LETTERS

2004 Nature Publishing Group http://www.nature.com/naturemedicine

Tissue plasminogen activator neurovascular toxicity is controlled by activated protein C

Dong Liu1,3, Tong Cheng1,3, Huang Guo1,3, Jos A Fernndez2, John H Griffin2, Xiaomei Song1 & Berislav V Zlokovic1
Although thrombolytic effects of tissue plasminogen activator (tPA) are beneficial, its neurotoxicity15 is problematic. Here, we report that tPA potentiates apoptosis in ischemic human brain endothelium and in mouse cortical neurons treated with N-methyl-D-aspartate (NMDA) by shifting the apoptotic pathways from caspase-9 to caspase-8, which directly activates caspase-3 without amplification through the Bid-mediated mitochondrial pathway6. In vivo, tPA-induced cerebral ischemic injury in mice was reduced by intracerebroventricular administration of caspase-8 inhibitor, but not by caspase-9 inhibitor, in contrast to controls in which caspase-9 inhibitor, but not caspase-8 inhibitor, was protective. Activated protein C (APC), a serine protease with anticoagulant, anti-inflammatory and antiapoptotic activities7, which is neuroprotective during transient ischemia8,9 and promotes activation of antiapoptotic mechanisms in brain cells by acting directly on endothelium911 and neurons12, blocked tPA vascular and neuronal toxicities in vitro and in vivo. APC inhibited tPA-induced caspase-8 activation of caspase-3 in endothelium and caspase-3dependent nuclear translocation of apoptosis-inducing factor in NMDA-treated neurons and reduced tPA-mediated cerebral ischemic injury in mice. Data suggest that tPA shifts the apoptotic signal in stressed brain cells from the intrinsic to the extrinsic pathway which requires caspase-8. APC blocks tPAs neurovascular toxicity and may add substantially to the effectiveness of tPA therapy for stroke. Despite an urgent need to improve the effectiveness of tPA therapy for ischemic stroke, mechanisms mediating the toxicity of tPA within the neurovasculature are not fully understood13. We found that exposure of human brain endothelial cells (BEC) to tPA plus hypoxia (low oxygen and glucose deprivation)9 substantially increases lactate dehydrogenase (LDH) release (Fig. 1a). tPA alone did not affect LDH release from normoxic BEC, but increased the number of TUNEL-positive hypoxic BEC to 100% from 60% seen for hypoxia alone at 4 h (Fig. 1b). Human plasma-derived APC blocked the apoptosis-enhancing effects of tPA in a dose-dependent manner (Fig. 1b,c), consistent with APCs direct cytoprotective effects912.
1Frank

High levels of cyclohexamide did not affect BEC subjected to hypoxia plus tPA (Fig. 1a,b), suggesting that the toxicity of tPA may not depend on protein synthesis. Because the NMDA receptor (NMDAR) in brain endothelium is involved in mediating glutamate toxicity14, we explored whether tPA potentiates BEC apoptosis through NMDAR. NMDA alone or NMDA plus tPA (Fig. 1d) did not affect the viability of normoxic BEC. At a very high concentration, NMDA transiently increased hypoxic BEC injury (Fig. 1d), which was blocked by MK-801, a noncompetitive NMDAR antagonist (Fig. 1e). But tPA plus NMDA did not potentiate hypoxic BEC apoptosis compared to tPA and hypoxia alone (Fig. 1e). Moreover, MK-801 did not affect tPA-induced apoptosis in hypoxic BEC (Fig. 1e). Consistent with a recent report4, tPA did not directly cleave the NR1 NMDAR subunit in BEC (not shown) or neuronal cell membranes (Fig. 1f). The difference between the present findings and a recent study suggesting that the toxicity of tPA does not require its direct action on NMDAR4 versus an earlier study reporting the NR1 cleavage by tPA2 could be explained by findings showing that tPA-mediated NR1 cleavage requires the presence of plasminogen and is secondary to the proteolytic action of plasmin4. Next, we showed a sustained threefold increase in caspase-3 (Fig. 2a) and caspase-8 (Fig. 2b) activities in BEC subjected to hypoxia plus tPA compared to hypoxic BEC, whereas caspase-9 activity was not affected by tPA (Fig. 2c). In BEC treated with hypoxia plus tPA, we confirmed the enhanced formation of the active forms of caspase-3 and caspase-8, but not of caspase-9 from procaspase-9 by immunoblotting (Fig. 2d). Studies with caspase-8 and caspase-9specific inhibitors z-IETD-fmk and z-LEDH-fmk12,15, respectively, in hypoxic BEC treated with tPA have shown that inhibition of caspase-8, but not caspase-9, blocks caspase-3 activation (Fig. 2e) and apoptosis (Fig. 2f) by >80%. These findings are consistent with a report showing that immunodepletion of procaspase-9 does not affect caspase-8induced activation of procaspase-3 (ref. 16). In contrast, in the absence of tPA, blockade of caspase-9, but not caspase-8, effectively inhibited caspase-3 activation (Fig. 2e) and apoptosis (Fig. 2f), suggesting that tPA shifts the apoptotic pathways from the initiator caspase-9 to caspase-8.

