Ava i l ab l e on l i ne a t www.sc i enced i r ec t . com

www.e l sev i e r . com/ loca te / sc r

Stem Cell Research (2012) 8, 285–291

Hypoxic preconditioning involves systemXc − regulation in mouse neural stem cellsBrian Sims a, b, c,⁎, Melinda Clarke a, Ludwig Francillion a, Elijah Kindred a,Elana Shuford Hopkins a, Harald Sontheimer c

a Department of Pediatrics, University of Alabama at Birmingham, USAb Department of Cell Biology, University of Alabama at Birmingham, USAc Center for Glial Biology in Medicine, University of Alabama at Birmingham, USA

Received 27 January 2011; received in revised form 8 August 2011; accepted 19 September 2011

Available online 28 September 2011

Abstract In animals, hypoxic preconditioning has been used as a form of neuroprotection. The exact mechanism involved inneuroprotective hypoxic preconditioning has not been described, yet could be valuable for possible neuroprotective strategies.The overexpression of the cystine-glutamate exchanger, system Xc−, has been demonstrated as being neuroprotective (Shih,Erb et al. 2006). Here, using immunohistochemistry, we demonstrate that C57BL/6 mice exposed to hypoxia showed an in-crease in system Xc− expression, with the highest level of intensity in the hippocampus. Western Blot analysis also showedan almost 2-fold increase in system Xc− protein in hypoxia-exposed versus control mice. The mRNA for the regulatory subunitof system Xc−, xCT, and the xCT/actin ratio were also increased under hypoxic conditions. Experiments using hypoxia-induciblefactor (HIF-1α) siRNA showed a statistically significant decrease in HIF-1α and system Xc− expression. Under hypoxic condi-tions, system Xc− activity, as determined by cystine uptake, increased 2-fold. Importantly, hypoxic preconditioning was atten-uated in neural stem cells by pharmacological inhibition of system Xc− activity with S4-carboxyphenylglycine. These dataprovide the first evidence of hypoxic regulation of the cystine glutamate exchanger system Xc−.© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Hypoxic preconditioning has been investigated as a neuro-protective strategy although the exact details are not wellunderstood. Known mechanisms in this pathway involve theupregulation of the hypoxia-inducible factor (HIF-1α) anderythropoietin (Epo) (Kumral et al., 2003; Grimm et al.,

⁎ Corresponding author at: Department of Pediatrics, Division ofNeonatology, UAB, 619 20th Street South, 525 New Hillman Building,Birmingham, AL 35294, USA. Fax: +1 205 934 3100.

E-mail address: bsims@peds.uab.edu (B. Sims).

1873-5061/$ - see front matter © 2011 Elsevier B.V. All rights reserved.doi:10.1016/j.scr.2011.09.002

2005; Liu et al., 2005; Grimm et al., 2006), both of whichhave been shown to be neuroprotective agents (Digicayliogluand Lipton, 2001; Buemi et al., 2002; Juul, 2002; Yu et al.,2002; Kumral et al., 2003; Hasselblatt et al., 2006). The hyp-oxia signaling cascade appears to be vital in preconditioning,and several studies have demonstrated the importance ofHIF-1α. For example, inactivation of HIF-1α increased braininjury (Baranova et al., 2007), and blocking HIF-1α activa-tion with low-dose cadmium attenuated myocardial hypoxicpreconditioning (Belaidi et al., 2008). Additionally, otheragents such as insulin growth factor 1 (IGF-1) and cobaltchloride have been shown to activate or increase HIF-1αand lead to preconditioning (Wang et al., 2004; Xi et al.,

286 B. Sims et al.

2004). Heat shock protein 27 (Hsp27) also increases HIF-1αand causes protection against retinal ischemia (Whitlock etal., 2005). In vitro studies in C6 glioma cells demonstratedthat increased HIF-1α decreased the insults caused by 3-nitroproprionic acid (Yang and Levison, 2006).

