MMC 3
MMC 3
MMC 3
Correspondence
[email protected] (S.-T.H.),
[email protected] (G.K.),
[email protected] (T.L.W.)
In brief
Exercise increases new neuron
generation in the hippocampus, a brain
region with important learning and
memory functions. Here, Leiter et al.
report that this process is mediated by an
increase in systemic selenium transport.
Dietary selenium supplementation
restored neurogenesis and reversed
cognitive decline in aging and in a model
of hippocampal injury, suggesting
potential therapeutic relevance.
Highlights
d Selenium mediates the exercise-induced increase in adult
hippocampal neurogenesis
Article
Selenium mediates exercise-induced adult
neurogenesis and reverses learning deficits
induced by hippocampal injury and aging
Odette Leiter,1,2,3,8 Zhan Zhuo,1,4,8 Ruslan Rust,2,3,10 Joanna M. Wasielewska,2,3 Lisa Grönnert,2,3,11 Susann Kowal,2,3
Rupert W. Overall,2,3 Vijay S. Adusumilli,2,3 Daniel G. Blackmore,1 Adam Southon,5 Katherine Ganio,6
Christopher A. McDevitt,6 Nicole Rund,2,3 David Brici,1 Imesh Aththanayake Mudiyan,1 Alexander M. Sykes,7
€ nker,2,3 Sara Zocher,2,3 Scott Ayton,5 Ashley I. Bush,5 Perry F. Bartlett,1 Sheng-Tao Hou,4,*
Annette E. Ru
Gerd Kempermann,2,3,9,* and Tara L. Walker1,2,3,9,12,*
1Queensland Brain Institute, The University of Queensland, Brisbane, Australia
2CRTD – Center for Regenerative Therapies Dresden, Technische Universita €t Dresden, Dresden, Germany
3German Center for Neurodegenerative Diseases (DZNE) Dresden, Dresden, Germany
4Brain Research Centre and Department of Biology, Southern University of Science and Technology, Shenzhen, China
5The Melbourne Dementia Research Centre, The Florey Institute of Neuroscience and Mental Health, University of Melbourne, Melbourne,
Australia
6Department of Microbiology and Immunology, The Peter Doherty Institute for Infection and Immunity, The University of Melbourne,
Melbourne, Australia
7Max Planck Institute for Molecular Cell Biology and Genetics, Dresden, Germany
8These authors contributed equally
9These authors contributed equally
10Present address: Institute for Regenerative Medicine (IREM), University of Zurich, Zurich, Switzerland
11Present address: Pharmaceutical Research and Early Development, Ocular Technologies, I2O, Roche Innovation Center Basel, F.
SUMMARY
Although the neurogenesis-enhancing effects of exercise have been extensively studied, the molecular
mechanisms underlying this response remain unclear. Here, we propose that this is mediated by the exer-
cise-induced systemic release of the antioxidant selenium transport protein, selenoprotein P (SEPP1). Using
knockout mouse models, we confirmed that SEPP1 and its receptor low-density lipoprotein receptor-related
protein 8 (LRP8) are required for the exercise-induced increase in adult hippocampal neurogenesis. In vivo
selenium infusion increased hippocampal neural precursor cell (NPC) proliferation and adult neurogenesis.
Mimicking the effect of exercise through dietary selenium supplementation restored neurogenesis and
reversed the cognitive decline associated with aging and hippocampal injury, suggesting potential therapeu-
tic relevance. These results provide a molecular mechanism linking exercise-induced changes in the sys-
temic environment to the activation of quiescent hippocampal NPCs and their subsequent recruitment
into the neurogenic trajectory.
INTRODUCTION curs at a mechanistic level. Although the two major adult neural
stem cell niches in the hippocampal dentate gyrus (DG) and the
Adult hippocampal neurogenesis is an evolutionarily conserved subventricular zone (SVZ) of the lateral ventricles share many
process that provides a particular type of structural plasticity common features, it is becoming increasingly clear that they
to the brain, allowing life-long flexibility that supports effective are under distinct regulatory control. One key difference be-
learning and memory. Although there is strong evidence that tween the neural precursor cells (NPCs) of these two neurogenic
this process is found in humans where it is thought to underlie niches is their susceptibility to the induction of proliferation in
cognitive processes critical for adaptive behavior (Boldrini response to physiological stimuli, such as physical exercise
et al., 2018; Moreno-Jiménez et al., 2019; Spalding et al., (Adusumilli et al., 2021; Brown et al., 2003). The broad range of
2013; Tobin et al., 2019), we rely on animal studies to learn effects of exercise across the body makes it likely that systemi-
how the activity-dependent regulation of adult neurogenesis oc- cally released factors in the blood could serve as systemic
408 Cell Metabolism 34, 408–423, March 1, 2022 ª 2022 The Authors. Published by Elsevier Inc.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
ll
Article OPEN ACCESS
mediators of the neurogenesis-promoting effect. In support of generated from primary DG cells grown in the presence of 0.1
this, a communication route via vessel-associated apical pro- or 0.5 mM sodium selenite, compared with cells grown in the
cesses of radial glia-like stem cells has recently been described, absence of selenium (Figure 1C). Selenium supplementation
which facilitates the direct access of systemic blood-borne fac- also significantly increased the number of neurospheres gener-
tors exclusively into hippocampal neural stem cells (Licht ated from primary cells isolated from the SVZ, the other major
et al., 2020). neurogenic region in the adult brain (Figure 1D). Neurosphere
In this study, we sought to identify systemic mechanisms by size is indicative of the proliferative potential of individual precur-
which exercise regulates adult hippocampal neurogenesis. As sor cells, with larger neurospheres being generated from earlier
part of a previous study to identify proteins that are systemically precursors and small neurospheres being derived from more
released in response to exercise, we performed a proteomic restricted progenitor cells (Walker et al., 2008). In addition to
screen on the blood plasma from mice that were housed in either increasing the neurosphere number, a significant increase in
standard cages or with a running wheel for 4 days (Leiter et al., the average neurosphere size and a greater prevalence of larger
2019). We identified 68 proteins with significant running-induced neurospheres (>150 mm in diameter) was observed following se-
changes in their plasma levels. Of these, selenoprotein P lenium supplementation (0.1 mM in the DG and between 0.05 and
(SEPP1) was one of the most significantly upregulated proteins, 1 mM in the SVZ; Figures 1E and 1F). Culturing cells in the pres-
present in the plasma of running mice at more than twice the ence of the organic form of selenium, seleno-L-methionine, pro-
level of controls (Figure 1A). Twenty-five mammalian selenopro- duced similar results, increasing both the number and size of the
teins, all of which contain selenocysteine residues, have been neurospheres generated (Figures 1G–1J). The effect of selenium
identified. However, among these SEPP1 is the most important supplementation was direct, as precursor cells isolated using a
for maintaining selenium levels in the brain. Unlike other seleno- lysophosphatidic acid receptor (LPA1-GFP) reporter mouse
proteins that carry single selenocysteines, SEPP1 carries 10, one line, which can be used to efficiently isolate a pure population
at its N-terminal domain with redox activity and nine at its C-ter- of hippocampal precursor cells (Walker et al., 2016), showed a
minal domain, which supply selenium to the brain via its receptor similar response in the absence of other niche cells (Figures 1K
low-density lipoprotein preceptor-related protein 8 (LRP8) at the and 1L). The total number of NPCs in the hippocampal stem
blood-brain barrier (Hill et al., 2007; Kurokawa et al., 2014; Burk cell niche is a result of both the direct control of cell proliferation
et al., 2014). The activity-dependent increase in blood SEPP1 and the controlled elimination of daughter cells, many of which
levels suggested a potential role of selenium in the activation die within the first 4 days after division of the mother cell. This
of quiescent NPCs. cell death occurs in two waves, an early phase 1–4 days after
Here, we investigated whether a SEPP1-mediated increase in cell birth and a later phase at 1–3 weeks. We first confirmed
selenium transport underlies the exercise-induced increase in that, although apoptosis does occur, selenium treatment did
adult hippocampal neurogenesis. We also determined whether not alter the percentage of apoptotic (annexinV+) NPCs in either
mimicking the effect of exercise by dietary selenium supplemen- adherent hippocampal precursor cultures (Figure S1A) or pri-
tation could restore neurogenesis and reverse the cognitive mary DG cells (Figure S1B). To confirm that selenium treatment
decline associated with aging and hippocampal injury. was not toxic at the concentrations used in this study, we next
used the Resazurin cell viability assay to quantify the number
RESULTS of viable adherent NPCs following treatment with 5 nM to 1 mM
sodium selenite or seleno-L-methionine. This analysis revealed
Selenium increases NPC proliferation ex vivo a small but significant increase in the number of viable cells in
Having previously identified SEPP1 as one of the most signifi- the adherent monolayer cultures supplemented with 0.5 or
cantly upregulated proteins in the blood of exercising mice 1 mM sodium selenite (Figure 1M) or seleno-L-methionine
(Leiter et al., 2019), we first used an ELISA-based approach to (Figure 1N).
confirm that the 4-day running paradigm used in our initial Following the induction of differentiation, we also found that
mass-spectrometry-based proteomic screen resulted in a signif- significantly more bIII-tubulin+ neurons were formed from pri-
icant increase in plasma SEPP1 levels (Figure 1B). We also mary DG-derived neurospheres that were generated under
extended this analysis to reveal that a more acute bout of exer- high-selenium conditions (0.1 and 0.5 mM) compared with those
cise (2-day running or 2 days of running followed by 2 days of generated in the absence of selenium (Figure 1O). Similarly, se-
standard housing) was similarly sufficient to increase plasma lenium treatment increased the number of bIII-tubulin+ neurons
SEPP1 levels (Figure 1B). generated from SVZ-derived neurospheres at selenium concen-
Having established that systemic SEPP1 levels are signifi- trations of 0.05 and 0.1 mM (Figure 1P). Taken together, these
cantly increased following exercise, we next investigated findings indicate that selenium supplementation significantly in-
whether the treatment of NPCs with exogenous selenium also re- creases NPC proliferation and neuronal lineage differentiation
sulted in a subsequent increase in the proliferation of these cells potential without affecting apoptotic cell death.
using the in vitro neurosphere assay. This assay is a readout of
proliferating precursor cells (Reynolds and Weiss, 1992). As Selenium targets activatable NPCs ex vivo
the medium normally used to culture neurospheres from NPCs We have previously identified two populations of precursor cells
contains basal levels (30 nM) of the inorganic form of selenium, that can be activated by either potassium chloride (KCl) or
sodium selenite, we prepared the medium free of sodium sele- norepinephrine (Jhaveri et al., 2010; Walker et al., 2008). As pre-
nite but otherwise as originally described (Reynolds et al., viously described, at basal selenium levels (30 nM), both KCl
1992). We found that significantly more neurospheres were and norepinephrine significantly increased the number of
Exercise increases plasma Sepp1 Figure 1. Selenium acts directly on NPCs ex vivo
(A) A mass-spectrometry-based proteomic screen re-
A B vealed that selenoprotein P (Sepp1) is the most signif-
icantly upregulated protein in the plasma of mice after
4 days of wheel running.
