| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
* INSERM, U862 and University of Bordeaux, Bordeaux, France;
Department of Molecular Neurobiology, University of Groningen, Haren, The Netherlands; and
Department of Neurobiology and Physiology, Northwestern University, Evanston, Illinois, USA
2Correspondence: Centre de Recherche INSERM U862, Physiopathologie de la Plasticité Neuronale, 146 Rue Léo Saignat, 33077 Bordeaux Cedex, France. E-mail: koehl{at}bordeaux.inserm.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
Key Words: running • opioids • cell survival • adult neurogenesis • homeostatic balance
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
, 2)
, and exerts antidepressant effects in mildly to moderately depressed patients (3
4
5)
. This has prompted the development of animal models, and it was shown that voluntary running enhances spatial learning in young and aged mice (6
, 7)
and exerts antidepressant effects in rodent models of depression (8
, 9)
. It was further shown that exercise increases plasticity in the hippocampus (1
, 10)
, a key structure for cognitive function and stress-related pathologies such as depression (11)
.
One aspect of hippocampal plasticity that has received considerable attention is adult neurogenesis: neural progenitors located in the subgranular zone of the adult dentate gyrus (DG) proliferate, differentiate, mature, and integrate with the granular cell layer, where they become functional granular neurons (12)
. This adult neurogenesis has been associated with cognition and emotion (12)
and is hypothesized to mediate, at least in part, the beneficial effects of exercise. Indeed, running enhances hippocampal neurogenesis (6
, 7
, 13
14
15
16)
, which, in turn, has been associated with improved memory and enhanced long-term potentiation (LTP), a physiological model of memory (17)
. Furthermore, the mechanism of antidepressant pharmacotherapy has been linked to increased neurogenesis (8
, 18)
, and recent evidence suggests that antidepressant effects of exercise may be related to improved neurogenesis as well (8
, 18)
.
Altogether, this indicates that increased neurogenesis may be a key factor in the beneficial effects of exercise on cognition and mood. This has prompted research on the mechanisms by which exercise increases adult hippocampal neurogenesis, which are not yet fully understood. One key candidate is the endorphin system, and more specifically the opioid β-endorphin, whose levels are affected by physical exercise. Indeed, investigations over the last 20 years have reported an increase in blood plasma levels of β-endorphin following physical activity (19
20
21
22)
. Beta-endorphin is primarily synthesized in the pituitary gland and cleaved from the proopiomelanocortin (POMC) precursor molecule. It can be released into the circulation or can project into areas of the brain through nerve fibers, where it acts preferentially on µ opioid receptors (MOR) distributed throughout the brain, including the hippocampus (23
, 24)
. Importantly, in vitro studies have shown that hippocampal progenitors release β-endorphin, which can bind to opioid receptors present on these cells (25)
. Furthermore, proliferation of hippocampal progenitors is increased by β-endorphin treatment and MOR activation (26)
, while decreased by blockade of opioid receptors (25
, 27)
.
Altogether, these results suggest that β-endorphin could play a role in cell proliferation in vivo and that it could be involved in the promotion of hippocampal neurogenesis by exercise. To address these issues, we compared cell proliferation, cell survival/cell death, and cell differentiation in sedentary wild-type and β-endorphin-deficient mice, as well as in animals exposed to short-term (10 days) and long-term (5 wk) wheel-running.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
. In brief, β-endorphin-deficient homozygous mice carry a mutant POMC allele (POMC*4) in which the codon for the amino-terminal tyrosine of β-endorphin was changed to a premature translational stop codon, which results in a complete loss of β-endorphin expression (28)
.
For the present study, male and female homozygous wild-type C57BL/6J and β-endorphin-deficient mice (C57BL/6-Pomctm1Low) were acquired from The Jackson Laboratory (Bar Harbor, ME, USA). With these mice, a breeding colony was started. Homozygous β-endorphin-deficient mice, maintained on a C57BL/6J background, were crossed with homozygous wild-types to obtain heterozygous offspring. This offspring was crossed, and the third generation was used in the experiments described hereafter. This breeding method was chosen to prevent the β-endorphin-deficient offspring from gaining more weight than wild-type mice, as has been reported when animals are raised by homozygous β-endorphin-deficient mothers (Jackson Laboratory). Animals were genotyped from DNA samples obtained from mouse tails by polymerase chain reaction (PCR) following the procedure originally described by Rubinstein et al. (28)
.