To study this complex neuroprotective pathway, a commonendpoint needs to be identified. A potential pathway involvesthe regulation of themajor cellular antioxidant glutathione. Glu-tathione biosynthesis, which beginswith the import of cystine di-mers through the cystine glutamate exchanger (System Xc−), is apivotal step in cellular metabolism, and cysteine availability isthe rate limiting to this process.

System Xc− is an important transmembrane protein responsi-ble for the intracellular transport of cystine and the export ofglutamate. System Xc− is composed of two subunits, the regula-tory subunit (4F2hc-CD98) and the catalytic subunit (xCT). Thecatalytic subunit belongs to a family of transmembrane proteinsthat have 12 transmembrane spanning regions. CD98 binds toother amino acid transporters while xCT is considered the func-tional portion of system Xc−. Because of this, xCT and systemXc− will be used interchangeably.

There are several studies showing the importance of glu-tathione and system Xc−. For example, the HIV-1 Tat proteinled to a decrease in glutathione and a concomitant increasein system Xc− activity, producing a cell-state more vulnera-ble to oxidative stresses (Bridges et al., 2004). High extra-cellular glutamate concentrations inhibited system Xc− andreduced glutathione levels in HT22 cells (Rossler et al.,2004), and diethyl maleate increased the xCT gene at theblood–brain barrier, causing an increase in L-cystine trans-port (Hosoya et al., 2002). Increased xCT has been demon-strated in retinal ganglion cells exposed to oxidative stress(Dun et al., 2006), and inhibition of system Xc− activity incultured microglia cells prevented glutamate excitotoxicityand facilitated neuroprotection (Qin et al., 2006). Thestress-induced transcription factor, NF-E2 related factor(Nrf2) is considered a defense against oxidative stress andhas been shown to upregulate xCT, as well as other glutathi-one synthesis/release components (Pacchioni et al., 2007).Hypoxic preconditioning protects brains from future insultsbut its exact mechanism has not been elucidated. We haverecently demonstrated that Epo-induced neuroprotectionrequires system Xc− activity (Sims et al., 2010), and othershave shown that overexpression of xCT is neuroprotectiveagainst oxidative stress (Shih et al., 2006).

Given the possible importance of system Xc− in the neuro-protective mechanism of hypoxic preconditioning, the cur-rent study examined the role of system Xc− activity andprotein expression during hypoxia. Understanding the regu-lation of system Xc− activity in vivo could prove to be anovel neuroprotective strategy.

2. Results

To examine the expression of the cystine glutamate exchanger,postnatal day 7 (PND 7) C57BL/6 mice were exposed to eitherhypoxia (10% oxygen, treatment group) or normal room air (con-trol group) for 24 h. Animals were sacrificed and brain sectionswere stained for system Xc−. Fig. 1A shows that system Xc− ex-pression increased in animals placed in hypoxia compared tothose exposed to normal room air. This increase was most

apparent in the hippocampus. Whole brain lysates were collect-ed from both groups of mice for Western Blot analysis of xCT andactin using a standard densitometer (Fig. 1B). The xCT/actinratio increased by almost 2-fold in the hypoxic mice (0.89±0.1in control versus 1.62±0.31 in hypoxia animals).

To test if the increase in system Xc− expression was due to anincrease in translational or post-translational regulation of sys-temXc−mRNA, total RNAwas collected fromwhole brain of con-trol and hypoxic C57BL/6 mice for real-time PCR analysis. Therewas an 8-fold increase in xCT mRNA in brains of hypoxic micecompared to those of control mice (Fig. 1D).

