[33 Copper Zinc Superoxide Dismutase Transgenic Brain in Neonatal Hypoxia Ischemia

By R. Ann Sheldon, Lynn Almli, and Donna M. Ferriero Introduction

Although significant progress has been made in characterizing the mechanisms of cellular injury in models of adult stroke, the neonatal brain is less well understood. Often, discoveries in the adult brain are either not duplicated in the developing brain or the results are quite different. We found this to be the case for copper/zinc superoxide dismutase (SOD1). Whereas research on adult animals has shown reduced cerebral ishemic injury in SODl-overexpressing mice,1-3 we found the opposite in a model of hypoxia-ischemia in neonatal animals of the same transgenic strain.4

Hypoxia-Ischemia as Model of Neonatal Stroke

The Vannucci model is a standard technique for neonatal stroke.5,6 It consists of a permanent, unilateral ligation of the common carotid artery combined with a period of hypoxia. The combination of ischemia and hypoxia produces an area of infarct ipsilateral to the ligation. Neither ligation nor hypoxia alone produces cell death discernable by standard histological techniques. Postnatal day 7 (P7) animals are used, as the brain development at this age is considered to be comparable to a term newborn human.


P7 mouse pups are removed from the dam as a litter and kept on a 37° heating pad. Pups are individually removed from the group and anesthetized with a continuous flow of 2.5% halothane in 40% oxygen, balance nitrogen. When the animal does not respond to a tail or foot pinch, an incision is made in the midline

1 H. Kinouchi, C. J. Epstein, T. Mizui, E. Carlson, S. F. Chen, and P. H. Chan, Proc. Natl. Acad. Sci. U.S.A. 88, 11158(1991).

2 G. Yang, P. H. Chan, J. Chen, E. Carlson, S. F. Chen, P. Weinstein, C. J. Epstein, and H. Kamii, Stroke 25, 165 (1994).

3 P. H. Chan, M. Kawase, K. Murakami, S. F. Chen, Y. Li, B. Calagui, L. Reola, E. Carlson, and C. J. Epstein, J. Neurosci. 18, 8292 (1998).

4 i. S. Ditelberg, R. a. Sheldon, C. J. Epstein, and D. M. Ferriero, Pediatr. Res. 39,204 (1996).

5 J. E. D. Rice, R. C. Vannucci, and J. B. Brierley, Ann. Neurol. 9, 131 (1981).

6 R. Vannucci, J. Connor, D. Mauger, C. Palmer, M. Smith, J. Towfighi, and S. Vannucci, J. Neurosci. Res. 55, 158(1999).

of the neck, no longer than 3-4 mm. The right common carotid artery is then exposed with the use of two curved forceps and permanently ligated with a bipolar coagulator (Codman & Schurtleff, Randolph, MA). The incision is closed and the pup is returned to the heating pad. The ligation procedure should last no longer than 5 min. When all pups are ligated, they are weighed and returned to the dam, where they remain for 1.5 hr for recovery and feeding.


Pups are again removed from the dam and placed in airtight containers partially submerged in a 37° water bath. Humidifed 8% oxygen, balance nitrogen flows through the containers after passing a flowmeter (Manostat; Barnant, Barrington, IL), which maintains a constant rate of flow. We have determined that the ultimate degree of injury is based not only on the duration of hypoxia, but also on the strain of mouse being used. Hence, the duration of hypoxia used is dependent on the strain.7

Superoxide Dismutase-Overexpressing Mouse in Hypoxia-Ischemia

The first application of the Vannucci procedure in mice was in our study of the effect of SOD overexpression in transgenic animals, developed by Epstein et al.8 We modified the original rat model of hypoxia-ischemia for use in mice, which allowed us to investigate the effect of SOD in this model by utilizing transgenic mice that overexpress SOD.4 Anesthesia and the ligation procedure are the same as for rats, only scaled down for pups about one-quarter the size of rats. The incision is not sutured in mice, but rather pinched together, as the dam will chew out the sutures. A small drop of methyl methacrylate adhesive may also be used to close the incision. We determined that the ideal duration of hypoxia for SOD-overexpressing pups is 90 min. This produces a severe degree of injury. We have subsequently found that 45 min of hypoxia produces a moderate degree of injury, while minimizing mortality, in the background strain CD1.