(B) A significant increase in plasma Sepp1 levels after
4 days of exercise was confirmed using ELISA.
(C) Primary DG cells formed significantly more neuro-
spheres when cultured in medium containing either 0.1
or 0.5 mM sodium selenite compared with medium
devoid of selenium.
(D) Primary SVZ cells formed significantly more neu-
rospheres when cultured in medium containing be-
tween 0.05 and 1 mM sodium selenite compared with
C D G H medium devoid of selenium.
(E–J) The addition of selenium resulted in a significant
increase in the average DG (E) and SVZ (F) neurosphere
diameter. Primary DG and SVZ cells both formed more
(G and H) and larger (I and J) neurospheres when
cultured in medium containing increasing concentra-
tions of seleno-L-methionine compared with medium
devoid of selenium.
(K and L) LPA1-GFP+ cells responded directly to sele-
nium treatment in the absence of other niche cells.
(M and N) Hippocampal adherent monolayer cultures
displayed significantly increased proliferation following
E F I J treatment with either sodium selenite (M) or seleno-L-
methionine (N).
(O) Hippocampal neurospheres generated in the pres-
ence of 0.1 or 0.5 mM sodium selenite generated
significantly more bIII-tubulin+ neurons following dif-
ferentiation.
(P) Similarly, sodium selenite increased the neuronal
differentiation potential of SVZ-derived neurospheres.
(Q) Sodium selenite treatment activated the latent
norepinephrine (N)-responsive, but not the KCl (K)-
responsive stem cell population.
One-way ANOVA with Dunnett’s post hoc test (B–D, G,
H, and M–P) or Tukey’s post hoc test (L and Q), Stu-
dent’s t test (A). *p < 0.05, **p < 0.01, ***p < 0.001, and
K L M N ****p < 0.0001.
O P Q
neurospheres generated. When primary DG cells were cultured of selenium treatment was labeled on day 8 using a second
in high levels of sodium selenite, the addition of 15 mM KCl led thymidine analog, 5-iodo-20 -deoxyuridine (IdU). We found
to a further significant increase in the neurosphere number (Fig- that selenium treatment resulted in approximately 50% more
ure 1Q). In contrast, treatment with 10 mM norepinephrine in surviving CldU+ cells 3 weeks after the end of this treatment
addition to high selenium resulted in no further increase in the (Figure 2N). Phenotyping of these cells revealed that the
neurosphere number, suggesting that norepinephrine and sele- majority of the newborn cells had differentiated into NeuN+
nium are activating the same population of precursor cells. neurons (Figure 2O). Short-term selenium infusion, therefore,
resulted in an approximately 55% net increase in neurogene-
Selenium induces NPC proliferation and neurogenesis sis (Figure 2P). The data from the short-term infusion (Fig-
in vivo ure 2B) show that 7 days of selenium infusion increased the
Having shown that selenium increases NPC numbers ex vivo, we number of proliferating NPCs by approximately 300%. In
next sought to determine whether the same effect could be this experiment, we found that labeling this already expanded
observed in vivo. To investigate this, we infused sodium selenite NPC population with IdU after 7 days of selenium treatment,
(1 mM) directly into the hippocampus for 7 days using micro-os- followed by a 3-week chase period during which the labeled
motic pumps and labeled the proliferating cells with the thymi- cells could undergo additional divisions before becoming
dine analog BrdU, 2 h prior to perfusion (Figure 2A). We found post-mitotic, resulted in a greater than 6-fold increase in the
that selenium treatment resulted in a greater than 3-fold increase number of surviving IdU+ cells (Figure 2Q). Together these re-
in the number of proliferating cells in the subgranular zone of the sults indicate that, in addition to its effect on the number of
DG (Figure 2B). In contrast, selenium had no effect on the num- proliferating cells, selenium exerts an additional effect on
ber of proliferating cells in the SVZ (Figure 2C). To confirm this ef- newborn neuron survival.
fect, we performed a second experiment using the same infusion
paradigm but with a lower dose of sodium selenite (200 nM). Selenium preferentially recruits quiescent NPCs in vivo
Again, selenium infusion significantly increased the number of There are several cellular mechanisms that could potentially
proliferating precursor cells in the DG (Figure 2D) with no differ- explain the selenium-mediated increase in neurogenesis. These
ence observed in the SVZ proliferation (Figure 2E). To confirm include increased survival of proliferating cells, cell-cycle short-
that the DG-specific effect was not due to the location of the can- ening, increased number of cell divisions, or recruitment of
nula, we also infused 200 nM sodium selenite directly into the quiescent stem cells (Overall et al., 2016). We have previously
lateral ventricle (Figure 2F). Strikingly, this resulted in a significant shown that the alteration of the cell-cycle length is not sufficient
increase in the number of proliferating precursor cells in the DG to explain the increased number of proliferating precursor cells
(Figure 2G), but not the SVZ (Figure 2H), confirming that selenium observed after exercise (Fischer et al., 2014). We and others
has a specific in vivo effect only on the precursor cells of the DG. have instead proposed that physical activity increases the num-
Finally, we also quantified the number of activated caspase-3+ ber of proliferating precursors by recruiting quiescent, activat-
cells in the DG of the selenium- and saline-infused mice. This re- able stem cells into proliferation (Adusumilli et al., 2021; Lugert
vealed that, although apoptosis did occur, infusion of sodium et al., 2010). To determine whether acute selenium treatment
selenite (200 nM) into the DG for 7 days had no effect on the total causes an increase in proliferation by increasing the survival of
number of apoptotic (activated caspase3+) cells in vivo the already dividing cells or by recruiting quiescent cells, we
(Figure S1C). used a dual thymidine analog (CldU and IdU) labeling paradigm
To determine whether selenium infusion could, in addition to and a single intraperitoneal injection of selenium (Figure 2R). We
increasing the number of proliferating precursor cells, in- found no difference in the percentage of either CldU+IdU cells
crease net neurogenesis, we next labeled proliferating cells (cells that were already proliferating prior to the selenium injec-
with the thymidine analog 5-chloro-20 -deoxyuridine (CldU) tion but had exited the cell cycle prior to the first IdU injection
then infused sodium selenite (200 nM) into the DG for 24 h later) or CldU+IdU+ cells (cells that were already proliferating
28 days (Figure 2I). We found that selenium treatment resulted prior to the selenium injection and remained proliferative 24 h
in a significant increase in net hippocampal neurogenesis later when IdU was injected; Figure 2S). We did, however,
in vivo, with more new neurons (CldU+NeuN+) surviving observe a significant increase in the percentage of CldUIdU+
4 weeks after the start of selenium infusion (Figure 2J). The cells (representing the cells that were non-dividing prior to the
number of proliferating (Ki67+) cells remained increased at selenium treatment and entered proliferation during the 24-h se-
the end of the 28-day infusion period (Figure 2K). Similar to lenium treatment time window; Figures 2S and 2T). To determine
its effects on proliferation, 28 days of selenium infusion did the identity of these CldUIdU+ cells recruited by selenium treat-
not alter net neurogenesis levels in the olfactory bulb to which ment, we phenotyped them using two sets of antibodies to iden-
the SVZ projects via the rostral migratory stream (Figure 2L). tify early progenitors (CldU/IdU/Sox2/Tbr2) and late progenitors
We also investigated whether selenium treatment was (CldU/IdU/Tbr2/DCX). Our results revealed that the majority of
required for the entire 28-day period in order to produce an in- the CldUIdU+ cells were early progenitors that were either
crease in newborn neurons. In this experiment, sodium sele- Sox2+Tbr2, Sox2+Tbr2+, or Sox2Tbr2+ (Figures 2U and 2V).
nite (200 nM) was infused for 7 days, followed by a 3-week Selenium treatment did not significantly alter the proportion of
chase period (Figure 2M). The cells that were proliferating cells within these classes. Together, these findings suggest
prior to the onset of selenium treatment were labeled with that selenium treatment, similar to physical activity, increases
CldU on day 1, immediately after surgery, and the pool of proliferation through the recruitment of quiescent cells into the
proliferating cells that had expanded in response to 7 days neurogenic trajectory.