For system Xc− regulation to be a practical tool for neuro-protection, it must respond in a timely fashion. However,the time course of system Xc− expression during hypoxia isnot known. The current experiment was designed to addressif system Xc− expression would increase after a short expo-sure to hypoxia and the time course of this response. PND 7mice were placed in 10% hypoxia for 45 min, and sampleswere collected from 0 to 24 h. Animals were sacrificed,and the right cortex was immediately dissected on ice andplaced in lysis buffer at the specified time point. Sampleswere frozen at −20 °C until all time points were collected.Fig. 2A shows a representative Western Blot of both xCTand actin at 0 (1), 4 (2), 8 (3), 12 (4) and 24 (5)h. Fig. 2B isthe xCT/actin ratio for all samples. The xCT/actin ratiowas 5.64±0.51 (p≤0.01) at 4, 4.61±0.69 (p≤0.01) at 8,6.49±1.36 (p≤0.01) at 12, and 5.08±1.37 (p≤0.05) at24 h. These data show that there was a persistent elevationin xCT after a short exposure to hypoxia.

The mechanism of the system Xc− increase may be linkedto the hypoxia signaling cascade. Hypoxia increases HIF-1αwhich leads to an increase in multiple genes. In order todemonstrate the role of HIF-1α signaling in xCT regulationwe probed blots from 10% hypoxia treated cells with HIF-1α siRNA for both HIF-1 α (Fig. 3A) and xCT (Fig. 3B). ΗIF-1α siRNA was added at 10 or 20 ng/ml to B104 cells and theeffect of HIF-1α knockdown on HIF-1α (Fig. 3A) and xCT(Fig. 3B) was examined. HIF-1α siRNA resulted in a 53% and73% decrease in HIF-1α band intensity at 10 ng/ml(p≤0.05) and 20 ng/ml (p≤0.05), respectively (Fig. 3A). Inaddition, xCT band intensity at 20 ng/ml was significantlydecreased by 34% (p≤0.05, Fig. 3C), suggesting that xCT islinked directly to this pathway downstream of HIF-1α.

To investigate the involvement of system Xc− expressionin hypoxic preconditioning, we performed a hypoxic precon-ditioning experiment using differentiated mouse neural stemcells. Cells were placed in a 10% hypoxia chamber for 4 hthen placed in room air under excitotoxic conditions (3 mMglutamate) overnight. Control cells at room air exposed toglutamate had 61% cell loss (p≤0.01, Fig. 4 (gray bars)). Incontrast, cells exposed to hypoxia pretreatment before glu-tamate exposure showed similar cell counts to cells not ex-posed to glutamate (Fig. 4 (black bars)). In the presence ofS4-CPG (500 μM), a system Xc− inhibitor, cells that receivedhypoxic preconditioning had total cell counts similar to thenon-hypoxic cells treated with S4-CPG. To further determineif hypoxic preconditioning involved glutathione biosynthesis,we inhibited glutathione biosynthesis with buthionine sul-foximine (BSO). Hypoxic preconditioning had no protectiveeffect if S4-CPG or BSO was used.

Expression of system Xc− under hypoxic conditions has beenestablished but the change in activity has not. Differentiated

Control

xCT

Actin

Control Hypoxia

Mouse BrainB

C

HypoxiaA

0

0.5

1

1.5

2

2.5

1 2

xCT

/act

in r

atio

*

0

1

2

3

4

5

6

7

8

9

Control Hypoxia

xCT

mR

NA

Fo

ld In

crea

se

D

50kDa

36kDa

Figure 1 System Xc− expression in neonatal mouse brain. A) Immunohistochemistry of mouse hippocampus under hypoxic and roomair conditions. B) Western Blot of whole mouse brain under control and hypoxic conditions. C) xCT/actin ratio of three Western Blots.Real time PCR of system Xc− in control versus hypoxic whole mouse brain samples. D) xCT mRNA fold increase. *p≤0.05 determinedby paired Student's t test.

287Hypoxic preconditioning involves system Xc− regulation in mouse neural stem cells

neural stem cells were placed in 10% hypoxia for 4 h then placedovernight and collected the next day for cystine uptake mea-surement. Cystine uptake experiments showed an increase influorescence after hypoxic exposure. Fluorescence increasedby about two-fold in hypoxia (373 fluorescence units/mg pro-tein) versus control (182 fluorescence units/mg protein)(p≤0.05, Fig. 5A). These data suggest that system Xc− activityis regulated by hypoxia. Fig. 5B describes amodel of hypoxic reg-ulation of System Xc.

