treatment with the preferential MR agonist CORT, young animals showed
no such changes (Fig. 4B
). The changes in the older animals
were accompanied by a reduced incidence of apoptosis after CORT
treatment (cf. Fig. 2
). In stark contrast, the levels of
bcl-xL transcripts were reduced by DEX in
young, but not in old, animals (Fig. 4B
). As in the case of
the bax gene, age is an important contributory factor to the
final response (Fig. 4B
). Observations on the
bcl-xL mRNA levels were paralleled by
changes in the expression of the corresponding protein detected by
Western blot analysis (Fig. 5)
.
The pattern of bcl-2 expression after the various CS
manipulations in young and old rats closely paralleled that observed
for bcl-xL expression. Thus, basal levels
of bcl-2 gene expression were low in young rats and
increased only in the old animals after exposure to CORT
(P<0.05); both age groups responded to the GR agonist DEX
with suppressed expression of the bcl-2 gene (Fig. 4C
). Although Bcl-2 protein could not be measured by Western
blotting, a parallelism between changes in bcl-2 expression
and cell survival (CORT) or death (DEX) was apparent (cf. Figs. 2
, 4C
).
Ratio between expression of pro- and antiapoptotic gene expression
as a function of age and CS treatment
As compared to saline-treated age-matched controls, young and old
rats treated with the preferential MR agonist CORT exhibited reduced
bax:bcl-xL and bax:bcl-2
expression ratios (P<0.05; Fig. 6
), a finding that matches the reduced incidence of apoptosis after this
treatment (Fig. 2)
. Taking these ratios as a measure, the older animals
proved to be slightly more sensitive to CORT than their younger
counterparts. Although it was found (Fig. 2
; also see ref
3
) that younger rats are less sensitive to the
apoptosis-promoting effects of the GR ligand DEX, we found the younger
animals to have an increased bax:bcl-xL
expression ratio (P<0.05; Fig. 6
) as compared to the older
animals. Further, both age groups showed a similar relative increase
(P<0.05; Fig. 6
) in the ratio of bax to
bcl-2 gene expression.
|
Role of Bax in mediating the apoptotic actions of DEX
The mouse bax gene consists of six exons assembled into
multiple alternatively spliced transcripts encoding a 21 kDa membrane
() and two forms of cytosolic protein (ß and )
(31)
. To generate mice lacking all isoforms, we targeted
the first four exons, which encompass most of the protein coding
sequence. The bacterial lacZ gene, under the transcriptional
regulatory sequences of the bax gene, was inserted into the
construct to serve as a genetic marker of the mutation. The targeting
strategy is detailed in Materials and Methods and Fig. 1
.
Wild-type mice treated with the GR agonist DEX showed an abundance of
apoptotic cells in the dentate gyrus (Fig. 7a, b
). Whereas the relative number of apoptotic cells by far
exceeded that found in rats (cf. Fig. 2
), the spatial distribution of
such cells in the two species showed close resemblance. In contrast to
the results obtained in wt controls, bax knockout mice
(bax -/-; Fig. 7c, d
) failed to
respond to DEX treatment with increased apoptosis of dentate neurons.
As shown in Fig. 7e
, mice with only partial knockout of
bax (bax -/+) showed an apoptotic response to
DEX that was lower than that seen in wt animals. These differences in
the occurrence of DEX-induced apoptosis corresponded well with the
amounts of Bax protein found by Western blotting (Fig. 7
, lower
left-hand panel), indicating a relationship between bax gene
dose and incidence and susceptibility to apoptosis.