P. Smith Laboratory for Neuroscience and Neurosurgical Research, Department of Neurosurgery and Division of Neurovascular Biology, University of Rochester Medical Center, 601 Elmwood Avenue, Box 645, Rochester, New York 14642, 2Department of Molecular and Experimental Medicine, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037, USA. 3These authors contributed equally to this work. Correspondence should be addressed to B.Z. (berislav_zlokovic@urmc.rochester.edu). Published online 31 October 2004; doi:10.1038/nm1122

NATURE MEDICINE VOLUME 10 | NUMBER 12 | DECEMBER 2004

1379

LETTERS a
LDH release (% total)
120 100 80 60 40 20 0 0 2 4 6 8

b
Hypoxia + tPA

* *

Hochest

+ CHX

Hypoxia

TUNEL
Normoxia + tPA
10 12 14 16

Hypo xia

2004 Nature Publishing Group http://www.nature.com/naturemedicine

Time (h) LDH release (% total)

Hypo xia + tPA

Hypo xia Hypo xia + tPA + APC + tPA + CHX

c
TUNEL-positive cells (%)
100 80 60 40 20 0 0

d
tPA + Vehicle

100 80 60 40 20 0 0 2

Hypoxia + NMDA 1.0

0.6

Hypoxia 0.3

** *

*
100

*
200

tPA + APC
400

Normoxia + NMDA + tPA

4 6 8 10 12 14 16

*
600

Figure 1 tPA toxicity to human BEC and cytoprotection by human APC. (a) LDH release from hypoxic and normoxic BEC without (closed circle and closed triangle, respectively) and with tPA (open circle and open triangle, respectively), and from hypoxic BEC with tPA + cyclohexamide (CHX, open diamond). (b) Hoechst and TUNEL immunostaining of hypoxic BEC exposed to tPA APC or CHX at 4 h. Bar, 10 m (two panels, left); 40 m (two panels, right). (c) Effects of APC (closed circle) and vehicle (open circle) on tPA-induced apoptosis in hypoxic BEC at 4 h. (d) LDH release from hypoxic and normoxic BEC without (closed triangle and closed circle, respectively) and with NMDA (0.3, 0.6, 1.0 mM; open circle, open triangle and open diamond, respectively); tPA + NMDA (1 mM) in normoxic BEC (open square). (e) LDH release from hypoxic BEC in the presence of NMDA, MK-801 and tPA at 4 h. (f) Western blot analysis of NR1 and NR2A subunits in cell membrane fractions 24 h after tPA. Mean s.e.m.; n = 35. *P < 0.01, **P < 0.05.

APC (nM)
120

Time (h)

e
LDH assay (% total)

