Temporal profile of ultrastructural changes in cortical neurons after a photochemical lesion

Journal of Neuroscience Research 77:901–912 (2004)

Temporal Profile of Ultrastructural Changes in Cortical Neurons After a Photochemical Lesion Benita Andersson,1,2* Xingchen Wu,1 Bo¨rje Bjelke,1 and Eva Sykova´2,3 1

Department of Clinical Neuroscience, Karolinska Institute, Stockholm, Sweden Institute of Experimental Medicine ASCR, Prague, Czech Republic 3 Department of Neuroscience, Second Medical Faculty, Charles University, Prague, Czech Republic 2

A photochemical lesion was induced in the right sensory motor cortex of rat brains. We examined at various time points the occurrence of different types of neuronal death with respect to a potential therapeutic window. The lesion appearance was documented by magnetic resonance imaging, and functional recovery was evaluated by behavioral tests showing recovery at 48 hr after lesioning. At 0.5, 1, 3, 6, 12, 24, 48, and 72 hr postlesion, cortical layers IV and V were examined by light and electron microscopy. Ultrastructural changes, which corresponded well to light microscopy findings, were found in both hemispheres. In the lesioned area, the neuropil appeared disorganized at 0.5 hr, and apoptotic and necrotic cell death was found at 0.5–3 hr. After 3 hr, the tissue was disintegrated. On the contralateral side, chromatin clumping appeared at 0.5–3 hr. At 3 hr, ruptured membranes were found, a sign of irreversible cell death. At 6 –72 hr, the membranes were intact, and the chromatin was not clumped but heterogeneously distributed. The nuclei contained dispersed nucleoli at 48 –72 hr. The morphology correlated well with magnetic resonance images and functional behavior. Our study demonstrates that a photochemical lesion is a useful model for studying morphological changes in injured cells. It results in a permanent infarction within 3 hr. In that the morphology on the contralateral side drastically changed between 3 and 6 hr, the cellular alterations at these time points might represent a break point at which cells either progress toward cell death or recover. © 2004 Wiley-Liss, Inc.

Key words: stroke; ischemia; sensory motor cortex; cell death; ultrastructure

Cytoprotection within 24 hr after cerebral infarction does not preserve brain functionality (Reese et al., 2000). However, in recent studies, functional recovery after focal stroke lesions in the sensory motor cortex has been explained by neuronal network modulation on the contralateral side of the brain (Abo et al., 2001). Given the neurological outcome for stroke patients, it is of great importance to find therapeutic windows for neuromodu© 2004 Wiley-Liss, Inc.

latory recovery mechanisms. Relevant therapeutic windows, based on ultrastructural evidence, have not yet been defined. Neither neuronal death pathways in ischemic brain tissue nor locations for potential recovery have been studied over time at the ultrastructural level. Cell death is conventionally classified into two forms, necrosis and apoptosis (Kerr et al., 1972). Necrosis, with karyolysis and cell swelling (Studzinski, 1999), is an accidental cell death resulting from circumstances outside the cell, whereas apoptosis, with karyorrehxis and pycnosis (Studzinski, 1999), is a process that involves an inherent cellular program leading to cell death. The apoptotic process is energy consuming and can be initiated by internal signals or by agents in the environment. Importantly, there are similarities in the ultrastructural criteria for necrosis and apoptosis. Cell membrane “blebbing” or “budding,” swollen mitochondria, and cytoplasm condensation combined with vacuolization occur in both forms of cell death (Studzinski, 1999). Necrosis and apoptosis are triggered in parallel in ischemic brain tissue (Lee et al., 1999). Three types of ischemic cell death have been described: swollen, “pale” neurons with edematous cell changes (ECC; Kalimo et al., 1977, 1982); condensed, “dark” neurons with ischemic cell changes (ICC; Brown and Brierly, 1972; Brown, 1977); and disrupted, “ghost” neurons with homogenizing cell changes (HCC; Garcia et al., 1993, 1995). All three differ morphologically from the two conventional death pathways. Contract grant sponsor: Stiftelsen for Strategisk Forskning; Contract grant number: ramanslag: R98:022; Contract grant sponsor: Academy of Sciences of the Czech Republic; Contract grant number: AVOZ 5039906; Contract grant sponsor: Grant Agency of the Czech Republic; Contract grant number: 304-03-1189. *Correspondence to: Benita Andersson, Institute of Experimental Medicine, ASCR, Videnska´ 1083, 140 20 Prague 4, Czech Republic. E-mail: [email protected] Received 23 January 2004; Revised 17 May 2004; Accepted 18 May 2004 Published online 1 July 2004 in Wiley InterScience (www. interscience.wiley.com). DOI: 10.1002/jnr.20217

902

Andersson et al.