3. Discussion

Our current study was designed to understand a specific mecha-nism involved in the neuroprotective effects of hypoxic precon-ditioning. System Xc− is pivotal in glutathione biosynthesis and

an increase in system Xc− activity has been associated with neu-roprotection (Shih et al., 2006). We therefore hypothesized thatthis pathway could be critical for neuroprotection. Herewe showthat, in vivo, hypoxia increased both system Xc− protein andmRNA expression, and the increased xCT could persist for up to24 h after only a brief hypoxic exposure. We also demonstratein vitro that xCT may be linked to HIF-1α, since knocking downHIF-1α in B104 cells inhibited xCT expression. Further, differen-tiated mouse neural stem cells showed hypoxia-induced in-creases in system Xc− activity and were protected againstglutamate-induced excitotoxicity when exposed to hypoxic con-ditions. This hypoxia-related protection was lost when systemXc− was inhibited or glutathione synthesis was blocked.

Hypoxic preconditioninghasbeen showntobeneuroprotective.For example, in ischemic rat brain, embryonic stem cells trans-planted after being exposed to hypoxic preconditioning protected

P7 Right Cortex

0123456789

1 2 3 4 5

****

***

xCT

/act

in r

atio

xCT

Actin

50kDa

36kDa

50kDa

64kDa

0

1 2 3 4 5

4 8 12 24 hrs

A

B

Figure 2 Immunoblot of hypoxic right cortex. A) Time course:0, 4, 8, 12 and 24 h. B) xCT/actin ratio. *p≤0.05 and **p≤0.01determined by paired Student's t test.

xCT

Ban

d In

ten

sity

(No

rmal

ized

to

act

in)

A

B

0

0.2

0.4

0.6

0.8

1

1.2

0

0.2

0.4

0.6

0.8

1

1.2

Control 10ng HIF 1 20ng HIF 1Scrambled siRNA siRNA siRNA

Control 10ng HIF 1 20ng HIF 1Scrambled siRNA siRNA siRNA

*

HIF

1-α

Ban

d In

ten

sity

(No

rmal

ized

to

act

in)

*

*

Figure 3 Effect of HIF-1α knockdown on xCT expression inB104 cells. A) HIF-1α normalized band density after HIF-1 αknockdown. B) xCT normalized band intensity after HIF-1αknockdown. * p≤0.05 determined by paired Student's t test.

0

5

10

15

20

25

30

35

1 2 3 4

To

tal C

ells

* 1

05

1= Control2= 3mM glutamate3= 3mM glutamate +

500μM S4-CPG4= 3mM glutamate +

200μM BSO

** *

*

*

Control Hypoxia

Figure 4 System Xc− pharmacological inhibition during hyp-oxic preconditioning in differentiated neural stem cells. Dif-ferentiated neural stem cells under room air conditions (gray):1, control; 2, 3 mM glutamate; 3, 3 mM glutamate+500 μM S4-carboxyphenylglycine; 4, 3 mM glutamate+200 μMbuthionine sul-foximine. Differentiated neural stem cells under hypoxic condi-tions (black): 1, control; 2, 3 mM glutamate; 3, 3 mM glutamate+500 μM S4-carboxyphenylglycine; 4, 3 mM glutamate+200 μMbuthionine sulfoximine. * p≤0.05 determined by paired Student'st test.

288 B. Sims et al.

the ischemic brain much more than did control cells, with 30–40%less cell death (Theus et al., 2008). Peroxide is considered a poten-tial cause of brain injury in hypoxic ischemic encephalopathy bycausing an increase in glutathione peroxidase, which changes dur-ing preconditioning (Sheldon et al., 2007). Peroxidemay also havesome protective effect at certain levels (Furuichi et al., 2005).However, no studies have investigated the role of system Xc− inthis process. In the current study,mouse brain slices showed a dra-matic increase in system Xc− immunoreactivity, and Western Blotanalysis also showedanalmost 2-fold increase in systemXc− immu-noreactivity inwholebrain duringhypoxia. Interestingly, therewasan even more robust increase in mRNA. Collectively, these datasuggest that there may be a need for system Xc− expression/activity during hypoxia that may provide neuroprotection.