* * * *

100 80 60 40 20 0

Hypoxia NMDA MK-801 tPA APC

_ _ _ _ _

+ _ _ _ _

+ + _ _ _

+ + + _ _

+ + _ _ +

+ _ _ + _

+ _ + + _

+ + + + _ + + + _ _

+ + _ + +

We confirmed that the tPA-induced apoptotic signal does not require amplification through the intrinsic pathway by showing that Bid, a proapoptotic promoter of cytochrome c release and a caspase-8 substrate6, is not involved in hypoxic BEC apoptosis with or without tPA (Fig. 2g), in contrast to neurons subjected to simultaneous oxygen and glucose deprivation in vitro or focal ischemia in vivo6,17. Thus, it is conceivable that BEC subjected to tPA and hypoxia can produce active caspase-8 in amounts sufficient to directly cleave and activate caspase-3. Human plasmaderived APC abolished tPA-induced increases in caspase-3 (Fig. 2a,d) and caspase-8 (Fig. 2b,d) activation in hypoxic BEC, and inhibited caspase-9 (Fig. 2c, 2d), caspase-3 (Fig. 2a,d) and caspase-8 (Fig. 2b,d) activation as a result of hypoxia alone, consistent with its ability to block caspase-3 activation regardless of the initiator caspase9,12. Neuronal apoptosis caused by NMDAR overstimulation is implicated in stroke and neurodegeneration, but does not involve caspase8 (refs. 12, 15). tPA doubled the number of apoptotic neurons exposed to NMDA, as reported4, and this effect of tPA was inhibited by >80% by the mouse APC (Fig. 3a). tPA substantially increased caspase-3 activity relative to NMDA alone (Fig. 3b), and, as in hypoxic BEC, produced a robust activation of caspase-8 (Fig. 3c,e), but did not enhance caspase-9 activity (Fig. 3d,e). tPA alone did not affect neuronal cells or the activation of caspases 8, 9 or 3 (data not shown). Both caspase-8 and caspase-3 activation in neurons treated

with the combination of NMDA and tPA were substantially reduced (>80%) by APC (Fig. 3b,c,e). In NMDA-treated neurons, APC reduced caspase-9 (Fig. 3d,e) and caspase-3 (Fig. 3b,e) activation by >75% and 85%, respectively. Consistent with a lack of effect on caspase-9 activation (Fig. 3d,e), tPA did not increase p53 and Bax expression, or decrease Bcl-2 expression (data not shown), all of which participate in NMDA-activated intrinsic apoptotic cascade12,15. A specific caspase-3 inhibitor, Ac-DEVD-CHO12,15, and the caspase-8 inhibitor, but not the caspase-9 inhibitor, blocked caspase-3 activation in cells treated with the combination of tPA and NMDA by >80% (Fig. 3f). In contrast, in cells treated with NMDA alone, the caspase-9 inhibitor reduced (>80%) caspase-3 activation (Fig. 3f), whereas caspase-8 inhibitor was ineffective, suggesting that tPA shifts the apoptotic mechanisms during NMDA-mediated neuronal apoptosis to the extrinsic pathway with critical contributions from activated caspase-8. As in hypoxic BEC, we did not show the role of Bid in NMDA-treated neurons with or without tPA (Fig. 3g). tPA significantly increased (P < 0.05) nuclear translocation of apoptosis-inducing factor (AIF) in NMDA-treated neurons by 50% (Fig. 3h). Both caspase-8 and caspase-3specific inhibitors blocked AIF nuclear translocation in the presence of tPA, whereas caspase-9 inhibitor was not effective. Several recent studies have shown the critical role of caspases in NMDA-mediated apoptosis12,18,19, suggesting that AIF nuclear translocation is downstream of the caspases12,20. The differences with an earlier study suggesting that caspase-independent AIF translocation21 could be related to the different amplitude and duration of NMDA signaling, which is likely to influence the prevailing apoptotic cascade. Consistent with the inhibition of caspase-8 activation (Fig. 3c), APC blocked AIF translocation (Fig. 3h). Next, we studied whether our findings in vitro translate in vivo by using a mouse model of 45 min middle cerebral artery (MCA) occlusion followed by 24 h reperfusion. In stroke models with robust secondary brain thrombosis22 or thromboembolism23, beneficial

1380

VOLUME 10 | NUMBER 12 | DECEMBER 2004 NATURE MEDICINE

LETTERS
Figure 2 tPAs apoptotic signal in hypoxic human BEC requires caspase-8 and is blocked by APC. (ac) Caspase-3 (a), caspase-8 (b) and caspase-9 (c) activities in hypoxic cells tPA (open circle and closed circle) and APC (open triangle and closed triangle); control caspase-3, 8 and 9 activities were <0.3 optical density (OD)/mg protein. (d) Immunoblotting for caspase3, 8 and 9 in hypoxic cells tPA and APC. (ef) Effects of caspase-8 (z-IETD-fmk) and caspase-9 (z-LEDH-fmk)specific inhibitors and APC on caspase-3 activation (e) and apoptosis (f) in hypoxic BEC tPA at 4 h. (g) Immunoblotting for Bid in hypoxic cells tPA. Mean s.e.m.; n = 36 measurements per point. *P < 0.01.

a
Caspase-3 activity

4 3 2 1 0 0 2

* *

Caspase-8 activity

Hypoxia + tPA

3.5 3.0 2.5 2.0 1.5 1.0 0.5 0 0

* *

Hypoxia + tPA

Hypoxia Hypoxia + tPA + APC

Hypoxia Hypoxia + tPA + APC

4 6 8 10 12 14 16

Time (h)

