nature medicine  volume 15 | number 12 | december 2009 1399

a r t i c l e s

Stroke and brain trauma are leading causes of disability and mortality; however, effective measures to minimize damage and improve recov-ery are, at present, lacking. Evidence from a number of studies suggests that neuronal excitotoxicity caused by overactivation of NMDARs is a primary neuropathological process contributing to neuronal injury after stroke and brain trauma1,2. However, several large-scale clinical trials have failed to find the expected efficacy of NMDAR antagonists in reducing brain injury after stroke3–5. Possibly, this is the result of untolerated side effects as a consequence of interference with crucial physiological NMDAR function and a narrow therapeutic window, but the causes are probably multifactorial. Thus, new NMDAR-based stroke therapies are urgently needed.

Several recent studies have suggested that some of these limita-tions may be circumvented by targeting excitotoxic signaling pathways downstream of NMDARs6. Activation of NMDARs has been linked to the modulation of a number of transcription factors, with either pro–neuronal survival or pro-death activity7–13, suggesting that alteration of transcription factor activity may crucially contribute to excitotoxic neuronal injuries after stroke. We reasoned that targeting pro-death NMDAR-dependent transcription may therefore present a promising avenue to reducing excitotoxic neuronal death with a much prolonged therapeutic window while preserving physiological NMDAR function. In this study, using a nonbiased transcription factor screening assay, we discovered an NMDAR-dependent activation of SREBP-1, a trans-cription factor best known for its role in regulating genes required for cholesterol, fatty acid, triglyceride and phospholipid biosynthesis14,15. We characterize a mechanism whereby SREBP-1 mediates neuronal cell death in response to ischemia and present a therapeutic strategy to reduce stroke-induced neuronal damage.

RESULTSSREBP-1 is activated in an NMDAR-dependent mannerTo systemically study changes in transcriptional regulation after excitotoxic insults, we isolated nuclear extracts from cortical neuronal cultures 4 h after a 60-min challenge with 50 µM NMDA and screened 345 transcription factors against their characterized transcriptional binding sites (Fig. 1a). We identified 21 that showed a twofold or greater increase in binding activity in NMDA-treated neurons in comparison with untreated controls and 16 with a twofold or more decrease (confirmed by electrophoretic mobility shift assay, results not shown). Among those transcription factors elevated more than twofold upon NMDA-induced excitotoxicity, we were particularly interested in further characterizing SREBP-1 for several reasons: SREBP-1–regulated lipid products, including phospholipids, have a key role in mediating ischemia-induced cell damage16,17, SREBP-1 activation has recently been linked to glucolipotoxic cell death in pancreatic beta cells18–22 and activation of SREBP-1 has also been found in response to anaerobic and hypoxic stress in fission yeast, Cryptococcus neoformans and Caenorhabditis elegans23–25. Activation of SREBP-1 requires the cleavage of the inactive membrane inte-gral precursor protein of 130 kDa by two dedicated proteases in the Golgi apparatus, leading to the release and nuclear translocation of the soluble mature N-terminal SREBP-1 (nt-SREBP-1) of 68 kDa for transcriptional activity14,15. Stimulation of neuronal cultures with excitotoxic NMDA produced a marked increase in active nt-SREBP-1 and a corresponding decrease in full-length precursor SREBP-1 6 h later (Fig. 1b). Using a similar NMDA stimulation we followed the time course of nt-SREBP-1 appearance and found an approximate 4.5-fold increase in expression of mature nt-SREBP-1 after 6 h (Fig. 1c).

Role of NMDA receptor–dependent activation of SREBP1 in excitotoxic and ischemic neuronal injuriesChangiz Taghibiglou1,3, Henry G S Martin1,3, Ted Weita Lai1, Taesup Cho1, Shiv Prasad1, Luba Kojic1, Jie Lu1, Yitao Liu1, Edmund Lo1, Shu Zhang1, Julia Z Z Wu1, Yu Ping Li1, Yan Hua Wen1, Joon-Hyuk Imm1, Max S Cynader1 & Yu Tian Wang1,2

Excitotoxic neuronal damage caused by overactivation of N-methyl-d-aspartate glutamate receptors (NMDARs) is thought to be a principal cause of neuronal loss after stroke and brain trauma. Here we report that activation of sterol regulatory element binding protein-1 (SREBP-1) transcription factor in affected neurons is an essential step in NMDAR-mediated excitotoxic neuronal death in both in vitro and in vivo models of stroke. The NMDAR-mediated activation of SREBP-1 is a result of increased insulin-induced gene-1 (Insig-1) degradation, which can be inhibited with an Insig-1–derived interference peptide (Indip) that we have developed. Using a focal ischemia model of stroke, we show that systemic administration of Indip not only prevents SREBP-1 activation but also substantially reduces neuronal damage and improves behavioral outcome. Our study suggests that agents that reduce SREBP-1 activation such as Indip may represent a new class of neuroprotective therapeutics against stroke.

1Brain Research Centre and Department of Medicine, Coastal Health Research Institute, University of British Columbia, Vancouver, Canada. 2Center for Neuropsychiatry and Graduate Institute for Immunology, China Medical University Hospital, Taichung, Taiwan. 3These authors contributed equally to this work. Correspondence should be addressed to Y.T.W. (ytwang@interchange.ubc.ca).

Received 13 April; accepted 28 October; published online 22 November 2009; doi:10.1038/nm.2064

©20

09 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

1400  volume 15 | number 12 | december 2009 nature medicine

a r t i c l e s

This activation is specifically mediated by NMDARs, as it was fully prevented when we stimulated neurons with NMDA in the presence of NMDAR antagonist AP5 (Fig. 1d). Emerging evidence suggests that activation of NMDARs could lead to pro-survival or pro-death signal-ing pathways in neuronal cells, depending on the subunit composi-tion, subcellular location or both of the receptors26–28. We found that the NMDA-induced SREBP-1 activation is resistant to NVP-AAM077, an NR2A subunit–containing NMDAR–specific antagonist27,29,30, but blocked by Ro25-6981, an NR2B subunit–containing NMDAR–specific

antagonist27,31 (Fig. 1d). This suggests that NMDAR-induced SREBP-1 activation is primarily mediated by NR2B subunit–containing NMDARs and hence may specifically contribute to the mediation of excitotoxic neuronal injuries.