Only β-endorphin wild-type (WT, n=29) and β-endorphin-deficient (KO, n=29) homozygous male mice of 3 to 4 mo of age were used. They were individually housed for 3 wk in standard cages under a 12 h/12 h light-dark cycle with ad libitum access to food and water, after which they were divided into nonrunners (NR) and runners (R). Animals from both groups were transferred into individual cages equipped with running wheels (diameter: 12.5 cm) that were locked by a rod for the nonrunners. Running distance (in kilometers) was calculated from computerized recordings of wheel rotations (Chronobiology Kit, Stanford Software Systems, Stanford, CA, USA). All experimental protocols were approved by the appropriate institutional animal care and use committees at our institute, followed the guidelines in the Public Health Service Guide for the Care and Use of Laboratory Animals, and minimized both the suffering and number of animals used.
5-bromo-2'-deoxyuridine injections
Mice were given daily intraperitoneal 5-bromo-2'-deoxyuridine (BrdU) injections (50 mg/kg in 10 ml NaCl; Sigma, St. Louis, MO, USA) at the beginning of the active phase for 9 days starting the day after exposure to the running wheels.
Histological procedure
Mice were sacrificed either 1 day or 4 wk after the last BrdU injection. They were deeply anesthetized with pentobarbital and perfused transcardially with 40 ml of heparinized phosphate-buffered saline (PBS, pH=7.3) followed by 30 ml of 4% paraformaldehyde in 0.1 M PBS (pH=7.3). Brains were postfixed in paraformaldehyde, and sequential 30-µm coronal sections were cut on a vibratome. For immunoperoxidase labeling of BrdU, phosphorylated histone H3 (pH3), Ki67, and doublecortin (DCX), 1 in 10 free-floating sections were processed as described previously (29)
, using a rat monoclonal anti-BrdU (1:2000; Accurate Scientific, Westbury, NY, USA), a rabbit polyclonal anti-pH3 (1:2000; Upstate, Lake Placid, NY, USA), a rabbit polyclonal anti-Ki67 (1/1000, Novocastra, Newcastle, UK), or a goat polyclonal anti-DCX (1:1000; Santa Cruz Biotechnology, Santa Cruz, CA, USA), visualized with the biotin-streptavidin technique (ABCkit, DAKO, Copenhagen, Denmark) with 3,3-diaminobenzidine as chromogen.
Quantitative evaluation of staining
The number of immunoreactive (IR) cells in the DG was estimated using the optical fractionator method, in which every tenth section through the rostral/caudal extent of the DG was examined (bregma –1.3 to –3.8 mm). Numbers of BrdU-marked cells were categorized according to their localization into the GCL, including the subgranular zone (SGZ), defined as a two nucleus-wide band below the apparent border between the GCL and the hilus, or into the hilus proper, as defined previously (30
, 31)
. All cells, excluding those in the uppermost focal plane, were counted at x1000. Cell death was quantified on a thionin-counterstained series by counting the number of pyknotic cells, characterized by a condensed nucleus of smaller size, and of karyorrhexic cells displaying chromatin clumps (32)
. The number of granular cells was determined on the counterstained sections using the optical fractionator method (Stereo Investigator software, Microbrightfield, Williston, VT, USA). For each 1-in-10 section, granular cells were counted in 10 µm3 optical probes at evenly spaced x-y intervals of 100 by 100 µm. The volume of the granular cell layer (GCL) was determined on counterstained sections at x400 with the StereoInvestigator software. All results are expressed as the total number of cells in the whole DG (left and right hemispheres).