10

12

14 16

Time (h)

c
3

d
pro-casp-3 casp-3 pro-casp-8

2004 Nature Publishing Group http://www.nature.com/naturemedicine

thrombolytic effects of tPA prevail. In contrast, in models with minimal brain thrombosis1 such as the present model, the neurotoxic tPA effects may prevail1. Indeed, tPA infusion doubled the volume of cerebral injury (Fig. 4a) and worsened the motor neurological score (Fig. 4b). As in vitro, intracerebroventricular administration of caspase-8 inhibitor produced modest effects in controls, but significantly reduced brain infarction (Fig. 4a) and motor neurological score (Fig. 4b) in tPA-treated animals. Conversely, caspase-9 inhibitor substantially reduced brain injury in controls, but was much less effective in animals treated with tPA. Mouse recombinant APC reduced the volume of tPA-induced cerebral injury by 80% (Fig. 4a) and minimized the motor neurological deficit in tPA-treated mice (Fig. 4b) irrespective of whether it was infused simultaneously with tPA or after tPA infusion. The neuroprotective effects of APC on tPA-induced ischemic brain injury were dose dependent (Fig. 4a,b), as were the effects of APC alone in a mouse stroke model8,9. Treatment with tPA or APC increased the postischemic cerebral blood flow by 2025% compared to controls (Fig. 4c) and prevented ischemic fibrin deposition (Fig. 4d), but did not improve the postischemic cerebral blood flow compared to each single treatment. Although APCs anticoagulant and profibrinolytic effects7 may contribute to its neuroprotection during ischemia in vivo, both work in vivo and in cultured brain cells indicated that neuroprotection by APC does not require its anticoagulant and profibrinolytic properties9,12. Triple staining of microvessels, neurons and TUNEL indicated that 14% of vessels and 32% of neurons were TUNEL-positive in the penumbral area in control mice. In contrast, the number of TUNELpositive vessels and neurons in mice receiving tPA were 50% and 58%, respectively (Fig. 4eg). Although unequivocal human data are lack-

Caspase-9 activity

casp-8

Hypoxia + tPA
1

pro-casp-9 casp-9

Hypoxia
0 0 2 4

-actin

Hypoxia + tPA + APC

6 8 10 12 14 16

Hypoxia (h) tPA APC

_ _ _

2 _ _

2 + _

4 _ _

4 + _

8 _ _

8 + _

4 + _

4 + +

Time (h)

e
Caspase-3 activity
5 4 3 2 1 0

f
* * * * *
_ _ _ _ _

TUNEL-positive cells (%)

100 80 60 40 20 0

* * * *

Hypoxia tPA z-IETD-fmk z-LEHD-fmk APC

+ + _ _ _

+ + + _ _

+ + + + + + + + _ _ _ + _ _ + + + _ _ _ _ _ + _ _

+ _ _ + _

+ _ + + _

Hypoxia tPA z-IETD-fmk z-LEHD-fmk APC

_ + _ + _ _ _ _ _ _

_ _

+ + +

_ _ + _ _ +

+ +

+ +

_ _ _ _ + _ _ _ + _ _ _

Bid

-actin

Hypoxia (h) tPA

_ _

2 _

2 +

4 _

4 +

8 _

8 +

ing13, the present study suggests tPA critically enhances endothelial and neuronal ischemic injury, whereas APC substantially reduces tPA-mediated vascular and neuronal injury in vivo. (Fig. 4eg). A purely neurocentric view argues that only neurons are affected by ischemia, whereas a current concept states that injury and cell death of BEC, astrocytes and neurons are equally important13, as supported by the present data and findings showing massive BEC apoptosis in vivo in response to normoxic deadaptation after hypoxia24.
Figure 3 tPA shifts NMDA-induced apoptosis in mouse cortical neurons to caspase-8 which is blocked by mouse APC. (a) TUNEL-positive neurons treated with NMDA and tPA in the absence and presence of APC at 24 h. (bd) Caspase-3 (b), caspase-8 (c) and caspase9 (d) activities in neurons exposed to NMDA tPA (closed circle and open circle) and APC (closed triangle and open triangle); control caspase-3, 8 and 9 activities were <0.3 optical density (OD)/mg protein). (e) Immunoblotting for caspase-3, 8 and 9 in NMDA-treated cells tPA and APC. (f) Caspase-3 activity in neurons treated with NMDA tPA in the presence of caspase-8 (z-IETD-fmk), 9 (z-LEDH-fmk) and 3 (Ac-DEVD-CHO) inhibitors and APC at 12 h. (g) Immunoblotting for Bid in neurons treated with NMDA and tPA. (h) Western blot analysis of AIF in nuclear extracts from cells treated with NMDA and tPA in the presence of APC and caspase-8, 9 and 3 inhibitors at 24 h. Mean s.e.m.; n = 35. *P < 0.05, **P < 0.01.