Adults of the same SOD-overexpressing mouse strain were found to have reduced injury in a model of focal cerebral ischemic injury, leading to the reasonable conclusion that, in these animals, SOD is protective, and that the superoxide radicals generated by ischemia are, at least in part, responsible for pathogenesis.1 With neonatal hypoxia-ischemia (HI), however, we found that the SOD animals were not only not protected, they were more frequently and more severely injured than nontransgenic littermates.4 Assuming that superoxide radicals were indeed

7 R. A. Sheldon, C. Sedik, and D. M. Ferriero, Brain Res. 810, 114 (1998).

8 C. J. Epstein, K. B. Avraham, M. Lovett, S. Smith, O. Elroy-Stein, G. Rotman, C. Bry, and Y. Groner, Proc. Natl. Acad. Sci. U.S.A. 84, 8044 (1987).

responsible for initiating pathogenesis, and that the superoxide dismutase reaction generates hydrogen peroxide (H2O2), we would expect an accumulation of H2O2 after HI. Under normal conditions, endogenous glutathione peroxidase (GPx) and catalase convert H2O2 to oxygen and water. We looked at how these enzymes were affected after HI in the SOD1 transgenic pups, as well as levels of H2O2.9 Activity of both catalase and glutathione peroxidase was the same in the SOD overexpres-sors and wild-type P7 mice. In response to HI, GPx activity decreased significantly by 24 hr posthypoxia in both the transgenic and wild-type pups. In addition, an increase in H2O2 accumulation was seen at 24 hr in the transgenic pups only. Catalase was unchanged in both transgenic and wild-type mice. We also examined the role of iron in oxidative stress due to HI. The iron chelator deferoxamine (DFO; Sigma, St. Louis, MO) has been shown to be neuroprotective in the rat model of HI.10 Using the same experimental paradigm (DFO at 100 mg/kg 10 min after HI and again at 24 hr) in SOD overexpressors, we showed that wild-type mice were protected by iron chelation, but that there was no significant protection in the SOD transgenic mice.11

Histological Methods

Seven days after hypoxia-ischemia, mice are injected with a lethal dose of pentobarbital (50 mg/kg) and perfused transcardially with cold 4% (w/v) paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Brains are then removed and postfixed in the same fixative for 4 hr. The forebrain is sectioned on a Vibratome (Ted Pella, Redding, CA), and alternate 50-¡xm sections are collected for cresyl violet and enhanced Perl's iron stain.12-14 Analysis of brains with both stains increases the ability to visualize dead and dying cells microscopically. For quantification of the degree of injury, we have developed a scoring system for mice, which reflects the greater degree of injury seen in the hippocampus. Eight regions are given a score from 0 to 3: 0, no injury; 1, mild with focal areas of cell loss; 2, moderate, such as a majority of cells dead in the pyramidal cell layer of the hippocampus, or columnar areas of cell loss in the cortex; and 3, cystic infarction. The regions scored are the anterior, middle, and posterior cortex; the striatum/caudate putamen

9 H. J. Fullerton, J. S. Ditelberg, S. F. Chen, D. P. Sarco, P. H. Chan, C. J. Epstein, and D. M. Ferriero, Ann. Neurol. 44, 357 (1998).

10 C. Palmer, R. L. Roberts, and C. Bero, Stroke 25, 1039 (1994).

11 D. Sarco, J. Becker, C. Palmer, R. A. Sheldon, and D. M. Ferriero, Neurosci. Lett. 282,113 (2000).

12 C. Palmer, S. L. Menzies, R. L. Roberts, G. Pavlick, and J. R. Connor, J. Neurosci. Res. 56, 60 (1999).

13 J. R. Connor, G. Pavlick, D. Karli, S. I. Menzies, and C. Palmer, J. Comp. Neurol. 355,111 (1995).

14 J. Nguyen-Legros, J. Bizot, M. Bolesse, and J. P. Pulicani, Histochemistry 66, 239 (1980).

cortex cortex

Fig. 1. Schematic representation of a coronal section of a mouse brain, showing areas of injury (dots). CA1, CA2, and CA3 are regions of hippocampus that are commonly injured in this model. DG, Dentate gyrus.

(as a whole); and CA1, CA2, CA3, and dentate gyrus of the hippocampus. Taken together, a score of 0-24 is possible (Figs. 1 and 2).

