NMDA blockade attenuates caspase-3 activation and DNA fragmentation after neonatal hypoxia–ischemia

NEUROREPORT

DEVELOPMENTAL NEUROSCIENCE

NMDA blockade attenuates caspase-3 activation and DNA fragmentation after neonatal hypoxia±ischemia Malgorzata Puka-Sundvall,1,2,CA Ulrika Hallin,2 Changlian Zhu,2 Xiaoyang Wang,2 Jan-Olof Karlsson,1 Klas Blomgren2,3 and Henrik Hagberg4 Perinatal Center, Department of Physiology and Department of Anatomy and Cell Biology, GoÈteborg University, GoÈteborg, Sweden; Departments of 1 Anatomy and Cell Biology, 2 Physiology, 3 Pediatrics and 4 Obstetrics and Gynecology, GoÈteborg University, Box 432, S-405 30 GoÈteborg, Sweden CA

Corresponding Author and Address

Received 26 April 2000; accepted 22 June 2000

The aim was to study the effects of an NMDA receptor antagonist on caspase-3 activation and DNA fragmentation after hypoxia±ischemia (HI) in 7-day-old rats. Animals were treated with vehicle or MK-801 (0.5 mg/kg) directly after HI and sacri®ced 8, 24 or 72 h later. MK-801 reduced injury (by 53%), cells positive for active caspase-3 (by 39%) and DNA fragmentation (by 79%) in the cerebral cortex. Furthermore,

MK-801 signi®cantly decreased caspase-3 activity, and Western blots revealed a tendency towards decreased proteolytic cleavage of the caspase-3 proform. The data imply that NMDA receptors are involved in the activation of apoptotic processes in the immature brain after HI. NeuroReport 11:2833±2836 & 2000 Lippincott Williams & Wilkins.

Key words: Apoptosis; Brain damage; Caspase-3; Hypoxia±ischemia; NMDA receptor; Neonatal

INTRODUCTION

Excitatory amino acid (EAA) receptors of the NMDA subtype appear to be critical for development of hypoxicischemic (HI) injury in the immature brain [1,2]. Excitotoxic and HI cell injury may exhibit some apoptotic features [3], but have both been characterized as a predominantly necrotic type of cell death, being distinctly different morphologically from developmental apoptosis [4]. A newly discovered family of cysteine proteases, caspases, has been found to be critical in the process of apoptosis in many cell types [5]. Activation of caspase-3 leads to cleavage of the inhibitor of caspase-activatedDNase (ICAD) which triggers the activation of caspaseactivated DNase (CAD) and subsequent DNA fragmentation [6]. Caspase-3 is markedly activated in the immature brain after HI: the activity increases, pro-caspase-3 (32 kDa) decreases parallel to the accumulation of caspase-3 fragments (29 kDa, 17 kDa) and the appearance of cells immunopositive for active caspase-3 at 3±24 h of reperfusion [7,8]. Furthermore, administration of caspase-3 inhibitors reduces brain injury even if applied several hours after the insult [7]. These data suggest that caspase-3-dependent, apoptotic mechanisms are important for development of brain injury after HI, which appears contradictory considering that the morphology has been described to be necrotic and distinctly different from apoptosis as it occurs in the immature CNS [4]. Furthermore, the NMDA recep-

0959-4965 & Lippincott Williams & Wilkins

tor antagonist MK-801, given in a dose of 0.5 mg/kg 3 3, was recently shown to induce widespread apoptotic cell death in the brain of immature rats, reaching a maximum 24 h after administration [9]. The aim of the present study was to evaluate the extent to which caspase-3 activation and DNA fragmentation after HI are affected by a neuroprotective dose of the NMDA receptor antagonist MK-801 administered after HI [2].

