Extended role of necrotic cell death after hypoxia–ischemia-induced neurodegeneration in the neonatal rat

www.elsevier.com/locate/ynbdi Neurobiology of Disease 27 (2007) 354 – 361

Extended role of necrotic cell death after hypoxia–ischemia-induced neurodegeneration in the neonatal rat Silvia Carloni, a Andrea Carnevali, b Mauro Cimino, a and Walter Balduini a,⁎ Istituto di Farmacologia e Farmacognosia, Università degli Studi di Urbino “Carlo Bo”, Via S. Chiara 27, 61029 Urbino, Italy Unità Operativa di Anatomia Patologica, Azienda Ospedaliera S. Salvatore, Pesaro, Italy

a

b

Received 28 February 2007; revised 24 April 2007; accepted 4 June 2007 Available online 18 June 2007 The relative contribution of apoptosis and necrosis after neonatal hypoxia–ischemia (HI) is still a matter of debate. Here we determined the time course of necrotic cell death after neonatal HI and its relationship to caspase-3 activation and apoptotic cell death. Necrosis was evaluated by intracerebroventricular injection of propidium iodide (PI) before sacrificing the animal and processing brain sections for caspase-3 immunohistochemistry and TUNEL assay. PI-positive cells were found starting from 30 min after HI and increased rapidly in different brain areas. PI co-localized with the neuronal-specific nuclear marker NeuN but not with GFAP indicating that the dye label neurons with damaged plasma membrane but not reactive astrocytes. In the cerebral cortex 24 h after HI, the superficial layers showed cells with strong caspase-3 and TUNEL staining and with nuclei having apoptotic morphology whereas the deep layers of the cortex and the hippocampus showed cells with necrotic features. At later times, cells of the superficial layers were positive to PI, caspase-3, TUNEL and cathepsin-B. These data indicate that necrosis has an extended role in the progression of brain injury after neonatal HI and that a different spectrum of suicidal programs can be activated in the same cell. The extended period of caspase-3 activation in PI-positive necrotic cells supports the possibility that the apoptotic-to-necrotic continuum may ensue as the result of an incomplete execution of the apoptotic program. © 2007 Elsevier Inc. All rights reserved. Keywords: Ischemia; Apoptosis; Necrosis; Newborn rat; Caspases

Introduction Hypoxic–ischemic brain injury is a well-established cause of mortality and neurological morbidity in infants. The mechanisms responsible for the extensive neurodegeneration which occurs after Abbreviations: GFAP, glial fibrillary acidic protein; HI, hypoxia– ischemia; NeuN, neuron-specific nuclear protein; PI, propidium iodide; PN, postnatal day; TUNEL, terminal dUDP Nick-end labeling. ⁎ Corresponding author. Fax: +39 0722 303521. E-mail address: [email protected] (W. Balduini). Available online on ScienceDirect (www.sciencedirect.com). 0969-9961/$ - see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.nbd.2007.06.009

hypoxia–ischemia (HI) during brain development are not completely understood but there is a general agreement that progression of cell death takes place with a first rapid phase of necrotic cell death followed by a second delayed phase, which last from hours to days, characterized by inflammation and apoptotic cell death (Northington et al., 2001; Balduini et al., 2004). Apoptosis is an energy-dependent delayed process that involves activation of cell-suicide genes and is a physiological mechanism actively present in the brain during its maturation. The apoptotic death program is executed by caspases and caspase-3, the main executioner of this program, is strongly activated after HI in neonatal rats. Necrosis, in contrast, can occur as an unregulated response or as a result of activation of specific signal transduction pathways (Golstein and Kroemer, 2007) and is characterized by failure of osmotic regulation, depletion of ATP and rapid permeabilization of plasma membrane. Its morphological features are mitochondrial and nuclear swelling, dissolution of organelles, condensation of chromatin around the nucleus and degradation of DNA and is common in a wide variety of pathological conditions, including neonatal HI (Martin et al., 1998; Northington et al., 2001; Nieminen, 2003). Necrosis is associated with an inflammatory response that may facilitate wound healing. Because of the propensity of the neonatal brain to activate programmed cell death pathways following HI (Gill et al., 2002), the prolonged expression of caspase-3 and the persistence of cells with apoptotic morphology after HI-induced brain damage (Nakajima et al., 2000), apoptosis is considered as a prominent type of cell death in neonatal HI. In general, however, necrotic cell death predominates in the ischemic core whereas apoptosis occurs mainly in areas with milder ischemic injury (Nakajima et al., 2000; Northington et al., 2001). Increasing evidence, however, indicates that cell death phenotypes are more heterogeneous. The death of neurons, indeed, can also be a hybrid of apoptosis and necrosis with overlapping characteristics, the apoptosis-to-necrosis continuum or aponecrosis and the classical necrosis and apoptosis could be only the extreme of these hybrid forms of cell demise (Formigli et al., 2000; Nakajima et al., 2000; Northington et al., 2005; Wei et al., 2006). In severe ischemic situations, necrosis could simply substitute for failed apoptosis, as observed in vitro