Cresyl Violet

Cresyl violet is a standard histological stain for neurons. Sections are first mounted onto gelatin-coated slides and dried overnight. Neonatal brains do not need to be delipidized, and after a rinse in H2O slides are immersed in stain for 3-5 min. Three components are made and then mixed: 0.3 g of cresyl echt violet (Roboz) in 50 ml of H20, 3.48 ml of glacial acetic acid in 300 ml of H20, and 5.44 g of sodium acetate in 200 ml of H20. Slides are then rinsed twice in H20, differentiated in 70% (v/v) ethanol with a few drops of acetic acid, followed by dehydration in graded ethanols and two changes of xylene, and coverslipped with Depex (Biomedical Specialties, Santa Monica, CA).

Perl's Iron Stain

Perl's iron stain is enhanced with diaminobenzidine (DAB).13 Free-floating sections are incubated for 30 min in 2% (w/v) potassium ferrocyanide (Sigma) mixed 1:1 with 2% (v/v) hydrochloric acid, rinsed three times in H20, and

Fig. 2. Three different representative brains from SODl-overexpressing mice showing potential injury outcome. Cresyl violet stain (A) and iron stain (B) of a brain with small focal areas of.cell loss (arrowheads). Injury score, 4. Cresyl violet stain (C) and iron stain (D) of a brain with greater cell loss in the pyramidal cell layer and general shrinkage of the hippocampus and prominent injury to the cortex (arrowheads). Injury score, 16. Cresyl violet stain (E) and iron stain (F) of a severely injured brain. Note the large area of cystic infarction to the cortex (asterisk). Injury score, 23. There is relative sparing of the dentate gyrus in all three brains. Scale bars: 0.25 mm.

Fig. 2. Three different representative brains from SODl-overexpressing mice showing potential injury outcome. Cresyl violet stain (A) and iron stain (B) of a brain with small focal areas of.cell loss (arrowheads). Injury score, 4. Cresyl violet stain (C) and iron stain (D) of a brain with greater cell loss in the pyramidal cell layer and general shrinkage of the hippocampus and prominent injury to the cortex (arrowheads). Injury score, 16. Cresyl violet stain (E) and iron stain (F) of a severely injured brain. Note the large area of cystic infarction to the cortex (asterisk). Injury score, 23. There is relative sparing of the dentate gyrus in all three brains. Scale bars: 0.25 mm.

then reacted with DAB [20 mg/10 ml of phosphate buffer, 13.3 /xl of 30% (v/v) H2O2]. When DAB is visibly deposited in injured areas of tissue, the reaction is stopped by rinsing the sections three times in H2O. Sections are then mounted onto gelatin-coated slides, allowed to dry, dehydrated in graded ethanols, and cover-slipped.

Enzymatic Assays

Right and left cortices and hippocampi are dissected and flash frozen on dry ice.15

Glutathione Peroxidase

GPx is assayed according to Roveri et al.,16 with minor modifications. Samples are homogenized in phosphate buffer (0.1 M potassium phosphate, 2 mM sodium azide, pH 7.0), and centrifuged at 5000g for 5 min. GPx activity is determined kinetically in duplicate samples by monitoring the decrease in absorbance of NADPH at 340 nm.15 GPx activity is expressed as units per milligram of protein. One unit is defined as 1 nmol of NADPH consumed per minute.


Catalase is assayed by previously described methods, again with minor modifications.17-19 Brain samples are homogenized in 0.2 ml of cold phosphate buffer-0.1 mMEDTA-0.1 % (v/v) Triton X-100 (pH 7.8,4°), and purified by addition to glass wool columns resting in microcentrifuge tubes and centrifugation at 5000g for 5 min at 4°. Samples of cortex are diluted 1:10 with buffer, and hippocampus samples are diluted 1:5 with buffer. Catalase activity is determined kinetically in duplicate samples by monitoring the decrease in absorbance of a known amount of H202 at 240 nm.19 Catalase activity is expressed as units per milligram of protein. One unit is defined as 1 ¿imol of H202 reduced per minute. Values are normalized to an internal control consisting of a homogenate of several cortices or hippocampi from naive animals.