MATERIALS AND METHODS

Neonatal HI was induced in rats at postnatal day (PND) 7 [2]. Wistar rats (Moellegaard, Denmark), of either sex were anesthetized with en¯uran (3% for induction, 1.5% for maintenance) in nitrous oxide/oxygen (1:1). The left common carotid artery was ligated and cut and HI was induced for 55 min by exposure to 7.7  0.01% oxygen in nitrogen in a humidi®ed chamber at 368C. Immediately after hypoxic exposure animals were injected i.p. with 0.5 mg/kg MK-801 or vehicle (ICN Biomedicals Inc. Ohio, USA, Chemicon AB). All animal experiments were approved by the local Ethical Committee in GoÈteborg (NR: 267/96, 183/99). Animals (n ˆ 97) were sacri®ced by decapitation 8, 24 and 72 h post-HI. Control animals were sacri®ced 24 h after MK-801 administration. The brains were rapidly removed and frozen in isopentane and dry ice and stored at ÿ808C. Cortical tissue rostral to the hippocampus (80 mg) was

Vol 11 No 13 11 September 2000

2833

NEUROREPORT dissected out at ÿ108C. The tissue was homogenized by sonication in ice-cold 50 mM Tris±HCl (pH 7.3), containing 5 mM EDTA, aliquoted and stored at ÿ808C. The protein concentration was determined according to Whitaker and Granum [10], adapted for microplates. Immunoblotting: Homogenate samples (75 ìg total protein) were run on 8±16% Tris±glycine gels (Novex, San Diego, CA, USA) and transferred to reinforced nitrocellulose (Optitran, 0.2 ìm, Schleicher and Schuell, Dassel, Germany) or PVDF (Hybond-P, Amersham, UK) membranes. The antibody against caspase-3 (H-277, Santa Cruz Biotechnology, Santa Cruz, CA, USA) was used diluted 1:1000. Immunoreactive bands were visualized using a peroxidase-conjugated secondary antibody (Vector Laboratories, Burlingame, CA, USA), Super Signal West PICO or DURA (Pierce, Rockford, IL, USA) and a LAS 1000 cooled CCD camera (Fuji®lm, Tokyo, Japan). Bands were quanti®ed using the Image Gauge software (version 3.3, Fuji®lm, Tokyo, Japan). Fluorometric assay of caspase-3-like activity: Samples of homogenate (50 ìl) were assayed for DEVD-AMC-cleaving activity [7], and results were expressed as AMC released/ min/mg protein. Every sample was analyzed 3±4 times and the average value was used as n ˆ 1. Immunohistochemistry and hybridization: Paraf®n sections (5 ìm) from control and 24 h post-HI animals (n ˆ 32) were used for MAP2 (Sigma, clone HM-2), hairpin probe (HPP) or caspase-3 staining. The caspase-3 antibody was raised against the p17 fragment (residues 176±277, Pharmingen, 67341A). The biotinylated oligonucleotide HPP with one nucleotide overhang in the 39 end was synthesized by Scandinavian Gene Synthesis (KoÈping, Sweden) according to Didenko et al. [11,12], but with three biotins instead of ®ve [8]. All secondary antibodies and Vectastain ABC were from Vector Laboratories (Burlingame, CA). The stainings were performed according to Zhu et al. [8]. Brain damage: Cerebral damage at 24 h post-HI was evaluated as loss of MAP2 immunoreactivity using NIH Image 1.62. Hemisphere damage was de®ned as: [1ÿ(MAP2-positive area in the ipsilateral hemisphere/total area of the contralateral hemisphere)] 3 100%, and cortical damage was de®ned as: [1ÿ(MAP2-positive area in the ipsilateral cortex/total area of the contralateral cortex)] 3 100%. HPP and caspase-3 positive cells: HPP and caspase-3positive cells were counted in one visual ®eld (0.14 mm2 ) in four de®ned areas: parietal cortex, CA1, CA3 and dentate gyrus (DG), and cell density was expressed per mm2 . Statistics:

The Mann±Whitney U-test was used.