Rising intracellular Ca2+ concentration and subsequent activa-tion of Ca2+-dependent cell death signaling steps, such as activation of calpain, has been considered a major downstream step leading to NMDAR-mediated excitotoxicity1,2,32. We found that preincuba-tion of neurons with either the Ca2+ chelator BAPTA-AM (Fig. 1e)

or the calpain inhibitors calpeptin or MDL-28170 (Fig. 1f) effectively inhibited NMDAR- mediated SREBP-1 activation, lending further support to a potential role of SREBP-1 acti-vation in NMDAR-mediated excitotoxicity. Caspase-3 activation has also been implicated

–2.0

–1.5

log 2

ratio

–1.0

–0.5

0

0.5

1.0

1.5

2.0

ATF

/CR

E

CR

EB

CO

UP

-TF c-

Rel

GA

G

MR

E PPA

R

Sm

ad3/

4

ER

E

FAS

T-1

ME

F-1

Myo

DN

F-E

1 (Y

Y1)

NF

-E2

NF

-κB

Pax

-5

Pur

-5

SR

EB

P1

Sta

t1 p

84/p

91S

tat3

Sta

t4ets

2.5

3.0a c dbCon

trol

NMDA

Full-lengthSREBP1

nt-SREBP1

Nonspecificband

β-tubulin

nt-SREBP1

β-tubulin

1.2

0.8

nt-S

RE

BP

1/tu

bulin

0.4

0Control 0 1

Time after NMDA (h)3 6

**

nt-SREBP1

β-tubulin

0.8

nt-S

RE

BP

1/tu

bulin

0.4

0

Contro

l

NMDA

NMDA +

NVP

NMDA +

Ro

NMDA +

AP5

** **

e

nt-SREBP1

β-tubulin

DMSO

Contro

l

NMDA

BAPTA

BAPTA +

NM

DA f

nt-SREBP1

β-tubulin

Contro

l

NMDA

Calpep

tin

Calpep

tin +

NM

DA

MDL

MDL

+ NM

DAFigure 1 SREBP-1 is activated in a time- and calcium-sensitive calpain-dependent manner in response to NMDA insult in cortical neuronal cultures. (a) Bar graph of protein-DNA microarray for transcription factors showing greater than twofold change in attachment to identified transcriptional binding sites (TBS) after NMDA-induced excitotoxicity (50 µM, 1 h, followed by 4-h washout). (b) SREBP-1 immunoblot for full-length precursor (immature) and active (mature) nt-SREBP-1 after excitotoxic NMDA stimulation (50 µM, 20 min, followed by 6-h washout). (c) Time-dependent nt-SREBP-1 immunoblot after NMDA excitotoxicity (n = 4). (d) Effect of various NMDAR antagonists (pan-NMDAR AP5, 50 µM; NR2B subunit–specific antagonist Ro25-6981, 0.5 µM; NR2A subunit–specific NVP-AAM077, 0.4 µM) on SREBP-1 activation in response to NMDA-induced excitotoxicity (n = 4). (e) Effect of Ca2+ chelation by BAPTA-AM (10 µM, 1 h pretreatment) on SREBP-1 activation after NMDA insult. (f) Effect of calpain inhibition (calpeptin; 25 µM or MDL; 50 µM, 15 min pretreatment) on SREBP-1 activation in response to NMDA treatment. Error bars represent s.e.m. **P < 0.01.

NMDAAP5

NuclearSREBP1

TBP

0 2.5

Time after NMDA (h)

Contro

l

NMDA

NMDA +

Ro

NMDA +

NVP

NMDA +

AP5

5

––

+–

++

––

+–

++

––

+–

++

NuclearSREBP1

SREBP1

Control

NMDA

NMDA + AP5

Hoechst33358

Lamin A/C

R 2 = 0.5417

8 × 106

6 × 106

4 × 106

2 × 106

0 4 × 106

Integrated nuclear SREBP1

8 × 106 12 × 106

Inte

grat

ed H

oech

st s

igna

l

a

b

d

c

**

0

Contro

l

NMDA

NVP RoAP5

NMDA +

NVP

NMDA +

Ro

NMDA +

AP5

4.0

Nuc

lear

SR

EB

P1 8.0

**

Figure 2 Nuclear translocation of nt-SREBP-1 is induced by NMDA insult and is associated with neuronal apoptosis. (a,b) Purified nuclear fractions from cortical cultured neurons probed for the active form of SREBP-1 (Nuclear SREBP-1) on western blots. Nuclear fraction identity and loading control confirmed by nuclear protein TBP (a) or lamin A/C (b) reprobe of membranes (bottom blots). (a) Nuclear SREBP-1 immunoblot after excitotoxic NMDA treatment (50 µM, 20 min) with indicated recovery time. (b) NMDAR antagonist effect on nuclear SREBP-1 immunoblot with NMDA excitotoxicity (n = 4). (c,d) SREBP-1 immunofluorescent image after NMDA insult (20 µM, 1 h, followed by 4-h washout) in the presence of NMDAR antagonist AP5 counterstained with Hoechst 33258 (c). Correlation between nuclear SREBP-1 immunofluorescence and Hoechst 33258 stain for apoptosis progression. Scatter plot of integrated nuclear fluorescence from individual neurons after NMDA insult (n = 300, R2 = 0.54) (d). Scale bar, 40 µm. Error bars represent s.e.m. **P < 0.01.

©20

09 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

nature medicine  volume 15 | number 12 | december 2009 1401

a r t i c l e s

in SREBP-1 activation33; however, we found that NMDA-induced SREBP-1 activation under our conditions was insensitive to caspase-3 inhibition (Supplementary Fig. 1).

After activation, mature nt-SREBP-1 translocates into the nucleus to exert its function by regulating the expression of genes containing sterol regulatory elements. Consistent with this mode of activation, we observed a time-dependent accumulation of SREBP-1 in isolated cortical nuclei after NMDA treatment (Fig. 2a). Like its activation, this nuclear accumulation of SREBP-1 is sensitive to NR2B- specific antagonist, but not NR2A antagonists (Fig. 2b). We also found SREBP-1 concentrated in the nuclei of neurons immunofluorescently labeled for SREBP-1 after NMDA-induced excitotoxicity, which was sensitive to AP5 (Fig. 2c). To confirm that nuclear SREBP-1 is func-tional, we assessed transcription of the confirmed gene target malic enzyme34 using real-time PCR (Supplementary Fig. 2a). NMDA induced a time-dependent activation of the gene encoding malic acid dehydrogenase, whereas, conversely, the SREBP-2 gene target, 3-hydroxy-3-methylglutaryl-CoA reductase35, showed no NMDA-induced alteration (Supplementary Fig. 2b).

NMDA excitotoxicity requires SREBP-1 activationThe requirement for NR2B activation and calpain activity by both SREBP-1 activation and excitotoxic neuronal injuries suggests that SREBP-1 may causatively relate to NMDAR-mediated excitotoxic neuronal injuries. Indeed, by measuring the integrated nuclear intensity of SREBP-1 signal

and Hoechst 33258 staining (as an index of nucleus condensation and thus progression of apoptosis), we found a strong positive correlation between the two factors, implying that SREBP-1 may have a causative role in mediating NMDAR-mediated excitotoxicity (Fig. 2d). To directly investigate a causative role of SREBP-1 activation in mediating NMDAR-mediated neuronal death, we inhibited SREBP-1 activation during excito-toxicity with two mechanistically distinct strategies: inhibiting SREBP-1 activation by increasing extracellular cholesterol concentrations and knockdown of SREBP-1 expression with SREBP-1-specific siRNA.

SREBP activation is closely tied to the cholesterol concentration inside the cell. During conditions of cholesterol scarcity, the proteolytic cleavage of SREBP is promoted; however, under sterol-rich conditions cholesterol directly inhibits SREBP activation15. Therefore, we tested whether supplementing our cultured cortical neurons with a high con-centration (100 µM) of cholesterol could inhibit NMDAR-dependent SREBP-1 activation and hence exert neuroprotection. Although lower doses were ineffective at reducing NMDA-induced SREBP-1 activation or excitotoxicity, we found that a 12-h high cholesterol concentration application lowered both NMDA-dependent activation of SREBP-1 and neuronal death (Fig. 3a,b). These results are consistent with a role of SREBP-1 activation in mediating NMDA-induced cell death.