S. Carloni et al. / Neurobiology of Disease 27 (2007) 354–361

when caspase inhibitors are applied following chemical hypoxia, which cause cells to shift from aponecrosis to necrosis (Formigli et al., 2000). To gain insight on the interaction between apoptosis and necrosis after neonatal HI, we have investigated the time course of necrotic cell death after HI in the neonatal rat and its relationship to caspase-3 activation and apoptotic cell death. Necrotic cells were identified in vivo by intracerebroventricular (icv) injection of the fluorescent dye propidium iodide (PI), a dye unable to pass intact lipid membranes and commonly used to detect necrosis in cell culture experiments. PI has been successfully used as a marker of disrupted plasma membrane integrity of necrotic cells after icv administration in adult animals subjected to focal cerebral ischemia (Unal Cevik and Dalkara, 2003; Unal-Cevik et al., 2004). We report here that necrosis has a prolonged role in the progression of brain injury and may represent the prevalent type of cell death after neonatal HI. Many PI-positive necrotic cells also express activated caspase-3 and cathepsin-B indicating that a different spectrum of suicidal programs can be activated in the same cell. The extended period of caspase-3 activation in PI-positive necrotic cells supports the possibility that the apoptotic-to-necrotic continuum may ensue as the result of an incomplete execution of the apoptotic program.

355

1:500; Chemicon); glial fibrillary acidic protein (GFAP, rabbit polyclonal antibody; 1:300; DAKO); and cathepsin-B (rabbit polyclonal antibody; Upstate). In some experiments, Hoechst-33258 was added to the mounting medium to counterstain the nuclei. Fluorescein isothiocyanate-conjugated horse anti-rabbit IgG (1:200; Santa Cruz Biotechnology, California) or Alexa Fluor 350 goat anti-rabbit IgG1 were used to demonstrate immunoreactivity as green or blue fluorescence, respectively. Sections were incubated first with primary antibodies overnight at 4 °C and then with secondary antibodies (1:200; FITC goat anti-mouse or anti-rabbit IgG) for 60 min. In some experiments, GFAP immunoreactivity was detected by means of a biotinylated goat anti-rabbit antibody (1:200) and visualized using avidin–biotin peroxidase solution (Elite ABC kit, Vectastain Vector, USA). Peroxidase activity was evidenced by 0.05% 3,3′-diaminobenzidine tetrahydrochloride (DAB) and 0.03% H2O2 at the appropriate stage. The specificity of immunoreactivity was tested by omitting each primary antibody from the incubation medium. Terminal dUDP Nick-end labeling (TUNEL)

Materials and methods

DNA fragmentation was detected using a terminal dUTP nickend labeling apoptosis detection kit (Roche Diagnostics, Germany) according to the manufacturer’s instruction.

Cerebral hypoxia–ischemia (HI)

Western blot analysis

All surgical and experimental procedures were carried out in accordance with Italian regulations for the care and use of laboratory animals. On postnatal day 7 (PN7), Sprague–Dawley pup rats (Charles River, Italy) were anesthetized with ether and subjected to ligation of the right common carotid artery followed by 2.5 h hypoxia (92% nitrogen and 8% oxygen) (Balduini et al., 2001).

Samples (50 μg protein; Bradford dye-binding procedure, BioRad Laboratories, Italy) were separated onto SDS–polyacrylamide membrane and probed with the polyclonal anti-caspase-3 antibody (1:1000, Cell Signaling Technology, Beverly, MA, USA), which recognizes the 17-kDa cleaved form. Calpain activation was evaluated by cleavage of α-spectrin using an antibody (monoclonal, 1:2000, MAB 1622, Chemicon International, Tamecula, CA), which recognizes the intact α-spectrin (280 kDa) as well as the fragments generated by calpain-mediated cleavage (150- and 145-kDa fragments) and caspase-3-mediated cleavage (150- and 120-kDa fragment) (Carloni et al., 2006). A monoclonal antibody against βactin (1:4000, Sigma, Italy) was used as a control for protein gel loading. Blots were analyzed using the NIH-Image software. Data were normalized to those of β-actin and expression as OD integration.

Propidium iodide labeling and tissue preparation Half of a microliter of propidium iodide (PI, 1 mg/mL in distilled water, Sigma) was injected into the right lateral ventricle. Twenty minutes after injection, pups were deeply anesthetized with ether and perfusion-fixed with 4% paraformaldehyde in 0.1 mol/L PBS. Brains were rapidly removed on ice and processed for antigen retrieval by immersing overnight in 10 mmol/L sodium citrate buffer (pH 6.0, 4 °C) and boiling in the same buffer for 3 min. After boiling brains were cryoprotected with 30% sucrose/PBS (72 h, 4 °C). Preliminary experiments have shown that the boiling procedure does not affect the distribution of the dye. The brains were cut on a cryostat into coronal sections (thickness 20 μm) and then processed for immunohistochemical analyses. In initial experiments, we also injected the dye in the right lateral ventricle of both control (neither hypoxic nor ischemic; n = 3) and sham-operated hypoxic animals (subjected only to the hypoxic procedure; n = 2) and no dye detection was found.