S T U V
H I J K
Selenium decreases intracellular ROS levels in the ure 3A). We found that treatment with 0.5 mM sodium selenite
hippocampal neurogenic niche significantly reduced the number of DG cells in the high ROS
We have previously shown that fluctuations in the levels of intra- fraction (Figure 3B). Notably, the ROS reduction appeared to
cellular reactive oxygen species (ROS) precede the exercise- be specific for the ROS high cells (and to a lesser extent to those
induced activation of quiescent NPCs (Adusumilli et al., 2021). with medium ROS levels), which we have recently shown to be
Having found that SEPP1 levels are significantly increased in primarily NPCs (Adusumilli et al., 2021). In contrast, treatment
the blood following exercise, and given that selenium and sele- of primary SVZ-derived cells with sodium selenite had no effect
noproteins are known antioxidants, we next investigated on the distribution of the cells among the ROS classes (Figures
whether selenium could reduce the levels of intracellular ROS 3C and 3D). To determine whether in vivo selenium treatment
in our cells. To investigate this, primary DG cells were treated also led to decreases in cellular ROS in the DG, we infused sele-
with 0.5 mM sodium selenite for 16 h, after which intracellular nium directly into the hippocampus for 3 days using mini osmotic
ROS levels were measured by flow cytometry. The gates were pumps, after which we dissected the DG and directly determined
set based on cells cultured without selenium, with the top the median ROS content of these primary cells. Importantly, we
approximately 8% of cells being designated as ‘‘ROS high’’ (Fig- again observed a decrease in intracellular ROS following 1 and
3 days of selenium infusion (Figures 3E and 3F). To confirm that that primary SVZ-derived cells had an over 2-fold higher expres-
selenium supplementation could reduce the levels of intracellular sion of the Sepp1 transcript than the primary cells of the DG (Fig-
ROS specifically in the precursor cell population, we turned to ure 4A), suggesting that SVZ precursor cells partially depend on
the adherent monolayer culture system, which consists of a the baseline selenium provided by SEPP1 and potentially ex-
near pure population of these cells. We found that following plaining why these cells do not respond to exogenously supplied
either 1 or 3 days of selenium treatment in vitro, adherent hippo- selenium. In line with this explanation, we found that primary SVZ
campal precursor cells had significantly lower levels of intracel- cells had significantly lower levels of Lrp8 expression than those
lular ROS than untreated cells (Figure 3G). Although mitochon- of the DG (Figure 4B), strengthening our conclusion that the
drial respiration is a major source of cellular ROS (Zorov et al., NPCs of the SVZ are less responsive to selenium fluctuations.
2014), we have previously shown that exercise does not affect Using the neurosphere assay, we then investigated whether
mitochondrial ROS levels in these cells (Adusumilli et al., 2021). Sepp1 depletion affected the responsiveness of the NPCs to
Similarly, we observed no significant effect of selenium treat- exogenous selenium. As expected, the primary DG cells isolated
ment on mitochondrial ROS levels (Figures 3H and 3I), indicating from wild-type littermate brains were activated by selenium,
that the selenium-mediated ROS changes are due to changes in leading to a significant increase in the number of neurospheres
cellular, not mitochondrial ROS. (Figure 4C). Although the effect size was much smaller, the pri-
mary DG cells of Sepp1 knockout mice were also responsive
The selenium-induced activation of NPCs is Nox2- to selenium activation (Figure 4D). No difference in the size of
dependent the neurospheres generated from Sepp1 wild-type or knockout
NADPH oxidase (Nox) enzymes are the major source of cellular mice was observed (Figure 4E).
ROS in many tissues, including the neurogenic regions of the To investigate whether effective selenium transport to the
brain (Le Belle et al., 2011; Nayernia et al., 2017). We have pre- brain is required to regulate the exercise-induced increase in hip-
viously demonstrated that although Nox2 activity is not required pocampal neurogenesis, we gave Sepp1 and Lrp8 knockout
to maintain baseline levels of adult neurogenesis, cell-autono- mice and their wild-type littermates access to running wheels
mous Nox2-dependent ROS changes are required for the exer- for 7 days and compared them with standard-housed mice.
cise-induced activation of quiescent neural stem cells and the Although they ran similar distances per night (Figure 4F),
subsequent increase in their proliferation (Adusumilli et al., 10 days of wheel running did not increase NPC proliferation in
2021). To determine whether the selenium-induced changes in the Sepp1 knockouts as it did in their wild-type littermates (Fig-
intracellular ROS were dependent on Nox2 activity, neurosphere ure 4G). However, we observed no difference in baseline prolif-
assays were performed on primary DG or SVZ cells isolated from eration (Figure 4H) or the levels of neurogenesis (Figure 4I) in
Nox2/ mice and cultured in a medium containing either 0, 0.03, the hippocampus of Sepp1-deficient mice. Phenotyping of the
or 0.5 mM selenium. Similar to our previous results in wild-type proliferating Ki67+ cells in the Sepp1 knockout and wild-type hip-
mice, significantly more neurospheres were generated in SVZ- pocampus revealed that most of these cells were Sox2+Tbr2+,
derived precursor cell cultures of Nox2/ mice treated with Sox2+Tbr2, or Sox2Tbr2+ NPCs, with less than 3% of the
0.5 mM selenium (Figure 3K). In contrast, the DG-derived precur- proliferating cells being negative for both markers (Figure 4H).
sor cells of the Nox2 mutant mice did not respond to selenium The relative proportion of the cells expressing each of the marker
treatment (Figure 3J), indicating that selenium-induced NPC combinations was not changed by either genotype or exercise.
activation in the DG is Nox2-dependent. Lrp8 knockout mice also ran a similar distance to their wild-
type littermates (Figure 4J). However, in contrast to the Sepp1-
Sepp1-Lrp8-mediated selenium transport is required for deficient mice, Lrp8 knockout mice had a significant reduction
the exercise-induced increase in NPC proliferation in baseline proliferation (Figure 4K) and neurogenesis (Fig-
Having determined that selenium increases neurogenesis spe- ure 4M). Similar to the Sepp1-null mice, the Lrp8 knockout
cifically in the DG through the recruitment of quiescent neural mice did not respond to the running stimulus (Figure 4K) and
stem cells, we next sought to identify the mechanism by which phenotyping of the proliferating cells revealed that neither geno-
peripheral selenium enters the brain. Although there are 25 type nor exercise changed the relative proportion of NPCs within
known selenoproteins, SEPP1 is of particular interest in the each group of mice (Figure 4L). Together, these data support the
context of adult neurogenesis as it is required to transport sele- idea that, in the hippocampus, effective selenium transport is
nium across the blood-brain barrier by binding to its receptor particularly required for the activity-dependent increase in the
LRP8 on brain capillary endothelial cells. In mice, Sepp1 deletion precursor cell number.
leads to a large decrease in brain selenium levels, but only a
moderate decrease in most other tissues, confirming that Selenium compensates for the age-related reduction in
SEPP1 is essential for adequate brain selenium supply (Hill neurogenesis and associated decrease in cognitive
et al., 2003). In addition, both Sepp1 and Lrp8 knockout mice function
develop severe neurodegeneration when exposed to selenium- The results presented earlier demonstrate that selenium in-
deficiency levels that are tolerated by wild-type mice (Burk creases neurogenesis in the young adult (8- to 10-week-old)
et al., 2014). In addition to its function at the blood-brain barrier, hippocampus. However, neural precursor proliferation and neu-
SEPP1-LRP8-mediated selenium transport is also important rogenesis decrease significantly with age (Kuhn et al., 1996),
within the brain for the transport of selenium into cell popula- concomitant with a decline in associated learning and memory
tions. We first examined whether Sepp1 is expressed to the function (Gil-Mohapel et al., 2013). It has previously been
same extent in the two neurogenic niches. Our results revealed shown that the age-related decrease in plasma selenium levels
A B C D E
F G H I
J K L M
Figure 4. Sepp1-Lrp8-mediated selenium transport is required for the exercise-induced increase in NPC proliferation
(A) Expression of the Sepp1 transcript was significantly lower in primary cells of the DG compared with the SVZ.
(B) In contrast, significantly higher levels of Lrp8 expression were observed in the DG compared with the SVZ.
(C and D) Primary DG cells isolated from both wild-type (C) and Sepp1/ (D) mice could be activated by selenium treatment.
(E) No difference in neurosphere diameter between wild-type and Sepp1/ mice was observed.
(F) Sepp1 wild-type and knockout mice ran a similar number of kilometers per night.
(G) Although no deficit in baseline proliferation was observed, Sepp1 knockout mice did not respond to 7 days of wheel running.
(H) Phenotyping of the Ki67+ cells revealed that most of the proliferating cells were either Sox2+Tbr2, Sox2+Tbr2+, or Sox2Tbr2+ NPCs, the relative percentages
of which were not altered by genotype or exercise.
(I) Sepp1 deletion had no effect on the number of DCX+ newborn neurons.
(J) Lrp8 wild-type and knockout mice ran a similar number of kilometers per night.
(K) Lrp8 knockout mice had lower baseline proliferation levels and did not respond to the running stimulus.
(L) Phenotyping of the Ki67+ cells revealed that most of the proliferating cells were either Sox2+Tbr2, Sox2+Tbr2+, or Sox2Tbr2+ NPCs, the relative percentages
of which were not altered by genotype or exercise.
(M) Lrp8 deletion significantly decreased the number of DCX+ newborn neurons.
Student’s t test (A–F, I, J, and M) or one-way ANOVA with Tukey’s post hoc test (G, H, K, and L), *p < 0.05, **p < 0.01, and ***p < 0.001.
in humans positively correlates with the probability of cognitive In a series of pilot experiments, we first investigated various
decline (Akbaraly et al., 2007). We, therefore, next investigated selenium sources, concentrations, and delivery routes to deter-
whether selenium could rescue the dramatic decrease in neu- mine the optimal treatment paradigm. We found that unilateral
rogenesis observed in old (12- and 18-month-old) mice. infusion of sodium selenite into the DG of 12-month-old mice
A B
C D
F G
H I
J K
L M N
O
Figure 6. Dietary selenium supplementation rescues the age-related decline in hippocampus-associated learning and memory
(A and B) Experimental design for the active place avoidance (APA) test. Prior to selenium treatment, mice were pre-tested using the APA task and then divided
into two cohorts with equivalent learning.