The neuroprotective effect of hypoxic preconditioning inthe brain has been studied by several groups who found evi-dence for the involvement of HIF-1α and its downstream tar-gets and/or Epo (He et al., 2008) (Liu et al., 2005; Theuset al., 2008) (Prass et al., 2003) (Bernaudin et al., 2002).In hyperbaric preconditioning there is tolerance after hypox-ia that is linked to the upregulation of HIF-1α (Peng et al.,2008), and infusion of soluble EpoR reduced the protectiveeffect seen in mouse brain (Prass et al., 2003). Stresses in-volving excitotoxicity are attenuated by hypoxic precondi-tioning through the upregulation of NF-κΒ and antioxidativeenzymes (Wang et al., 2005). Indeed, the protective effectof Epo in hypoxic preconditioning was shown to be effectedthrough Jak2-Stat5 and NF-κB (Liu et al., 2005).

RoomAir

Hypoxic-preconditionedNSCs

C14-Cystine UptakeC

14 c

ysti

ne

up

take

(cp

m)/

mg

pro

tein

050

100150200250300350400450500

1 2

*

HIF 1-α

Hypoxic-Preconditioning

System Xc

Neuroprotection

A

B

Figure 5 A) C14-cystine uptake in room air and hypoxic pre-conditioned neural stem cells. B) Possible system Xc regulationschematic. *p≤0.05 determined by paired Student's t test.

289Hypoxic preconditioning involves system Xc− regulation in mouse neural stem cells

Given the involvement of HIF-1α, a plausible explanation tothe link between systemXc− and hypoxia is through the hypox-ia signaling cascade involving HIF-1α. As demonstrated by thelocalization, neurons tend to have a robust increase in systemXc− after hypoxic stimulation and this may involve HIF-1α.Thus, we targeted HIF-1α in B104 neuroblastoma cells with asiRNA probe designed to knockdown HIF-1α mRNA and subse-quently protein. The correlation demonstrates that systemXc− is a downstream target of HIF1-α signaling. This is thefirst direct evidence linking system Xc− expression with HIF-1α. The current study shows that system Xc− appears to beregulated through the HIF-1α pathway and, by our previous re-port (Sims et al., 2010), through Epo. These data beg the ques-tion: can system Xc− alone produce neuroprotection? Shihet al. (2006) showed that xCT overexpression caused neuro-protection, further suggesting that this mechanism may be anovel neuroprotective strategy.

We also determined if hypoxic preconditioning would occurif system Xc− was pharmacologically inhibited. Differentiatedneural stem cells were exposed to hypoxia for 4 h and thentreated with a toxic concentration of glutamate. In the pres-ence of hypoxic preconditioning, the cells survived whilethose not exposed to hypoxia showed significant cell loss.Moreover, when system Xc− was inhibited or glutathione bio-synthesis blocked, hypoxic preconditioning was not protectiveand cells did not survive the excitotoxic conditions. Thesedata support the hypothesis that system Xc− activity, and glu-tathione synthesis, is required for protective effects of hypox-ic preconditioning. Future studies will be aimed at looking atother regulators of this important protein which could givebetter insight into a possible protective pathway.

4. Materials and methods

All reagents were ordered from Sigma-Aldrich (St. Louis, MO)unless otherwise stated. Antibodies were ordered from Che-micon unless otherwise stated. C57BL/6 mice were obtainedfrom Charles River Laboratories (Wilmington, MA). All proto-cols were approved by the Institutional Animal Care and UseCommittee (IACUC) of UAB, and are consistent with the PHSpolicy on Humane Care and Use of Laboratory Animals (Of-fice of Laboratory Animal Welfare, Aug 2002) and theGuide for the Care and Use of Laboratory Animals (NationalResearch Council, National Academy Press, 1996). B104cells were a gift from Dr. Vince Gallo.