Cell counting Cell counting was conducted in the cerebral cortex and in the CA1 and CA2/CA3 regions of the hippocampus on 20× microscopic images using a BX-51 Olympus microscope. Positive cells were counted in three separate fields of each brain region in slices cut at the level depicted in Fig. 1B (A 3750 of the Koning and Klippel stereotaxic atlas). Five animals for each group, with the exception of the 96-h time point that is representative of two animals because the lack of the tissue in the ischemic side in 3 animals, were analyzed.

Immunohistochemistry Results PI-labeled cryosections were incubated with 1.5% normal blocking serum for 1 h at room temperature and then incubated overnight at 4 °C with the following antibodies: anti-caspase-3 (polyclonal, 1:50, Cell Signaling Technology, Beverly, MA, USA) that recognizes the 17-kDa cleaved form of caspase-3; neuronspecific nuclear protein (anti-NeuN, mouse monoclonal antibody;

Time course of PI labeling PI is a fluorescent dye utilized in cell culture experiments to stain necrotic cells because it is unable to cross not-damaged lipid membranes. PI has been recently used in vivo to detect necrotic cells

356

S. Carloni et al. / Neurobiology of Disease 27 (2007) 354–361

S. Carloni et al. / Neurobiology of Disease 27 (2007) 354–361

after ischemia in adult mice (Unal Cevik and Dalkara, 2003; UnalCevik et al., 2004). In agreement with these studies, we found that administration of PI 20 min before sacrifice efficiently labeled necrotic cells after HI in neonatal rats. PI-positive cells were found in the damaged but not in the contralateral side. We were also unable to detect positive cells in the contralateral side even when PI was injected directly in the contralateral ventricle (data not shown). Fig. 1 shows that only a fraction of the cells were PI-positive and labeling occurred in the whole cell and not only in the nucleus (Fig. 1A, panels c and d) (Unal Cevik and Dalkara, 2003; Unal-Cevik et al., 2004). In this experiment, Hoechst-33258 was used to counterstain the nuclei (Fig. 1A, panel b). PI-positive cells were observed starting from 30 min after HI in the CA1 area of the hippocampus. The number of positive cells rapidly increased in the injured brain areas, and 6 h after HI PI-positive cells were also detectable in the CA2/ CA3 region of the hippocampus, in the deep layers of the cerebral cortex and in the striatum. The number of PI-positive cells increased up to 12 h in the analyzed fields but a comparable amount of positive cells was found at later times (Fig. 1B, red dots, and Fig. 1C). However, PI-positive cells expanded over time especially to the superficial layers of the parietal and piriform cortex (not shown). It should be noted that PI-positive cells were also detectable 96 h after HI. At this time, however, we could analyze only two of the five animals sacrificed, because three of them already lost most of the tissue in the damaged side of the brain indicating that the progression of brain injury can differ significantly among animals. Co-localization of PI labeling and NeuN but not GFAP-positive cells To investigate the cell types in which PI staining was localized, brain sections preloaded with PI were subjected to immunofluorescence staining with antibodies against cell-type-specific antigens. Staining with the neuronal-specific nuclear marker NeuN revealed a strong co-localization of PI and NeuN within the CA1 and CA2/ CA3 regions of the hippocampus (Fig. 1D, panels G–L) as well as within the deep layers of the cerebral cortex (not shown). In contrast, no co-localization was found between GFAP, a marker for astrocytes, and PI in the 5 animals examined. GFAP-positive cells were found in the hippocampus and in the surrounding corpus callosum indicating a marked astrogliosis after HI (Fig. 1D, panels A and F). The absence of co-localization in reactive astrocytes further indicates that PI labels damaged cells. PI, caspase-3, cathepsin-B and TUNEL labeling after HI The time course of PI-labeling was compared with the expression of activated caspase-3. Fig. 2A shows representative photo-