There is no well-accepted morphological definition of the point at which a cell dies (Lipton, 1999). The moment when the cell becomes unable to recover its normal morphology (Lipton, 1999) is the most satisfactory definition. At the ultrastructural level, irreversible nerve cell death is verified by dense flocculent mitochondria and membrane ruptures (Garcia et al., 1978; Kalimo et al., 1977). In this study at the light microscopic (LM) as well as the electron microscopic (EM) level, we correlated morphological changes in ischemic rat brain tissue with functional recovery. A photochemical lesion in the brain results in complete ischemia. The extent of the lesion was evaluated by using diffusion-weighted magnetic resonance imaging (DW-MRI) to determine the apparent diffusion coefficient of water (ADCW), relaxation times (T2), and proton density (PD) maps. Regions of abnormal water molecule motion, occurring very early after ischemia onset, have been identified by DW-MRI and can be correlated with the EM picture (Pierpaoli et al., 1993). To examine functional recovery, motor function after lesion was studied by a beam-walking test with a seven-point rating scale (Feeney et al., 1981). Histological evaluation at the LM level shows the tissue character and submorphological neuronal changes. However, the morphological criteria for neuronal death are subtle, so it is difficult at the optical level to distinguish between different types of cell death. The EM technique makes it possible to identify the ultrastructural criteria for irreversible cell death. We studied the time course of cell death in ischemic brain tissue in the area of the lesion as well as the time course of morphological changes on the contralateral side, the proposed location of the mechanisms underlying functional recovery (Abo et al., 2001). The correlation of the temporal profile of the morphological features in damaged neurons with functional improvement and DW-MRI allows for the possibility of studying neuronal network reorganization. MATERIALS AND METHODS Experimental Design This study was performed according to animal care legislation and was approved by the Animal Ethical Committee (Stockholms norra djurfo¨rso¨ksetiska na¨mnd, So¨dra Roslags Tingsra¨tt, Stockholm, Sweden) and the Institute of Experimental Medicine (Prague, Czech Republic). The study included 53 male Sprague-Dawley rats (BW 250 g), housed under standardized conditions of day/night cycle, temperature, and humidity, with free access to water and food pellets. Anesthesia was induced with a mixture of 3% isoflurane/air and maintained during surgical procedures and DW-MRI recordings at 1.5%. In total, 53 animal brains were evaluated at eight time points, namely, 0.5, 1, 3, 6, 12, 24, 48, and 72 hr postlesion. Six brains in each group were evaluated, four for LM examination and two for EM. In addition, three control animals were used for LM and two for EM. After 2 days of beam-walking pretraining (Abo et al., 2001), a focal photochemical lesion was

induced in the sensory motor cortex (Watson et al., 1985). After the lesioning, animals were scored daily in their sensory motor ability, by using beam walking (Feeney et al., 1981). DW-MRI analysis was performed at each evaluation time point. Thereafter, 2.0 ml pentobarbital i.p. was given as terminal anesthesia, the animals were transcardically perfused, and samples for LM and EM examinations were taken from the lesioned cortical area and from the corresponding contralateral brain region. After 3 hr postlesion, the lesioned tissue was disintegrated. Consequently, there were problems in keeping the tissue intact during preparation, so only samples taken at 0.5–3 hr were examined from the lesioned hemisphere. The control animals were subjected to the surgical procedure but not to the lesion. Photochemically Induced Ischemic Lesion A unilateral, focal photochemical lesion was induced in the sensory motor cortex in the right hemisphere (Watson et al., 1985). The center of the lesion was set to bregma –1.0 mm, 2.0 mm lateral to the midline according to the brain atlas of Paxinos and Watson (1998). 80 mg/kg BW of the photosensitive dye rose bengal was infused at a concentration of 160 ml/mg i.v. at 60 ␮l/min. The dye was equilibrated in the blood pool for 90 sec before illumination for 15 min. A circular light beam, 8.0 mm in diameter, was focused at the lesion center, and a homogenous exposure was obtained by a slow 360° rotation of the beam during illumination. During the operation, body temperature was kept at 37.0 –38.0°C, and constant cardiac and respiratory frequency was maintained. Behavioral Testing To measure motor function after lesioning, animals were subjected to beam walking, including training and testing (Abo et al., 2001). Motor function was evaluated in a test with a seven-point rating scale (Feeney et al., 1981). The animals were trained for 2 days before lesioning and tested every day starting at 24 hr postlesion. The rating was based on the number of foot slips of the hind limbs. When scoring impairment, a score of 7 corresponded to no more than two foot slips and a score of 1 was given when the animal fell off the beam, thus indicating major impairment. Training and testing sessions were conducted during the morning hours in a quiet room with dim lightning. MRI DW-MRI recordings were performed with a 4.7-Tesla magnet (Biopspec Avance 47/40 Bruker). Images were constructed from a standard set of maps, namely, ADCW, T2, and PD maps. ADCW maps offered information about changes in extra- and intracellular water homeostasis. These diffusion maps were spin echo images along the read-out gradient direction. From T2 maps, the proton relaxation times, which indicate local molecular motion, were measured. Matrix dimension parameters for T2 maps were 256 ⫻ 256, and the slice thickness was 1.0 mm. To determine the number of water molecules, PD maps were calculated from the T2 maps. Incisors and ear bars stereotaxically fixed the animal’s head during DW-MRI recordings, and the measurements were made in the center of the lesion.