4.1. In vivo hypoxia

Cells were cultured in plexiglass hypoxia chamber at 10% for4 h. Animals were placed in larger plexiglass chamber for 4 hat desired oxygen concentration. The oxygen percentagewas monitored throughout experiment.

4.2. Western Blot

Cells were triturated and pelleted then lysed with RIPA buff-er [150 mM NaCl, 1% Nonidet P-40, 0.1% SDS and 50 mM Tris(pH 8.0)]. Twenty-five micrograms of protein was collectedand run on a Bio-Rad Ready Gel Tris 10% HCL. Sampleswere run at 120 V for 45 min on ice and transferred to nitro-cellulose using the Trans Blot Semi-Dry apparatus for 15 minat 15 V and probed for the appropriate antibody. Primary an-tibodies were used at 1:1000 dilutions in 3% goat serum. Sec-ondary antibodies were also used at 1:1000 dilution in TBST(8 g NaCl, 0.2 g KCl, 3 g Tris, 1 ml of Tween 20 in 1 L pH7.4) for 1 h on a rotator. Blots were developed using thePierce Super signal and standard film with a Kodakdeveloper.

4.3. Real time polymerase chain reaction

Total RNA was prepared using the Trizol reagent and protocol.One microgram of total RNA was used for each experiment.GAPDH and xCT primers (Integrated DNA Technology) with thefollowing sequencewere used: xCT, forward 5′-CCTGGCATTTG-GACGCTACAT-3′, reverse 5′-TGAGAATTGCTGTGAGCTTGCA-3′;and GAPDH, forward 5′-TTGTCTCCTGCGACTTCAAC-3′, reverse5′-ACCAGGAAATGAGCTTGACA-3′. ABI Prism 7000 was used forall real time PCR experiments and data were analyzed usingthe Pfaffel method (2ΔΔCT).

4.4. Immunohistochemistry

Primary antibodies for immunohistochemistry included Actin(MAB1501-Chemicon), System Xc (Novus Biological), GFAP(NB300-141-Novus Biological), and NeuN (MAB377-Chemicon)and were used at 1:1000 overnight. Sections were stainedby immunofluorescence using FITC and Cy3-conjugated sec-ondary antibodies at 1:1000 for 2 h. Nonspecific IgG andomission of primary antibody were used as controls for thestaining specificity. To avoid bias, observers estimating

290 B. Sims et al.

immunohistochemistry staining were blind to treatment con-ditions and the same exposure time was used for all slides.

4.5. siRNA production

Using primers designed with the Block-It siRNA program fromInvitrogen for HIF-1α, siRNA was made using the Ribomax T7RNA Polymerase System. Cells were then transfected withsiRNA using the CodeBreaker siRNA Transfection Reagentprotocol. Cells were grown for 24–48 h and then lysed forWestern Blot analysis. RNA was normalized using theQuant-it Ribogreen RNA quantification kit.

4.6. Cortical neural stem cells

Cryopreserved Mouse Cortical Neural Stem Cells wereobtained from Chemicon (Catalog Number SCR029). Polyor-nithine–laminin coated plates were made using concentra-tions of polyornithine (10 mg/ml) and laminin (5 μg/ml).Plates were incubated with polyornithine overnight,washed, and laminin solution was added and plates wereallowed to incubate overnight before being used directly orstored at −80 °C. In brief, cells were grown in Neural StemCell Expansion Medium (Chemicon Neural Stem Cell Medium,20 ng/ml Fibroblast Growth Factor 2 (FGF-2), 20 ng/ml Epi-dermal Growth Factor (EGF), and 2 μg/ml heparin) and pas-saged every 3–5 days. Cells were allowed to grow to 80%confluence then split 1:3 into T25cm flasks. Differentiationwas done after passage 3. Differentiation was induced bywithdrawing EGF, FGF-2, and heparin, and cells wereallowed to reach 70% confluence.