357

micrographs of experiments performed 24 h after HI. Both PIpositive cells and cells expressing cleaved caspase-3 were found in the cerebral cortex, hippocampus and striatum of ischemic animals. In the cerebral cortex, differently from PI labeling, caspase-3 immunoreactivity was present in higher amount in the superficial layers. Merged images showed a strong co-localization of caspase-3 and PI in the hippocampus, in the striatum and in the deep layers of the cortex whereas the superficial layers of the cortex showed double labeling only in few cells. However, at later times (48 h and 96 h) most cells of the superficial layers were positive to both PI and caspase-3. Caspase-3-positive cells were found starting from 6 h after HI, their number increased up to 24 h and remained almost the same at the following time points. The distribution of caspase-3positive cells in the injured cerebral cortex was comparable to that found in the TUNEL assay (n = 3). Only few TUNEL-positive cells were found 6 h after HI and their number increased over time (not shown). Twenty-four hours after HI, the superficial layers of the cortex showed many cells positive to the TUNEL assay (Fig. 2B, panel m) and cells displayed mostly apoptotic features (darkly stained chromatin nuclei; Fig. 2B, panel n). The number of TUNEL-positive cells in the deep layers of the cortex and in the CA1 region of the hippocampus was much lower (Fig. 2B, panels m, o, p). These cells also showed necrotic features (Fig. 2B, panel o) and only few cells were also PI positive. It should be noted that the number of PI-positive cells was much higher than that of TUNELpositive cells (Fig. 2B, panels p and q). In addition, only few positive cells were found in the CA2/CA3 area (not shown), where a large number of PI and caspase-3-positive cells were found (Fig. 2A). Forty-eight hours after HI, the superficial layers of the cortex displayed strong PI and TUNEL staining (Fig. 2B, panels 48h-a and 48h-b). We also tested whether PI and caspase-3-positive cells showed increased cathepsin-B activation, which is considered a marker of loss of lysosomal membrane integrity and necrotic cell death (Yamashima, 2000, 2004). Fig. 2C shows representative pictures of PI, cathepsin-B and caspase-3 labeling in brain sections 24 h after HI, whereas the percentage of cells labeled by the different markers is reported in Table 1. The number of cells showing diffuse cytoplasmic cathepsin-B release increased over time (not shown) and cells were found in higher amount in the superficial layers of the cortex. It should be noted that PI did not co-localize with cathepsin-B alone since cathepsin-B/PI-positive cells were also positive to caspase-3 (Fig. 2C; Table 1). PI/cathepsin-B/caspase-3 labeling was higher in the hippocampus whereas the superficial layers of the cortex showed many cells that were positive to cathepsin-B and caspase-3 but not to PI (Table 1). The time course of PI labeling and caspase-3 immunoreactivity observed in the experiments described above was similar to that

Fig. 1. (A) Photomicrographs showing cells labeled with propidium iodide (PI; red) and Hoechst-33258 (blue) in the CA1 region of the hippocampus 24 h after hypoxia–ischemia (HI). The merged image (C) shows that only a fraction of cells are labeled with PI and that the whole cell fluoresces red and not only the nucleus (arrows in panel D). (B) Diagrams showing the time course of propidium iodide (PI) labeling. Diagrams do not reflect quantitative values but represent a rough reconstruction of images obtained from five different animals analyzed at the selected time points with the exception of the 96 h that is representative of 2 animals because the lack of the tissue in the ischemic side in 3 animals. (C) Counts of PI-positive cells in the cerebral cortex and CA1 and CA2/CA3 regions of the hippocampus. Counts of PI-positive cells were performed as described in Materials and methods and data were reported as mean ± SE. PI-positive cells were observed 30 min and 6 h after hypoxia–ischemia (HI), particularly in the CA1 region of hippocampus, and increased significantly at 12 h. PI-positive cells were also detectable 96 h after HI. (D) Representative photomicrographs of experiments performed 24 h after hypoxia–ischemia (HI) showing propidium iodide (PI)positive cells (A, D, G, J) and GFAP (B, E) and NeuN (H, K) immunoreactivity in the CA1 (A, B and G, H) and in the CA2 (D, E and J, K) regions of the hippocampus. Merged images (C, F, I, L) illustrate that there is co-localization of PI with NeuN (I and L; yellow) but not of PI with GFAP (C and F). The arrow show the part of the picture enlarged in the inset. Panels B and C show images of immunohistochemical experiments where GFAP immunoreactivity was detected using DAB. Scale bars = 50 μm.

358

S. Carloni et al. / Neurobiology of Disease 27 (2007) 354–361

found for calpain and caspase-3 immunoblotting (Fig. 3). Calpains are Ca2+-sensitive proteases mainly implicated in necrotic cell death. Calpain activation can be evaluated by studying the cleav-

age of α-spectrin, a cytoskeletal protein that is cleaved by both calpains and caspase-3 into specific fragments that can be resolved by Western blot (Wang, 2000). At 30 min and 6 h after HI, there

S. Carloni et al. / Neurobiology of Disease 27 (2007) 354–361 Table 1 Percentage of cells labeled and co-labeled by PI, cleaved caspase-3 and cathepsin-B, in neonatal rats 24 h after hypoxia–ischemia Brain area

PI Caspase-3 Cathepsin-B PI/caspase-3 PI/cathepsin-B Caspase-3/cathepsin-B PI/caspase-3/cathepsin-B

Hippocampus

Cerebral cortex

CA1

CA2/CA3

Deep layers

Superficial layers

53.3 ± 0.9 9.6 ± 1.2 0.4 ± 0.2 16.7 ± 0.6 0.0 ± 0.0 2.9 ± 1.1 16.9 ± 3.9

51.1 ± 4.5 5.0 ± 1.2 0.5 ± 0.5 22.9 ± 3.3 0.0 ± 0.0 14.7 ± 2.7 5.5 ± 1.1

76.3 ± 2.3 2.2 ± 0.6 3.2 ± 1.2 4.8 ± 0.8 0.0 ± 0.0 12.8 ± 0.6 0.5 ± 0.2

22.9 ± 5.6 9.7 ± 1.8 13.9 ± 1.4 0.2 ± 0.0 0.0 ± 0.0 52.6 ± 5.9 0.6 ± 0.3

Cell counting was performed in the CA1 and CA2/CA3 regions of the hippocampus and in the deep and superficial layers of the cerebral cortex of 4–5 ischemic animals as described in Materials and methods. Positive cells were counted in three separate fields of each brain region and values are reported as percentage of cells labeled and co-labeled by the different markers. The average number of cells labeled and co-labeled in the different brain area was 159.3 ± 6.0, 151.7 ± 16.7, 297.3 ± 16.8, 88.7 ± 5.6 for the CA1, CA2/CA3 and the deep and superficial layers of the cerebral cortex, respectively. Each value represents the mean ± SE.