Ultrastructure in Lesion

Fig. 1. A fixture made of polystyrene used for standardized sample collection. On the bottom plate (a), a supporting wall (b; thick arrow) is inclined 90°. A central mark (thin arrow) is outlined on the bottom plate at a 90° angle to the supporting wall. Bilateral help lines (arrowhead) are marked on the bottom plate in order to find the sensory motor cortex at distances of 2 and 3 mm from the central mark.

Morphology The animals were transcardically perfused at a pressure of 0.3 bars, initially with 50 ml 0.1 M phosphate-buffered saline (PBS) at pH 7.4 with 500 IU heparin added, followed by 500 ml fixative solution for LM or EM preparation, respectively. LM. For LM preparation, the brains were perfused with 4% paraformaldehyde in 0.1 M PBS at pH 7.4. Thereafter, the brains were rapidly removed from the skull and stored at 4°C in fixative solution with 10% sucrose added. From the lesion area at bregma – 0.8 mm to 1.2 mm, coronal sections (12 ␮m) were cut on a Microm HM 400 low-temperature microtome. Every fourth section was processed for Nissl staining, and at each evaluation time point 40 –50 tissue sections were examined by LM. EM. For EM preparation, 3.0% glutaraldehyde in 0.1 M PBS at pH 7.4 was used as a fixative solution. Immediately after perfusion, the brains were dissected and stored overnight at 4°C in the fixative. Tissue samples were taken supported by a fixture (Fig. 1) according to a standardized procedure. The brains were cut into two parts close to the lesion center at bregma – 0.5 mm, and the caudal cutting face was placed tightly against the fixture wall (Fig. 1b). Coronal slices (⬃1 mm thick at bregma – 0.5 mm to –1.5 mm) were placed on the fixture plate (Fig. 1a) with the brain midline aligned to the central mark. Tissue samples were selected bilaterally, according to the fixture marks, between 2 and 3 mm lateral to midline, and, thereafter, these two samples were further cut into three specimens, which represented cortical layers I–II, III–IV, and V–IV. Unbiased sampling was obtained by randomized sample orientation during further EM preparation. The specimens were postfixed for 2 hr in 1% osmium tetroxide in 0.1 M PBS, dehydrated in increasing concentrations of ethanol, embedded in resin Agar 100, and polymerized in blocks at 60°C. The blocks were cut, with an LKB V ultramicrotome, into semithin sections (1 ␮m), and the sections were stained with toloudine blue. Areas of interest in cortical layers IV and V were selected by LM. Ultrathin sections (⬃60 nm) were cut from the selected areas, stained with uranyl acetate and lead

903

citrate, and examined in transmission electron microscopes (Zeiss CEM 900 and Philips Morgagni 268D). To obtain a semiquantitative estimation, we developed a standardized evaluation protocol considering both quantitative and qualitative parameters. The evaluation was based on the following structural parameters: nuclear condensation, chromatin clumping or margination, cytoplasm condensation and vacuolization, organelle changes, and nuclear as well as plasma membrane “blebbing” and rupture. The evaluation also included the counting of different cells. The section areas were measured and the cell densities per square millimeter calculated. Pyramid cells, interneurons, astrocytes, and oligodendrocytes, affected as well as unaffected, were counted separately. Furthermore, the status of the neuropil, considering the profile of dendrites and axons, was qualitatively evaluated. Electron micrographs were made in the areas of interest, and the final evaluation was made from these micrographs.