Using a vector-based approach, we designed an shRNA construct directed at SREBP-1 to knock down SREBP-1 expression. Expression of this construct in HEK293 cell lines reduced native full-length SREBP-1 expression by approximately 40%, whereas a scrambled control was

Cholestrol 0 1

Control

10 100

nt-SREBP1

β-tubulin

0 1

NMDA

10 100 (µM)

Full-lengthSREBP1

GFP SREBP1

shRNA

Contro

l

shRNA

β-tubulin

a

c

1.2

1.0

0.8

Cel

l dea

th in

dex

0.6

0.40 10

Control

Cholesterol(µM)

100 0 10

NMDA

** *

100

b

1,800

1,600

1,400

**1,200

Mea

n S

RE

BP

1im

mun

oflu

ores

cenc

e

1,000

800

GFP

SREBP1

shRNA

Contro

l

shRNA

Per

cent

age

apop

totic

cel

ls

0

20

40

60

80

100

GF

P

SR

EB

P1

shR

NA

Con

trol

shR

NA

Control NMDA

GF

P

Em

pty

vect

or

SR

EB

P1

shR

NA

Con

trol

shR

NA

**

****

Untreated

GFPalone

SREBP1shRNA

ControlshRNA

NMDA + 4 h

e

f

GFPalone

GFPFlag–

nt-SREBPHoechst33358

SREBP1shRNA

ControlshRNA

d

Figure 3 Reduced mature nt-SREBP-1 protects against NMDA-induced excitotoxicity. (a,b) Cholesterol treatment inhibits NMDA-induced activation of SREBP-1 and reduces excitotoxicity in cultured neurons. (a) nt-SREBP-1 immunoblot 6 h after NMDA-induced excitotoxicity with cholesterol treatment (12-h pretreatment) at the indicated concentrations. (b) Cell death determined by LDH release after NMDA-induced excitotoxicity (20 µM, 1 h, followed by 24-h recovery) with the indicated concentrations of cholesterol (n = 3). (c) SREBP-1 immunoblot after 3 d expression of SREBP-1–specific shRNA knockdown vector or with scrambled control shRNA vector in HEK293-transfected cells. (d) Immunofluorescent evaluation of (mature) Flag–nt-SREBP-1 expression in transfected neurons (identified by GFP transfection), cotransfected with SREBP-1–specific shRNA or control shRNA. Nuclear flag-ntSREBP-1 identified by Hoechst 33258 counterstain and indicated by an arrow. (e) Bar graph of mean soma SREBP-1 endogenous immunofluorescence from SREBP-1–specific shRNA–expressing or scrambled control shRNA–expressing cells transfected with GFP. (f) Assessment of NMDA-induced apoptosis (20 µM NMDA, 1 h, followed by 4-h washout) in SREBP-1 shRNA–expressing neurons determined by Hoechst 33258 nuclear staining. Transfected cells (identified by GFP fluorescence indicated by arrow) scored for apoptotic nuclear condensation are shown in the bar chart (n = 10 coverslips from three independent experiments). Error bars represent s.e.m. Scale bar, 10 µm; *P < 0.05, **P < 0.01.

©20

09 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

1402  volume 15 | number 12 | december 2009 nature medicine

a r t i c l e s

ineffective at changing SREBP-1 expression (Fig. 3c). In our cultured neurons, we concurrently expressed Flag-tagged nt-SREBP-1 with the shRNA construct and looked for changes in SREBP-1 expression. Treatment with shRNA specific for SREBP-1 dramatically lowered Flag–nt-SREBP-1 expression 5 d after transfection, whereas control shRNA had no effect (Fig. 3d). Likewise, quantification of endogenous SREBP-1 from SREBP-1 shRNA–transfected neurons showed reduced SREBP-1 immunofluorescence compared to controls (Fig. 3e). Using the SREBP-1– specific shRNA construct we transfected hippocampal cultured neurons and looked for NMDA-induced excitotoxicity (Fig. 3f). Using nuclear DNA condensation as a marker of apoptosis in cells, we found that in neurons transfected with SREBP-1 shRNA cell death was significantly reduced compared to control shRNA. We then took advantage of a direct siRNA approach using the same gene region of SREBP-1. Treatment of cortical cultures with SREBP-1 siRNA resulted in a 40% reduction in SREBP-1 expression 6 days after transfection (Supplementary Fig. 3a). However, more importantly, NMDA-induced cell death as determined by LDH release was significantly reduced in SREBP-1 siRNA-treated cultures compared to control siRNA-treated neurons (Supplementary Fig. 3b). Together with the aforementioned results with cholesterol treatments, our results strongly support the idea that SREBP-1 has a causative role in NMDA-induced excitotoxicity.

Insig-1 degradation mediates excitotoxic SREBP-1 activationUnder basal unstimulated conditions, immature SREBPs form a stable complex with SREBP cleavage–activating protein (SCAP)36

and are retained in the endoplasmic reticulum (ER) by the interaction between SCAP and the ER membrane–resident protein Insig-1 (refs. 37,38) (Fig. 4a). When cholesterol is low, Insig-1 protein is rapidly ubiquitinated on Lys156 and Lys158 and is degraded by the proteasome39–41. SCAP is then released from the ER membrane and chaperones SREBP to the Golgi apparatus for proteolytic processing, thereby producing the active mature nt-SREBPs (Fig. 4a). Because cellular stress such as hypotonic shock and ER stress also activate SREBPs via rapid turnover of Insig-1 (ref. 42), and high cholesterol treatment inhibits NMDA-induced SREBP-1 activation (Fig. 3a), we hypothesized that NMDAR activation may produce the mature nt-SREBP-1 via a similar Insig-1–dependent mechanism. On the basis of this hypothesis, we predict that Insig-1 depletion might increase NMDAR-dependent cell death. Treatment of neuronal cultures with Insig-1–specific siRNA resulted in a 40% reduction in Insig-1 and a corresponding increase in active nt-SREBP-1 (Supplementary Fig. 4a). Challenging these neurons with excitotoxic NMDA stimuli resulted in increased cell death compared to controls (Supplementary Fig. 4b). These results support a crucial role for the degradation of Insig-1 in mediating NMDAR-dependent SREBP-1 activation and excitotoxicity and also suggest that inhibition of Insig-1 degradation may be neuroprotective.