Phenotype analysis
The phenotype of the newly born cells was examined by immunofluorescence double-labeling using a rat anti-BrdU (1:1000; Accurate Scientific) revealed with a Cy3 goat anti-rat antibody (1:1000; Jackson) combined with markers for neurons (rabbit polyclonal anti-calbindin D28K, 1:500; Chemicon) or glia (rabbit polyclonal anti-S100β, 1:2500; Sigma) revealed with an Alexa 488 goat anti-rabbit (1:400 or 1:1000; Molecular Probes, Eugene, OR, USA). The percentage of BrdU-labeled cells expressing calbindin or S100β was determined throughout the DG of 4 to 5 animals per group. For each animal, 50 BrdU-positive cells were randomly selected and analyzed for the coexpression of BrdU and calbindin or BrdU and S100β. Using a confocal microscope (DMR TCS SP2; Leica Microsystems, Rueil-Malmaison, France) equipped with an x63 PL APO oil objective, each BrdU-positive cell was analyzed in its entire z axis using 1-µm intervals.
General procedures
The experimental design of our study was modeled after the initial reports of running-enhanced neurogenesis (6
, 13
, 33)
.
First experiment: impact of short-term running
The aim of this experiment was to study the influence of short-term running on cell proliferation (Fig. 1
). Wild-type (WT) and knockout (KO) mice were individually housed for 9 days in cages equipped with a running wheel. Four groups were thus created: WT nonrunners (WT-NR, n=9), WT runners (WT-R, n=6), KO nonrunners (KO-NR, n=7), and KO runners (KO-R, n=8). Animals were injected daily with BrdU and were sacrificed on day 10. Adrenal glands were collected at the time of sacrifice in order to evaluate the effects of short-term exercise on the hypothalamic-pituitary-adrenal (HPA) axis activity (34
35
36
37)
. Cell proliferation was assessed by measuring the number of BrdU-labeled cells and the numbers of phosphorylated histone H3 (pH3) and Ki67-labeled cells. Doublecortin (DCX) was used to evaluate the number of immature neurons.
|
Second experiment: impact of long-term running
In the next step, we examined the influence of long-term running on cell survival and differentiation (Fig. 1)
. Animals of the different experimental groups (n=7/group) were injected with BrdU as described previously and continued to live in their respective experimental condition for 4 more weeks after the last BrdU injection before sacrifice. This time point was chosen as 4-wk-old newborn cells are mature and express neuronal markers (38)
. Adrenal glands were collected at the time of sacrifice in order to evaluate the effects of long-term exercise on the HPA axis activity. The long-term influence of running was examined on cell proliferation (using Ki67 staining) and on the number of immature neurons (using DCX staining). Neuronal and astrocytic differentiations were assessed by evaluating the percentage of newborn cells expressing the neuronal marker calbindin or S100β, respectively.
Statistical analysis
Data are expressed as mean ± SE values. Whenever two groups were compared, an unpaired t test was applied. Otherwise, a two-way ANOVA (genotypexday or genotypexexercise) was used for comparing the time course of running and the effects of running, respectively, in WT and KO mice. Whenever appropriate, a Newman-Keuls post hoc test was performed.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
Impact of short-term running on cell proliferation
As a first evaluation of cell proliferation, BrdU labeling over 9 days was assessed in mice sacrificed 1 day after the last injection (Fig. 1)
. The BrdU-IR cells were small, irregularly shaped, and found mainly in the subgranular layer of the DG (Fig. 2
A–D). Running increased BrdU cell counts in the GCL of the DG (Fig. 3
A; F1,26=27.72; P<0.001), but only in wild-type animals (genotypexexercise interaction, F1,26=9.62, P<0.005; WT-R>all other groups at P<0.001). The distribution of BrdU-IR cells in all groups differed significantly along the longitudinal axis of the GCL with a peak at bregma –3.10 to –3.40 mm (Fig. 3B
; bregma effect, F8,208=28.78, P<0.0001), but this distribution was unaffected by genotype or exercise (bregmaxgenotype interaction, F8,208=1.48, p=ns; bregmaxexercise interaction, F8,208=1.88, p=ns). To verify that the blockade of running-induced increase in cell proliferation in KO mice was specific for the neurogenic zone of the DG, cells were counted in the hilar region of the DG (Table 1
). Running increased BrdU-IR cell numbers in the hilus of both WT and KO mice, indicating that the blockade of running-induced increase in cell proliferation in KO mice was specific to the GCL and was not due to differences in BrdU availability or metabolism.