RESULTS Behavioral Testing and MRI The animals had the expected functional impairment at 24 hr postlesion, scoring 1 or 2 in the beam-walking test, indicating an adequate lesion. At 48 and 72 hr, the beam-walking score was 3 or 4, which indicated functional improvement. DW-MRI showed reduced ADCW values in the lesion area at 0.5– 6 hr (Fig. 2a– c), whereas T2 maps showed increased relaxation times (Fig. 2e– h). The corpus callosum and the external capsule were found as bright areas in ADCW and T2 maps, starting at 6 hr (Fig. 2c,g) and becoming more pronounced at 24 hr (Fig. 2d,h). A brain midline shift toward the contralateral hemisphere was observed at 3 hr in ADCW maps (Fig. 2b) as well as in PD maps (Fig. 2k) and at 6 hr in T2 maps (Fig. 2g). No changes were observed with DW-MRI in the contralateral hemisphere. Morphology at the LM Level On LM, the control material (Fig. 3a) showed a well-organized neuropil with the expected distribution of homogeneously stained cells in layers IV and V. The neuropil in the lesioned area was already disorganized at 0.5 hr, and the disorganization was more pronounced at 1–3 hr (Fig. 3b– d). Pycnotic cells were found at 0.5–3 hr in layers IV and V (Fig. 3b– d), and at 3 hr intensely stained cells (Fig. 3d) were found most frequently. At 6 –72 hr, the tissue in the lesion area was disintegrated, and individual cells were not possible to identify. On the contralateral side, the neuropil was well organized at 0.5–72 hr. However, large hyperchromatic cells (Fig. 3e,f) were found in layer V already at 0.5–1 hr. At 3 hr, intensely stained cells (Fig. 3g) were present in layers IV and V. Together with weakly stained cells (Fig. 3h), similar cells were found at 6 hr. Weakly stained as well as intensely stained cells were all small. Also, at 12 hr, some hyperchromatic cells were found in layer V, although most of the cells contained large, light nuclei surrounded by a thin rim of cytoplasm (Fig. 3i). At 24 hr, a few pycnotic cells (Fig. 3k) were noticed in layer V. At 48 –72 hr, most

904

Andersson et al.

Fig. 2. MR images showing ADCW, T2, and PD maps from coronal sections of lesioned brain at 0.5, 3, 6, and 24 hr postlesion. The images are compared with Nissl-stained cortical sections showing the contralateral side and the lesioned area at the same time points. In ADCW maps, reduced values are seen as dark areas in the lesion (a– d), and. in

T2 maps, increased values are seen as bright areas (e– h). The corpus callosum was observed as a bright area at 6 and 24 hr in ADCW maps (c,d) and T2 maps (g,h). In PD maps, midline shifts (arrowheads) are observed (i–m). In the Nissl sections, the tissue in the lesioned area can be seen to disintegrate gradually between 0.5 and 24 hr (n– q).

cells were of normal size and were homogeneously stained (Fig. 3l,m).

number, and calculated the cell densities at each postlesion time. Cell densities. The cell density in the lesioned area (Fig. 5A) was 11.93, 10.56, and 18.36 cells/mm2, at 0.5, 1, and 3 hr, respectively. At each time point, approximately half of the cells were EDN cells. Only a few EWN cells were observed at 0.5–1 hr, whereas, at 3 hr, the EWN cell density was 10.01 cells/mm2. Compared with 218.92 cells/mm2 in control tissue, the cell density in the contralateral hemisphere (Fig. 5 B) decreased from 218.82 cells/mm2 at 0.5 hr to 177.36 cells/mm2 at 6 hr. At 12, 24, 48, and 72 hr, the cell density was 120.99, 131.78, 82.29 and 118.27 cells/mm2. EWN cell density in total decreased from 173.03 cells/mm2 at 0.5 hr to 10.43 cells/ mm2 at 6 hr, whereas EDN cell density increased from 10.18 cells/mm2 to 36.52 cells/mm2 at the same time points. At 12–72 hr, a mixed cell population was present. Ultrastructure. In the lesioned area, apoptotic cells containing distinct condensed chromatin, intact membranes, and apoptotic bodies (Fig. 6a) were noted at 0.5 hr. Neuropil appeared disorganized already at 0.5 hr postlesion (Fig. 6b). The ultrastructure of the neuropil was severely affected at 0.5–3 hr, with swollen dendrites and large extracellular vacuoles (Fig. 6c– h), whereas individual cell bodies were easily identified as separate entities. How-