To directly inhibit Insig-1 degradation, we designed an Insig-1–derived interference peptide (Indip) that contains the Lys156 and Lys158 ubiquitination sites39 on the basis of the assumption that the peptide would competitively block Insig-1 ubiquitination and

**

****

β-tubulin

Insig-1

β-tubulin

nt-SREBP-1

Cytosol

Cytosol

Cytosol

Basal

Low cholesterol andexcitotoxicity

High cholesterol and Indip peptide+ excitotoxicity

Golgi lumen

Golgi lumen

Degradation

Degradation

Golgi

Golgi

ER lumen

ER lumen

ER lumen

LegendSCAP SREBP-1 Insig-1

E3 UbiquitinIndip peptide

IndipK-R control peptide

Activation

Contro

l

NMDA

NMDA +

Indip NM

DA +

Indip

K-R

(Ubq)n-Insig-1

Insig-1

0

Contro

l

NMDA

NMDA +

Indip

NMDA +

Indip

K-R

0.4

Insi

g-1

/ tub

ulin

0.8

1.2

0

Contro

l

Indip

K-R

Indip

NMDA

NMDA +

Indip

K-R

NMDA +

Indip

0.4

0.8

nt-S

RE

BP

1 / t

ubul

in

1.2

1.6

Cntl

No peptide Indip IndipK-R

Sim Cntl Sim Cntl Sim

Insig-1

β-actin

b

ed

ca

Figure 4 Indip reduces NMDA-dependent SREBP-1 activation by decreasing Insig-1 ubiquitination and degradation. (a) Cartoon showing the rationale behind Indip peptide design and proposed function during low cholesterol and excitotoxicity. Immature SREBP-1 forms a SREBP-1–SCAP complex and is retained in the ER by the interaction between SCAP and the ER-anchoring protein Insig-1. Low cholesterol or excitotoxicity leads to ubiquitination and degradation of Insig-1, causing the release of SREBP-1–SCAP from the ER for SREBP-1 cleavage and activation in the Golgi. Interference peptide Indip contains the ubiquitination sites of Insig-1 (green), which are mutated from lysine (K) to arginine (R) in control peptide IndipK-R and a TAT cell-membrane transduction domain (red). High levels of cholesterol or Indip, but not IndipK-R, are predicted to inhibit Insig-1 degradation and hence stabilize the Insig-1 SCAP interaction, thereby preventing the release of and activation of SREBP-1. (b) Immunoblot for myc–Insig-1 degradation after cholesterol depletion with simvastatin (10 µM, 24 h) in HEK293 cells transiently expressing myc–Insig-1 in the presence of Indip or control IndipK-R (2 µM, 30-min pretreatment). (c) NMDA-induced ubiquitination of Insig-1 in cultured hippocampal neurons in the presence of Indip. Detection of ubiquitinated Insig-1 from p62 precipitated total ubiquitinated protein mass after NMDA-induced excitotoxicity (50 µM, 20 min, followed by 30-min washout) in the presence of peptides. (d) Insig-1 immunoblot for total protein after NMDA insult in the presence of peptides (n = 4). (e) nt-SREBP-1 immunoblot 6 h after excitotoxic NMDA insult in the presence of Indip or control IndipK-R (n = 4). Error bars represent s.e.m. **P < 0.01.

©20

09 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

nature medicine  volume 15 | number 12 | december 2009 1403

a r t i c l e s

hence its degradation by the proteasome (Fig. 4a). We rendered Indip membrane permeable by fusing the cell-membrane transduction domain of the HIV-1 Tat pro-tein (YGRKKRRQRRR)43 (Fig. 4a). We had considerable confidence in the Tat fusion peptide strategy, as it had previously been successfully used in the systemic delivery of peptides into neurons in the brain for stroke therapies44 and for regu-lating neuronal signaling pathways45,46. In parallel, we designed a con-trol peptide, IndipK-R, where the active lysine residues were switched to arginine, which is predicted to be ineffective at blocking Insig-1 ubiquitination, but maintains a similar structure. To first validate the rationale behind the Insig-1 ubiquitination–blocking peptide, we tested the effect of Indip in cell lines with reduced cholesterol syn-thesis. Blocking cholesterol biosynthesis with simvastatin induced a decrease in transfected myc–Insig-1 mass after 24 h (Fig. 4b). However, in cells treated with Indip, we observed no loss of myc–Insig-1—a result not seen with IndipK-R (Fig. 4b). Consistent with our hypo-thesis, NMDA-induced excitotoxicity increased Insig-1 ubiquitination, an effect blocked by Indip, but not IndipK-R, treatment (Fig. 4c). We observed no change in total cellular ubiquitination with Indip (Supplementary Fig. 5). After probing whole-cell lysates from NMDA-treated neurons for total Insig-1, we found that NMDA treat-ment resulted in a marked loss of Insig-1 (~40%) (Fig. 4d). More notably, the NMDA-induced degradation of Insig-1 was blocked in neurons pretreated with Indip (Fig. 4d).

After pretreating neurons with Indip, we challenged cells with NMDA-induced excitotoxicity and probed for SREBP-1 activation. As described above, we saw a strong activation of SREBP-1 in response to NMDA insult (Fig. 4e). However, in the presence of Indip, but not IndipK-R, the abundance of nt-SREBP-1 was reduced to the levels observed in unstimulated neurons (Fig. 4e). Confirming a block of activation, we failed to observe the translocation of SREBP-1 to the nucleus in response

to NMDA in the presence of Indip (Supplementary Fig. 6). Together, these results suggest that inhibition of Insig-1 degradation by Indip is an effective strategy to block excitotoxic SREBP-1 activation.

With the effective block of SREBP-1 activation, we asked whether Indip might prove effective at lowering NMDAR-dependent excito-toxicity. After pretreating cortical cultures with Indip or IndipK-R, we assessed NMDA-induced excitotoxicity 24 h later by measuring lactate dehydrogenase (LDH) release (Fig. 5a). NMDA alone caused a large increase in cell death; however, neuronal cultures treated with Indip, but not IndipK-R, showed a marked reduction in NMDA-induced exci-totoxicity, a result further confirmed by assessing the apoptotic pro-duction of mono- and oligonucleosomes (Supplementary Fig. 7a). Indip was also effective at reducing cell death in response to acute high-concentration NMDA-induced excitotoxicity (Supplementary Fig. 8). To further confirm a neuron-specific function, we repeated our NMDA-induced excitotoxicity stimulation in the presence of Indip or IndipK-R and assessed cell death by examining propidium iodide uptake. Using NeuN as a neuronal marker counterstain, we saw a decrease in propidium iodide–positive neurons in Indip-treated cultures compared to NMDA alone–treated and IndipK-R- treated dishes (Fig. 5b). Notably, using Hoechst 33258 stain to follow NMDA-induced apoptotic progression, we found Indip protection was most apparent at later time points, at which SREBP-1 activation would otherwise be maximal (4 h; Supplementary Fig. 7b).

Together, the above results strongly support a major role for SREBP-1 activation in mediating excitotoxic neuronal injuries.

Figure 5 Indip protects against NMDA-induced excitotoxicity and OGD-induced neuronal death in neuronal cultures. (a) Effect of Indip (2 µM, 20-min pretreatment, n = 12) treatment on NMDA-induced neuronal damage determined by LDH release (20 µM NMDA, 1 h followed by 24-h washout). (b) Propidium iodide (PI)–determined neuronal death in NMDA-treated hippocampal cultures in the presence of Indip or control IndipK-R peptides. Neuronal nuclei identified by NeuN immunofluorescence and Hoechst 33258 counterstain (indicated by arrow). Bar chart represents portion of PI-positive neurons on each coverslip (n = 8). (c) Active nt-SREBP-1 immunoblot after 1 h OGD challenge, followed by indicated recovery periods (n = 4). (d) OGD-induced SREBP-1 activation in the presence of subunit-specific NMDAR antagonists. Immunoblot for nt-SREBP-1 after 1 h OGD followed by 6-h recovery with antagonists (AP5; 50 µM, Ro25-6981; 0.5 µM; NVP-AAM077; 0.4 µM, n = 4). (e) Effect of Indip treatment on OGD-induced neuronal damage. LDH release determined cortical neuron cell death in response to 1-h OGD followed by 24-h recovery in the presence of Indip (n = 12). Error bars represent s.e.m. Scale bar, 10 µm. **P < 0.01.