(legend continued on next page)
Cell Metabolism 34, 408–423, March 1, 2022 417
ll
OPEN ACCESS Article
We corroborated these results using a second spatial learning APA tests (Figure 7A). Mice that received sham surgery per-
task, the Barnes maze. In this task, the mice were placed in the formed significantly better than lesioned mice 3 days and
center of a brightly lit table with 32 evenly spaced holes around 5 weeks after surgery (Figure S5A). We found that dietary se-
its circumference (Figure 6K). The mice were tested for their abil- leno-L-methionine supplementation, either 4 weeks pre-treat-
ity to use the visual cues placed around the room to locate the ment or beginning immediately following the lesion, exhibited
one hole with an escape chamber located beneath it. As in the no adverse effects on overall health (Figure S5B) and exploratory
APA task, 16-month-old mice were pre-tested for their baseline behavior measured using the open-field test (Figures S5F–S5H)
learning ability using the Barnes maze and divided into two co- but could completely rescue the spatial learning and memory
horts of equal learning ability (Figure S4D). The mice then deficits observed in the APA task 3 days (Figures 7B and 7C)
received either saline or selenium in their drinking water for and 5 weeks (Figures S5C–S5E) post-lesion.
8 weeks after which they were re-tested (Figure 6J). No effect Given that the ET-1-induced hippocampal lesion leads to a
of the selenium treatment on body weight (Figure S4E) or velocity significant loss of DCX+ newborn neurons, we next examined
(Figure S4F) was observed. Similar to the APA task, we found DG-dependent behavioral pattern separation using the novel ob-
that selenium treatment significantly improved Barnes maze per- ject location (NOL) task (Figure 7D). On the habituation day, all six
formance. Although there was no significant difference between groups of mice spent a similar amount of time exploring the two
the selenium-treated and control mice in the primary latency to objects (Figure S6D). Twenty-four hours later, on the test day, the
the goal box or the escape latency, a Sidák post hoc analysis re- ET-1-lesioned mice receiving standard drinking water spent
vealed that only the selenium-treated mice showed a significant significantly less time exploring the object presented in a novel
decrease in their primary latency to the goal location on days 3 location compared with sham mice (Figure 7E), whereas the
and 4 compared with those on day 1 (Figure 6L). In contrast, mice receiving selenium-supplemented water, either immedi-
the primary latency of the control mice was not significantly ately after the lesion or for the previous 4 weeks, performed as
improved over the 4-day testing period. The selenium-treated well as the controls (Figure 7E). There was no difference in the
mice also learned to enter the escape box faster than the control overall exploratory time of the mice between groups on the
mice (Figure 6M). Whereas the control mice displayed no signif- test day (Figure S6E). This result suggests that selenium treat-
icant improvement in the escape latency over the 4-day testing ment rescued the ET-1-induced behavioral pattern separation
period, the selenium-treated mice entered the escape box signif- deficit. In a separate cohort of mice, following the same selenium
icantly faster on days 2, 3, and 4 compared with those on the first treatment and ET-1-induced hippocampal lesion, we tested per-
day (Figure 6M). Further analysis comparing the post-treatment formance in the Y-maze, novel object recognition (NOR) test,
learning with pre-treatment learning revealed an overall signifi- and contextual fear conditioning test to evaluate spatial explora-
cant positive effect of selenium on escape latency (Figures 6N tion behavior, long-term recognition, and contextual memory,
and 6O). Together, these findings indicate that selenium supple- respectively (Figure 7F). The NOR test was performed to assess
mentation can restore age-related deficits in hippocampal recognition memory in mice based on their inherent preference
function. for novelty (Figure 7G). In line with the results of the NOL test,
all six groups of mice spent a similar amount of time exploring
Selenium rescues the hippocampal learning and the two objects on the habituation day (Figure S6F). However,
memory deficits induced by an endothelin-1-induced on the test day, ET-1-lesioned mice spent significantly less
hippocampal lesion time exploring the object in a novel location compared with
Unilateral stereotaxic injection of the vasoconstrictor endothelin- sham mice, suggesting an impairment in long-term recognition
1 (ET-1) into the hippocampal fissure results in a stroke-like memory due to the lesion (Figure 7H). Selenium treatment, start-
lesion that manifests as a significant reduction in the number ing 4 weeks prior to the lesion, successfully rescued this deficit
of DCX+ immature neurons and a decline in cognitive function (Figure 7H). Following the NOR test, mice were assessed using
that can be rescued by physical exercise (Codd et al., 2020). contextual fear conditioning consisting of an 8-min exposure to
To investigate whether dietary selenium supplementation pro- a novel environment, with 1 s, 0.8 mA foot shocks delivered at
vides similar beneficial effects in this ischemic injury model to 2, 4, 6, and 8 min (Figure 7I). Mice from all six groups exhibited
those observed with physiological aging, mice were first pre- similar acquisition of conditioned fear (Figure S6H). However,
tested using the APA task 5 weeks prior to ET-1 lesion to deter- significantly less freezing was observed in ET-1-lesioned mice
mine their baseline learning ability (Figure 7A). All animals receiving standard drinking water compared with sham mice
showed a progressive decrease in the number of shocks during the recall test 24 h after fear conditioning (Figure 7J),
received as they learned over the 5 days of testing (Figure S5A). revealing an impairment in long-term contextual memory. This
The animals were subsequently evenly divided into 6 groups deficit could be rescued by selenium supplementation, with
based on this performance and subjected to two additional the lesioned mice receiving the selenium-supplemented water
(C–I) The selenium-treated mice performed significantly better at the APA task than control mice, with fewer shocks (C–E), and a greater time to both first (F and G)
and second (H and I) entrance into the shock zone observed at day 4.
(J–N) Experimental design for the Barnes maze test (J and K). Selenium treatment significantly improved Barnes maze performance, decreasing primary latency
to the goal location (L) and escape latency (M). A significant positive effect of selenium treatment on the percentage improvement in escape latency between the
pre-treatment and post-treatment learning was observed (N).
Two-way repeated measures (RM) ANOVA with Sidák post hoc test (C, F, H, L, and M) and two-way RM ANOVA (N) (selenium effect F(1,25) = 5.05; p = 0.03). *p <
0.05, **p < 0.01.
and S6N). We confirmed the results mentioned earlier in a sepa- in vivo selenium supplementation increased hippocampal but
rate experiment in which mice were treated as described earlier not SVZ NPC proliferation.
but were assessed using the Barnes maze. We again found that In addition to differential exposure to the systemic environ-
DT-induced ablation of DCX+ newborn cells markedly reduced ment, another factor contributing to the contrasting response
the learning capacity of selenium-treated, ET-1-lesioned mice, of the two NPC populations to exercise is their unique transcrip-
as evidenced by a significant reduction in primary latency to tional signatures (Adusumilli et al., 2021). In contrast to the NPCs
the goal location (Figures 7O and 7P). Together, these results of the SVZ, where the most enriched gene ontology pathways
indicate that, here, selenium is functioning in a neurogenesis- are related to cell division and transcriptional regulation, the
dependent manner. most highly enriched pathway in the DG is redox regulation,
with the NPCs of the DG being maintained in a more highly
oxidized state than those of the SVZ (Adusumilli et al., 2021).
DISCUSSION We have identified redox regulation as the critical mediator of
the exercise-dependent recruitment of quiescent hippocampal
Although the neurogenesis-enhancing effects of exercise have NPCs into the proliferative state (Adusumilli et al., 2021). We sub-
been known for more than two decades (van Praag et al., sequently identified the antioxidant SEPP1 as one of the most
1999a, 1999b), the molecular mechanisms underlying this significantly increased systemic proteins released in response
response have remained largely unclear. Taken together, our to exercise (Leiter et al., 2019). In this study, we found that exog-
data show that an exercise-induced increase in antioxidant sele- enous selenium treatment lowers cellular ROS levels in both pri-
nium transport activates quiescent hippocampal NPCs, resulting mary cells of the DG niche and pure populations of adherent hip-
in increased NPC proliferation and adult neurogenesis. Further- pocampal NPCs and triggers the recruitment of quiescent
more, we show that mimicking the exercise-induced increase hippocampal NPCs into the neurogenic trajectory in vivo. Fitting
in systemic selenium transport by dietary selenium supplemen- with this finding and highlighting the key role of SEPP1 in this
tation can restore neurogenesis and reverse the cognitive process, we also observed that transgenic SEPP1 or LRP8 dele-
decline associated with aging and hippocampal injury. tion abolishes the exercise-induced increase in NPC prolifera-
The hippocampal NPC niche is highly vascularized, allowing tion. We, therefore, propose that exercise-induced SEPP1
for direct contact between the circulatory system and the release contributes to our previously identified ROS-mediated
NPCs. Studies using heterochronic parabiosis, in which the cir- mechanism of exercise-induced activation of quiescent NPCs.
culatory systems of young and old mice are connected, were In contrast to the DG, we found that, under baseline conditions,
the first to suggest the involvement of the systemic environment primary SVZ cells had an over 2-fold higher expression level of
as a central mediator of adult hippocampal neurogenesis and Sepp1 and significantly lower levels of Lrp8 transcript than the
physiological brain aging (Villeda et al., 2011, 2014). More primary cells of the DG. We speculate that the lower cellular
recently, systemic blood factors have been shown to transfer ROS content, mediated by higher basal levels of antioxidant
the beneficial effects of exercise to the aged hippocampus (Hor- SEPP1, may explain why the SVZ precursor cells do not respond
owitz et al., 2020). In their plasma proteomic screen, Horowitz to exogenously supplied selenium (or exercise).
et al. did not identify SEPP1 as being increased in the blood An interesting facet of this issue is the presence of multiple
following exercise. Interestingly, however, another selenopro- subpopulations of non-dividing hippocampal NPCs that are
tein, glutathione peroxidase 3, was identified as one of the responsive to different stimuli (Jhaveri et al., 2015; Walker
most significantly upregulated proteins. This difference is likely et al., 2008). In this study, we found that selenium activates the
due to differences in the age of the mice and the running para- same population of quiescent NPCs as norepinephrine, further
digms used (6 weeks of running in aged mice in the Horowitz characterizing this quiescent NPC population. Our finding that
study compared with 4 days of running in young mice in selenium transport mediates the exercise-induced NPC activa-
our study). tion, therefore, identified the norepinephrine-responsive popula-
In further support of the systemic mediation of adult hippo- tion as the one that is activated in response to exercise. This is in
campal neurogenesis, basement membrane-reduced zones agreement with the earlier finding that GABA, a known inhibitor
located between radial glial-like processes of the hippocampal of NPC activation and exercise-induced neurogenesis (Dumitru
NPCs and endothelial cells have recently been discovered (Licht et al., 2017), selectively inhibits the activation of the norepineph-
et al., 2020), revealing a communication route that allows for rine-responsive but not the KCl-responsive subpopulation of
direct access between the systemic environment and the hippocampal NPCs (Jhaveri et al., 2015).