EM Evaluation In EM, the neuropil in control tissue had a normal profile of myelinated axons, axon terminals, and dendrites (Fig. 4a). Homogeneously distributed nuclear chromatin, distinct mitochondria cristae, and distinctly double-lined nuclear and plasma membranes were present as previously described (Cragg, 1976; Fig. 4b). Condensed cells were occasionally noticed at the EM level. Injured neurons. After evaluation of all the experimental samples, ultrastructural features were used to define two categories of damaged neurons, as follows. 1) “Electron-dense neurons” (EDN) were highly electrondense nerve cells with condensed nuclei. Nuclear chromatin appeared clumped or marginated. In the vacuolated cytoplasm, dense organelles were occasionally visible, and the nuclear and plasma membranes were intact (Fig. 4c). 2) “Electron-weak neurons” (EWN) were nerve cells with enlarged round nuclei. The diluted cytoplasm showed peripheral clearing, and expanded or normal organelles were clustered around the nuclei. Nuclear chromatin appeared flocculent with some margination, and the nuclear as well as the plasma membranes were intact (Fig. 4d). We characterized these neurons, made an estimation of their

Fig. 3. Photomicrographs of Nissl-stained sections from cortical layers IV and V of rat brains. In the upper panels, control tissue from a nonlesioned brain (a) and tissues from the lesioned area at 0.5–3 hr (b– d) are shown. A pyramidal neuron with an oval nucleus and homogeneously stained cytoplasm (a) is marked in layer V. At 0.5 hr, triangular pycnotic cells (b) were noticed in both layers in the lesioned area. These cells were more frequent at 1 hr (c). At 3 hr, intensely stained, small cells were observed (d). In the two lower rows the contralateral side is shown. At 0.5–1 hr, some intensely stained, large cells (e,f) were found in layer V. At 3 hr, the tissue showed a different

appearance, with a large number of small, condensed cells (g) in each layer. Together with these cells, weakly stained, small cells (h) were noted at 6 hr. At 12 hr, a few large, condensed cells (i) were found in layer V. Most cells at this time point had a light appearance (i), containing large nuclei surrounded by thin cytoplasm rims. At 24 hr, similar cells (k) were noted; however, they were smaller, and, together with these cells, some small dark, condensed cells (k) were found. At 48 –72 hr, the tissue showed a more normal appearance with homogeneously stained cells of different sizes and a few condensed cells (l,m) in layer V. Scale bars ⫽ 50 ␮m.

906

Andersson et al.

Fig. 4. In control tissue, the neuropil (a) showed a normal profile of myelinated axons (thick arrow), dendrites (arrowhead), and axon terminals (thin arrow). The cells in control tissue contained oval nuclei with homogeneously distributed chromatin (asterisk), distinct organelles (thin arrow), and intact membranes (arrowheads) as shown in b. EDN cell (c) showing clumped chromatin (thin arrows), cytoplasm

vacuoles (thick arrow), and intact membranes (arrowhead). In d, an EWN cell is shown with a large, round nucleus containing slightly disorganized chromatin (asterisk). Clusters of organelles (thin arrow) surrounded the nucleus, and a peripheral cytoplasm clearing (thick arrow) was noticed. The membranes were intact (arrowheads). Scale bars ⫽ 5 ␮m.

ever, at 0.5–3 hr, most cells showed varying degrees of cytoplasm condensation and chromatin margination (Fig. 6c– h). Nuclear membrane ruptures were noted at 0.5 hr (Fig. 6d) and lamellar formations at 1 hr (Fig. 6f) and 3 hr (Fig. 6h), which indicated necrotic cell death. In the contralateral hemisphere, the ultrastructure of the neuropil was well organized between 0.5 and 72 hr postlesion. However, EWN and EDN cells were found at all time points. At each time point, additional cells with different morphology were present. Compared with controls (Fig. 6i– k), some cells in both layer IV and layer V, at 0.5 hr, showed nuclear chromatin clumping and margination (Fig. 6l). However, EWN cells (Fig. 6m) were

most frequent at 0.5–1 hr. Also, at 1 hr, cytoplasm vacuoles (Fig. 6n) and expanded organelles (Fig. 6o) were found. The chromatin margination was more pronounced at 3 hr (Fig. 7a,b). At this time point, most cells showed a changed ultrastructure: Highly marked chromatin clumping and margination was noted as well as reduced cell size. This gave the cell an electron-dense appearance. In some of the cells, ruptured nuclear membranes (Fig. 7a) were found, which is one of the criteria for irreversible cell death. Diluted cytoplasm with expanded organelles and membrane “blebbings” were also seen in these cells (Fig. 7b). At 6 hr, the nuclear chromatin in most cells, though heterogeneously distributed, showed only slight clumping