a
Apoptotic cells (%)
50 40 30 20 10 0 NMDA tPA APC

b
Caspase-3 activity

c
Caspase-8 activity

d
4 3 2 1 00
Caspase-9 activity

** ** **
+ + + + _ _ _ + + + _ + _ + _

5 4 3 2 1 00

*
5

* *
10

NMDA + tPA

NMDA

NMDA + tPA + APC NMDA + APC

* *

* *
5 10

3 2 1 0

NMDA + tPA

*
25

NMDA + tPA + APC NMDA

*
0 5

* * *
10

NMDA + tPA

NMDA NMDA + tPA + APC NMDA + APC

15

20

25

15

20

15

20

25

Time (h)

Time (h)

Time (h)

e
hour procasp-3 casp-3 procasp-8 casp-8 procasp-9 casp-9 -actin
NMDA tPA APC

f
Caspase-3 activity
0 3 12

*
5 4 3 2 1 0
_ _ _ _ _ _

g hour
Bid -actin
NMDA tPA

_ _

+ _

+ +

+ _

+ +

+ +
_

h
+ + _ _ _ _ + + + _ _ _ + + + + + + + _ _ _ _ _ + _ _ _ _ + _ _ _ _ + _ + + _ _ _ + + _ _ _ + _ _ _ _ _ _ _

AIF
_ _ _ _ _ _

Histone
NMDA tPA z-IETD-fmk z-LEHD-fmk Ac-DEVD-CHO APC

_ _ _

+
_ _

+ +

+ + +

_ _

+ +
_

+ + +

NMDA tPA z-IETD-fmk z-LEHD-fmk Ac-DEVD-CHO APC

+ _ _ _ _ _

+ + + + + + + + _ + _ _ _ _ + _ _ _ _ + _ _ _ _

+ _ + + _ _ _ _ _ _ + _

NATURE MEDICINE VOLUME 10 | NUMBER 12 | DECEMBER 2004

1381

LETTERS a
80

Volume (mm3)

60 40 20 0
_ _ _ _ _ _ + _ _ _

* * *

** *

Vehicle

2004 Nature Publishing Group http://www.nature.com/naturemedicine

tPA z-IETD-fmk z-LEHD-fmk APC APC post-tPA

_ _ + _ _

_ _ _ + _

+ _ _ _ _

+ + _ _ _

+ _ + _ _

+ _ _ _

+ _ _ _

+ _ _ _

+ _ _ _

+ _ _ 1

1 0.2 0.04 0.02 _

tPA

b
5

Score

4 3 2 1

** *

* * **

tPA + APC
_ _ _ _ _ _ + _ _ _ _ _ + _ _

tPA z-IETD-fmk z-LEHD-fmk APC APC post-tPA

CBF (% of baseline)

_ _ _ + _

+ _ _ _ _

+ + _ _ _

+ _ + _ _

+ _ _ _

+ _ _ _

+ _ _ _

+ _ _ _

+ _ _ 1

Figure 4 tPA-induced cerebral injury in a mouse stroke model and protection by mouse APC. (a,b) Injury volume (a) and motor neurological score (b) in the absence and presence of tPA, caspase-8 (z-IEDTD-fmk) and caspase-9 (z-LEDH-fmk) inhibitors and APC. tPA (10 mg/kg) was infused intravenously during the last 10 min of MCA occlusion and within 20 min of reperfusion. APC was infused either simultaneously with tPA or after tPA infusion (APCpost-tPA). Caspase inhibitors were administered intracerebroventricularly 30 min after MCA occlusion. Values represent mg/kg of APC. (c) Cerebral blood flow (CBF) in the presence of vehicle (closed circle), tPA (closed traingle) and tPA + APC infusion (open circle). (d) Brain fibrin deposition. (e) Triple staining for microvessels (collagen IV, red), neurons (MAP2, green) and TUNEL (white) in brain sections of mice treated with 45 min MCA occlusion and 24 h reperfusion and vehicle (top), tPA (middle) or tPA + APC (bottom); bar, 10 m. (f,g), TUNEL-positive vessels (f) and neurons (g) in the presence of tPA and tPA + APC. Values are mean s.e.m.; n = 412 per group; APC (mg/kg). *P < 0.01 compared to controls. **P < 0.05.