NPCs of the DG. Our results suggest that exercise increases Selenium is important for maintaining normal brain function
the transport of selenium from the systemic environment to and its deficiency has been linked to a number of age-related
NPCs of the DG by elevating the systemic level of SEPP1, which neurodegenerative disorders, including Alzheimer’s disease,
binds to LRP8 expressed on endothelial cells. Interestingly, a Parkinson’s disease, and Huntington’s disease (Cardoso et al.,
similar route of access between the systemic environment and 2015). Selenium status also declines naturally with age (Akbaraly
the other major neurogenic region of the brain, the SVZ, does et al., 2007). In this study, we showed that selenium supplemen-
not seem to exist (Licht et al., 2020). We propose that this finding tation can rescue the decreased hippocampal neurogenesis and
provides an explanation for why blood components that are associated learning and memory deficits in animal models of
known to be changed by exercise exclusively affect hippocam- physiological aging and ET-1-induced hippocampal lesion,
pal neurogenesis without influencing proliferation in the SVZ. models in which the cognitive deficits have also been shown to
Fitting with this hypothesis, we found that similar to exercise, be rescued by physical exercise (Codd et al., 2020; Horowitz
et al., 2020). It has been reported that, during aging, increased d RESOURCE AVAILABILITY
quiescence rather than stem cell depletion accounts for the B Lead contact
observed decrease in neurogenesis (Kalamakis et al., 2019; Lu- B Materials availability
gert et al., 2010). In support of this, we found that although NPC B Data and code availability
proliferation was greatly reduced in the aged hippocampus, the d EXPERIMENTAL MODEL AND SUBJECT DETAILS
magnitude of selenium-induced activation was greater in older B Mice
mice than young mice. This is similar to our previous observation d METHOD DETAILS
in aged mice following KCl-induced hippocampal NPC activation B Neurosphere culture
(Walker et al., 2008). Determining whether long-term activation B Adherent monolayer cultures
of quiescent NPCs by selenium supplementation (or exercise) re- B Resazurin assay
sults in their depletion or whether these NPCs divide asymmetri- B Immunostaining of neurospheres
cally to continually self-renew following selenium-induced acti- B In vivo infusions
vation will provide insight into the life-long maintenance of B Intracellular ROS measurements
NPC populations in the DG. B MitoSOX measurements
Overall, our finding that selenium metabolism is involved in B BrdU and Ki67 immunohistochemistry and quantifica-
mediating the exercise-induced increase in adult hippocampal tion of proliferation
neurogenesis demonstrates how systemic or environmental fac- B CldU and NeuN fluorescence immunohistochemistry
tors can regulate adult neurogenesis and hence plasticity in the and quantification of neuronal survival
DG. This could have far-reaching implications, as the activity-de- B BrdU and NeuN fluorescence immunohistochemistry
pendency of adult hippocampal neurogenesis is one of its key and quantification of neuronal survival in aged animals
features and central to modern concepts of how adult-generated B CldU and IdU fluorescence immunohistochemistry
neurons provide life-long adaptability to the hippocampus in B Quantitative real-time PCR
both health and disease. The identification of the mechanism un- B Inductively coupled plasma-mass spectrometry
derlying the exercise-induced increase in adult neurogenesis B Hippocampal lesion surgery
could facilitate the discovery of novel therapeutic interventions B Running distance tracking
(including dietary selenium supplementation), which could be B Behavioral testing
used to mimic the beneficial effects of exercise on cognitive B Active place avoidance test
function. Given that selenium is a cheap, readily available dietary B Barnes maze
supplement that is found in a number of commonly eaten foods, B Open field test
such as nuts, grains, and dairy products, it could easily be B Novel object recognition test
boosted in the diet of elderly people. This is particularly impor- B Novel object location test
tant for the treatment of individuals who are unable to exercise B Contextual fear conditioning
due to advanced age, frailty, or disability. B Y-maze
d QUANTIFICATION AND STATISTICAL ANALYSIS
Limitations of study
In this study, we have identified systemic SEPP1 as a mediator of SUPPLEMENTAL INFORMATION
from mouse brain to neural differentiation of patient derived iPSC. Redox Biol van Praag, H., Kempermann, G., and Gage, F.H. (1999b). Running increases
13, 82–93. cell proliferation and neurogenesis in the adult mouse dentate gyrus. Nat.
Overall, R.W., Walker, T.L., Fischer, T.J., Brandt, M.D., and Kempermann, G. Neurosci. 2, 266–270.
(2016). Different mechanisms must be considered to explain the increase in Villeda, S.A., Luo, J., Mosher, K.I., Zou, B., Britschgi, M., Bieri, G., Stan, T.M.,
hippocampal neural precursor cell proliferation by physical activity. Front. Fainberg, N., Ding, Z., Eggel, A., et al. (2011). The ageing systemic milieu nega-
Neurosci. 10, 362. tively regulates neurogenesis and cognitive function. Nature 477, 90–94.
Perez-Riverol, Y., Csordas, A., Bai, J., Bernal-Llinares, M., Hewapathirana, S., Villeda, S.A., Plambeck, K.E., Middeldorp, J., Castellano, J.M., Mosher, K.I.,
Kundu, D.J., Inuganti, A., Griss, J., Mayer, G., Eisenacher, M., et al. (2019). The Luo, J., Smith, L.K., Bieri, G., Lin, K., Berdnik, D., et al. (2014). Young blood re-
PRIDE database and related tools and resources in 2019: improving support verses age-related impairments in cognitive function and synaptic plasticity in
for quantification data. Nucleic Acids Res. 47, D442–D450. mice. Nat. Med. 20, 659–663.
Pollock, J.D., Williams, D.A., Gifford, M.A., Li, L.L., Du, X., Fisherman, J., Orkin, Vukovic, J., Borlikova, G.G., Ruitenberg, M.J., Robinson, G.J., Sullivan, R.K.,
S.H., Doerschuk, C.M., and Dinauer, M.C. (1995). Mouse model of X-linked Walker, T.L., and Bartlett, P.F. (2013). Immature doublecortin-positive hippo-
chronic granulomatous disease, an inherited defect in phagocyte superoxide campal neurons are important for learning but not for remembering.
production. Nat. Genet. 9, 202–209. J. Neurosci. 33, 6603–6613.
Reynolds, B.A., and Weiss, S. (1992). Generation of neurons and astrocytes Walker, T.L., and Kempermann, G. (2014). One mouse, two cultures: isolation
from isolated cells of the adult mammalian central nervous system. Science and culture of adult neural stem cells from the two neurogenic zones of individ-
255, 1707–1710. ual mice. J. Vis. Exp. (84):, e51225.
Reynolds, B.A., Tetzlaff, W., and Weiss, S. (1992). A multipotent EGF-respon- Walker, T.L., White, A., Black, D.M., Wallace, R.H., Sah, P., and Bartlett, P.F.
sive striatal embryonic progenitor cell produces neurons and astrocytes. (2008). Latent stem and progenitor cells in the hippocampus are activated by
J. Neurosci. 12, 4565–4574. neural excitation. J. Neurosci. 28, 5240–5247.
Spalding, K.L., Bergmann, O., Alkass, K., Bernard, S., Salehpour, M., Huttner, Walker, T.L., Overall, R.W., Vogler, S., Sykes, A.M., Ruhwald, S., Lasse, D.,
H.B., Boström, E., Westerlund, I., Vial, C., Buchholz, B.A., et al. (2013). Ichwan, M., Fabel, K., and Kempermann, G. (2016). Lysophosphatidic acid re-
Dynamics of hippocampal neurogenesis in adult humans. Cell 153, ceptor is a functional marker of adult hippocampal precursor cells. Stem Cell
1219–1227. Rep 6, 552–565.
Tobin, M.K., Musaraca, K., Disouky, A., Shetti, A., Bheri, A., Honer, W.G., Kim, Willis, E.F., MacDonald, K.P.A., Nguyen, Q.H., Garrido, A.L., Gillespie, E.R.,
N., Dawe, R.J., Bennett, D.A., Arfanakis, K., and Lazarov, O. (2019). Human Harley, S.B.R., Bartlett, P.F., Schroder, W.A., Yates, A.G., Anthony, D.C.,
hippocampal neurogenesis persists in aged adults and Alzheimer’s disease et al. (2020). Repopulating microglia promote brain repair in an IL-6-dependent
patients. Cell Stem Cell 24, 974–982.e3. manner. Cell 180, 833–846.e16.
Trejo, J.L., Carro, E., and Torres-Aleman, I. (2001). Circulating insulin-like Yamaguchi, M., Saito, H., Suzuki, M., and Mori, K. (2000). Visualization of neu-
growth factor I mediates exercise-induced increases in the number of new rogenesis in the central nervous system using nestin promoter-GFP transgenic
neurons in the adult hippocampus. J. Neurosci. 21, 1628–1634. mice. Neuroreport 11, 1991–1996.
van Praag, H., Christie, B.R., Sejnowski, T.J., and Gage, F.H. (1999a). Running Zorov, D.B., Juhaszova, M., and Sollott, S.J. (2014). Mitochondrial reactive ox-
enhances neurogenesis, learning, and long-term potentiation in mice. Proc. ygen species (ROS) and ROS-induced ROS release. Physiol. Rev. 94,
Natl. Acad. Sci. USA 96, 13427–13431. 909–950.