Ultrastructure in Lesion

Fig. 5. Stacked column graphs showing the cell density of EDN and EWN cells as well as the cell density of the remaining cells in the tissue. In control tissue, a mixed cell population was found with only a few EDN cells and no EWN cells. A shows the cell density in the lesioned area compared with control tissue. Greatly decreased cell density was found at 0.5, 1, and 3 hr. B shows the cell density on the contralateral side. At 0.5 hr, the EWN cell is the dominant cell type. EWN cells decrease in number at 1–3 hr but remain a frequent cell type throughout all recorded time points.

and margination (Fig. 7c,d), which gave the cell a less electron-dense appearance. All membranes were intact at this point. The ultrastructural picture was dominated at 12 hr by EWN cells (Fig. 7f) and at 24 hr by EDN cells (Fig. 7g). At 48 –72 hr, cytoplasm vacuolization (Fig. 7i,m) was occasionally observed, but not to the same extent as at 1 hr postlesion. Some nuclear chromatin margination (Fig. 7m) was noticed, although most nuclei contained homogeneously distributed chromatin. In a few cells, dispersed nucleoli were noted (Fig. 7i–l). We found that, in the lesion, the neuropil was disorganized at the ultrastructural level as early as 0.5 hr

907

postlesion. The ultrastructural changes correlated well with the DW-MRI, which showed reduced ADCW values in the lesion area already at 0.5 hr. Individual cells were easily identified at 0.5–3 hr. Although the brain tissue was severely affected in the lesioned area, apoptotic as well as necrotic cells were identified. Cellular changes on the contralateral side, such as chromatin clumping and margination, were identified as early as 0.5 hr postlesion. The changes were most pronounced at 3 hr, when ruptured nuclear membranes were also found, a sign of irreversible cell death. Insofar as all membranes were intact at 6 –72 hr and the chromatin was not clumped, though heterogeneously distributed, 3– 6 hr after lesioning was the time period during which the cell structure on the contralateral side was altered. Also, at 6 hr, DW-MRI showed extensive water movement in the corpus callosum. We found that both apoptotic and necrotic cell death contributed to the total tissue damage in the lesioned area within 3 hr and that there was an early response to the insult on the contralateral side, which could be observed only by EM. At 3– 6 hr after lesioning, the cell structure on the contralateral side was altered. DISCUSSION Our study shows that the photochemical lesion is a useful ischemia model for studying morphological changes in injured cells and that the insult results in permanent infarction in the lesioned area. The LM results showed a homogenous morphological appearance within the studied groups in the sampled regions, so we considered two animals for EM examination (n ⫽ 2) to be representative for each group. EM images represent very selective events in time and space (Lloreta-Trull, 2000). For this reason, we developed a standardized sampling procedure for unbiased selection of the tissue. Because the sample orientation was randomized, we considered the final ultrathin sections to be representative of the lesion. EM showed that the neuropil on the lesion side had lost its integrity already at 0.5 hr, while cells were easily identified as separate cell bodies. The early tissue disorganization might be explained by a greatly increased amount of metabolites, such as lactate, and electrolytes released into the extracellular space (Ginsberg, 1990). Glial cells, also affected early by the lesion, have been proposed to have impaired metabolism following injury (Kraig and Chesler, 1990). Therefore, their endocytotic functions are diminished, resulting in an increase of metabolites in the extracellular space. This excess of metabolites leads to more acid extracellular pH and either the functional loss of membrane ion pumps (Chesler, 1990) or cellular swelling (Sykova´, 1997). As a consequence, cell homeostasis will be greatly changed, followed by metabolic disturbance, e.g., decreased protein synthesis; therefore, the cell structure loses its integrity at a later time point. Also, increased glutamate release from affected astrocytes produces irreversible cell damage within a few hours. The lesioned area was observed already at 0.5–24 hr as a dark or bright area in ADCW or T2 maps, respectively. ADCW values showed changes in the water homeostasis in the intra- and extra-

908

Andersson et al.

Figure 6.

