Abstract
Exposure of teleost fish to predator stress has been shown to elevate plasma cortisol
levels. Consequently, we examined the effects of predator stress (albino oscars) on
neurogenesis in adult zebrafish. To find out how stress affects adult neurogenesis in the
zebrafish, we subjected zebrafish to acute or chronic predator stress and assessed the
proliferation of newly born cells in the optic tectum and cerebellum. These regions are
known to have measurable rates of cell proliferation. Stereologic analyses of the number
of 5-bromo-2' deoxyuridine (BrdU) -labeled cells revealed that 1 exposure of zebrafish
to a predator (acute stress) did not change cell proliferation in the optic tectum or the
cerebellum. However, three day repeated predator stress (chronic stress) significantly
reduced cell proliferation, by 37% in the optic tectum, while having no significant effect
in the cerebellum. A comparison between non-stressed zebrafish optic tectum and
cerebellum revealed no significant regional difference in cell proliferation rate. These
findings suggest that stressful experiences may down-regulate adult neurogenesis in
certain regions of the zebrafish brain, similar to stress induced neurogenesis effects in
mammals and birds. This study presents the zebrafish as a unique new model for
understanding adult neurogenesis and its underlying molecular mechanisms.
Introduction
Neurogenesis in the adult CNS is an intriguing phenomenon, both because of its
unexpected discovery by the scientific community and its implications for providing
insight into the cellular mechanisms and treatment of depression, PTSD, and other mood
disorders. Teleosts' potential to produce new neurons during adulthood in the central
nervous system was established decades ago (Rahmann, 1968; Johns, 1977; Meyer,
1978). This discovery came in parallel with studies on teleosts' unique ability to
regenerate axons after injury. Most of these investigations were conducted early on
before modern BrdU labeling of dividing cells, in the electric gymnotiform fish.
Around the same time, a number of major discoveries were made in the neurogenesis
field, causing most biologists to focus on non-fish neurogenesis models.
The first histological evidence of neurogenesis in the dentate gyrus region of adult
rat hippocampi was found in 1965 and over the subsequent three decades, adult generated
neurons have become a major focus in stem cell and developmental biology,
pharmacology, and the study of mental illness (Altman and Das, 1965). Neurogenesis in
adult hippocampal cells has also been found in mice and humans, as well as a number of
non-human primates (Kempermann, 1997; Eriksson, 1998; Gould et al., 1997; Gould et
al., 1999). Rats and non-human primates, because of their pharmacological similarities to
humans, became the models of choice for studying neurogenesis and its regulation. In a
short period of time, regulation of neurogenesis in the hippocampus was linked to
numerous stimuli: 1) It is up-regulated by exercise, an enriched environment, learning,
estrogen, and antidepressant drugs (RS Duman, 2001), and 2) It is down-regulated by
stress, glucocorticoids, age, opiates, and excitatory amino acids (Gould et al., 1999;
Gould et al., 2000; Malberg et al., 2000a; Duman et al., 2001). However, despite such
major discoveries, the molecular mechanisms behind neurogenesis and its regulation are
still largely unknown because the current model systems and techniques for studying
neurogenesis are logistically complicated and too insensitive.
In contrast, zebrafish are ideal for examining the phenomenon of adult
neurogenesis on the cellular and molecular levels. Generally, they are a prominent model
because of their short generation time, which has allowed them to become a common
model for genetic analyses, developmental embryology, and cell biology. In regards to
neurogenesis, much is already known about embryonic neurogenesis in zebrafish. Most
importantly, however, a number of excellent protocols for administering and labeling
BrdU in zebrafish have already been developed and it appears BrdU labeling is much less
complicated in zebrafish than in mammals. In mammals, one must administer BrdU via
intraperitoneal injection, while in zebrafish, it's possible to administer BrdU via the gills,
with adequate blood-brain barrier uptake (Byrd, 2001).