STAR+METHODS
Continued
REAGENT or RESOURCE SOURCE IDENTIFIER
Neurobasal medium Thermo Fisher Scientific Cat# 21103049
L-(-)-noradrenaline(+)-bitartrate salt Merck Cat# A9512
monohydrate
Laminin Roche Cat# 11243217001
MitoSOX Red mitochondrial superoxide Thermo Fisher Scientific Cat# M36008
indicator
Neo-Clear Merck Cat# 109843
Neo-Mount Merck Cat# 109016
Propidium Iodide Thermo Fisher Scientific Cat# P3566
Poly-D-lysine (PDL) Sigma-Aldrich Cat# P7280
Progesterone Sigma-Aldrich Cat# P0130
Putrescine Sigma-Aldrich Cat# 51799
Potassium chloride Sigma-Aldrich Cat# P9333
Resazurin Sigma-Aldrich Cat# R7017
Sodium selenite Sigma-Aldrich Cat# 214485
Seleno-L-methionine Sigma-Aldrich Cat# S3132
0.05 % Trypsin-EDTA Thermo Fisher Scientific Cat# 25300054
Trypsin inhibitor Sigma-Aldrich Cat# 10109886001
Critical commercial assays
Neural Tissue Dissociation Kit (P) MACS Miltenyi Biotec Cat# 130-092-628
VECTASTAIN Elite ABC-HRP Kit Vector Laboratories Cat# PK-6100
SEPP1/Selenoprotein P ELISA Kit LifeSpan BioSciences Cat# LS-F24254
Experimental models: Organisms/strains
Mouse: LPA1-GFP (Tg(Lpar1-EGFP) Mutant Mouse Resource & Research Tg(Lpar1-EGFP)GX193Gsat/Mmucd
GX193Gsat/Mmucd) Centre
Mouse: C57BL/6JRj Janvier Labs, France Cat# SC-C57J-F
Mouse: B6.129-Selenoptm1Rfb/J The Jackson Laboratory Cat# 008201
Mouse: B6;129S6-Lrp8tm1Her/J The Jackson Laboratory Cat# #003524
Mouse: Nestin-GFP (Yamaguchi et al., 2000) N/A
Software and algorithms
Freeze Frame Actimetrics version 4
Ethovision XT Noldus version XT14
Graph Pad Prism 6 for Mac OS X N/A version 6.0f
FlowJo BD Bioscience v10
Other
Osmotic pumps (7d) Alzet Cat# 1007D
Osmotic pumps (28d) Alzet Cat# 1004
RESOURCE AVAILABILITY
Lead contact
Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Dr. Tara
Walker ([email protected]).
Materials availability
This study did not generate new unique reagents.
Mice
Mice were maintained on a 12/12 h light/dark cycle at between 21-22 C and housed in pairs with access to standard mouse chow (in
Dresden: Ssniff R/M-H; catalogue # V1534 and in Brisbane: Specialty Feeds, young mice catalogue # SF00-100 and aged mice cata-
logue # SF00-105) and autoclaved water ad libitum and inspected daily by responsible animal caretakers at either the Centre for
Regenerative Technologies Dresden or the University of Queensland. All animals were 8 weeks old at the time of the experiment,
unless otherwise stated. C57BL/6JRj mice were purchased from Janvier Labs as required, LPA1-GFP mice (Tg(Lpar1-EGFP)
GX193Gsat/Mmucd; (Gong et al., 2003; Walker et al., 2016) were purchased from the Mutant Mouse Resource & Research Centre
and maintained as a hemizygous breeding colony. Nox2 knockout mice (B6.129S-Cybbtm1Din/J; Pollock et al., 1995) were initially
purchased from The Jackson Laboratory and maintained as a heterozygous breeding colony. Sepp1 knockout mice (B6.129-Sele-
noptm1Rfb/J) were purchased from The Jackson Laboratory and maintained as a heterozygous breeding colony, Lrp8 knockout mice
(B6;129S6-Lrp8tm1Her/J) were purchased from The Jackson Laboratory and maintained as a heterozygous breeding colony, DCXDTR
mice were previously generated in our laboratory and were maintained as a homozygous breeding colony (Vukovic et al., 2013). All
transgenic mice were maintained on a C57BL/6J background and wildtype littermates were used for all experiments involving the
Sepp1, Lrp8 and Nox2 lines. All C57BL6/J mice were female and transgenic mice were mixed sex. All experiments were conducted
in accordance with the applicable European or Australian regulations and approved by the responsible authority (Landesdirektion
Sachsen or the University of Queensland Animal Ethics Committee).
METHOD DETAILS
Neurosphere culture
8 week-old C57BL/6 mice were killed and their brain was immediately removed. The DG and SVZ were microdissected as described
previously (Walker and Kempermann, 2014). Briefly, the brain was placed on its ventral surface and bisected along the longitudinal
fissure then rotated dorsal side up and cut coronally at the level of the optic chiasm. To dissect the SVZ, the rostral portion of the brain
was placed with the cut side upwards and the septum removed. The SVZ was then dissected by cutting the approximately 1 mm of
tissue adjacent to the ventricle with a pair of curved forceps. To dissect the DG, the cerebellum and diencephalon were removed from
the caudal portion of the brain and a 27-gauge needle was used to cut along the border of the DG. The SVZ tissue was dissociated by
mincing the tissue with a scalpel and transferring to a tube containing 1 ml of 0.05% Trypsin-EDTA and incubating for 7 min at 37 C.
The enzymatic reaction was stopped by adding an equal volume of trypsin inhibitor containing DNaseI and centrifugation at 300 x g
for 5 mins. The DG tissue was enzymatically digested using the Neural Tissue Dissociation Kit (P) (Miltenyi) according to the manu-
facturer’s instructions. Following a final wash in Hank’s balanced salt solution (HBSS) (PAA; GE Healthcare) the pellet was resus-
pended in 1 ml of growth medium and filtered through a 40 mm cell sieve (Falcon; BD Biosciences). Given that commercially pur-
chased hormone supplements contain sodium selenite, we customised a hormone mix as described by Reynolds and colleagues
(Reynolds et al., 1992). The culture medium consisted of insulin (25 mg/ml), apotransferrin (100 mg/ml), progesterone (20 nM), putres-
cine (60 mM), glucose (0.6 %), sodium bicarbonate (3 mM), HEPES (5 mM), bovine serum albumin (2 %), Penacillin/ streptomycin
(50 U/ml), Heparin (2 mg/ml), basic fibroblast growth factor (20 ng/ml) and epidermal growth factor (20 ng/ml) in DMEM/F12. The
DG and SVZ cells were diluted to a final volume of 20 ml in growth medium and plated into a 96-well plate. Cells were seeded
into 96-well plates at a density of approximately 5 cells/ml for SVZ and 7.5 cells/ml for DG-derived cells, a density which has previously
been shown to result in the formation of clonally derived neurospheres. Compounds were added to the growth medium at the time of
plating and included sodium selenite (0.005–1 mM; Sigma-Aldrich), seleno-L-methionine (0.005 –1 mM; Sigma-Aldrich), KCl (15 mM;
Sigma-Aldrich) or norepinephrine (L-(-)-noradrenaline(+)-bitartrate salt monohydrate; 10 mM; Merck). Plates were incubated for 7 d
(SVZ) or 14 d (DG) at 37 C with 5% CO2, and the resulting neurospheres were counted and sized using an inverted light microscope
fitted with an ocular graticule.
For differentiation, neurospheres were plated onto poly-D-lysine (PDL; 10 mg/ml) and laminin (5 mg/ml)-coated coverslips in neuro-
sphere growth medium without growth factors. The neurospheres were allowed to differentiate for 7 d in humidified 5 % CO2 until
flattened and adherent. The differentiated neurospheres were then fixed with 4 % paraformaldehyde (PFA; Sigma-Aldrich) in PBS
at room temperature for 15 min.
Resazurin assay
Initially, a stock solution of 0.5 mM was prepared by dissolving resazurin dye (7-hydroxy-3H-phenoxazin-3-one 10-oxide) in
dH2O and subsequently sterile filtered. Approximately 104 cells/cm2 passage 9–13 hippocampal adherent cells were plated
on coated 24-well plates in proliferation medium. After 24 h, the medium was exchanged and supplemented with sodium sele-
nite in a concentration range of 0 mM–1 mM. The cytotoxicity assay was performed after another 24 h. Resazurin solution was
added (10 % of cell culture volume) to a final working concentration of 0.05 mM and the cells were incubated for 2 h at 37 C.
Samples were analysed fluorometrically using a spectrophotometer (Infinite 7200; Tecan) by monitoring the emitted fluores-
cence signal at a wavelength of 590 nm when excited at 535 nm. Background signals obtained from cell-free wells were sub-
tracted from each sample.
Immunostaining of neurospheres
The differentiated neurospheres were fixed with 4 % PFA in 0.1 M phosphate buffer at room temperature for 20 min. After washing
with PBS, the cells were incubated in blocking solution (10 % normal donkey serum in PBS containing 0.2 % Triton X-100) for 60 min
at room temperature. They were then incubated in fresh blocking solution containing mouse monoclonal bIII-tubulin antibody
(1:1000; Promega) and rabbit GFAP antibody (1:500; Dako Cytomation) for 60 min at room temperature. The cells were washed three
times with PBS and incubated in fresh blocking solution containing donkey anti-mouse Cy3 antibody (1:1000; Jackson ImmunoR-
esearch), DyLight 488 donkey anti-rabbit antibody (1:1000; Dianova) and 40 ,6-diamidino-2-phenylindole (DAPI; 1:5000; Sigma-Al-
drich) for 30 min at room temperature. Following another three PBS washes, the slides were mounted using fluorescence mounting
medium (Dako Cytomation) before being viewed at 20x magnification on a Zeiss Apotome microscope.