RESULTS

Treatment with MK-801 signi®cantly reduced hemispheric brain damage (by 47%; p , 0.05) and cerebrocortical injury (by 53%, p , 0.05) after HI. Very few HPP-positive cells were found in the cerebral cortex (Fig. 1) and hippocampus

2834

Vol 11 No 13 11 September 2000

M. PUKA-SUNDVALL ET AL.

(a)

HPP-positive cells in parietal cortex sample (cells/mm2) 1400

**

1200

vehicle

1000

MK-801 0.5 mg/kg

800 600 400 NS 200 0 1400

Caspase-3-positive cells in parietal cortex sample (cells/mm2)

(b) **

1200 1000 800 600 NS 400 200 0 control

HI 1 24 h

Fig. 1. HPP-positive (a) and caspase-3-positive (b) cell density (per mm2 ) was evaluated in a de®ned area of the ipsilateral parietal cortex in controls and 24 h after HI. The difference between vehicle- and MK-801treated groups was compared by unpaired Mann±Whitney U-test:  p , 0.01.

of non-HI controls in both the MK-801 and the vehicle group. Occasional caspase-3-positive cells were found in the hippocampus and in parietal cortex (Fig. 1). The number of both HPP- and caspase-3-positive cells was considerably higher in HI animals than in controls (Fig. 1). In the parietal cortex of vehicle-treated animals 24 h after HI, the density of HPP- and caspase-3-positive cells was almost identical (924  173 and 979  122, respectively). MK-801 administration decreased the number of both HPP- and caspase-3-positive cells (to 189  78 and 600  101, respectively, p , 0.01). A similar pattern was seen in the hippocampus 24 h after HI. MK-801 decreased the number of HPP- and caspase-3-positive cells in the CA1 and CA3 subregions (Fig. 2). Changes in caspase-3 detected by immunoblotting in the ipsilateral hemisphere were expressed both as loss of uncleaved (32 kDa) and increase of the cleaved forms (29 kDa and 17 kDa) of caspase-3 (Fig. 3). In control animals, there was some detectable immunoreactivity of both cleaved forms: 29 kDa and 17 kDa (4.2  0.3 and 0.05  0.02%, respectively). MK-801 administration did not affect the 32 kDa or 29 kDa bands, but slightly increased the level of the 17 kDa form (0.05  0.02 vs 0.11  0.03%; Fig. 3).

NEUROREPORT

EFFECTS OF MK-801 ON CASPASE-3/DNA BREAKAGE

32 kDa band (% of contralateral side)

(a)

HPP-positive cells in hippocampus (cells/mm2)

vehicle

160

MK-801 0.5 mg/kg

vehicle

(a)

MK-801 0.5 mg/kg

150

140 ** 120

*

100

100 50

80 60

0

40

29 kDa band (% of 32 1 29 1 17 kDa)

(b)

17 kDa band (% of 32 1 29 1 17 kDa)

(c)

15

20 0 Caspase-3-positive cells in hippocampus (cells/mm2)

(b)

160

10

5

140 120

**

0

100 80

4

60

3

40

2

20 1

*

0 CA1

CA3

DG

Fig. 2. HPP-positive (a) and caspase-3-positive (b) cells were counted in one visual ®eld (magni®cation 3400) in de®ned areas of the ipsilateral hippocampus: CA1, CA3 and dentate gyrus (DG) 24 h after HI. The difference between vehicle- and MK-801-treated groups was compared using unpaired Mann±Whitney U-test:  p , 0.05,  p , 0.01.

HI ‡ 24 h of recovery induced a decrease of the 32 kDa band and elevations of the 29 kDa and 17 kDa fragments (to 11.4  3.11 and 2.70  1.07%, respectively; Fig. 3). MK801 did not affect the loss of pro-caspase 3 (32 kDa band) but tended to attenuate accumulation of active (17 kDa) caspase-3 (1.20  0.30% compared to 2.70  1.07%), although the difference was not statistically signi®cant. Caspase-3 activity (8.7  0.3 pmol AMC formed/min/mg protein) was detectable in the cerebral cortex of controls. Administration of MK-801 slightly decreased caspase-3 activity in brains of control animals. HI evoked a marked increase of caspase-3 activity at 8 h (49.3  5.4), reaching its peak at 24 h (158.4  9.5), and returned to near normal at 72 h (18.7  2.1; Fig. 4). MK-801 treatment reduced the activity of caspase-3 at 24 h (114.7  41.5; p , 0.01, compared with vehicle-injected group) but not at 8 h and 72 h of reperfusion (Fig. 4).