Age-related differences in CS-mediated changes in p53 expression
Western blot analysis revealed that the tumor suppressor protein
p53 is expressed at very low levels in the hippocampus of
untreated young rats; treatment with the GR agonist DEX caused a marked
increase in p53 (including its phosphorylated forms) protein levels,
whereas treatment with MR-activating doses of CORT resulted in an
almost complete abolition of p53 expression (Fig. 5)
. In contrast,
whereas basal p53 expression in aged rats was higher than that in
younger animals, levels of this protein failed to respond to treatment
with DEX and CORT in a quantitative manner (Fig. 5)
.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
. We found that the preferential MR agonist, CORT, and
the specific GR agonist DEX produced opposite effects on dentate cell
survival. Whereas CORT treatment diminished the number of
TUNEL-positive (apoptotic) cells, treatment with DEX induced a
significant degree of apoptosis. These findings are consistent with our
earlier unbiased stereological observations that GR, but not MR,
activation leads to a reduction in total dentate neuronal numbers
(48)
. The opposing actions of CORT and DEX on apoptosis
applied equally to both young and aged rats. However, the older animals
showed appreciably higher basal levels of apoptosis and proved more
vulnerable to the apoptotic effects of GR occupation. In both age
groups, DEX-induced apoptosis occurred predominantly within the
subgranular zone (the area where granule cells are formed before
migrating into the granule cell layer; see refs 19
, 21
, 94) and in the suprapyramidal blade of the granule cell layer.
In general, the above findings are consistent with results from our
previous experiments where CORT and DEX were administered on a single
occasion (3)
. In that study, CORT simply maintained
apoptosis at a rate equal to that found in controls. In this context,
our present observations are of particular interest since they further
indicate that chronic administration of low, MR-activating doses of
CORT can reverse cell death occurring under basal conditions. Although
we did not identify the specific types of cells that are prevented from
entering apoptosis by CORT(e.g., neurons vs. glia), it is likely that
they represent neurons, as previous studies have shown that CORT is
essential for both granule cell survival and proliferation (7
, 9
, 50)
. Further, the predominant localization of TUNEL-positive
cells in the subgranular zone suggests that either newly born granule
cells, or their immediate precursors, are the target of GR-mediated
apoptosis.
Expression of representative members of the bcl-2
family is influenced by changes in the CS milieu
Influence of age
Our observations on age-related changes in the expression of cell
death inhibitor/promoter genes agree with previous reports
(25
26
27
28
29)
. Specifically, we found that expression levels of
the cell death promoter gene bax were similar in both young
and old saline-treated control rats. In contrast, increasing age was
associated with a striking decrease in the expression of the cell death
inhibitor gene bcl-2; expression of
bcl-xL was also significantly lower in old
vs. young animals.
Differential control of bax, bcl-x, and bcl-2 expression by MR and
GR ligands
Gene targeting experiments have established the importance of Bax
for the occurrence of neuronal death (51
, 52)
; the
death-inducing actions of Bax can be largely antagonized by the
presence of Bcl-2 or Bcl-xL (31
, 53
, 54)
. In the present study, a dose of CORT sufficient to
preferentially activate MR was found to reduce bax gene
expression in the hippocampus of both young and old rats. This finding
may be causally related to the ability of CORT to lower the apoptotic
index in the dentate gyrus and is consistent with reports that MR
agonists promote granule cell survival (7
, 9
, 50)
. In
contrast to the results obtained with CORT, treatment with the GR
agonist DEX led to a marked increase in the steady-state levels of
bax mRNA in both young and aged rats, a result that
correlates well with the increased apoptotic index displayed by
DEX-treated animals. The essential role of Bax in mediating the
apoptotic actions of DEX was demonstrated by the fact that this GR
agonist did not lead to apoptosis in bax -/- mice.
Collectively, the above results are consistent with the proapoptotic
nature of Bax (11
, 3
, 33
, 51)
and indicate that the
expression of this molecule can be regulated by CS in the central
nervous system. In addition, the Bax response (which correlates with
the ultimate decision of a cell to die or not) after a particular
hormonal manipulation seems to be critically dependent on the age of
the subject insofar that the availability of MR and GR varies with age
(5)
.
When CORT was administered at a dose to preferentially activate MR, the
hippocampi of old, but not young, rats showed an increase in the
relative number of bcl-xL mRNA transcripts.