1 0.2 0.04 0.02 _

f
Apoptotic vessels (%)
60

Reagents. We used human recombinant * 120 * tPA (Alteplase, Genentech) and human c plasmaderived APC and mouse recombinant 40 * * * 80 APC9. NMDA and MK-801 were from Sigma. We used polyclonal antibodies against human AIF 20 40 (1:1000, 1 mg/ml; Chemicon), bovine histone (crossreacts with mouse histone; 1:1000, 0 0 0 15 30 45 60 75 90 0.6 mg/ml; US Biological), human -actin (1:2500, X 0.2 mg/ml; Santa Cruz Biotechnology), human Reperfusion Occlusion Time (min) caspase-3 (1:500, 100 g/ml; Cell Signaling) and g human caspase-8 (1:2000, 1 g/l; BD 80 d * 0.3 Ischemic Pharmingen), which recognize the respective * * 60 mouse forms, human-specific caspase-9 and 0.2 human caspase-9, which recognizes mouse 40 caspase-9 (1:1000, 100 g/ml, Cell Signaling), * 0.1 20 * * human Bid (crossreacts with mouse Bid; 1:500, Nonischemic 1 g/l; R & D system), rat NR1 (crossreacts with 0 0.0 human NR1; 1:1000, 1 g/l; Chemicon), rat NR1 _ + _ + _ _ + _ + tPA tPA + + _ _ 0.2 0.2 _ _ 0.2 0.2 (crossreacts with mouse NR1; 1:500, 1mg/ml; APC _ _ APC + Upstate Biotechnology), mouse NR2A (1:500, 1mg/ml; Upstate Biotechnology), mouse type IV collagen (1:500, 22.4 mg/ml; Karlan), rat microtubule-associated protein (MAP2) (crossreacts with mouse MAP2; 1:500, 4 In summary, we found that tPA shifts the apoptotic signal in mg/ml; Sigma). The peptides, Ac-DEVD-CHO (caspase-3, Sigma), z-IETDstressed brain cells in vitro and in vivo from the intrinsic to the fmk (caspase-8, Sigma) and z-LEHD-fmk (caspase-9, Calbiochem), were extrinsic (death receptormediated25) pathway, which activates selected on the basis of substrate specificity12,17,18.
Fibrin ( g/0.1 mg)

METHODS

caspase-8 in amounts that directly cleave and activate caspase-3, therefore bypassing the need for mitochondrial amplification through Bid or the NMDAR-mediated intrinsic cascade. Exactly how tPA influences the death-effector domain proteins, receptors and ligands and how it connects the Fas-associated death-domain protein to caspase-8 remain to be investigated. APC, an FDA-approved drug for severe sepsis7, blocks tPA vascular and neuronal toxicities in vitro and in vivo. Thus, the combination of tPA and APC, in which tPA reopens the occluded blood vessels and dissolves clots and APC acts to diminish direct vascular and neuronal toxicities of tPA, may hold a great promise for ischemic stroke therapy.

Apoptotic neurons (%)

tPA and hypoxia in human BEC. We maintained primary human BEC9 in serum-free DMEM and exposed them to tPA (20 g/ml) for 116 h under normoxic conditions (20% oxygen, 5 mM glucose), hypoxia (<2% oxygen, no glucose), tPA + hypoxia, tPA + hypoxia + cyclohexamide (25 M), NMDA (0.3, 0.6, 1.0 mM), NMDA + tPA, NMDA + hypoxia, or NMDA + tPA + hypoxia. The effects of MK-801 (100 M) on NMDA (1 mM) tPA were studied at 4 h. We induced hypoxia using an anaerobic chamber (Forma Scientific)9. The levels of O2 were monitored by O2 Fyrite (Forma Scientific). We applied z-IETD-fmk (10 M) and z-LEHD-fmk (15 M) 2 h before tPA + hypoxia treatment. We added human APC (400 nM) at the time of tPA and hypoxia treatment. We determined caspase-3, 8 and 9 activation within 16 h.