was a significant increase in α-spectrin cleavage detected as an increased expression of the p-150/145 fragments of the protein. On the other hand, the p120 caspase-3-specific fragment was absent at 30 min and barely detectable at 6 h. This fragment increased significantly at 24 h after HI. Caspase-3 immunoblots confirmed the delayed activation of the caspase that overlapped what observed in immunohistochemical experiments.

Discussion In the present study, we used the icv administration of PI, a dye unable to pass intact lipid membranes and commonly used to detect necrosis in cell culture experiments, to assess the time course of necrotic cell death after neonatal HI. According to Unal Cevik et al., who used in vivo PI administration to detect necrotic cells after ischemia in adult mice (Unal Cevik and Dalkara, 2003; Unal-Cevik et al., 2004), we found that when injected directly in the brain of neonatal rats PI enters into cells and the whole cell fluoresces red (Fig. 1A). At variance with what observed when PI is directly applied to fixed brain sections, where it labels all cells (Liu et al.,

359

2004), in vivo administration of PI allows to detect only cells with damaged plasma membrane. This assertion is supported by the following observations: (a) the number of Hoechst-33258-positive cells, used to counterstain the nuclei, remarkably exceeded that of PI-positive cells (Unal Cevik and Dalkara, 2003; Unal-Cevik et al., 2004); (b) at short times after HI, PI-positive cells were not TUNEL positive; (c) PI-positive cells were found in the injured but not in the contralateral hemisphere or after PI administration in the brain of control or sham-operated hypoxic pup rats (not shown); and (d) PI co-localized with NeuN but not with GFAP indicating that the dye labels neurons with damaged plasma membrane but not reactive astrocytes. In the CA1 area of the hippocampus PIpositive cells were found starting from 30 min after the ischemic insult. The appearance of PI-positive cells was parallel with an increased expression of α-spectrin fragments, indicative of an increased activity of calpain, a Ca2+-activated protease thought to be involved in necrotic cell death (Wang, 2000). At this time, there was no evidence of caspase-3 activation in both immunohistochemical and Western blot experiments. The time course of calpain activation was like that previously reported (Han et al., 2002; Carloni et al., 2006). The number of PI-positive cells increased rapidly in the injured brain areas and was parallel with caspase-3 activation. Twelve hours after the hypoxic–ischemic insult, most cells in the hippocampus and the striatum were positive to both PI and caspase-3 but many double positive cells were also found at 6 h. In the cerebral cortex, in contrast, 24 h after HI the superficial layers showed many cells positive to caspase-3 and TUNEL but not to PI, whereas the deep layers showed many cells positive to PI but only few positive to TUNEL or caspase-3. However, in the cerebral cortex only the progression of injury is delayed compared to the other areas but probably occurs with similar mechanisms, since at later times most cells of the superficial layers were positive to both PI and/or caspase-3 and TUNEL. Furthermore, many caspase-3-positive cells were also cathepsin-B positive denoting also a loss of lysosomal membrane integrity but yet a preserved plasma membrane integrity since they are not PI positive. The high number of PI and PI/caspase-3-positive cells observed 24 h after the insult in the hippocampus and in the deep layer of the cortex also correlate with the intensity of injury observed in these areas. It is conceivable that PI- and caspase-3-positive cells are cells that started their apoptotic program but undergoes to plasma membrane permeabilization as well, or secondary necrosis (i.e. “necrosis after apoptosis in which the bioenergetic consequences of mitochondrial permeability transition cause plasma membrane disruption after the activation and action of proteases and nucleases”; Yamashima, 2000, and references therein). Secondary necrosis is frequently