0.4

Contro

lIn

dip

NMDA

** **

NMDA +

Indip

NMDA +

Indip

K-R

0.6

Cel

l dea

th in

dex

0.8

1.0

1.2a

****

0.4

Contro

lOGD

OGD +

Indip OGD +

Indip

K-R

Cel

l dea

th in

dex

0.6

0.8

1.0

1.2e

nt-SREBP1

β-tubulin

1.2

0.8

nt-S

RE

BP

1 / β

-tub

ulin

0.4

0

Contro

l 0 1 2

Time after OGD (h)

3 4 6

*

**

c

NeuN

Control

NMDA

NMDA +Indip

NMDA +IndipK-R

Hoechst33358

Propidiumiodide

Contro

l0

0.1

0.2

Pro

port

ion

PI-

posi

tive

neur

ons

0.3

0.4

0.5

0.6

NMDA

NMDA +

Indip

NMDA +

Indip

K-R

****

b

nt-SREBP1

β-tubulin

0

Contro

lOGD

OGD + N

VP

OGD + R

o

OGD + A

P5

****

1.2

0.8

nt-S

RE

BP

1 / β

-tub

ulin

0.4

d

©20

09 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

1404  volume 15 | number 12 | december 2009 nature medicine

a r t i c l e s

Because NMDA-mediated excitotoxicity is thought to be a primary event leading to neuronal injuries after ischemic brain insults1,2, and SREBP-1 activation has recently been shown after hypoxic stimula-tion23–25, we predicted that SREBP-1 activation may also mediate neuronal injury after ischemic insults such as stroke. To examine this prediction, we first used a well-characterized in vitro stroke model: oxygen and glucose deprivation (OGD)6,47. Treatment of cortical neuronal cultures with OGD, similar to NMDA stimula-tion, resulted in time-dependent activation of SREBP-1 (Fig. 5c), and the OGD-induced activation could be blocked by either AP5 or NR2B antagonist, but not NR2A selective antagonist (Fig. 5d). Consistent with a crucial role for SREBP-1 activation in OGD-induced neuronal injuries, pretreatment of the neurons with Indip prevented OGD-induced increase in LDH release assayed 24 h after OGD challenge, whereas IndipK-R was ineffective at reducing cell death (Fig. 5e).

Indip reduces stroke SREBP-1 activation and neuronal damageWe next investigated whether SREBP-1 activation has a role in neuro-nal injuries after focal ischemic stroke in vivo and whether Indip might be used as a potential neuroprotective agent. Rats subjected to 90-min transient middle cerebral artery occlusion (MCAo)27,48 received either Indip or saline control 45 min before stroke onset. Immunoblotting

of tissue from the infarct area showed reduced Insig-1 mass 1 h after ischemia (Fig. 6a). However, with Indip treatment, the loss of Insig-1 mass was effectively smaller (Fig. 6b). Nuclear extraction of SREBP-1 24 h after MCAo showed a marked increase in SREBP-1 activation and translocation into the nucleus in the occluded hemisphere, which was substantially reduced in Indip-treated rats (Fig. 6c). These results show not only that SREBP-1 is activated during stroke in vivo, but also that Indip is effective at blocking this ischemia-induced activation. To see whether blocking SREBP-1 activation translates into in vivo neuroprotection, we assessed the effect of Indip in preventing neuro-nal damage. Pretreatment with Indip 45 min before MCAo reduced the size of the infarct compared to saline-injected rats (Fig. 6d). Furthermore, Indip pretreatment markedly improved the combined behavioral outcome 24 h post-ischemia (Fig. 6e). Thus, SREBP-1 activation has a crucial role in neuronal injuries following stroke.

The above results demonstrate the 24-h acute efficacy of Indip when given prophylactically 45 min before the onset of focal ischemic stroke. However, to examine clinical relevance, we were interested in exploring the effectiveness of Indip as a therapy after stroke and whether it offers sustained neuroprotection rather than simply delay-ing the injury process. Rats subjected to 90-min transient MCAo were given Indip, IndipK-R, or vehicle 120 min after the onset of ischemia (Fig. 6f). One week after ischemia, we assessed neuronal damage with

Figure 6 Indip blocks MCAo-induced SREBP-1 activation and protects against ischemia-induced neuronal damage and behavioral deficits. (a) Schematic showing timeline of tissue collection for immunoblot analysis of Insig-1 mass and SREBP-1 activation in rats subjected to 90-min transient MCAo. (b) Left, image showing brain areas removed for tissue collection. This brain section was stained with 2,3,5-triphenyltetrazolium chloride (TTC) to facilitate the identification of the infarct area. Right, immunoblot for Insig-1 mass in cortical tissues collected from the occluded hemisphere (Occ) and contralateral hemisphere (Cntl) in rats injected with either saline or Indip (8.4 mg per kg body weight, i.v.) with tubulin as loading control. (c) SREBP-1 immunoblot from purified nuclear fraction from the occluded hemisphere and contralateral hemisphere in rats injected with either saline or Indip, with NeuN as nuclear marker (n = 4). (d) Infarct volume assessment of brain damage 24 h after ischemia with Indip pretreatment as per scheme in a. Left, representative images of brain slices stained with TTC (which stains healthy tissue red and leaves infarct tissue pale white). Right, line chart showing infarct areas from sequential slices from occluded hemisphere in tissues collected from saline-treated (n = 9) and Indip-treated (n = 12) rats. Inset quantification of the infarct volume. (e) Behavioral outcome assessed 24 h after ischemia with Indip pretreatment as per scheme in a. (f) Schematic showing a timeline for stroke Indip treatment after stroke. (g) Left, representative fluorescent images of brain slices stained with Fluoro-Jade (a cell type–specific fluorescent stain for degenerating neurons) in rats injected with saline, Indip or IndipK-R after treatment as shown in f. Right, line chart showing the Fluro-Jade–positive area from slices at indicated distance from bregma (n = 8 for saline and Indip groups and n = 7 for IndipK-R group). Inset, quantification of the Fluoro-Jade–positive volume. Error bars represent s.e.m. *P < 0.05, **P < 0.01.

0.75 h

CntlCntl

Saline Indip

Saline

Saline

Indip

IndipSaline Indip

Saline

*

**

*

Indip

OccOcc Cntl Occ

Cntl

**

Occ Cntl Occ

l.v. peptideinjection

Occlusion(ischemia)

Insig-1immunoblot

nt-SREBP1

nt-S

RE

Bp1

/ N

euN

Hemisphere

b-tubulin

Insig-1

NeuN1.2

0.8

0.4

12

10

8

6

100

75

50

25

0

4

0

100

80

400

200

0

Saline

Indip60

Intr

act a

rea

(mm

2 )

Flu

oro-

Jade

–pos

itive

area

(m

m2 )

Flu

oro-

Jade

–pos

itive

volu

me

(mm

3 )

Intr

act

volu

me

(mm

3 )

Beh

avio

ral d

efic

it sc

ore

40

20

15

10

5

5 2.5 –2.5 –50Distance from bregma (mm)

0

06 4 2 0

Distance from bregma (mm)1 h after MCAo 24 h after MCAo–2 –4 –6 –8

SREBP1immunoblot

1.5 h

1.5 h

Occlusion(Ischemia)

Reperfusion I.v. peptideinjection

Kill rat

Saline

SalineIndip

Indip IndipK-R

Saline

Indip

Indip

K-RIndipK-R

0.5 h 7 d

1 h 24 ha c

b

d

f

g

e

©20

09 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

nature medicine  volume 15 | number 12 | december 2009 1405

a r t i c l e s

the neuron-specific death marker Fluoro-Jade49. Similar to the acute pretreatment efficacy study above, rats that received after-stroke treat-ment of Indip showed significantly reduced brain infarct volume compared to IndipK-R-treated and saline-treated rats (Fig. 6g). Thus, our results strongly indicate that activation of SREBP-1 is crucial in mediating ischemic neuronal damage after focal ischemic stroke in vivo and that Indip, by inhibiting SREBP-1 activation and reducing ischemic neuronal damage, may have neuroprotective therapeutic potential for treatment after stroke.