Previous stereologic studies have demonstrated that zebrafish produce several
thousand new neurons in the periventricular zone (PVZ) of the optic tectum (TeO), as
well as the olfactory bulb (Byrd, 2001; Mueller and Wullimann, 2002). Other studies
have shown that stress decreases cell proliferation in various cortical and neo-cortical
brain regions (Gould et al., 1997; Gould et al., 1999; Gould et al., 2000). The causative
factors for this stress induced change in cell proliferation are presently unknown, but
evidence has linked it to an adrenal steroid-dependent mechanism (Sapolsky, 1996). In
rats and tree shrews, exposure to predator odor is known to activate the hypothalamic-
pituaitary-adrenal axis, elevating glucocorticoids, in conjunction with decreasing cell
proliferation (Gould et al., 2000). Exposure of zebrafish to a predator also appears be a
natural stressor that leads to elevation of cortisol-glucocorticoid levels for teleosts
(Kagawa, 2000; Kagawa and Mugiya, 2002). To determine whether predator exposure is
alters cell proliferation in the zebrafish brain we examined the numbers of proliferating
cells in the TeO and cerebellum (CB) of the zebrafish after single and multiple exposures
to predators.
Materials & Methods
Animal Treatments
Adult male zebrafish from the Thompson lab were used in the following
experiments. Albino oscar fish were obtained from the local aquarium supplier.
Zebrafish and oscars were acclimated separately to experimental conditions in 10 gallon
plastic tanks with filtration at 68°C. Throughout the acclimation period, fish were fed
pellet or flake food put on a LD 12:12. Fish were not fed on the day(s) of the experiment.
Stress Treatment Protocol
Acute Stress. Six zebrafish were acclimated in a single transparent tank with a mesh
divider, dividing them into 2 groups of 3 fish. The oscar fish were acclimated in an
identical tank, also able to be divided in half. Äfter 2 weeks of acclimation, at 10AM, 3
zebrafish were transferred from their home-tank into the identically divided oscar tank,
separated from the oscars by a mesh divider. The mesh divider allowed for the
circulation of water between both sides of the tank. Following 4 hours of exposure to the
oscars, the zebrafish were removed, placed into BrdU, and then sacrificed. The group of
3 zebrafish not exposed to the oscars was used as a control.
Chronic Stress. Six zebrafish were acclimated in a single transparent tank with a mesh
divider, dividing them into 2 groups of 3 fish. The oscar fish were acclimated in an
identical tank, also capable of being divided in half. After 2 weeks of acclimation, at
IOAM, 3 zebrafish were transferred from their home-tank into the identically divided
oscar tank, separated from the oscars by a mesh divider. The mesh divider allowed for
the circulation of water between both sides of the tank. Following 4 hours of exposure to
the oscars, the zebrafish were removed and placed into BrdU and then back into their
home-tank. This was repeated for two more days and after the third day of oscar
exposure, the fish were sacrificed following BrdU exposure. The other group of
zebrafish was not exposed to the Oscars and sacrificed after the third day of BrdU
exposure to be used as a control.
Labeling New Cells with 5-bromo-2'-deoxyuridine (BrdU)
In both experiments, the zebrafish were placed into BrdU (Sigma, St. Louis, MO)
immediately following oscar exposure to label mitotically active cells. 1% aqueous¬
BrdU was made in distilled water and placed in 60ml aliquots into Petri dishes.
Immediately after oscar exposure, the 3 acutely stressed zebrafish and 3 controls were
each placed into separate Petri dishes for 2 hours of BrdU exposure. To allow for
adequate S-phase incorporation of BrdU, fish were placed back into their home-tank for
1 hour before being sampled. The chronically stressed fish were placed in 60 ml, 1%
BrdU Petri dishes daily for 1 hour, immediately following exposure to the oscar fish.
After their last BrdU exposure, all of the fish were anaesthetized in cold water,
decapitated, and then heads were placed in 4.0% paraformaldehyde for 24 hrs.
Histology and Immunostaining
Following rehydration and dissection, for each brain, 100um-200um cross
sections through the entire brain were cut and bathed in 70% Hanks solution (Zebrafish
Book, Univ of Oregon). It was recorded what division of the brain, telencephalon
(forebrain), mesencephalon (optic tectum), or rhombencephalon (cerebellum) each cross
section was from. For fluorescent immuno-labeling, sections were placed onto well
containing slides, wash 2x in 2N HCl and then denatured at room temperature for 1 hour
in 2N HCl. The sections were then rinsed in 0.1% PBS-Tw and blocked in 5% Normal
Goat Serum + PBS-D-Tw + 1% BSA for 10 minutes. The sections were then rinsed
(PBS-Tw) and incubated overnight at room temperature in mouse monoclonal antibody
raised against BrdU (1:400 dilution in 5% NGS blocking solution, Sigma, St. Louis,
MO). The sections were then rinsed several times and washed 4x, 15 minutes each in
PBS-D-Tw (1% DMSÖ + 0.1% Tween-20). Next, they were incubated in biotinylated
secondary antibody(exact name?) (Sigma, St. Louis, MO) for 7-9 hours. Äfter several
rinses and 4x-15min washes in PBS-D-Tw, sections were individually mounted and
cover-slipped on glass slides in glycerol.