|
|
|
Because multiple BrdU injections were performed over a period of 9 days, proliferation and short-term survival of the nascent cells might have been confounded. Therefore, to verify that exercise-induced cell proliferation was indeed blocked in KO mice, we stained for pH3 and Ki67, two endogenous molecules that only label dividing cells at the moment of sacrifice (Fig. 2I, J
). Confirming the results obtained with BrdU staining, the number of both pH3-labeled and Ki67-labeled cells was increased specifically in wild-type runners (Fig. 3C, D
; genotypexexercise interaction, F1,26=4.01 and 4.05 for pH3 and Ki67, respectively; both P<0.05; WT-R different from all other groups at P<0.02 for both markers). As was observed for BrdU, the number of Ki67-immunoreactive cells present in the hilus was increased by exercise in both wild-type and KO mice (Table 1)
.
Impact of short-term running on neurogenesis
To assess the consequence of this blockade of cell proliferation on neurogenesis, the number of immature neurons expressing DCX was determined (39)
. DCX-labeled cells had their soma in the deepest region of the granular cell layer at the interface of the hilus, and their dendrites were mostly observed in the molecular layer (Fig. 2E-H
). As expected, running increased the number of DCX-IR cells (Fig. 3E
; F1,26=25.23, P<0.001). Unexpectedly, the number of new immature neurons was similarly increased in WT-R and KO-R mice (genotype effect, genotypexexercise interaction, both P>0.1). This increased neurogenesis was not accompanied by any modification of the GCL volume (Table 1)
.
Impact of short-term running on cell death
In order to explain why immature DCX-positive neuron production was increased while cell proliferation was not in KO-R mice, we hypothesized that the survival of newborn cells was increased in these mice. To test this hypothesis indirectly, we assessed cell death in the same animals by counting the number of pyknotic and karyorrhexic cells that are characterized by condensed clumps of chromatin (Fig. 2K
). We found a trend for KO-R mice to have less apoptotic cells than KO-NR mice (Fig. 3F
), although the differences were not significant (t(13)=1.46; P=0.16) and were not accompanied by any modification of cell number in the granular cell layer (KO-NR, cell number=799200±24329; KO-R, cell number=891771±53500; t(13)=–1.498; P>0.1). This result thus prompted us to examine more directly the effect of β-endorphin deficiency on cell survival by sacrificing animals 4 wk after the last BrdU injection (Fig. 1)
.
Impact of long-term running on cell proliferation
We first verified in this experiment that the running-induced increase in cell proliferation was still blocked in β-endorphin-deficient mice after 5 wk of exercise. In this long-term running condition, cell proliferation evaluated by Ki67 staining was still enhanced in WT mice while it was unchanged in KO animals (Fig. 4
A; genotypexexercise interaction, F1,24=7.30, P<0.05, WT-R different from all other groups at P<0.02). Similarly to what was observed after a short period of running, cell proliferation was enhanced in the hilus of long-term runners whatever their genotype (Table 1)
, further indicating that the blockade of exercise-induced cell proliferation in β-endorphin-deficient mice is restricted to the neurogenic zone of the hippocampus.
|
Impact of long-term running on neurogenesis
We then verified in this batch that neurogenesis was also increased in KO runners despite the lack of increase in cell proliferation. As in the short-term experiment, the number of immature neurons expressing DCX was similarly increased in WT and KO runners compared to sedentary mice after 5 wk of exercise (Fig. 4B
; exercise effect, F1,24=54.84, P<0.0001; genotype effect, genotypexexercise interaction, both P>0.2).
Given that running-induced neurogenesis was not affected in KO mice, we hypothesized that an increase in cell survival and/or differentiation toward a neuronal phenotype may compensate for the absence of running-induced increase in cell proliferation.