Ultrastructure in Lesion

cellular space. Also, as a consequence of function loss of membrane ion pumps, e.g., sodium/potassium exchange pumps, sodium enters the cell, the sodium gradient across the membrane is diminished, and water enters the cell by osmosis. This results in cellular edema, and, subsequently, the architecture of the extracellular space is changed (Nicholson and Sykova´, 1998). However, the severe structural changes in the neuropil were observed only at the ultrastructural level. On the lesion side, the morphological changes in most cells fit into the two conventional pathways leading to cell death. Although apoptosis was most prominent at 0.5 hr, apoptotic and necrotic changes were seen simultaneously in the tissue at 0.5–3 hr, which was also shown in earlier studies (Lipton, 1999). At 0.5 hr, ATP levels within the cell probably are still high enough for the energy-demanding apoptotic process (Studzinski, 1999). At later time points, the energy levels are decreased, allowing only the necrotic cell death pathway. The final stages of apoptosis in ischemic tissue are very similar to the end stages of necrosis (Mac Manus and Linnik, 1997), which also complicates the differentiation between apoptosis and necrosis. In our study, both apoptotic and necrotic cell death contributed to total tissue damage in the lesioned area within 3 hr. EWN cells, which were present in large numbers on the contralateral side at 0.5 and 1 hr, were at the LM level most likely equivalent to the large cells with weakly stained nuclei found at the same time points. Their large, pale nuclei indicate a hypermetabolic state, in that nuclei containing an active form of chromatin, e.g., euchromatin, represent high protein synthesis (Ghadially, 1975). The metabolically active state is probably a consequence of inhibited aerobic energy metabolism resulting from the

909

lesion. Therefore, a compensatory energy metabolism, e.g., anaerobic glycolysis, is present, also contributing to the active state (Alberts, 1989). The use of cytoplasmatic glycogen in anaerobic glycolysis could lead to a glycogen deficit, resulting in the observed diluted cytoplasm. The early injury caused by the insult might be mediated by rapid axonal signalling, and primary damage is thus also present on the contralateral side. The metabolic activity of the cells suggests an attempt by the cells to modulate their morphology and ultimately survive. The remarkable cell changes noted at 3 and 6 hr on the contralateral side might represent a breakpoint for structural alteration. Showing membrane ruptures, which is a sign of irreversible cell death, the most frequent cells at 3 hr could represent the ultimate stage in a fast death process starting already at 0.5 hr. Although the cells had a dark appearance and were hyperchromatic on LM and slightly electron dense on EM, the cells showed neither a condensed morphology nor mitochondria with flocculent densities. These are the main criteria of irreversible cell death (Garcia et al., 1978). Therefore, the most frequent cells at 3 hr were not considered as dying cells but as highly affected but still viable cells. At 6 hr, no cells on the contralateral side showed signs of irreversible cell death. Insofar as the EDN cells, defined as condensed but not dying cells, were increased in number at 6 hr, some electron-dense cells at 3 hr might have changed in structure and been identified as EDC cells at 6 hr. Containing irregularly clumped chromatin and a condensed cytoplasm, the cells at 3 hr had characteristics similar to those of ischemic cell changes in ICC neurons. The characteristics of ICC neurons, which are pycnosis and cytoplasm shrinkage, fit into the definition of the electron-dense EDN cells in this study. In earlier studies of

Š

Fig. 6. Electron micrographs showing the ultrastructure in cortical layers IV and V of the lesioned hemisphere. A few cells with an apoptotic appearance shown in a were found already at 0.5 hr. Within the highly disintegrated cells, a separate body, probably an apoptotic body, with clumped chromatin was noted (asterisk). Marked chromatin marginations and large chromatin clumps (thin arrow) dominated the nucleus. The neuropil (b) was disorganized at 0.5 hr with large vacuoles (thin arrow) and swollen dendrites (thick arrow). At 0.5 hr, the ultrastructure was severely changed in most of the cells. Some of the cells appeared with a distinct nucleus (c). The chromatin was marginated (thin arrow) to the double-lined nuclear membrane (arrowhead). Within the disorganized cytoplasm, organelles, surrounding the nucleus, were identified as separate units (thick arrow). In d, the nucleus is still identifiable as a separate unit, and the cell body can be outlined (thin arrow). The chromatin is electron dense and clumped (asterisk), and the nuclear envelope is in some parts observed as double-lined and in some parts ruptured (arrowheads and insets). At 1 hr, the damage was more pronounced. Large chromatin clumps (e; asterisk) and marked chromatin marginations (thin arrows) were noted in most of the cells. A few highly disintegrated cells (f) were found at this time. Small nuclei with intact nuclear envelopes (arrowhead) were still visible, and lamellar formations (thin arrow) were observed within plasma compartment. At 3 hr, most of the cells showed a highly disintegrated ultrastructure. The nuclear envelope (g; arrowhead) was identifiable as an expanded double-lined membrane, the nuclear chromatin was