Data Collection
Sections on each slide were optically dissected along the Z axis into 2 um sections
using a Fluoview confocal microscope and software (Olympus). All of the images were
taken in grayscale at similar contrast-LUT settings, with a confocal aperture of 3 on a 40x
UPLFL dry lens. The regions of interest were the TeO, from the periventricular zone
(PVZ) to the outer edge (Fig 1A) and the CB, from the rostral ventricle (RV) to the outer
edge (Fig 1B). For purposes of comparison, images were taken around similar
morphological markers in each section. The general brain region, forebrain, midbrain, or
hindbrain was labeled during section slicing, but the specific brain section was identified
before image collection. To ensure unbiased counting, the digital images of regions of
interest were saved as extended views under a code number, which was not broken until
data analysis was complete.
Quantification of Cell Proliferation & Statistical Analysis
Total cell count for each coded image was manually quantified using Image-Pro
Plus (Silver Spring, Md) and strictly defined cell-criteria based on previous studies of
BrdU labeling of cells (Zupanc and Horschke, 1995; Byrd, 2001; Mueller and
Wullimann, 2002). Each slide was counted twice and an average cell number per slide
was used in statistical analysis. Äfter all slides were assigned cell counts, the code was
broken and total cell count between brain regions in stressed and non-stressed zebrafish
was compared using Student's t tests with one exception: because there were 3 different
brain regions of interest, two one-factor ANOVAs were used to examine cell number
differences between brain regions in controls and experimental fish separately. Because
of small sample size, analyses of the forebrain region were omitted and will be
considered in the discussion of results.
Results
The two regions of interest, TeO and CB were coded and counted at the same
time with no indication of which specific region each image came from. Immediately
following counting, the mean cell number of control TeO was compared to that of the
control CB and no significant difference was detected (t-0.226, P-0.8248, df-12).
Observation of CB slices showed stained nuclei throughout the entire image but at a
relatively higher density near the RV (Fig. 2A-F). The nuclei in tectal slices were in
relatively high-density clusters, generally near the tectal ventricle (Fig. 3A-C).
Acute Stress
Proliferation was examined by sacrificing the animals 2 hours after the BrdU
injection and viewing fluorescently stained nuclei under a confocal microscope. A single
4-hour predator stress exposure has no significant effect (P20.10) on the cell proliferation
rate in the examined regions of the zebrafish brain (Fig. 4). Two hours after BrdU
immersion, zebrafish exposed to predators exhibited the same mean cell number per slice
(100um-200um thick) as control fish. In the TeO, both control and experimental
zebrafish had similar numbers of cells clustered throughout. The control group cell
mean was 29.600 and the experimental group cell mean was 21.186 (t=1.132; P=0.1378;
df=15) (Fig. 4). A similar non-significant difference was seen in the CB. The control
mean cell number was 31.25 and the experimental group cell mean was 26.167(t=0.547;
P=.2996; df-8) (Fig. 4).
Chronic Stress
Proliferation was examined by confocal viewing of brain slices 1.5 hours after a
third BrdU injection. Chronic predator stress exposure of 4-hours per day for 3 days had
varying, region dependent effects on cell proliferation. In the TeO, chronic stress
resulted in a significant decrease in the number of BrdU-labeled nuclei per slice in adult
zebrafish. In the control group, the mean cell number was 39.538 (Fig. 5). In the
experimental group, a cell mean of 24.947 (Fig. 5) reflected a 36.7% decrease (Fig. 6) in
BrdU labeled cells (t=1.883; P-0.034; df-30). Generally, in the TeO, the density of and
total number of cell clusters increased with nearness to the PVZ and decreased toward the
tectal outer edge. In the CB, cells were clustered throughout relatively evenly with a
slight increase in cluster density near the RV.