The changes in the older animals were accompanied by a reduced
incidence of apoptosis after CORT treatment, thus conforming to the
existing view of MR agonists as promoters of neuronal survival
(7
, 9
, 50)
. However, an unexpected finding was that the
number of bcl-xL transcripts was reduced by
the GR agonist DEX in the younger animals only. In light of our present
(Fig. 2)
and earlier observations (3)
, demonstrating that
older subjects are more susceptible to the apoptosis-inducing actions
of DEX, the latter finding indicates that the absolute levels of
bcl-xL expression do not correlate directly
with the degree of apoptosis induced by this particular stimulus. Later
we consider the importance of the relative levels of apoptotic vs.
antiapoptotic molecules within the cell that may provide a plausible
explanation for this apparent discrepancy. At this point, it is
pertinent to note, however, that aged animals show significant
reductions in basal CS secretion and MR and GR abundance (3
, 5)
; these alterations may contribute to the present observations
on age-related differences in bcl-xL
expression after treatment with MR and GR agonists. An interesting
feature of our results is that they show that aged subjects have the
potential to respond to neuroprotective (MR agonists) as well as
neurodegenerative (GR agonists) stimuli with increases in the
production of antiapoptotic molecules, and it is highly plausible that
older animals are hypersensitive to MR agonist treatment. The fact that
the neurons of aged rats are more susceptible to death after GR
stimulation suggests that the ‘defensive response’ is nevertheless
inadequate to counteract other death-precipitating factors.
The neuroprotective role of Bcl-2 in the adult hippocampus is best
illustrated by the findings that 1) ischemia results in the
induction of the bcl-2 gene in neurons that survive the
hypoxic insult (CA3 pyramidal and dentate granular cells, cf. CA4
pyramidal cells that eventually die; see refs 24
, 47
, 55
, 56
), and 2) certain areas of the adult brain,
including the dentate gyrus, respond to traumatic injury with increases
in bcl-2 gene transcription and translation
(57)
. A ‘defensive’ role for this protein is also
suggested by the observation that Bcl-2 levels are increased in the
basal ganglia of patients with Parkinson’s disease or incidental Lewy
body disease (58)
. In the present study, bcl-2
gene expression was reduced in young animals and increased in old
animals after treatment with an MR-activating dose of CORT. However,
both age groups responded to the GR agonist DEX with suppressed
expression of the bcl-2 gene. It deserves mention that a
parallelism between changes in bcl-2 expression and cell
survival (CORT) or death (DEX) was evident (cf. Fig. 2
and 4
). The
apparently greater inducibility of bcl-2 in older rats than
in younger ones may be related to the previously mentioned differences
in basal CORT levels in the two age groups, i.e., the preexisting
corticosteroid milieu may exert an important influence on the
regulation of the bcl-2 gene.
Age and CS effects on neuronal apoptosis in the context of the
balance between expression of pro- and antiapoptotic molecules
Although the present experiments revealed a clear correlation
between the extent of apoptosis and expression of the proapoptotic gene
bax, such a simple relationship was not discernible for the
antiapoptotic molecules bcl-xL and
bcl-2. Although there is evidence showing that members of
the Bcl-2 family can also regulate apoptosis independently of
interactions with each other (12
, 59)
, a ‘rheostat
model’ has also been proposed. According to this model, the
competitive dimerization between selective pairs of pro- and
antiapoptotic molecules is a major determinant of whether a cell will
ultimately survive or die, i.e., death will not ensue if sufficient
amounts of repressor proteins are available to neutralize the
death-inducing proteins (11
, 18)
. This prompted us to
reevaluate our data in terms of the ratio of expression of the death
agonist bax to that of the death antagonists
bcl-xL and bcl-2 (cf. ref
31
).
In the present experiments, treatment of either young or old rats with
the MR agonist CORT led to reduced
bax:bcl-xL and bax:bcl-2
expression ratios (Fig. 5)
, findings that match the reduced incidence
of apoptosis after this treatment (Fig. 2)
and support our suggestion
that MR agonists help maintain dentate gyrus cell survival by
inhibiting Bax synthesis. Taking these ratios as a measure, the older
animals proved to be slightly more sensitive to CORT than their younger
counterparts, a fact that may be related to the relatively lower basal
levels of CORT to which older animals are physiologically exposed (see
above). On the other hand, in view of our present (see Fig. 2
) and
earlier (3)
findings that younger rats are less sensitive
to the apoptosis-promoting effects of the GR ligand DEX, it was
interesting to find that DEX caused a relatively greater increase in
the ratio of bax:bcl-xL expression in the
younger vs. older animals and that this treatment caused similar
relative increases in the ratio of bax to bcl-2
gene expression in both age groups.