1382

VOLUME 10 | NUMBER 12 | DECEMBER 2004 NATURE MEDICINE

LETTERS
tPA and NMDA treatmentinduced apoptosis in neurons. We treated 14-d-old primary neuronal cultures12 for 324 h with tPA (20 g/ml) alone, NMDA (25 M)/5 M glycine and/or tPA and NMDA. We applied Ac-DEVDCHO (50 M), z-IETD-fmk (20 M) and z-LEHD-fmk (5 M) 3 h before tPA and NMDA treatment. Mouse APC (20 nM) was applied 30 min after tPA and NMDA. We determined caspase-3, 8 and 9 activation and AIF translocation within 24 h. Detection of cell injury and apoptosis. We performed the Hoechst (Hoechst 33,342, Sigma) and in situ terminal deoxynucleotidyl transferase-mediated digoxigenin-dUTP nick-end labeling (TUNEL) (Phoenix Flow Systems) on acetone-fixed cultured BEC at 4 h and paraformaldexyde-fixed cultured neurons at 24 h. Apoptotic cells in tissue were visualized by TUNEL (Intergen Company) in paraffin-embedded 4-m thick brain sections. Cy3-conjugated donkey antibody to mouse IgG (1:500, Jackson ImmunoResearch) and Cy5conjugated donkey antibody to rabbit IgG (1:500, Jackson ImmunoResearch) were used as secondary antibodies to detect MAP2 and collagen IV, respectively. We analyzed immunofluorescent images using a confocal microscope (Olympus 1X80) and Fluorview (Olympus) software. We determined the number of TUNEL-positive microvessels and neurons in the penumbra using 50 randomly selected sections. Caspase-3, 8 and 9 activities. We incubated BEC and cortical neuron lysates at 37 C with caspase-3 (DEVD-pNA), caspase-8 (IETD-pNA) (ApoAlert caspase assay kit; Clontech) and caspase-9 (Ac-LEHD-pNA; Chemicon) substrate. Substrate hydrolysis was determined as absorbance change at 405 nm in a microplate reader12. Enzymatic activity was expressed in arbitrary units (optical density, OD) per mg of protein. Immunoblotting. We determined the presence of procaspases 3, 8 and 9 and their active forms from BEC and neuronal cell lysates9,12. We determined the presence of AIF in nuclear proteins and the NR1 and NR2a subunits in cell membrane proteins from BEC and neurons12. In vivo stroke model. Procedures were approved by the Animal Care Committee at the University of Rochester using the United States National Institutes of Health guidelines. We induced transient ischemia by 45 min MCA occlusion in C57Bl/6 mice with a siliconized filament. Vehicle, tPA (10 mg/kg, 10% bolus/90% infusion for 30 min) and/or tPA + mouse APC (0.021 mg/kg, 50% bolus/50% infusion for 30 min) were administered through the femoral vein. Caspase-8 (z-IETD-fmk, 100 ng in 0.5 L) and caspase-9 (z-LEDH-fmk, 400 ng in 0.5 L) were administered intracerebroventricularly 30 min after MCA occlusion. We monitored cerebral blood flow, arterial blood gases and physiological parameters8. We performed neurological and neuropathological examinations and fibrin deposition at 24 h as described8,9. Statistical analysis. ANOVA was used to determine statistically significant differences. Nonparametric data were compared by the Kruskal-Wallis test.
ACKNOWLEDGMENT This work was supported by the United States National Institutes of Health grant HL63290. COMPETING INTERESTS STATEMENT The authors declare competing financial interests (see the Nature Medicine web site for details).
Received 10 June; accepted 24 September 2004 Published online at http://www.nature.com/naturemedicine/