Impact of long-term running on cell survival
We examined the survival of 1-month-old BrdU-IR cells. We found that exercise increased the number of 1-month-old BrdU-IR cells, independently of the genotype (Fig. 4C
; exercise effect, F1,24=25.41, P<0.0001; genotype effect, genotypexexercise interaction, both P>0.3). To estimate the rate of cell survival, we then compared the number of 1-month-old BrdU-IR cells to the number of "proliferating" BrdU-IR cells labeled in the first experiment. Thus, whereas 
30% of BrdU-IR cells survived in sedentary mice (WT, 26%; KO, 34%), running increased this survival rate to 53% in WT mice and to 100% in KO mice (exercise effect, F1,24=17.70, P<0.001; genotypexexercise interaction, F1,24=3.50, P=0.07; KO-R>WT-R=WT-NR=KO-NR).
Impact of long-term running on cell differentiation
Finally, to determine whether β-endorphin deficiency influenced cell fate in sedentary and running animals, we phenotyped the 1-mo-old BrdU-labeled cells (Fig. 2L, M
). There was no effect of running or genotype on cell differentiation toward a neuronal fate (Fig. 4D
; exercise effect, genotype effect, exercisexgenotype interaction, all P>0.5). However, running decreased the astrocytic fate in both genotypes (Fig. 4E
; exercise effect, F1,16=9.30, P<0.01; genotype effect, genotypexexercise interaction, both P>0.5). The increase in neurogenesis was accompanied by an increase of the GCL volume (Table 1
; exercise effect, F1,24=3.91, P=0.05; genotype effect, genotypexexercise interaction, all P>0.5).
Impact of short-term and long-term running on HPA axis activity
The activity of the HPA axis, reflected by adrenal weight, was not modified after short-term or long-term running (Table 1)
.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
Effect of running on neurogenesis in wild-type mice
We report here that a 10-day period of voluntary exercise increased cell proliferation in wild-type mice, which is consistent with previous studies (13
, 14
, 16
, 40
, 41)
. When the period of exercise was extended to four additional weeks, this proliferative-promoting effect was still present, as measured by Ki67 immunoreactivity, an endogenous marker of dividing cells. This sustained effect of exercise on proliferation was not observed in other studies reporting that the proproliferative effect of voluntary wheel running was transient (40
, 42)
. In these studies, the lack of increase in cell proliferation observed in long-term runners has been related to the inhibitory effect of corticosterone on cytogenesis (13
, 14
, 16
, 40
, 41)
. Indeed, an increased activity of the HPA axis revealed by hypertrophy of the adrenal glands was observed in long-term (and not short-term) runners (40
, 43)
. In our experiment, neither short-term nor long-term running affected the weight of adrenal glands, which is in accordance with the original study by van Praag et al., reporting that blood corticosterone levels were similar in control and long-term runner mice after 53 days of exercise (6)
. This suggests that under our experimental conditions, the level of activity reached by our mice was not sufficient to be considered as a stressor for the organism and may explain why cell proliferation was still increased in wild-type mice after a 5-wk exposure to running wheels.
We also show here that the increased cell proliferation due to exercise is not associated with any change in apoptosis evaluated by morphological parameters, which complements a previous report using TUNEL as a marker of cell death (44)
. Furthermore, besides a proliferation-promoting effect, exercise increased the rate of survival of cells generated 4 wk earlier, which is in agreement with most studies on voluntary exercise (13
, 42
, 45)
. Finally, we report here that running did not affect neuronal differentiation, a result in accordance with some (41
, 46
, 47)
but not all previous studies (6
, 7
, 13
, 14
, 48)
, and that it decreased astrocytic fate, which is consistent with previous studies (6
, 13
, 48)
. Neuronal and glial fate is thus differentially affected by running, and altogether, wheel running led to a net induction of adult neurogenesis without alteration in the final number of astrocytes in wild-type mice.
Effect of β-endorphin deficiency on basal neurogenesis
Under sedentary, nonexercising conditions, none of the different phases that lead to neurogenesis (cell proliferation, differentiation, and survival/death) was modified in β-endorphin-deficient mice. This is consistent with the fact that this knockout did not lead to modification of endocrine functions and, in particular, stress responses, despite the large database concerning opiate effects in the hypothalamus (28)
. The absence of changes in these neurogenesis-related parameters in sedentary β-endorphin-deficient mice suggests that other factors maintain basal levels of neurogenesis. This is in line with previous studies showing that basal neurogenesis is not modified in FGF-2-deficient mice, whereas an environmentally induced increase in neurogenesis is blocked in these animals (49)
and that blockade of peripheral VEGF has no effect on basal neurogenesis in sedentary mice but prevents the running-induced amplification of neuroblasts (14)
.