Hydrogen Peroxide

H202 is measured indirectly by treatment of the animals with aminotria-zole (Sigma), which selectively and irreversibly inhibits catalase that is bound to H202, and is therefore directly proportional to H202 concentration at the time of inhibition.20'21 We first determined that there was no difference in brain aminotri-azole levels between transgenic and wild-type mice 2 hr after treatment. Amino-triazole levels were determined by diazorization of the aminotriazole with sodium

15 L. Flohe and W. A. Giinzler, Methods Enzymol. 105, 114 (1984).

16 A. Roveri, M. Maiorino, and F. Ursini, Methods Enzymol. 233, 202 (1994).

17 S. Przedborski, V. Jackson-Lewis, V. Kostic, E. Carlson, C.J. Epstein, and J. L. Cadet, / Neurochem. 58, 1760 (1992).

18 P. M. Sinet, R. E. Heikkila, and G. Cohen, J.Neurochem. 34, 1421 (1980).

20 E. Margoliash, A. Novogrodsky, and A. Schejter, Biochem. J. 74, 339 (1960).

21 P. Nicholls, Biochim. Biophys. Acta 59, 414 (1962).

nitrate, followed by coupling to a chromotropic acid (Sigma) to form a visible derivative that is determined by measuring absorbance at 595 nm.18,22 We then performed a time course of catalase activity on the basis of this information and determined that there was a 50% inhibition of catalase 2 hr after injection. Thus, mice are injected intraperitoneally with aminotriazole (200 mg/kg in saline) or saline 2 hr before they are killed and brains are dissected for assay. Catalase is then assayed as described above; greater inhibition of the catalase enzyme indicates a higher H202 concentration. Results are expressed as percent inhibition of catalase activity by aminotriazole.

Total protein in homogenates is determined for all assays with bicinchoninic acid (BCA) reagents by the BCA method (Pierce, Rockford, IL).

Oxidative Stress on Neurons in Vitro

We used primary neuronal cell cultures to further study the mechanisms of H202 toxicity. In cortical cultures of pure wild-type neurons, we found that immature neurons (6 days in vitro) are selectively vulnerable to H202 exposure and mature neurons (20 days in vitro) are relatively resistant.23 Cultures of hippocampal cells, which show greater injury in HI, showed the same vulnerability as immature cells, using similar methods.24'25 We also examined the impact of free metal ions in this system, as they are known to catalyze hydroxyl radical formation from H202 via the Fenton reaction, which is a one-electron nonenzymatic transfer reaction in which transition metals generate hydroxyl radicals from H202. When cultures are pretreated with the iron chelator deferoxamine (DFO), H202-mediated hippocampal cell death is attenuated, but treatment of cells at the same time as H202 exposure is ineffective. Consequently, the broad-spectrum metal chelator N,N,N',Af'-tetrakis(2-pyridylmethyl)ethylenediamine (TPEN) is used in a similar fashion to determine whether other metals are involved. TPEN added at the same time as H202 consistently reduces cell death by more than 50%.

Cell Culture Methods

The preparation of primary neuronal cultures is based on previously described methods,26 with some modifications. Hippocampal cells are isolated from fetal (embryonic day 16, E16) CD1 mice. Briefly, hippocampi are removed from adjacent cortices and meninges and enzymatically dissociated with trypsin (2 mg/ml; Sigma) in Hanks' buffered salt solution (HBSS) for 10 min at 37°. Dissociation is