1. Wang, Y.F. et al. Tissue plasminogen activator (tPA) increases neuronal damage after focal cerebral ischemia in wild-type and tPA-deficient mice. Nat. Med. 4, 228231 (1998). 2. Nicole, O. et al. The proteolytic activity of tissue-plasminogen activator enhances NMDA receptor-mediated signaling. Nat. Med. 7, 5964 (2001). 3. Yepes, M. et al. Regulation of seizure spreading by neuroserpin and tissue-type plasminogen activator is plasminogen-independent. J. Clin. Invest. 109, 15711578 (2002). 4. Matys, T. & Strickland, S. Tissue plasminogen activator and NMDA receptor cleavage. Nat. Med. 9, 371373 (2003). 5. Chen, Z.L. & Strickland, S. Neuronal death in the hippocampus is promoted by plasmin-catalyzed degradation of laminin. Cell 91, 917925 (1997). 6. Plesnila, N., et al. BID mediates neuronal cell death after oxygen/glucose deprivation and focal cerebral ischemia. Proc. Natl. Acad. Sci. USA 98, 1531815323 (2001). 7. Griffin, J.H. et al. Activated protein C and ischemic stroke. Crit. Care Med. 32, S247S253 (2004). 8. Shibata, M. et al. Anti-inflammatory, antithrombotic, and neuroprotective effects of activated protein C in a murine model of focal ischemic stroke. Circulation 103, 17991805 (2001). 9. Cheng, T. et al. Activated protein C blocks p53-mediated apoptosis in ischemic human brain endothelium and is neuroprotective. Nat. Med. 9, 338342 (2003). 10. Joyce, D.E., Gelbert, L., Ciaccia, A., DeHoff, B. & Grinnell, B.W. Gene expression profile of antithrombotic protein C defines new mechanisms modulating inflammation and apoptosis. J. Biol. Chem. 276, 1119911203 (2001). 11. Riewald, M., Petrovan, R.J., Donner, A., Mueller, B.M. & Ruf, W. Activation of endothelial cell protease activated receptor 1 by the protein C pathway. Science 296, 18801882 (2002). 12. Guo, H. et al. Activated protein C prevents neuronal apoptosis via protease activated receptors 1 and 3. Neuron 42, 563572 (2004). 13. Lo, E.H., Broderick, J.P. & Moskowitz, M.A. tPA and proteolysis in the neurovascular unit. Stroke 35, 354356 (2004). 14. Sharp, C.D., et al. Glutamate causes a loss in human cerebral endothelial barrier integrity through activation of NMDA receptor. Am. Physiol. Heart Circ. Physiol. 285, H2592H2598 (2003). 15. Budd, S.L., Tenneti, L., Lishnak, T. & Lipton, S.A. Mitochondrial and extramitochondrial apoptotic signaling pathways in cerebrocortical neurons. Proc. Natl. Acad. Sci. USA 97, 61616166 (2000). 16. Stennicke, H.R., et al. Pro-caspase-3 is a major physiologic target of caspase-8. J. Biol. Chem. 273, 2708490 (1998). 17. Zhang, W., et al. Fundamental role of the Rip2/caspase-1 pathway in hypoxia and ischemia-induced neuronal cell death. Proc. Natl. Acad. Sci. USA 100, 1601216017 (2003). 18. Du, Y.S., et al. Activation of a caspase 3-related cysteine protease is required for glutamate-mediated apoptosis of cultured cerebellar granule neurons. Proc. Natl. Acad. Sci. USA 94, 1165711662 (1997). 19. Okamoto, S., et al. Dominant interfering forms of MEF2 generated by caspase cleavage contribute to NMDA-induced neuronal apoptosis. Proc. Natl. Acad. Sci. USA 99, 39743979 (2002). 20. Penninger, J.M. and Kroemer, G. Mitochondria, AIF and caspasesrivaling for cell death execution. Nat. Cell Biol. 5, 9799 (2003). 21. Yu, S., et al. Mediation of poly (APD-Ribose) polymerase-1-dependent cell death by apoptosis-inducing factor. Science 297, 259263 (2002). 22. Tabrizi, P. et al. Tissue plasminogen activator (tPA) deficiency exacerbates cerebrovascular fibrin deposition and brain injury in murine stroke model: studies in tPAdeficient mice and wild-type mice on a matched genetic background. Arterioscler. Thromb. Vasc. Biol. 19, 28012806 (1999). 23. Jiang, Q. et al. Magnetic resonance imaging indexes of therapeutic efficacy of recombinant tissue plasminogen activator treatment of rat at 1 and 4 hours after embolic stroke. J. Cereb. Blood Flow Metab. 20, 2127 (2000). 24. Pichiule, P. & LaManna, J.C. Angiopoietin-2 and rat brain capillary remodeling during adaptation and deadaptation to prolonged mild hypoxia. J. Appl. Physiol. 93, 11311139 (2002). 25. Tibbetts, M.D., Zheng, L. & Lenardo, M.J. The death effector domain protein family: regulators of cellular homeostasis. Nat. Immunol. 4, 404409 (2003).

2004 Nature Publishing Group http://www.nature.com/naturemedicine

NATURE MEDICINE VOLUME 10 | NUMBER 12 | DECEMBER 2004

1383