DISCUSSION

This is the ®rst report addressing the effect of NMDA receptor antagonists on caspase-3 activation and DNA

0 control

HI 1 24 h

Fig. 3. Cleavage of caspase-3 in the cerebral cortex in controls and 24 h after HI. (a) Immunoreactivity of the uncleaved (32 kDa) procaspase-3 form was expressed as a percentage of the contralateral side. The cleaved 29 kDa (b) and 17 kDa (c) bands of caspase-3 were expressed as a percentage of total caspase-3 immunoreactivity in the sample (sum of 32 kDa, 29 kDa and 17 kDa fragments). Vehicle- and MK801-treated groups were compared using Mann±Whitney U-test:  p , 0.05.

fragmentation after HI. Previous studies in adult animals have demonstrated that MK-801 reduced brain injury after reversible focal ischemia without affecting the degree of DNA fragmentation [13]. Furthermore, a strong synergism between NMDA receptor blockers and inhibitors of caspase-3 [14] indicated separate mechanisms of action. The NMDA receptor was supposed to be involved in rapid necrotic cell death [15], whereas caspase-3 triggers delayed apoptotic cell death [7,13,16]. The situation may be completely different in immature rats as repeated doses (3 3 0.5 mg/kg) of the NMDA receptor antagonist MK-801 induced massive apoptotic cell death in the cerebral cortex of 7-day-old rats 24 h after administration [9]. Therefore, we expected that treatment with MK-801 might well lead to rebound activation of caspase-3 and apoptotic cell injury, which would counteract the cerebroprotection, exerted by NMDA receptor antagonists. Indeed, we found

Vol 11 No 13 11 September 2000

2835

NEUROREPORT

200

M. PUKA-SUNDVALL ET AL.

Caspase-3 activity (pmol AMC 3 min(21) 3 mg protein(21)) **

vehicle MK-801 0.5 mg/kg

150

100

50 *

fragmentation in this model. However, there are noncaspase-dependent pathways whereby endonucleases could be activated and gain access to the nucleus [22] and several other DNases may be involved (e.g. DNaseY, acidic endonucleases, endonucleases released from mitochondria) especially in the initial phase of DNA fragmentation [23,24]. The apparent discrepancy between the effect of MK-801 on DNA fragmentation as compared to caspaseactivation indicates that NMDA receptor activation may induce DNA degradation unrelated to caspase-3. Alternatively, DNA fragmentation is delayed relative to caspase-3 activation and some of the caspase-3-positive cells would eventually become HPP-positive.

0 control

8h

24 h

72 h

Fig. 4. Caspase-3-like activity in control rats and at 8, 24 and 72 h after HI was measured in the ipsilateral (hypoxic±ischemic) cerebral cortex as DEVD cleavage. Vehicle- and MK-801-treated groups were compared by Mann±Whitney U-test:  p , 0.05,  p , 0.01.

a tendency towards increased numbers of caspase-3 positive cells and higher levels of the active 17 kDa fragment of caspase-3 in Western blots of control animals treated with MK-801. These differences were small, however, and partly offset by a slightly lower caspase-3 activity in cortical homogenates of controls treated with MK-801. Contrary to what was expected, MK-801 reduced DNA fragmentation (by 80%) after HI and also decreased the number of caspase-3-positive cells in the cerebral cortex and hippocampus. Furthermore, caspase-3 activation after HI was signi®cantly attenuated by MK-801 and cleavage of procaspase tended to be reduced as well. These data imply that NMDA-receptor activation after HI may be partly involved in the activation of caspase-3 as well as other processes leading to DNA fragmentation. We recently found that calcium-activated cysteine proteases (calpains) cleave and activate caspases after HI in 7-dayold rats. It is also known that NMDA-receptor activation triggers activation of calpains following ischemia [17], offering a direct mechanism whereby NMDA receptors may activate caspases. Furthermore, NMDA receptor activation leads to disturbances of calcium homeostasis [15] and mitochondrial function [18,19] and triggers the production of oxygen free radicals and nitric oxide [20], which may all be part of caspase-3 activation [21]. Proportionately, MK-801 decreased DNA fragmentation to a greater extent than its effect on caspase-3. A biotinylated oligonucleotide hairpin probe (HPP) was used for speci®c detection of double strand breaks with one nucleotide 39 overhang [12], which is one of the types of cleavage considered to be characteristic for apoptotic DNA fragmentation [11]. We have previously found a close correspondence between cells positive for active caspase-3 and HPP after HI [8], supporting the hypothesis that caspase-3 is responsible for activation of CAD, resulting in DNA