In vivo infusions
Micro-osmotic pumps (Alzet, #1004; 28 d infusion at a flow rate of 0.11 ml/h or Alzet, #1007; 7 d infusion at a flow rate of 0.5 ml/h) were
loaded with sodium selenite (1 mM or 200 nM), or saline, and the cannula inserted to enable unilateral infusion directly into the hilus
region of the hippocampus (anterior/posterior -1.3 mm, dorsal/lateral +1.0 mm, dorsal/ventral -2.2 mm, relative to Bregma) or SVZ
(anterior/posterior +0.26 mm, dorsal/lateral +0.75 mm, dorsal/ventral -2.5 mm, relative to Bregma).
MitoSOX measurements
For each experiment, 2 x 8 week-old C57BL/6 mice were killed and their brains immediately removed. The DG were microdissected
and dissociated as described above. Following a final wash in Hank’s balanced salt solution (HBSS) (PAA; GE Healthcare) the pellet
was resuspended in 1 ml of growth medium and filtered through a 40 mm cell sieve (Falcon; BD Biosciences). Cells were plated in 1 ml
of neurosphere medium without sodium selenite into 2 wells of a 24-well plate at a density of one DG per well; 0.5 mM sodium selenite
was then added to one well, and the cells were incubated at 37 C for 16 h. MitoSOX red reagent (5 mM in DMSO stock) was added to
each well to a final concentration of 5 mM and incubated for 10 min at 37 C. The cells were transferred to a 15 ml tube and washed
with 5 ml of PBS. Following centrifugation, the supernatant was removed, the cells were resuspended in 200 ml PBS. Analysis was
performed using a BD LSRII cell analyser (BD Biosciences) and the BD FACS Diva software. Data were analysed with FlowJo V.10.1
software.
BrdU and NeuN fluorescence immunohistochemistry and quantification of neuronal survival in aged animals
To measure net adult neurogenesis in aged animals following a seven-day sodium selenite infusion into the hippocampus, mice were
given three i.p. injections of BrdU (50 mg/kg) spaced 6 h apart during the last day of sodium selenite treatment. After a three-week
chase period, the mice were perfused with 0.9% NaCl, their brains collected, and coronal sections cut as described above. Immu-
nohistochemistry and quantification of neuronal survival was performed as described above for CldU and NeuN, using mouse anti-
BrdU primary antibodies (1:500; Roche) and corresponding anti-mouse Alexa Fluor 568 secondary antibodies (1:1000; Jackson Im-
munoresearch Laboratories).
anti-BrdU (IdU, 1:1000 BD Biosciences), rabbit anti-Tbr2 (1:800; Abcam) and goat anti-Sox2 (1:500; R&D Systems), followed by the
secondary antibodies: donkey anti-rat Alexa Fluor 594, donkey anti-mouse Alexa Fluor 405, donkey anti-goat Alexa Fluor 488 and
donkey anti-rabbit Alexa Fluor 647 (all 1:1000). The second antibody combination consisted of the primary antibodies: rat anti-
BrdU (CldU, 1:500; AbD Serotec) and mouse anti-BrdU (IdU, 1:1000 BD Biosciences), rabbit anti-Tbr2 (1:800; Abcam) and guinea
pig anti-DCX (1:1000; Millipore), followed by the secondary antibodies: goat anti-rat Alexa Fluor 488, donkey anti-mouse Alexa Fluor
405, goat anti-guinea pig Alexa Fluor 568 and goat anti-rabbit Alexa Fluor 647 (all 1:1000). The staining was performed as described
above except for the first antibody combination donkey serum rather than goat serum was used.
Behavioral testing
All behavioral tests were performed during the light period between 9am and 6pm and the apparatus was carefully cleaned by wiping
with 70% ethanol after each trial to remove the olfactory cues. All mice were familiarized with the researchers and habituated to the
training room prior to the experiment. The animals to be tested were moved into the behavioral testing room 30 min before the start of
the trial, to limit olfactory cues. The bedding of the animal’s home cage was not changed during the period of the behavioral tests. To
minimize the difference made by experimenter handling, mice were handled by the same experimenter throughout all behavioral
experiments.
Barnes maze
The testing room was 2 m2 in size and the Barnes maze (32 holes and 1 m in diameter in size) was placed on a pedestal in the middle of
the room at a height corresponding to four high-contrast visual cues, one in the middle of each of its four walls. The light intensity in
the room was set to 1000 lux at the surface of the maze, encouraging the mice to find shelter in a darkened escape box, which was
located under one of the holes. During habituation to the maze, the animals were first placed into the escape box containing a fresh
tissue for 1 min. They were then placed into the center of the maze and were allowed to explore the area for 5 min or until they had
entered the escape box. Animals which did not locate the escape box were carefully placed into it and allowed to rest there for 20 s.
The mice were then placed back into their home cage. Between each mouse, the maze, as well as the escape box, were thoroughly
cleaned with 70% ethanol and the escape box equipped with a fresh tissue. On the first day of testing, the escape box was rotated 90
degrees from the location of the habituation and remained there for all testing days. Each testing day comprised three individual trials,
which were conducted at least 1.5 h apart. For each trial, the mice were gently placed into a darkened start chamber and placed into
the center of one of the four maze quadrants for 15 s. To minimize any location bias, the starting quadrant for each mouse was rotated
by 90 degrees in each subsequent trial. The trial commenced following removal of the start chamber. The mice were given 3 min to
locate the escape box. The trial finished once an animal had entered the escape box or after 3 min had elapsed. If an animal failed to
find the escape box, it was placed into the box and allowed to rest there for 20 s. After the trial, the mice were placed back into their
home cage. A tracking computer together with a ceiling-mounted video camera were used to collect visual data of the animals during
the trial. Data were analyzed using the Ethovision XT14 software to assess primary latency to the escape box, escape latency upon
entering the escape box and general movement parameters.
Y-maze
To assess long-term spatial memory with minimal stress, the forced alternation test was conducted using a Y-maze apparatus
composed of three arms made of transparent plastic joined in the middle to form a Y shape. The insides of each arm were identical,
providing no intramaze cues. Each arm of the Y-maze was 40 cm long, 9 cm wide, and 17 cm high. Unique visual cues were attached
to the outside of each arm, allowing mice to identify the difference between arms. This test is based on animals’ innate curiosity to
explore a novel area. To reduce anxiety in the animals, the light intensity in the centre of Y-maze was set to 100±5 lux.
Mice were placed into one of the arms (start arm) and allowed to freely explore the maze for 5 min with one of the arms closed by an
opaque guillotine door so that they could not observe the visual cue on the closed arm (training trial). After a 24 h interval, the guillotine
door was removed, and mice were returned to the Y-maze. The animals were placed into the start arm and allowed to access all three
arms of the maze for 5 min (test trial). The video was recorded by a monitor directly above the apparatus during the test trial. The total
distance travelled, average speed, number of entries into and time spent in each arm in the test trial were analyzed using the Noldus
video tracking system (Ethovision XT 14). If a mouse climbed on the maze wall, it was immediately returned to the maze arm. An arm
entry was recorded when 85% of a mouse’s body had entered the arm.
Data analysis was performed using Prism software (Version 9, GraphPad Software). Results were expressed as mean ± standard
error of the mean (SEM). Statistical significance was determined using a Student’s t test when the experiment contained two groups,
or ANOVA when comparing more than two groups. Post-hoc analysis was performed on the ANOVA using the Dunnett, Tukey’s or
Sidak post hoc tests as described in the figure legends. Prior to analyzing the statistical significance of differences among treatments,
we tested whether the variance was similar using either the F-test or Bartlett test. The level of conventional statistical significance was
set at p < 0.05. The statistical parameters can be found in the figures and the figure legends.
Supplemental information
A AnnexinV+PI- B AnnexinV+PI-
AnnexinV+PI+ AnnexinV+PI+
10 10
5 5
0 0
0 2 24 48
saline selenium
Selenium treatment (h)
C
160
NUmber caspase-3+ cells
120
80
40
0
saline selenium
Figure S1. Selenium treatment does not affect apoptotic cell death, Related to Figure 1. (A) Treatment of adherent
NPCs with sodium selenite for 2, 24 or 48 h did not alter the percentage of early apoptotic (annexinV+propidium iodide-),
late apoptotic (annexinV+propidium iodide+) or dead (annexinV+propidium iodide+) cells. (B) Primary DG cells did not alter
the proportion of annexinV+ or propidium podide+ cells. (C) Infusion of 200 nM sodium selenite into the DG for 7 d did not
affect the number of activated caspase-3+ cells in the DG.