Fig. 2. (A) Photomicrographs of experiments performed 24 h after hypoxia–ischemia (HI) showing propidium iodide (PI)-positive cells (a, d, g, j) and cells that express cleaved caspase-3 (b, e, h, k) in the cerebral cortex (a, b), hippocampus (CA2/CA3, d and e; CA1, g and h) and striatum (j, k) of ischemic animals. Merged images (c, f, i, l) illustrate co-localization (yellow) of PI fluorescence and cleaved caspase-3 immunoreactivity. PI labeling was mostly present in the deep layers of the cortex whereas caspase-3 immunoreactivity was diffused particularly in the superficial layers. At this time in the cerebral cortex, double labeling was found only in few cells (not shown in this panel). At variance, in the hippocampus and striatum most cells were double labeled with both PI and caspase-3. Scale bars = 50 μm. (B) Representative photomicrographs showing DNA fragmentation (TUNEL-positive cells) on coronal sections from ischemic neonatal rats 24 h after HI. The superficial layers of the cortex showed many cells positive to the TUNEL assay (panel m) and cells displayed mostly apoptotic features (darkly stained chromatin nuclei; dark arrows in panel n, and inset). The number of TUNEL-positive cells in the deep layers of the cortex and in the CA1 region of the hippocampus was much lower (panels m, o, p) and cells also showed necrotic features (arrows in panel o, and inset). It should be noted that the number of PIpositive cells, particularly in the hippocampus, was much higher as compared to TUNEL-positive cells (panels p and q). Forty-eight hours after HI, the superficial layers of the cortex displayed TUNEL staining (panel 48 h-a) which co-localized with PI (panel 48 h-b). Arrows shows examples of PI/TUNEL-positive cells. (C) Co-localization of PI, cathepsin-B and caspase-3. Panel r shows a merge image of PI and cathepsin-B in the deep layers of the cortex. Arrowheads show that examples of cells PI and cathepsin-B (green) co-localize. Panel s shows PI and caspase-3 co-localization (arrowheads), whereas panel t caspase-3 and cathepsinB co-localization (arrowheads). Note that some cells are PI but not cathepsin-B or caspase-3 positive (arrows). Panel u shows PI/cathepsin-B/caspase-3-positive cells (green arrowheads). Note that all PI/cathepsin-B-positive cells are also caspase-3 positive.

360

S. Carloni et al. / Neurobiology of Disease 27 (2007) 354–361

Fig. 3. α-Spectrin degradation (A) and activated caspase-3 expression (B) in the hippocampus of ischemic neonatal rats. The figure shows representative immunoblots and quantitative analysis of the cleaved p150/145 band of α-spectrin and of activated caspase-3 (bar graphs). A significant increase in αspectrin cleavage was detected at all time points evaluated, as indicated by the increased expression of the p-150/145 fragment. The caspase-3-specific p-120 fragment appeared at 6 h and increased significantly 24 h after hypoxia–ischemia (HI) in agreement with the data reported in panel B and in immunohistochemical experiments. Data (n = 5) are expressed as OD integration. C, control (sham); L: left side, contralateral; R: right side, ipsilateral to the occluded carotid artery.

found in in vitro models of apoptosis (Schwab et al., 2002) and in liver injury (Jaeschke and Lemasters, 2003). In vivo, apoptotic cells should be eliminated before their plasma membrane is disrupted, but in conditions of severe brain injury, as observed in this model of neonatal HI (Balduini et al., 2000, 2001), the scavenging system could not be sufficient to remove the large amount of dying cells, which then progress toward necrosis. The progression of a large amount of cells towards a necrotic type of cell death may explain the hybrid phenotypes of apoptosis and necrosis which is found in this model of neonatal HI (Martin, 2001; Northington et al., 2001; Liu et al., 2004; Northington et al., 2005) and in vitro after chemical hypoxia (Formigli et al., 2000). Necrotic cell death also contributes to the sustained and strong inflammatory reaction observed after HI in neonatal rats (Benjelloun et al., 1999) and can be functional for inflammatory cell recruitment (Sauter et al., 2000; Proskuryakov et al., 2003). However, it should be also considered the possibility that at least part of PI-positive cells might undergo autophagic cell death, a non-apoptotic form of cell death that is distinct from necrosis (Clarke and Clarke, 1996; Edinger and Thompson, 2004). Autophagy is a cellular process that causes degradation of long-lived proteins and recycling of cellular components to ensure survival during starvation. During autophagy, cells exhibit extensive internal membrane remodeling, engulfing portion of the cytoplasm in large double-membrane vesicles which dock and fuse with lysosomes and it cannot be completely excluded PI uptake during this processes. To our knowledge there are no studies that address the role of autophagy in this model of HI. However, autophagy is increased in adult mice after traumatic brain injury (Diskin et al., 2005) and after myocardial ischemia (Yan et al., 2006; Valentim et al., 2006). Furthermore, we cannot exclude that also macrophages could uptake PI as a result of their phagocytic activity. Experiments are in progress to clarify this points.

Several mechanisms have been proposed to switch towards different types of cell death such as osmotic imbalance, Ca2+ overload and ATP levels (Nicotera and Melino, 2004). ATP availability, in particular, may be crucial. Indeed, if ATP is preserved, at least in part, cytochrome c released after the mitochondrial permeability transition activates caspase-dependent apoptosis. Severe ATP depletion, in contrast, may not allow the ordered sequence of changes required for apoptotic cell death, and caspases may activate a subroutine(s) of the death program(s) leading to plasma membrane failure (Nicotera, 2003). Several findings are in line with this hypothesis. For example, caspases can cleave calpastatin, the endogenous calpain inhibitor (Wang et al., 1998), and caspase inhibitors reduce calpain activation and necrotic cell death in vitro and after traumatic brain injury (Knoblach et al., 2004). Aminocycline, a semisynthetic tetracycline which reduces brain injury when administered either before or after a hypoxic–ischemic insult, also causes a concomitant reduction of caspase-3 and calpain activation (Arvin et al., 2002). In addition, caspases were recently found to cleave plasma membrane calcium ATPases in this model of HI and this has been suggested as an important link between apoptosis and necrosis (Schwab et al., 2002). In conclusion, the present study demonstrates that in spite of the strong activation of caspase-3, necrotic cell death has a prolonged role in the progression of brain injury and represents the prevalent type of cell death after neonatal HI. PI-positive cells express activated caspase-3 and cathepsin-B indicating that a different spectrum of suicidal programs can be activated in the same cell. Likely, PI and caspase-3-positive cells are cells that started their apoptotic program(s) but undergoes to secondary necrosis. This may imply that therapies addressed at reducing apoptosis could shift to necrotic cell death (Northington et al., 2005), as already demonstrated in vitro (Formigli et al., 2000).