22 F. O. Green and R. N. Feinstein, Anal. Chem. 29, 1658 (1957).

23 R. E. Mischel, Y. S. Kim, R. A. Sheldon, and D. M. Ferriero, Neurosci. Lett. 231, 17 (1997).

24 L. M. Almli, S. E. G. Hamrick, A. Koshy, M. Tauber, and D. M. Ferriero, Dev. Br. Res. 132, 121 (2001).

25 L. M. Almli, P. H. Donohoe, M. Taeuber, and D. M. Ferriero, Soc. Neurosci. Abstr. 25,1522 (1999).

26 D. Choi, M. Maulucci-Gedde, and A. Kriegstein, J. Neurosci. 7, 357 (1987).

stopped by the addition of horse serum (10%, v/v) and the cell suspension is cen-trifuged at low speed (190g). The cells are resuspended in astrocyte conditioned medium (ACM) supplemented with 10% (v/v) horse serum. ACM is prepared from a custom base of minimum essential medium (MEM) with Earle's salts, without L-glutamine, which is supplemented with glucose (15 mM), glutamine (2 mM), and 10% (v/v) fetal bovine serum immediately before plating on a confluent layer of astrocytes in 75-ml flasks, 24 hr before use. After gentle trituration, the cells are diluted in ACM and plated at a density of 1.4 x 106 cells/ml onto 96-well plates (Falcon; Becton Dickinson Labware, Lincoln Park, NJ) that have been coated with poly-D-lysine (5 mg/100 ml of pyrogen-free water; Sigma). Thirty minutes after plating, half of the medium is removed and replaced with fresh ACM without horse serum, reducing the concentration of horse serum to 5% (v/v). Twenty-four hours later (1 day in vitro), astrocyte growth is inhibited by the addition of 10 mM cytosine arabinoside, (Ara-C; Sigma). At 2 days in vitro, medium is replaced such that the concentration of Ara-C is reduced to 3 mM. This procedure ensures an astrocyte population of less than 5%, which is confirmed in each experiment by immunocytochemistry for astrocytes, using glial fibrillary acidic protein (GFAP) antibody (ICN, Costa Mesa, CA) and neuron-specific enolase (NSE) antibody (Dako, Carpinteria, CA) according to standard methods. The astrocyte cultures are prepared from cortices isolated from postnatal day 1 (PI) mice and processed in a manner similar to the above-described procedure, except that they are plated onto untreated 75-ml flasks without the addition of Ara-C, and allowed to grow to confluency.


The neonatal brain in the SOD1 transgenic mouse accumulates H2O2 in response to HI, contributing to cell death, as H2O2 is known to be toxic to neurons as well as other cell types.27 Although some adult stroke models have shown protection with increased SOD1, we have found that the immature brain is particularly vulnerable to oxidative stress produced by excess SOD1. Similarly, we have seen that immature neurons (day 6 in vitro) are more susceptible to injury and death from H202 exposure than mature neurons (20 days in vitro).23 The mechanism of H202 toxicity is unclear, but one likely pathway is the conversion of H2O2 to the highly toxic hydroxyl radical in the presence of Fe2+ via the Fenton reaction.28 The immature brain has relatively high levels of iron and it has been shown that there is an accumulation of iron in response to HI by as early as 4 hr.12-29 We

27 T. Rando and C. Epstein, Ann. Neurol. 46, 135 (1999).

28 J. A. Imlay, S. M. Chin, and S. Linn, Science 240, 640 (1988).

29 A. J. Roskams and J. R. Connor, J. Neurochem. 63,709 (1994).

have shown that the iron chelator deferoxamine protects wild-type brains from HI, but not SOD overexpressors.11 Furthermore, our in vitro data with metal chelators indicate that the presence of iron and other free metal ions, such as zinc, and the subsequent generation of hydroxyl radicals contribute to neuronal death in culture.

The endogenous antioxidant enzymes glutathione peroxidase and catalase are low in neonates compared with mature animals, which may explain the increased accumulation of H2O2 in SOD1 overexpressors in the setting of HI, and subsequently the greater degree of injury seen in these animals. The fact that catalase is unchanged, and GPx decreases, after HI, further indicates that these enzymes are incapable of ameliorating the deleterious effects of an overproduction of H202.

SOD1 neurons in culture, taken from the same strain of mice that we have used, have been shown to be more vulnerable than wild-type neurons to the superoxidegenerating compounds menadione and paraquat.30 Like us, Ying et al. hypothesized that H2O2 production is the likely mechanism of this toxicity.

It has been shown clinically that SOD levels increase after stroke in adults, indicating that superoxide radicals are formed and a compensatory effort is made by SOD production.31 However, SOD activity in the sera of stroke patients has been shown to be reduced and is inversely correlated with infarct size, indicating a need for replacement of antioxidants in these patients.32 Our data indicate that babies may be even more susceptible to ischemic injury, and that injury could be exacerbated by SOD. Also, we need to understand the pathogenesis of hypoxic-ischemic injury in order to find therapies that can be administered after the initial insult, but before irreparable cellular injury has occurred. Although treatments that are administered after HI may shed light on the mechanisms of injury, they have limited clinical relevance. Targeting the downstream reactive oxygen species is a part of this goal of finding appropriate therapies.

30 W. Ying, C. M. Anderson, Y. Chen, B. A. Stein, C. S. Fahlman, J.-C. Copin, P. H. Chan, and R. A. Swanson, J. Cereb. Blood Flow Metab. 20, 359 (2000).

31 N. Gruener, B. Gross, O. Gozlan, and M. Barak, Life Sei. 54, 711 (1994).

32 M. Spranger, S. Krempien, S. Schwab, S. Donneberg, and W. Hacke, Stroke 28, 2425 (1997).

0 0

Post a comment