DISCUSSIONTogether, our results demonstrate a previously unappreciated event: NMDAR-mediated SREBP-1 activation in signaling pathways leads to excitotoxic neuronal injuries after ischemic insults. The NMDAR-mediated SREBP-1 activation is dependent on elevated intracellular Ca2+ concentration and shares a similar mechanism with lipid homeostasis regulation, involving rapid degradation of Insig-1. We provide several pieces of evidence supporting a causative role of SREBP-1 in mediating neuronal injuries and show that pharmacological inhibition of SREBP-1 activation protects against excitotoxic neuronal injuries.

Detailed mechanisms by which activation of SREBP-1 leads to enhanced neuronal damage remain to be established. SREBPs are the major transcription factors that regulate the expression of a large number of genes involved in cellular cholesterol and lipid biogenesis15,34, and metabolic alterations in some of these lipid products have recently been implicated in mediating neuronal damage after stroke insults16,17. Thus, it is very plausible to speculate that activation of SREBP-1 may exert its pro-neuronal death action by altering one or more of these lipid products. However, it is worth noting that SREBP-1 regulates the expression of a number of sterol response element–containing genes not directly involved in lipid metabolism, including G proteins50 and voltage-gated ion channels51. Thus, SREBP-1 may contribute to neuronal damage via a mechanism independent of alterations in lipid metabolism. It is also notable that an Akt-dependent SREBP-1 activation has recently been reported to have a role in mediating cell growth52. However, whether inactivation of SREBP-1 attenuates cell growth and thereby indirectly contributes to the neuroprotective effects observed herein remains to be determined.

Whatever the precise mechanism, our study establishes that acti-vation of SREBP-1 is a previously undescribed NMDAR downstream signaling step leading to excitotoxic neuronal death and also suggests that agents that reduce SREBP-1 activation such as Indip may be a new class of neuroprotective therapeutics against stroke with reduced side effects and improved therapeutic window. Given that excitotoxicity is thought to be a common neuropathology associated with a large number of neurological disorders ranging from acute brain insults such as stroke to chronic neurodegenerative disorders such as Huntington’s disease and Parkinson’s disease, our study may have broad implications beyond stroke and raises the potential for designing new therapeutics for clinical treatment of these neurological disorders.

METHODSMethods and any associated references are available in the online version of the paper at http://www.nature.com/naturemedicine/.

Note: Supplementary information is available on the Nature Medicine website.

ACKnoWLEdGMEnTSWe thank J. Ye (University of Texas Southwestern Medical Center) for myc–Insig-1 plasmid, T. Osborne (University of California–Irvine) for Flag–nt-SREBP-1 and

Y. P. Auberson (Novartis Pharma AG) for the generous gift of NVP-AAM077. We also thank J. Wang for his technical support. This work was supported by the Heart and Stroke Foundation of British Columbia and the Yukon, the Canadian Institutes of Health Research and CHDI (Cure Huntington’s Disease Initiative) Foundation. Y.T.W. is a Howard Hughes Medical Institute International Scholar and Heart and Stroke Foundation of British Columbia and the Yukon Chair in Stroke Research. C.T. was supported by post-doctoral fellowships from the Canadian Institutes of Health Research and Michael Smith Foundation for Health.

AUTHoR ConTRIBUTIonSC.T. initiated the study, wrote the manuscript and performed all biochemical experiments, H.G.S.M. performed most cell biology experiments and helped write the manuscript, T.W.L. performed the in vivo studies, T.C. and S.Z. contributed to the fluorescent microscopy, S.P., L.K. and Y.H.W. performed mRNA analysis and transcription factor screen, Y.L. aided in the in vivo studies, J.L. and J.Z.Z.W. designed and made Insig-1–specific antibody, E.L. aided in manuscript preparation, Y.P.L. and J.-H.I. prepared neuronal cultures and developed transfection methods, M.S.C. cosupervised the experiments of mRNA analysis transcription factor screen, and Y.T.W. designed the study and supervised the overall project.

Published online at http://www.nature.com/naturemedicine/. Reprints and permissions information is available online at http://npg.nature.com/reprintsandpermissions/.

1. Choi, D.W. Calcium-mediated neurotoxicity: relationship to specific channel types and role in ischemic damage. Trends Neurosci. 11, 465–469 (1988).

2. Lipton, S.A. Paradigm shift in neuroprotection by NMDA receptor blockade: memantine and beyond. Nat. Rev. Drug Discov. 5, 160–170 (2006).

3. Ikonomidou, C. & Turski, L. Why did NMDA receptor antagonists fail clinical trials for stroke and traumatic brain injury? Lancet Neurol. 1, 383–386 (2002).

4. Kemp, J.A. & McKernan, R.M. NMDA receptor pathways as drug targets. Nat. Neurosci. 5 (Suppl), 1039–1042 (2002).

5. Lee, J.M., Zipfel, G.J. & Choi, D.W. The changing landscape of ischaemic brain injury mechanisms. Nature 399, A7–A14 (1999).

6. Aarts, M. et al. Treatment of ischemic brain damage by perturbing NMDA receptor- PSD-95 protein interactions. Science 298, 846–850 (2002).

7. Camandola, S. & Mattson, M.P. NF-κB as a therapeutic target in neurodegenerative diseases. Expert Opin. Ther. Targets 11, 123–132 (2007).

8. Hardingham, G.E., Arnold, F.J. & Bading, H. Nuclear calcium signaling controls CREB-mediated gene expression triggered by synaptic activity. Nat. Neurosci. 4, 261–267 (2001).

9. Hetman, M. & Kharebava, G. Survival signaling pathways activated by NMDA receptors. Curr. Top. Med. Chem. 6, 787–799 (2006).

10. Rao, V.R. & Finkbeiner, S. NMDA and AMPA receptors: old channels, new tricks. Trends Neurosci. 30, 284–291 (2007).

11. West, A.E., Griffith, E.C. & Greenberg, M.E. Regulation of transcription factors by neuronal activity. Nat. Rev. Neurosci. 3, 921–931 (2002).

12. Zhang, S.J. et al. Decoding NMDA receptor signaling: identification of genomic programs specifying neuronal survival and death. Neuron 53, 549–562 (2007).

13. Zou, J. & Crews, F. CREB and NF-κB transcription factors regulate sensitivity to excitotoxic and oxidative stress induced neuronal cell death. Cell. Mol. Neurobiol. 26, 385–405 (2006).