Effect of β-endorphin deficiency on running-induced cell proliferation
In contrast to what was observed in sedentary mice, the absence of β-endorphin strongly affected the running-induced increase in cell genesis that was seen in wild-type mice, thus suggesting that β-endorphin released by exercise in wild-type mice is responsible for their increased cell proliferation. Levels of β-endorphin were not measured in the present experiment; however, many studies have previously reported increased levels of β-endorphin following physical activity (19
20
21
22)
. Furthermore, the level and duration of running of our exercising mice were similar to those reported in studies in which elevated levels of β-endorphin were observed as a function of running wheel activity (20
, 50)
.
Given that running has addictive properties (51)
and that the opioid system has been involved in the motivation to work for reinforcers (52)
, we considered that this system may also regulate the level of activity of the animals, and we thus verified that β-endorphin deficiency did not affect wheel running behavior. We found that the average distance per day was not different between wild-type and knockout mice. This is in accordance with a study showing that selected lines of high runners and low runners do not differ in response to naloxone and naltrexone with regard to wheel running, indicating that the opioid system probably does not play a key role in determining levels of voluntary physical activity (53)
.
One important outcome of our experiments was that the proliferation-promoting effect of exercise was completely blocked by β-endorphin deficiency as evidenced by the blockade of running-induced increase in cells expressing pH3, Ki67, or BrdU in animals sacrificed 1 day after the last BrdU pulse. Thus, these data suggest that β-endorphin is an important mediator of exercised-induced stimulation of hippocampal cell proliferation in vivo.
This running-induced and β-endorphin-mediated increase in cell proliferation may involve various different pathways. One possibility is that β-endorphin acts directly on the progenitor cell population. In support of this hypothesis, adult hippocampal progenitors were found to express MOR (25)
. Furthermore, blockade or activation of these receptors was found to decrease or increase, respectively, the number of mitotic progenitors in vitro (26)
, and in vivo (acute naltrexone treatment was reported to suppress cell proliferation in rats) (27)
. This proliferative direct action of the preferential receptor of β-endorphin is thus in accordance with our results. Alternatively, β-endorphin may modify cell proliferation in running animals via regulation of the GABA system. Indeed, GABAergic interneurons within the DG express high levels of MOR (54
, 55)
, and activation of these receptors strongly hyperpolarizes GABAergic cells, leading to an inhibition of GABA release (56
, 57)
. The lack of MOR activation in KO-R compared to WT-R may thus be associated with elevated GABA release in KO mice. Given that activation of GABAA receptors inhibits cell proliferation (58)
, the increased GABA release in KO mice may thus lead to a decrease in cell proliferation.
Effect of β-endorphin deficiency on running-induced neurogenesis
Unexpectedly, the numbers of DCX-labeled immature neurons were similar in wild-type and β-endorphin deficient running mice. There are several possible explanations for the lack of effect of β-endorphin deficiency on neurogenesis: 1) a compensatory proliferative response of the already existing DCX newborn cells, 2) a delayed compensatory increase in cell proliferation, 3) a compensatory increase in the proportion of cells that differentiated into neurons, and 4) a compensatory increased survival of newborn cells. The first two hypothesis can be ruled out given that even after 5 wk of voluntary running, proliferation measured by Ki67 staining was not increased in knockout mice (while still enhanced in wild-type mice). So even if DCX neurons are known to be capable of division (38
, 59
, 60)
, this mechanism does not seem to be at play in the present experiment. Given that cell differentiation was not influenced by β-endorphin deficiency, this cannot be invoked either. It rather seems that the blockade of exercise-induced increase in cell proliferation in knockout mice was compensated for by an increased survival of new cells. This hypothesis is partially confirmed by the increased cell death we found in β-endorphin-deficient runners, although it did not reach statistical significance most probably because of the short window of time in which dying cells can be detected in intact brains.