Sodium selenite infusion in 12-month old mice
A BrdU
sodium
selenite 12 months old DG infusion C D
40 200
20 100
0 0
I C I C I C I C
F G H I
Number BrdU+ cells / section
800
*
150 * DCX
0 0
PBS selenium PBS selenium
J K L M
Vehicle 50 nM NaSe 250 nM NaSe 500 nM NaSe
BrdU
BrdU
40 S T
2.5 μM SeMet 5 μM SeMet
20 BrdU Merge
0
0
0
0
00
20
50
10
seleno-L-methionine (µg/ml)
Figure S2. Infusion of sodium selenite into the DG of 18-month-old mice results in necrotic damage to the ipsilateral
hemisphere that is vastly reduced after seleno-L-methionine infusion, Related to FIgure 5. (A) Schematic of the experi-
mental protocol for infusion of sodium selenite (1 M) into the DG of 12-month-old mice for 7 d, followed by a 3-week chase
period. (B) Representative image of Ki67 staining (cyan; left panel) and of BrdU (cyan) and NeuN (magenta) staining (right
panel). (C) Infusion of sodium selenite into the DG of 12-month-old mice resulted in a significant increase in the number of prolif-
erating progenitor cells in both the ipsilateral and contralateral hemispheres. (D) Following a 3-week chase period, the number
of surviving newborn neurons (BrdU+NeuN+) was significantly increased in both hemispheres. (E) Schematic of the experimen-
tal protocol for infusion of sodium selenite (1 M) into the DG of 18-month-old mice for 7 d. (F,G) Infusion of sodium selenite into
the DG of 18-month-old mice resulted in a significant increase in the number of proliferating progenitor cells. (H,I) The number
of immature (DCX+) neurons also increased. Representative images of DAPI-stained coronal brain sections from mice that were
infused with vehicle (J), or 50 nM (K), 250 nM (L), or 500 nM (M) sodium selenite for 7 d. (N) Schematic of the experimental proto-
col for infusion of seleno-L-methionine (1 M) into the DG of 18-month-old mice for 7d. (O,P) Infusion of seleno-L-methionine into
the DG of 18-month-old mice for 7 d resulted in a significant increase in the number of proliferating progenitor cells. Representa-
tive images of DAPI-stained coronal brain sections from mice that were infused with vehicle (Q), or 1 M (R), 2.5 M (S) or 5
M (T) seleno-L-methionine for 7 d. No tissue damage was observed in the DG of brains infused with 1 M seleno-L-methionine
and only minimal damage was observed in the DG of the ipsilateral hemisphere of 2.5 M seleno-L-methionine-infused brains.
One-way ANOVA with Tukey’s post-hoc test (panels C,D), one-way ANOVA with Dunnett’s post-hoc test (panel O) and Student’s
t-test (panels F, H). * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001. Scale bars in B,G,I and P are 100 m and 500 m in
J-M and Q-T. Arrowheads indicate BrdU+ or DCX+ cells in higher magnification images.
A Se (SVZ)
B Fe (plasma)
C Fe (DG)
D Fe (SVZ)
1.0 800 30 40 saline
selenium
0.8
Se (mg/g protein)
30
Fe (mg/g protein)
Fe (mg/g protein)
600
20
Fe (μg/L)
0.6
400 20
0.4
10
200 10
0.2
0 0 0 0
8 week 18 month 8 week 18 month 8 week 18 month 8 week 18 month
Cu (mg/g protein)
30
Cu (mg/g protein)
2 2 100
Zn (μg/L)
Cu (μg/L)
20
1 1 50
10
0 0 0 0
8 week 18 month 8 week 18 month 8 week 18 month 8 week 18 month
I Zn (DG)
J Zn (SVZ) K Mn (plasma) L Mn (DG)
15 15 0.8 0.25
Mn (mg/g protein)
Zn (mg/g protein)
0.20
0.6
Zn (mg/g protein)
Mn (μg/L)
10 10
0.15
0.4
0.10
5 5
0.2 0.05
0 0 0 0
8 week 18 month 8 week 18 month 8 week 18 month 8 week 18 month
M N O P Ca (SVZ)
Ca (DG) 5
Mn (SVZ) Ca (plasma) 3
0.25 400
Ca (mg/g protein) 4
0.20
Ca (mg/g protein)
Mn (mg/g protein)
350
2 3
Ca (μg/L)
0.15
300 2
0.10 1
250 1
0.05
0
0 200 0 8 week 18 month
8 week 18 month 8 week 18 month 8 week 18 month
2000
Mg (mg/g protein)
60
Mg (mg/g protein)
60
Mg (μg/L)
1500
40 40
1000
500 20 20
0 0 0
8 week 18 month 8 week 18 month 8 week 18 month
Figure S3. Inductively coupled plasma mass spectrometry analysis of brain and plasma samples, Related to
FIgure 5. (A) ICP-MS analysis revealed a significant increase in selenium levels in the SVZ of both 8-week-old and
18-month-old mice following 4 weeks of dietary seleno-l-methionine supplementation. (B-S) Dietary selenium supplemen-
tation did not change the levels of the other elements measured in the SVZ, plasma or DG of 8-week-old or 18-month-old
mice.
Dietary seleno-L-methionine in aged (18-months old) mice
Active place avoidance
A B C
20 Number of shocks 35 Body weight 45 Distance travelled
(pre-treatment)
Number of shocks
Distance (m)
15 30
35
10 25
30
5 20 25
1 2 3 4 5 1 2 3 4 1 2 3 4
Day Week Day
Barnes maze
D Escape latency (BM1)
E F Velocity
45 8
200
Escape latency (s)
40
150
Velocity (cm/s)
7
Body weight (g)
35
100
30 6
50
25
5
0 20
1 2 3 4 control selenium 1 2 3 4
Day Day
Figure S4. Selenium supplementation significantly improves hippocampus-associated learning in aged mice,
Related to Figrue 6. (A) Baseline cognitive abilities of 18-month-old mice prior to selenium supplementation assessed
using the active place avoidance test. (B) Selenium treatment had no adverse effects on body weight. (C) Following seleni-
um supplementation, no difference in the total distance travelled was observed between treatment groups. (D)
16-month-old mice were pre-tested for their baseline learning ability using the Barnes maze, then divided into two cohorts
of equal learning ability. No effect of selenium treatment on body weight (E) or velocity (F) was observed.
Hippocampal lesion - active place avoidance
A B 24
20 Sham_control
*** * Pre-test Sham_4w Se
** ** *
15 22 Sham_Se
Lesion Lesion_control
10 Lesion_4w Se
20
Lesion_Se
5
18
0
d
1d
2d
3d
w6
k7
k8
k9
w 0
1
k1
-1
k1
w
w
w
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
w
APA3 APA3
C $ D *
##
150
* ** E
Discrimination ratio (%)
20 $ Sham_control
# * 100
Number of shocks
Sham_4w Se
15 50
Sham_Se
10 Lesion_control 0
Lesion_4w Se
Lesion_Se -50
5 Sham Lesion Lesion
-100 control control 4w Se
0
l
Se
Se
l
Se
Se
ro
ro
1 2 3 4 5
nt
nt
4w
4w
co
co
Figure S5. Selenium supplementation rescues the active place avoidance learning and memory deficits induced
by a hippocampal lesion, Related to Figure 7. (A) Prior to lesion all animals showed a progressive decrease in the
number of shocks in the APA1 test. Mice that received sham surgery performed significantly better than lesioned mice 3
days (APA2) and 5 weeks (APA3) after surgery. (B) Selenium supplementation had no effect on body weight. (C-E) Seleni-
um supplementation completely rescued the spatial learning and memory deficit observed 5 weeks after lesion. One-way
ANOVA with Bonferroni post-hoc test (D) and two-way repeated measures ANOVA with Bonferroni post-hoc test (B, C). *
p < 0.05, ** p < 0.01, *** p < 0.001. *: sham_control vs. lesion_control; # lesion_control vs lesion_Se; $ lesion_control vs
lesion_4w Se.
Open field test
A Total distance B Velocity C Time in centre
4 4 250
0 4w l 0 0
Se
Se
l
Se
Se
4w l
Se
Se
l
Se
Se
4w l
Se
Se
l
Se
Se
ro
ro
ro
ro
ro
ro
nt
nt
nt
nt
nt
nt
4w
4w
4w
co
co
co
co
co
co
Sham Lesion Sham Lesion Sham Lesion
0 0 0 0
4w l
Se
Se
l
Se
Se
4w l
Se
Se
l
Se
Se
4w l
Se
Se
l
Se
Se
4w l
ro
Se
Se
l
Se
Se
ro
ro
ro
ro
ro
ro
ro
nt
nt
nt
nt
nt
nt
nt
nt
4w
4w
4w
4w
co
co
co
co
co
co
co
co
60 40
3
Sham_Se
30
40 Lesion_control 2
Lesion_4w Se 20
20 Lesion_Se 1
10
0 0 0
4w l
Se
Se
l
Se
Se
4w l
Se
Se
l
Se
Se
ro
ro
0
ro
ro
12
24
36
48
nt
nt
nt
nt
4w
4w
co
co
co
co
Seconds Sham Lesion Sham Stroke
DCX-DTR
selenium drinking water
DT injection every 2nd day
K
APA1 APA2
perfusion
-11 -6 0 14 21
surgery
Days ns
ns
L M **** N
10000 ***
30 # ###
Lesion control
DCX cell number
Lesion control + DT *
*
20 Lesion selenium 6000
Lesion selenium + DT
4000
+
10
2000
0 0
1 2 3 4 5 L R L R L R L R
Day Cont Cont+DT Se Se+DT
Figure S6. Selenium supplementation rescues the learning and memory deficits induced by a hippocampal
lesion, Related to Figure 7. (A-C) Neither hippocampal lesion nor selenium supplementation affected exploratory
behaviour measured using the open field test. All animals spent a similar amount of time exploring the two objects during
the habituation phase of the novel object location test (D) and novel object recognition test (F). Similarly, all animals
spent a similar amount of time exploring the two objects during the test day of the novel object location test (E) and novel
object recognition test (G). (H) Mice from all six treatment groups exhibited similar acquisition of conditioned fear. (I) No
difference in the total distance travelled in the Y-maze was observed between groups. (J) The percentage of time spent
in the novel arm of the Y maze over first minute during the test session. (K) Schematic of the experimental design. (L)
Ablation of DCX+ cells blocked the beneficial effects of selenium supplementation. (M) DT injection significantly reduced
the number of DCX+ immature neurons in the DG of lesioned mice (N) Representative image of DCX staining. Scale bar
is 50 µm. One-way ANOVA with Bonferroni post-hoc test (A, B, C, D, E, F, G, I, J, M) and two-way repeated measures
ANOVA with Bonferroni post-hoc test (H, L). * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.001. *: sham_control vs.
lesion_control; # lesion_control vs lesion_Se.
Supplementary Table 1. Operating parameters for Agilent 8900 triple quadrupole ICP-
MS, Related to STAR Methods.
Instrument parameters All other elements (m/z Selenium (m/z 78 and 80)
56)