14. Espenshade, P.J. & Hughes, A.L. Regulation of sterol synthesis in eukaryotes. Annu. Rev. Genet. 41, 401–427 (2007).

15. Goldstein, J.L., DeBose-Boyd, R.A. & Brown, M.S. Protein sensors for membrane sterols. Cell 124, 35–46 (2006).

16. Adibhatla, R.M., Hatcher, J.F. & Dempsey, R.J. Lipids and lipidomics in brain injury and diseases. AAPS J. 8, E314–E321 (2006).

17. Siesjö, B.K. & Katsura, K. Ischemic brain damage: focus on lipids and lipid mediators. Adv. Exp. Med. Biol. 318, 41–56 (1992).

18. Sandberg, M.B. et al. Glucose-induced lipogenesis in pancreatic beta cells is dependent on SREBP-1. Mol. Cell. Endocrinol. 240, 94–106 (2005).

19. Takahashi, A. et al. Transgenic mice overexpressing nuclear SREBP-1c in pancreatic beta cells. Diabetes 54, 492–499 (2005).

20. Wang, H. et al. The transcription factor SREBP-1c is instrumental in the development of betacell dysfunction. J. Biol. Chem. 278, 16622–16629 (2003).

21. Wang, H., Kouri, G. & Wollheim, C.B. ER stress and SREBP-1 activation are implicated in beta cell glucolipotoxicity. J. Cell Sci. 118, 3905–3915 (2005).

22. Yamashita, T. et al. Role of uncoupling protein-2 up-regulation and triglyceride accumulation in impaired glucose-stimulated insulin secretion in a beta cell lipotoxicity model overexpressing sterol regulatory element–binding protein-1c. Endocrinology 145, 3566–3577 (2004).

23. Chang, Y.C., Bien, C.M., Lee, H., Espenshade, P.J. & Kwon-Chung, K.J. Sre1p, a regulator of oxygen sensing and sterol homeostasis, is required for virulence in Cryptococcus neoformans. Mol. Microbiol. 64, 614–629 (2007).

24. Todd, B.L., Stewart, E.V., Burg, J.S., Hughes, A.L. & Espenshade, P.J. Sterol regulatory element binding protein is a principal regulator of anaerobic gene expression in fission yeast. Mol. Cell. Biol. 26, 2817–2831 (2006).

©20

09 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

1406  volume 15 | number 12 | december 2009 nature medicine

a r t i c l e s

25. Taghibiglou, C. et al. Essential role of SBP-1 activation in oxygen deprivation induced lipid accumulation and increase in body width/length ratio in Caenorhabditis elegans. FEBS Lett. 583, 831–834 (2009).

26. Hardingham, G.E., Fukunaga, Y. & Bading, H. Extrasynaptic NMDARs oppose synaptic NMDARs by triggering CREB shut-off and cell death pathways. Nat. Neurosci. 5, 405–414 (2002).

27. Liu, Y. et al. NMDA receptor subunits have differential roles in mediating excitotoxic neuronal death both in vitro and in vivo. J. Neurosci. 27, 2846–2857 (2007).

28. Zhou, M. & Baudry, M. Developmental changes in NMDA neurotoxicity reflect developmental changes in subunit composition of NMDA receptors. J. Neurosci. 26, 2956–2963 (2006).

29. Liu, L. et al. Role of NMDA receptor subtypes in governing the direction of hippocampal synaptic plasticity. Science 304, 1021–1024 (2004).

30. Tigaret, C.M. et al. Subunit dependencies of N-methyl-d-aspartate (NMDA) receptor–induced α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor internalization. Mol. Pharmacol. 69, 1251–1259 (2006).

31. Mutel, V. et al. In vitro binding properties in rat brain of [3H]Ro 25–6981, a potent and selective antagonist of NMDA receptors containing NR2B subunits. J. Neurochem. 70, 2147–2155 (1998).

32. Mattson, M.P. Calcium and neurodegeneration. Aging Cell 6, 337–350 (2007).33. Wang, X. et al. Cleavage of sterol regulatory element binding proteins (SREBPs) by

CPP32 during apoptosis. EMBO J. 15, 1012–1020 (1996).34. Horton, J.D. et al. Combined analysis of oligonucleotide microarray data from

transgenic and knockout mice identifies direct SREBP target genes. Proc. Natl. Acad. Sci. USA 100, 12027–12032 (2003).

35. Horton, J.D. et al. Activation of cholesterol synthesis in preference to fatty acid synthesis in liver and adipose tissue of transgenic mice overproducing sterol regulatory element–binding protein-2. J. Clin. Invest. 101, 2331–2339 (1998).

36. Nohturfft, A., Brown, M.S. & Goldstein, J.L. Topology of SREBP cleavage-activating protein, a polytopic membrane protein with a sterol-sensing domain. J. Biol. Chem. 273, 17243–17250 (1998).

37. Sun, L.P., Li, L., Goldstein, J.L. & Brown, M.S. Insig required for sterol-mediated inhibition of Scap/SREBP binding to COPII proteins in vitro. J. Biol. Chem. 280, 26483–26490 (2005).

38. Yang, T. et al. Crucial step in cholesterol homeostasis: sterols promote binding of SCAP to INSIG-1, a membrane protein that facilitates retention of SREBPs in ER. Cell 110, 489–500 (2002).

39. Gong, Y. et al. Sterol-regulated ubiquitination and degradation of Insig-1 creates a convergent mechanism for feedback control of cholesterol synthesis and uptake. Cell Metab. 3, 15–24 (2006).

40. Lee, J.N., Gong, Y., Zhang, X. & Ye, J. Proteasomal degradation of ubiquitinated Insig proteins is determined by serine residues flanking ubiquitinated lysines. Proc. Natl. Acad. Sci. USA 103, 4958–4963 (2006).

41. Lee, J.N., Song, B., DeBose-Boyd, R.A. & Ye, J. Sterol-regulated degradation of Insig-1 mediated by the membrane-bound ubiquitin ligase gp78. J. Biol. Chem. 281, 39308–39315 (2006).

42. Lee, J.N. & Ye, J. Proteolytic activation of sterol regulatory element–binding protein induced by cellular stress through depletion of Insig-1. J. Biol. Chem. 279, 45257–45265 (2004).

43. Schwarze, S.R., Ho, A., Vocero-Akbani, A. & Dowdy, S.F. In vivo protein transduction: delivery of a biologically active protein into the mouse. Science 285, 1569–1572 (1999).

44. Borsello, T. et al. A peptide inhibitor of c-Jun N-terminal kinase protects against excitotoxicity and cerebral ischemia. Nat. Med. 9, 1180–1186 (2003).

45. Liu, X.J. et al. Treatment of inflammatory and neuropathic pain by uncoupling Src from the NMDA receptor complex. Nat. Med. 14, 1325–1332 (2008).

46. Wong, T.P. et al. Hippocampal long-term depression mediates acute stress-induced spatial memory retrieval impairment. Proc. Natl. Acad. Sci. USA 104, 11471–11476 (2007).

47. Goldberg, M.P. & Choi, D.W. Combined oxygen and glucose deprivation in cortical cell culture: calcium-dependent and calcium-independent mechanisms of neuronal injury. J. Neurosci. 13, 3510–3524 (1993).

48. Bederson, J.B. et al. Rat middle cerebral artery occlusion: evaluation of the model and development of a neurologic examination. Stroke 17, 472–476 (1986).