S. Carloni et al. / Neurobiology of Disease 27 (2007) 354–361

Acknowledgments We thank Claudia Scopa for excellent technical assistance. Research support was provided by a grant from the Ministero dell’Università e della Ricerca Scientifica e Tecnologica, Progetti di Ricerca di Interesse Nazionale to W. Balduini. References Arvin, K.L., Han, B.H., Du, Y., Lin, S.Z., Paul, S.M., Holtzman, D.M., 2002. Minocycline markedly protects the neonatal brain against hypoxic–ischemic injury. Ann. Neurol. 52, 54–61. Balduini, W., De Angelis, V., Mazzoni, E., Cimino, M., 2000. Long-lasting behavioral alterations following a hypoxic/ischemic brain injury in neonatal rats. Brain Res. 859, 318–325. Balduini, W., De Angelis, V., Mazzoni, E., Cimino, M., 2001. Simvastatin protects against long-lasting behavioral and morphological consequences of neonatal hypoxic/ischemic brain injury. Stroke 32, 2185–2191. Balduini, W., Carloni, S., Mazzoni, E., Cimino, M., 2004. New therapeutic strategies in perinatal stroke. Curr. Drug Targets CNS Neurol. Disord. 3, 315–323. Benjelloun, N., Renolleau, S., Represa, A., Ben-Ari, Y., CharriautMarlangue, C., 1999. Inflammatory responses in the cerebral cortex after ischemia in the P7 neonatal Rat. Stroke 30, 1916–1923. Carloni, S., Mazzoni, E., Cimino, M., De Simoni, M.G., Perego, C., Scopa, C., Balduini, W., 2006. Simvastatin reduces caspase-3 activation and inflammatory markers induced by hypoxia–ischemia in the newborn rat. Neurobiol. Dis. 21, 119–126. Clarke, P.G., Clarke, S., 1996. Nineteenth century research on naturally occurring cell death and related phenomena. Anat. Embryol. (Berl.) 193, 81–99. Diskin, T., Tal-Or, P., Erlich, S., Mizrachy, L., Alexandrovich, A., Shohami, E., Pinkas-Kramarski, R., 2005. Closed head injury induces upregulation of Beclin 1 at the cortical site of injury. J. Neurotrauma 22, 750–762. Edinger, A.L., Thompson, C.B., 2004. Death by design: apoptosis, necrosis and autophagy. Curr. Opin. Cell Biol. 16, 663–669. Formigli, L., Papucci, L., Tani, A., Schiavone, N., Tempestini, A., Orlandini, G.E., Capaccioli, S., Orlandini, S.Z., 2000. Aponecrosis: morphological and biochemical exploration of a syncretic process of cell death sharing apoptosis and necrosis. J. Cell. Physiol. 182, 41–49. Gill, R., Soriano, M., Blomgren, K., Hagberg, H., Wybrecht, R., Miss, M.T., Hoefer, S., Adam, G., Niederhauser, O., Kemp, J.A., Loetscher, H., 2002. Role of caspase-3 activation in cerebral ischemia-induced neurodegeneration in adult and neonatal brain. J. Cereb. Blood Flow Metab. 22, 420–430. Golstein, P., Kroemer, G., 2007. Cell death by necrosis: towards a molecular definition. Trends Biochem. Sci. 32, 37–43. Han, B.H., Xu, D., Choi, J., Han, Y., Xanthoudakis, S., Roy, S., Tam, J., Vaillancourt, J., Colucci, J., Siman, R., Giroux, A., Robertson, G.S., Zamboni, R., Nicholson, D.W., Holtzman, D.M., 2002. Selective, reversible caspase-3 inhibitor is neuroprotective and reveals distinct pathways of cell death after neonatal hypoxic–ischemic brain injury. J. Biol. Chem. 277, 30128–30136. Jaeschke, H., Lemasters, J.J., 2003. Apoptosis versus oncotic necrosis in hepatic ischemia/reperfusion injury. Gastroenterology 125, 1246–1257. Knoblach, S.M., Alroy, D.A., Nikolaeva, M., Cernak, I., Stoica, B.A., Faden, A.I., 2004. Caspase inhibitor z-DEVD-fmk attenuates calpain and necrotic cell death in vitro and after traumatic brain injury. J. Cereb. Blood Flow Metab. 24, 1119–1132. Liu, C.L., Siesjo, B.K., Hu, B.R., 2004. Pathogenesis of hippocampal neuronal death after hypoxia–ischemia changes during brain development. Neuroscience 127, 113–123.