How are cell death/survival signals evoked by the activation of
GR/MR transduced to regulatory molecules of the cell death machinery?
The results reported here provide the first analysis of how
members of the bcl-2 gene family may mediate CS effects on
neuronal survival and death in the dentate gyrus. The proteins encoded
by the bcl-2 gene family are associated with either
mitochondrial or nuclear membranes, and their effects include changes
in the plasma membrane, redox potentials, protease activities,
distribution of intracellular ions, and mitochondrial permeability
transition phenomena; free radical generation has been implicated to
underlie all these processes (13
, 16
, 60
, 61)
. We recently
found DEX-induced apoptosis in the dentate gyrus to involve the
generation of nitric oxide and possibly other reactive oxygen species
(M. G. Canteros and O. F. X. Almeida, unpublished
observations). However, the mechanisms by which changes in the CS
status regulate bax, bcl-xL, and
bcl-2 remain unknown. Since a GenBank database search failed
to reveal potential MR and GR responsive elements in the published
sequences of any of these genes, it is likely that the observed effects
on their expression are indirect. A very likely candidate for mediation
between CS receptors and Bcl-2 family members is the tumor suppressor
protein p53. There are reports of interactions between GR and p53
(34
, 62
, 63)
, a strong inducer of the death promoter
bax (36
, 37
, 64
, 65)
while at the same time
being a repressor of bcl-2 (36
, 37)
. Our
present data indicating that hippocampal levels of p53 (including its
phosphorylated forms) can be up-regulated by GR agonist treatment
therefore support a role for this molecule in CS-mediated apoptosis.
In addition to p53, interactions between GR and a number of other
transcription factors such as nuclear factor 
B and activator protein
1, which themselves show complex interactions, have been reported
(66
67
68)
. All of these transcription factors have been
implicated in cellular death and survival processes, depending on their
ability to influence the transcription of death inducer or repressor
genes (36
, 37
, 69)
. Another indirect mechanism whereby CS
may interact with Bcl-2 family members could involve the protein Bag-1;
Bag-1 binds to Bcl-2 to promote cell survival, and since Bag-1 is
negatively regulated by DEX (70
, 71)
, this mechanism could
conceivably predispose a cell to apoptosis. Last, the number of
possible permutations (CS x transcription or other factors x apoptotic genes) increase vastly in light of evidence suggesting
that MR and GR may heterodimerize (72)
.
Physiological and therapeutic significance
The data presented in this and our previous study (3)
clearly show dichotomous effects of MR and GR activation on dentate
cell survival. The effect of GR-induced neuronal apoptosis on brain
function can conceivably be of considerable importance, given that the
organism is frequently exposed to GR-activating levels of CS on both an
intermittent (e.g., during stress) and cyclic (during the diurnal rise
in adrenocortical activity) basis. Indeed, hippocampal neuronal loss is
still thought to lead to hypersecretion of CS, which in turn has been
causally related to disorders of mood and cognition (see refs 1
, 4
, 5
, 73
, 74
). These untoward effects on brain function are
exaggerated in aging subjects, which also frequently show endocrine
maladaptations to stress (1
, 73)
. However, the precise
physiological significance of GR-mediated neuronal death remains
unclear. Although previous studies have reported that stress leads to
hippocampal cell death in rodents and primates (75)
,
recent stereological analyses show that chronic unpredictable stress
does not reduce total cell numbers in either Ammon’s horn or dentate
gyrus (8)
. One interesting line that emerges from this and
our previous studies (3
, 5
; Wistar rats used in all
studies) is that basal (MR-selective) levels of CS decline with
increasing age. This alteration in adrenal activity may be a critical
factor in rendering aged subjects more susceptible to the
neurodegenerative processes that follow GR stimulation. Since our
experiments show that aged animals can respond to MR-selective doses of
CORT with increased production of antiapoptotic molecules, additional
studies should be aimed at examining the extent to which therapeutic
correction of MR ligand deficiency may lead to a reduction of
age-related hippocampal cell loss and the mood and cognitive
impairments associated with this phenomenon.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