Discussion
These findings demonstrate that repeated exposure to predatory oscar fish
decreases the number of BrdU-labeled cells in the TeO of adult zebrafish (Fig. 5). The
decrease in labeled cells most likely reflects a decrease of adult CNS precursor cell
proliferation in the TeO. After only 3-days, there was already a 36.7% decrease in
proliferation of cells in TeO, an impressively large number attesting to the extremely
deleterious effects on cell proliferation of chronic perturbations. Although the
invertebrate zebrafish model and vertebrate model are obviously different, it's interesting
to note that similarly to the zebrafish model, a 40% decrease in cell proliferation was
induced when rats were exposed to predator odor (Tanapat et al., 2001). In contrast, such
chronic stress did not affect baseline proliferative activity in the CB, even though
significant levels of cell proliferation were and have been observed in this brain region
for teleosts (Zupanc and Horschke, 1995; Byrd, 2001; Zupanc, 2001). Acute, single
predator exposures of zebrafish to oscar predator fish appear not to have negative effects
on neurogenesis in either ROI examined in this experiment. While the observed cell
mean in the TeO and CB was lower after a single stress, the difference was not
statistically significant.
Functional implications of effects of chronic predator stress on optic tectum
Three days of predator stress significantly reduced the number of proliferating cells in the
TeO, without changing the number of proliferating cells in the CB. However, increased
cortisol levels were measured in plasma of teleosts after predator stress, implying both
regions of the brain were exposed to the same elevated cortisol levels post stress (Kagawa
and Mugiya, 2002). If there is a glucocorticoid(GC)-neurogenesis link, as many previous
studies have suggested, a possible explanation for this observed regional difference is that
the TeO is a brain region of particularly high GC sensitivity (Sapolsky, 1996; Gould et
al., 2000). For example, the TeO may have a higher density of GC receptors or a higher-
affinity cortisol receptor as compared to the CB. Such richness in GC receptors would
make the TeO highly susceptible to insult-induced neuroendagerment. Another
compelling GC-related theory, different from the traditional steroid mechanism, stems
from recent studies that suggest glucocorticoids interact bi-directionally with a number of
enzymes, NMDA and non-NMDA receptors, and secondary hormones (IGF-1) during its
neuronal injury/endangerment cascade (Sapolsky, 1996; Aberg et al., 2000; Anderson et
al., 2002). These studies show that GCs are not always acting alone and on the contrary
may be regulating neurogenesis through a multi-leveled cascade. Moreover, it's possible
that an important receptor or signal in the cascade may be present in the TeO in
significantly higher density than in the CB.
Implications for the significant effects of chronic, but not acute stress
The two major schools in adult neurogenesis studies, the "glucocorticoid school" and
“the dendritic-remodeling school" would make the following arguments in regard to
acute versus chronic stress effect differences:
1) Repeated stress results in higher levels of GC (than a single stress), which
ultimately translates into deleterious effects of GC over the long-term (Sapolsky,
1996).
This theory has recently fallen out of favor since a number of studies have shown that
severe, acute stress (such as PTSD), or sustained stress both result in equally high levels
of GCs (Sapolsky, 1996).
2) After repeated stress exposures, stress induced dendritic atrophy results in
irreversible remodeling of structural plasticity (Pham, 2003).
This theory is gaining support, largely because it allows for the possibility of one or many
additional regulators of neurogenesis, including GCs. Additionally, recent studies have
suggested a role for dendrites in regulation of neurogenesis (paper?). Moreover, while
GC levels decrease shortly after the final the stressor, other irreversible changes in
neuronal plasticity persist (Sapolsky et al, 1985).
Clearly, both of these theories have merit from extensive supporting background
research. However, the mechanism for molecular regulation of neurogenesis is not so
clearly differentiated and probably lies somewhere between the two theories discussed,
which is why it's not yet possible to determine how chronic, but not acute stress down-
regulates TeO neurogenesis in this experiment.
References
Aberg MA, Aberg ND, Hedbacker H, Öscarsson J, Eriksson PS. 2000. Peripheral
infusion of IGF-I selectively induces neurogenesis in the adult rat hippocampus. )
Neurosci 20:2896-2903.
Altman J, Das GD. 1965. Autoradiographic and histological evidence of postnatal
hippocampal neurogenesis in rats. J Comp Neurol 124:319-335.