4.7. Cystine uptake

Cystine uptake was done using conditions described in Simset al. (2010). C14 Cystine (2.5 μCi/ml) was added to the sam-ple and a C14 detection program was used to determine totalradioactivity.

4.8. Statistical analysis

Statistical analysis was performed using the Student's t-testcomparing all samples to control. Data are presented asmean±standard deviation with each experiment beingdone in triplicate. Statistical significance was set at p≤0.05.

Acknowledgments

Funding for this project is provided by the Robert WoodJohnson Harold Amos Medical Faculty Development Awardand the CHRC K12 NICHD Training Grant.

References

Baranova, O., Miranda, L.F., et al., 2007. Neuron-specific inactivationof the hypoxia inducible factor 1 alpha increases brain injury in amouse model of transient focal cerebral ischemia. J. Neurosci.27 (23), 6320–6332.

Belaidi, E., Beguin, P.C., et al., 2008. Prevention of HIF-1 activationand iNOS gene targeting by low-dose cadmium results in loss of

myocardial hypoxic preconditioning in the rat. Am. J. Physiol.Heart Circ. Physiol. 294 (2), H901–H908.

Bernaudin, M., Tang, Y., et al., 2002. Brain genomic response followinghypoxia and re-oxygenation in the neonatal rat. Identification ofgenes thatmight contribute to hypoxia-induced ischemic tolerance.J. Biol. Chem. 277 (42), 39728–39738.

Bridges, C.C., Hu, H., et al., 2004. Induction of cystine-glutamatetransporter Xc− by human immunodeficiency virus type 1 transacti-vator protein tat in retinal pigment epithelium. Invest. Ophthalmol.Vis. Sci. 45 (9), 2906–2914.

Buemi, M., Cavallaro, E., et al., 2002. Erythropoietin and the brain:from neurodevelopment to neuroprotection. Clin. Sci. (Lond)103 (3), 275–282.

Digicaylioglu, M., Lipton, S.A., 2001. Erythropoietin-mediated neu-roprotection involves cross-talk between Jak2 and NF-kappaBsignalling cascades. Nature 412 (6847), 641–647.

Dun, Y., Mysona, B., et al., 2006. Expression of the cystine-glutamate exchanger (Xc−) in retinal ganglion cells and regula-tion by nitric oxide and oxidative stress. Cell Tissue Res. 324 (2),189–202.

Furuichi, T., Liu, W., et al., 2005. Generation of hydrogen peroxideduring brief oxygen-glucose deprivation induces preconditioningneuronal protection in primary cultured neurons. J. Neurosci.Res. 79 (6), 816–824.

Grimm, C., Hermann, D.M., et al., 2005. Neuroprotection by hypoxicpreconditioning: HIF-1 and erythropoietin protect from retinaldegeneration. Semin. Cell Dev. Biol. 16 (4–5), 531–538.

Grimm, C., Wenzel, A., et al., 2006. Hypoxic preconditioning anderythropoietin protect retinal neurons from degeneration. Adv.Exp. Med. Biol. 588, 119–131.

Hasselblatt, M., Ehrenreich, H., et al., 2006. The brain erythropoietinsystem and its potential for therapeutic exploitation in braindisease. J. Neurosurg. Anesthesiol. 18 (2), 132–138.

He, W., Qian Zhong, M., et al., 2008. Ginkgolides mimic the effectsof hypoxic preconditioning to protect C6 cells against ischemicinjury by up-regulation of hypoxia-inducible factor-1 alpha anderythropoietin. Int. J. Biochem. Cell Biol. 40 (4), 651–662.

Hosoya, K., Tomi, M., et al., 2002. Enhancement of L-cystine trans-port activity and its relation to xCT gene induction at the blood–brain barrier by diethyl maleate treatment. J. Pharmacol. Exp.Ther. 302 (1), 225–231.