361

Martin, L.J., 2001. Neuronal cell death in nervous system development, disease, and injury. Int. J. Mol. Med. 7, 455–478. Martin, L.J., Al-Abdulla, N.A., Brambrink, A.M., Kirsch, J.R., Sieber, F.E., Portera-Cailliau, C., 1998. Neurodegeneration in excitotoxicity, global cerebral ischemia, and target deprivation: a perspective on the contributions of apoptosis and necrosis. Brain Res. Bull. 46, 281–309. Nakajima, W., Ishida, A., Lange, M.S., Gabrielson, K.L., Wilson, M.A., Martin, L.J., Blue, M.E., Johnston, M.V., 2000. Apoptosis has a prolonged role in the neurodegeneration after hypoxic ischemia in the newborn rat. J. Neurosci. 20, 7994–8004. Nicotera, P., 2003. Molecular switches deciding the death of injured neurons. Toxicol. Sci. 74, 4–9. Nicotera, P., Melino, G., 2004. Regulation of the apoptosis–necrosis switch. Oncogene 23, 2757–2765. Nieminen, A.L., 2003. Apoptosis and necrosis in health and disease: role of mitochondria. Int. Rev. Cytol. 224, 29–55. Northington, F.J., Graham, E.M., Martin, L.J., 2005. Apoptosis in perinatal hypoxic–ischemic brain injury: how important is it and should it be inhibited? Brain Res. Brain Res. Rev. 50, 244–257. Northington, F.J., Ferriero, D.M., Graham, E.M., Traystman, R.J., Martin, L.J., 2001. Early neurodegeneration after hypoxia–ischemia in neonatal rat is necrosis while delayed neuronal death is apoptosis. Neurobiol. Dis. 8, 207–219. Proskuryakov, S.Y., Konoplyannikov, A.G., Gabai, V.L., 2003. Necrosis: a specific form of programmed cell death? Exp. Cell Res. 283, 1–16. Sauter, B., Albert, M.L., Francisco, L., Larsson, M., Somersan, S., Bhardwaj, N., 2000. Consequences of cell death: exposure to necrotic tumor cells, but not primary tissue cells or apoptotic cells, induces the maturation of immunostimulatory dendritic cells. J. Exp. Med. 191, 423–434. Schwab, B.L., Guerini, D., Didszun, C., Bano, D., Ferrando-May, E., Fava, E., Tam, J., Xu, D., Xanthoudakis, S., Nicholson, D.W., Carafoli, E., Nicotera, P., 2002. Cleavage of plasma membrane calcium pumps by caspases: a link between apoptosis and necrosis. Cell Death Differ. 9, 818–831. Unal Cevik, I., Dalkara, T., 2003. Intravenously administered propidium iodide labels necrotic cells in the intact mouse brain after injury. Cell Death Differ. 10, 928–929. Unal-Cevik, I., Kilinc, M., Can, A., Gursoy-Ozdemir, Y., Dalkara, T., 2004. Apoptotic and necrotic death mechanisms are concomitantly activated in the same cell after cerebral ischemia. Stroke 35, 2189–2194. Valentim, L., Laurence, K.M., Townsend, P.A., Carroll, C.J., Soond, S., Scarabelli, T.M., Knight, R.A., Latchman, D.S., Stephanou, A., 2006. Urocortin inhibits Beclin1-mediated autophagic cell death in cardiac myocytes exposed to ischaemia/reperfusion injury. J. Mol. Cell. Cardiol. 40, 846–852. Wang, K.K., 2000. Calpin and caspase: can you tell the difference? Trends Neurosci. 23, 20–26. Wang, K.K., Posmantur, R., Nadimpalli, R., Nath, R., Mohan, P., Nixon, R.A., Talanian, R.V., Keegan, M., Herzog, L., Allen, H., 1998. Caspase-mediated fragmentation of calpain inhibitor protein calpastatin during apoptosis. Arch. Biochem. Biophys. 356, 187–196. Wei, L., Han, B.H., Li, Y., Keogh, C.L., Holtzman, D.M., Yu, S.P., 2006. Cell death mechanism and protective effect of erythropoietin after focal ischemia in the whisker–barrel cortex of neonatal rats. J. Pharmacol. Exp. Ther. 317, 109–116. Yamashima, T., 2000. Implication of cysteine proteases calpain, cathepsin and caspase in ischemic neuronal death of primates. Prog. Neurobiol. 62, 273–295. Yamashima, T., 2004. Ca2+-dependent proteases in ischemic neuronal death: a conserved ‘calpain–cathepsin cascade’ from nematodes to primates. Cell Calcium 36, 285–293. Yan, L., Sadoshima, J., Vatner, D.E., Vatner, S.F., 2006. Autophagy: a novel protective mechanism in chronic ischemia. Cell Cycle 5, 1175–1177.