Conceptually, there are two possible explanations for the increased cell survival in β-endorphin-deficient mice. First, it may be directly related to a blockade of physiological effects of β-endorphin itself, which would indicate dissociation in the effects of β-endorphin on cell proliferation and cell survival. Such differential effects on one phase of neurogenesis has already been reported, for instance, when comparing the impact of enriched environment to that of voluntary exercise; indeed, enriched environment, which includes the usage of a wheel, is known to affect only cell survival whereas voluntary exercise by itself increases both cell proliferation and cell survival (for review, see ref. 61
). Second, this increased survival may be an indirect consequence, i.e., a response to the attenuated running-induced cell proliferation.
With respect to possible direct actions of β-endorphin, it might be that in physiological conditions, exercise-induced β-endorphin release may inhibit factors up-regulating survival or, alternatively, increase factors down-regulating cell survival. Interestingly, it was recently reported that knocking-out MOR enhances the survival of maturing neurons in adult hippocampal neurogenesis (62)
, which is in agreement with our results. Although the mechanisms for MOR-mediated down-regulation of cell survival are still hypothetical, acetylcholine (ACh), and norepinephrine (NE) appear as potential candidates since MOR agonists inhibit release of NE (63)
and ACh (64)
, which have both been implicated in the survival of adult-born hippocampal cells (65
, 66)
.
Alternatively, the increased survival of new cells in the exercising knockout mice may not be a direct effect of β-endorphin deficiency itself but, instead, a secondary and compensatory response to the attenuated cell proliferation in these animals. The condition of increased physical activity is normally associated with increased neurogenesis, as seen in the wild-type mice. If the latter increase is endangered by preventing the normal increase in cell proliferation, other mechanisms may become active to enhance survival of new cells and thereby fulfill the increased demand for new cells in the exercising brain. In this case, the production of new neurons in the hippocampus is homeostatically regulated, but in the exercising brain it is done so at a higher level than in the sedentary brain. Such a homeostatic regulation of new neurons is indeed supported by other studies. For example, mice deficient for mCD24-mouse cluster of differentiation 24, a glycophosphatidylinositol-anchored highly glycosylated molecule that is expressed on differentiating neurons during development- display an overproliferation, which is compensated for by an increased cell death (67)
. Reciprocally, enhanced survival of new granular cells was reported after hampering cell proliferation with methylazoxymethanol (68)
. Importantly, functional consequences have recently been attributed to this homeostatic control of neurogenesis, in particular, in the regulation of learning processes. Thus it has been shown that spatial learning is based both on the addition and removal of new neurons (32
, 69)
and that the blockade of one of this step (cell death) impairs learning-induced homeostatic regulation (i.e., learning-induced cell survival and proliferation) and leads to memory deficits (32)
. Altogether, these data imply that cell proliferation, cell survival, and cell death are interrelated events and belong to a homeostatic cascade regulating the final number of new neurons in the adult brain.
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
).
|
The relevance of this finding lies in the therapeutic potential of our knowledge about the ways exercise helps maintain brain health, and more specifically about the ways exercise can increase memory function and alleviate symptoms of depression.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication October 2, 2007. Accepted for publication January 17, 2008.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. J. Creer, C. Romberg, L. M. Saksida, H. van Praag, and T. J. Bussey Running enhances spatial pattern separation in mice PNAS, February 2, 2010; 107(5): 2367 - 2372. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. S. Yang, M. H. Vitaterna, A. D. Laposky, K. Shimomura, and F. W. Turek Genetic analysis of daily physical activity using a mouse chromosome substitution strain Physiol Genomics, September 9, 2009; 39(1): 47 - 55. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Haskell-Luevano, J. W. Schaub, A. Andreasen, K. R. Haskell, M. C. Moore, L. M. Koerper, F. Rouzaud, H. V. Baker, W. J. Millard, G. Walter, et al. Voluntary exercise prevents the obese and diabetic metabolic syndrome of the melanocortin-4 receptor knockout mouse FASEB J, February 1, 2009; 23(2): 642 - 655. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