49. Schmued, L.C., Albertson, C. & Slikker, W. Jr. Fluoro-Jade: a novel fluorochrome for the sensitive and reliable histochemical localization of neuronal degeneration. Brain Res. 751, 37–46 (1997).

50. Park, H.J. et al. Role of sterol regulatory element binding proteins in the regulation of Gαi2 expression in cultured atrial cells. Circ. Res. 91, 32–37 (2002).

51. Park, H.J. et al. Parasympathetic response in chick myocytes and mouse heart is controlled by SREBP. J. Clin. Invest. 118, 259–271 (2008).

52. Porstmann, T. et al. SREBP activity is regulated by mTORC1 and contributes to Akt-dependent cell growth. Cell Metab. 8, 224–236 (2008).

©20

09 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.

nature medicinedoi:10.1038/nm.2064

ONLINE METHODSGeneral antibodies and reagents. We obtained antibody to SREBP-1 from Santa Cruz Biotechnology. We raised rabbit Insig-1–specific antibody against the N-terminal region of rat Insig-1 (amino acids 1–80) using a GST fusion epitope. We obtained BAPTA-AM (10 µM) from Invitrogen. We obtained calpeptin (25 µM) and MDL-28170 (50 µM) from Calbiochem. Tat-fused Indip (YGRKKRRQRRR-GEPHKFKREW) and IndipK-R (YGRKKRRQRRR-GEPHRFRREW) were custom synthesized by Peptide Synthesizing Facility at the University of British Columbia. NR2B-specific antagonist Ro 25–6981 was from Sigma. We purchased AP5 from Tocris. We purchased p62-derived UBA domain agarose conjugate from Biomol and used as previously described53. We prepared shRNA knockdown constructs by subcloning shDNA coding structures into pSUPER vector (Oligoengine). We used a knockdown sequence correspond-ing to nucleotides 1152–1170 of human SREBP-1a and nucleotides 1062–1180 of human SREBP-1c derived from a validated Silencer Select siRNA (Ambion). We purchased Silencer Select Insig-1 siRNA from Ambion (s133883). We performed siRNA transfection of primary neuronal cultures with Genesilencer Reagent (Genlantis) as per the manufacturer’s instructions. We performed HEK293 cell line and primary neuronal DNA transfections with Lipofectamine 2000 (Invitrogen) as per the manufacturer’s instructions.

Primary culture of cortical neurons. We prepared dissociated cultures of rat cortical neurons from Sprague Dawley rat embryos collected from killed mothers 18 d after fertilization as described previously54. We used mature neurons (12–14 d in vitro (DIV)) for experiments. We used Lipofectamine 2000 (Invitrogen) for transfection 8–10 d in vitro. We performed knockdown experiments 4–5 d after transfection. We solubilized cholesterol in ethanol before culture treatment; all other drugs and peptides were first dissolved in water.

Experimental in vitro ischemia. We achieved OGD by transferring cortical cultures to an anaerobic chamber (Thermo EC) containing a 5% CO2, 10% H2 and 85% N2 (<0.01% O2) atmosphere6,47,54. We then washed cultured neurons three times with glucose-free bicarbonate-buffered solution and maintained them anoxic for 1 h at 37 °C. We terminated OGD by returning dishes to the original growth conditions.

Nuclear extract isolation and protein-DNA array assay. We isolated nuclear extracts from control and NMDA-treated cell cultures (1 × 108 cells) using the Panomics nuclear extraction kit as recommended by the manufacturer. We incubated biotin-labeled DNA binding oligonucleotides (Panomics) with 10 µg of nuclear extract for 30 min at 15 °C to allow the formation of protein-DNA complexes. We then separated the protein-DNA complexes from the free probes by column purification as described by the manufacturer’s recommenda-tions (Panomics). We then hybridized probes to the TranSignal Protein/DNA Combo Arrays overnight at 42 °C followed by post-hybridization washes. Each TranSignal Protein/DNA Combo Array contained 345 putative transcription factor nucleotide sequences. We incubated hybridized arrays with horseradish peroxidase–labeled streptavidin, and we detected bound signals with Amersham ECL detection system (GE Healthcare).

Assessment of neuronal death. We quantified neuronal death by measuring LDH release either 6 h or 20 h after treatments using a Cyto Tox 96 assay kit (Promega). We normalized raw absorbance readings without background sub-traction to NMDA or OGD-alone–treated dishes, and we expressed values as cell death index. For visualizing apoptotic neurons, Hoechst 33258 (1 µg ml–1) or dead neurons, we added propidium iodide (10 µg ml–1) to the culture medium after treatments and incubated for 30 min. We considered cells with condensed or fragmented chromatin apoptotic. We performed ELISA quantitative assess-ment of neuronal apoptosis 20 h after treatments using a Cell Death Detection ELISAPLUS Kit (Roche) as directed by the manufacturer.

Middle cerebral arterial occlusion. We performed all animal experiments were according to protocols approved by the University of British Columbia Committee on Animal Care. We used adult male Sprague-Dawley rats (330–360g, Charles River) in this study. They had free access to rat pellet chow and water throughout the study. We induced reversible MCAo with the suture-insertion method described previously6,27. Briefly, we inserted a nylon suture with a blunted tip through the right external carotid artery of anesthetized rats and advanced it slowly down the right internal carotid artery until the right MCA was occluded. We sutured the operation wound and allowed the rat to recover. After 90 min of occlusion, we reanesthetized the rat to facilitate the removal of the occlusion. We monitored body temperature and maintained it between 36.5 and 37.5 °C throughout the surgical procedure with a heating pad. Rats that showed subarachnoid hemorrhage after MCAo were excluded from the study (after treatment: saline; 0/8, Indip; 0/8, IndipK-R; 1/8 rats). Induction of stroke was highly reliable, and we observed a failure rate of <5%. We administered peptides (8.4 mg per kg body weight) or vehicle control (saline; 1 ml kg–1) as a single bolus injection via the femoral vein. After the recovery times indicated in Figure 6 we killed the rats and measured infarct volumes either with TTC or Fluoro-Jade fluorescent stain49.

Behavioral tests. We gave each rat behavior tests 1 h after occlusion to confirm successful stroke induction. The behavior tests used were described previously and included a postural reflex test and five forelimb placement tests6,27. Each test gave a score between 0 (no deficit) and 2 (complete deficit) for a total score between 0 and 12. We included only rats with a neurological score of 11 1 h after stroke in this study. We scored the rats again the next day to determine functional neurological outcome.

Statistical analyses. Data are expressed as means ± s.e.m. We used analysis of variance for comparison among multiple groups and Tukey’s least significant difference test to identify statistically significant groups. All in vivo and imaging studies were performed in a blinded manner. Statistical significance was defined as *P < 0.05, **P < 0.01.

53. Lee, S.H. et al. Synaptic protein degradation underlies destabilization of retrieved fear memory. Science 319, 1253–1256 (2008).

54. Mielke, J.G. & Wang, Y.T. Insulin exerts neuroprotection by counteracting the decrease in cell-surface GABA receptors following oxygen-glucose deprivation in cultured cortical neurons. J. Neurochem. 92, 103–113 (2005).

©20

09 N

atu

re A

mer

ica,

Inc.

All

rig

hts

res

erve

d.