Juul, S., 2002. Erythropoietin in the central nervous system, and itsuse to prevent hypoxic–ischemic brain damage. Acta Paediatr.Suppl. 91 (438), 36–42.

Kumral, A., Ozer, E., et al., 2003. Neuroprotective effect of eryth-ropoietin on hypoxic–ischemic brain injury in neonatal rats.Biol. Neonate 83 (3), 224–228.

Liu, J., Narasimhan, P., et al., 2005. Neuroprotection by hypoxicpreconditioning involves oxidative stress-mediated expressionof hypoxia-inducible factor and erythropoietin. Stroke 36 (6),1264–1269.

Pacchioni, A.M., Vallone, J., et al., 2007. Nrf2 gene deletion fails toalter psychostimulant-induced behavior or neurotoxicity. BrainRes. 1127 (1), 26–35.

Peng, Z., Ren, P., et al., 2008. Up-regulated HIF-1alpha is involvedin the hypoxic tolerance induced by hyperbaric oxygen precondi-tioning. Brain Res. 1212, 71–78.

Prass, K., Scharff, A., et al., 2003. Hypoxia-induced stroke toler-ance in the mouse is mediated by erythropoietin. Stroke 34 (8),1981–1986.

Qin, S., Colin, C., et al., 2006. System Xc− and apolipoprotein Eexpressed by microglia have opposite effects on the neurotox-icity of amyloid-beta peptide 1–40. J. Neurosci. 26 (12),3345–3356.

Rossler, O.G., Bauer, I., et al., 2004. Glutamate-induced cell deathof immortalized murine hippocampal neurons: neuroprotectiveactivity of heme oxygenase-1, heat shock protein 70, and sodiumselenite. Neurosci. Lett. 362 (3), 253–257.

291Hypoxic preconditioning involves system Xc− regulation in mouse neural stem cells

Sheldon, R.A., Aminoff, A., et al., 2007. Hypoxic preconditioning re-verses protection after neonatal hypoxia–ischemia in glutathioneperoxidase transgenic murine brain. Pediatr. Res. 61 (6), 666–670.

Shih, A.Y., Erb, H., et al., 2006. Cystine/glutamate exchangemodulatesglutathione supply for neuroprotection fromoxidative stress and cellproliferation. J. Neurosci. 26 (41), 10514–10523.

Sims, B., Clarke, M., et al., 2010. Erythropoietin-induced neuropro-tection requires cystine glutamate exchanger activity. Brain Res.1321, 88–95.

Theus, M.H., Wei, L., et al., 2008. In vitro hypoxic preconditioningof embryonic stem cells as a strategy of promoting cell survivaland functional benefits after transplantation into the ischemicrat brain. Exp. Neurol. 210 (2), 656–670.

Wang, C.H., Chang, A., et al., 2005. Kainic acid-induced oxidativeinjury is attenuated by hypoxic preconditioning. Ann. N. Y.Acad. Sci. 1042, 314–324.

Wang, X., Deng, J., et al., 2004. Potential role of IGF-I in hypoxiatolerance using a rat hypoxic–ischemic model: activation ofhypoxia-inducible factor 1alpha. Pediatr. Res. 55 (3), 385–394.

Whitlock, N.A., Agarwal, N., et al., 2005. Hsp27 upregulation byHIF-1 signaling offers protection against retinal ischemia inrats. Invest. Ophthalmol. Vis. Sci. 46 (3), 1092–1098.

Xi, L., Taher, M., et al., 2004. Cobalt chloride induces delayed cardiacpreconditioning in mice through selective activation of HIF-1alphaand AP-1 and iNOS signaling. Am. J. Physiol. Heart Circ. Physiol.287 (6), H2369–H2375.

Yang, Z., Levison, S.W., 2006. Hypoxia/ischemia expands the regen-erative capacity of progenitors in the perinatal subventricularzone. Neuroscience 139 (2), 555–564.

Yu, X., Shacka, J.J., et al., 2002. Erythropoietin receptor signallingis required for normal brain development. Development 129 (2),505–516.