CONCLUSION

The NMDA receptor antagonist MK-801 attenuated the activation of caspase-3 and markedly reduced the DNA fragmentation in the CNS after HI, suggesting involvement of NMDA receptors in the activation of apoptotic mechanisms in neonatal animals.

REFERENCES

1. Barks JD and Silverstein FS. Brain Pathol 2, 235±243 (1992). 2. Hagberg H, Gilland E, Diemer NH and Andine P. Biol Neonate 66, 205±213 (1994). 3. Martin LJ, Al-Abdulla NA, Brambrink AM et al. Brain Res Bull 46, 281±309 (1998). 4. Ishimaru MJ, Ikonomidou C, Tenkova TI et al. J Comp Neurol 408, 461±476 (1999). 5. Cohen GM. Biochem J 326, 1±16 (1997) . 6. Enari M, Sakahira H, Yokoyama H et al. Nature 391, 43±50 (1998). 7. Cheng Y, Deshmukh M, D'Costa A et al. J Clin Invest 101, 1992±1999 (1998). 8. Zhu C, Wang X, Hagberg H and Blomgren K. J Neurochem (2000) In Press 9. Ikonomidou C, Bosch F, Miksa M et al. Science 283, 70±74 (1999). 10. Whitaker JR and Granum PE. Analyt Biochem 109, 156±159 (1980). 11. Didenko VV and Hornsby PJ. J Cell Biol 135, 1369±1376 (1996). 12. Didenko VV, Tunstead JR and Hornsby PJ. Am J Pathol 152, 897±902 (1998). 13. Endres M, Namura S, Shimizu-Sasamata M et al. J Cerebr Blood Flow Metab 18, 238±247 (1998). 14. Ma J, Endres M and Moskowitz MA. Br J Pharmacol 124, 756±762 (1998). 15. Choi DW. J Neurobiol 23, 1261±1276(1992). 16. Gillardon F, Bottiger B, Schmitz B et al. Mol Brain Res 50, 16±22 (1997). 17. Minger SL, Geddes JW, Holtz ML et al. Brain Res 810, 181±199 (1998). 18. Beal MF, Hyman BT and Koroshetz W. Trends Neurosci 16, 125±131 (1993) . 19. Gilland E, Puka-Sundvall M, Hillered L and Hagberg H. J Cerebr Blood Flow Metab 18, 297±304 (1998). 20. Eliasson MJL, Huang Z, Ferrante RJ et al. J Neurosci 19, 5910±5918 (1999). 21. Kroemer G, Dallaporta B and Resche-Rigon M. Annu Rev Physiol 60, 619±642 (1998). 22. Peitsch MC, Muller C and Tschopp J. Nucleic Acids Res 21, 4206±4209 (1993). 23. Liu QY, Ribecco M, Pandey S et al. Ann N Y Acad Sci 887, 60±76 (1999). 24. Walker PR, Leblanc J, Carson C et al. Ann N Y Acad Sci 887, 48±59 (1999).

Acknowledgements: This work was supported by the Swedish Medical Research Council (09455, 12213), the Swedish Society for Medical Research, the Swedish Society of Medicine, the Sven Jerring Foundation, the AÊke Wiberg Foundation, the Ê hlen Foundation, the Magnus Bergvall Foundation and Frimurare Barnhus Laerdal Foundation for Acute Medicine, the A Foundation. The authors would like to thank Mrs Eva Cambert and Mrs Anna-Lena Andersson for technical assistance.

2836

Vol 11 No 13 11 September 2000