Anderson MF, Aberg MA, Nilsson M, Eriksson PS. 2002. Insulin-like growth factor-I
and neurogenesis in the adult mammalian brain. Brain Res Dev Brain Res
134:115-122.
Byrd C, Brunjes, P. 2001. Neurogenesis in the Olfactory Bulb of Adult Zebrafish.
Neuroscience 105:793-801.
Duman RS, Nakagawa S, Malberg J. 2001. Regulation of adult neurogenesis by
antidepressant treatment. Neuropsychopharmacology 25:836-844.
Gould E, McEwen BS, Tanapat P, Galea LA, Fuchs E. 1997. Neurogenesis in the dentate
gyrus of the adult tree shrew is regulated by psychosocial stress and NMDA
receptor activation. J Neurosci 17:2492-2498.
Gould E, Reeves AJ, Fallah M, Tanapat P, Gross CG, Fuchs E. 1999. Hippocampal
neurogenesis in adult Old World primates. Proc Natl Acad Sci U S A 96:5263.
5267.
Gould E, Tanapat P, Rydel T, Hastings N. 2000. Regulation of hippocampal neurogenesis
in adulthood. Biol Psychiatry 48:715-720.
Kagawa N, Mugiya Y. 2002. Brain HSP70 mRNA expression is linked with plasma
cortisol levels in goldfish (Carassius auratus) exposed to a potential predator.
Zoolog Sci 19:735-740.
Kagawa NM, Y. 2000. Exposure of Goldfish to Bluegills Enhances Expression of Stress
Protein 70 and mRNA in the Brains and Increases Plamsa Cortisol Levels.
Zoological Science:1061-1066.
Malberg JE, Eisch AJ, Nestler EJ, Duman RS. 2000a. Effects of antidepressant treatment
and learned helplessness training on hippocampal neurogenesis in adult rat. Soc
Neurosci Abstr 26:1044.
Mueller T, Wullimann MF. 2002. BrdU-, neuroD (nrd)- and Hu-studies reveal unusual
non-ventricular neurogenesis in the postembryonic zebrafish forebrain. Mech Dey
117:123-135.
Pham KN, J; Hof, P; McEwen, B. 2003. Repeated restraint stress suppresses
neurogenesis and indcue biphasic PSA-NCAM expression in the adult rat dentate
gyrus. European Journal of Neuroscience 17:879-886.
Sapolsky RM. 1996. Stress, Glucocorticoids, and Damage to the Nervous System: The
Current State of Confusion. Stress 1:1-19.
Sapolsky, R.M.; Krey, L.C.; McEwen, B. 1985. Prolonged glucocorticoid exposure
reduces hippocampal neuron number; implications for aging. J. Neurosci 5: 1222-
1227.
Tanapat P, Hastings NB, Rydel TA, Galea LA, Gould E. 2001. Exposure to fox odor
inhibits cell proliferation in the hippocampus of adult rats via an adrenal
hormone-dependent mechanism. J Comp Neurol 437:496-504.
Zupanc GK. 2001. Adult neurogenesis and neuronal regeneration in the central nervous
system of teleost fish. Brain Behav Evol 58:250-275.
Zupanc GK, Horschke I. 1995. Proliferation zones in the brain of adult gymnotiform fish:
a quantitative mapping study. J Comp Neurol 353:213-233.
Figures
e


0





P
TeO=optc tectum
CCe=corpus œerebelli
LCa=lobus caudalis cerebelli
CC-crista erebellans
0
16

cce
ee es




Fig. 1: Optic tectal (blue) & cerebellar (red) regions of interest
Fig. 2: Example of high cell density near the RV (top row) versu s low cell
density further away from the RV (bottom row).
Mean
Cell#
g 348 C hicheldenshouseceiner Edlone Teum
NStEs
Cerebelum
Optic lectum
Brain Region
Fig. 4. Acute. Stereological estimates of Mean Cell post 1x str ess split by ROl.
Cell:
Mean
Cell a
Fig. 5.
tress split by RC
-No Stress
-Stess
Obærved Decras
-Loss
In Cell Profferaton
from stress
Stress Level (Optic Tectum)
Fig. 6. Stress Induced % Reduction in Cell Proliferation the Op tic Tectum