highly clumped (thin arrow), and small vacuoles were seen within the plasma compartment (thick arrow). A few cells showed a complete disintegration (h). Only lamellar formations (thin arrow) could be identified in this cell. In control brain tissue, cells with oval nuclei containing homogeneously distributed chromatin (asterisk) are shown in (i,k). The nuclear membranes are distinctly double-lined (arrowhead), and the plasma compartments are well organized, containing normally shaped organelles without ruptures, densities, or expansions (thick arrows). l– o Show cells in cortical layers IV and V in the contralateral hemisphere. Already at 0.5 hr, the cells were affected. A few cells were small, as in l. An oval nucleus with slightly clumped chromatin (asterisk) and chromatin margination (thin arrow) at the intact nuclear membrane (arrowhead) was observed. Most of the cells had an expanded appearance (m) and were defined as EWN cells. The large oval nucleus contained flocculent chromatin (asterisk). Nuclear and plasma membranes were intact (arrowhead), and organelles appeared expanded in the diluted plasma compartment (thick arrow). At 1 hr, cells (n) with marked chromatin margination and clumping (thin arrows) were frequently observed. The nuclear membrane appeared amorphous, and the two layers could not be distinctly identified (arrowheads). The plasma compartment was highly vacuolated (thick arrow). Another neuronal cell type (o) containing organelles clustered around a large nucleus was observed. Some of the organelles were intact and some swollen (thick arrow). Scale bars ⫽ 5 ␮m.

910

Andersson et al.

Figure 7.

Ultrastructure in Lesion

ischemic tissue, ICC neurons, in spite of their triangular, condensed appearance, were not associated with cell death (Brown and Brierly, 1972; Petito and Pulsinelli, 1984). Furthermore, in hypoglycemic brain damage, “dark” neurons are compatible with cell survival (Auer et al., 1985a,b). Therefore, some electron-dense cells at 3 hr might survive as EDN cells at 6 hr. The most frequent cells at 6 hr, with irregular nuclei and a diluted cytoplasm, were similar to EWN cells, though smaller. Because EWN cells probably are metabolically active cells, the dominant cell type at 6 hr was considered as an active, viable cell. Thus, most cells at 6 hr seemed to be viable, although the most frequent cells at 3 hr were highly affected. These findings indicate a breakpoint for death vs. survival. At 24 hr, on the contralateral side, large numbers of affected cells as well as EDN cells were found. The injury might affect the synaptic as well as extrasynaptic “volume” transmission, e.g., the movement of neuroactive substances through the volume of the extracellular space (Sykova´, 1997; Nicholson and Sykova´, 1998). This information processing (Fuxe and Agnati, 1991) is probably damaged at 24 hr (Mazel et al., 2002), resulting in injured cells. The total cell density on the contralateral side at 12–72 hr (Fig. 5B) was reduced, suggesting increased Š

Fig. 7. Electron micrographs showing cortical layers IV and V in the contralateral hemisphere of a lesioned brain. At 3 hr, the cell ultrastructure was changed. Most cells were small, with a condensed appearance, as shown in a and b, with nuclei containing highly clumped and marginated chromatin (thin arrows). The double-lined nuclear envelope in a showed some “blebbings” and ruptures (arrowhead), and the cytoplasm in b appeared diluted (thick arrow). At 6 hr, most of the small cells (c,d) appeared diluted, with slightly disorganized and clumped chromatin (asterisk). The fold nuclear envelope was in c distinctly double-lined (arrowhead), whereas in d some “blebbings” (arrowhead) were noted. In c, large cytoplasm vacuoles and swollen organelles (thick arrow) were observed. The plasma compartment in d appeared diluted (thick arrow), without distinct organelles. At 12 hr, cells (e) similar to the most frequently observed cells at 6 hr were noticed. The chromatin was slightly clumped (asterisk) and marginated (thin arrow) to the ruptured nuclear membrane. However, many cells were defined as EWN cells (f) containing a large, round nucleus with flocculent chromatin (asterisk). Nuclear and plasma membranes were intact (arrowheads). In the diluted cytoplasm, remnants of organelles clustered around the nucleus were observed (thick arrow). At 24 hr, condensed disintegrated cells (g) defined as EDN cells were most frequently seen. The chromatin was highly condensed and clumped (asterisk), and the nuclear envelope was identified as a double layer (arrowhead). Small vacuoles within the plasma compartment were noted (thick arrow). Together with EDN cells, small cells with a “blown up” appearance (h) were observed. They were similar to EWN cells, but smaller. Also, distinctly from EWN cells, the nuclear envelope appeared amorphous (arrowhead). As in EWN cells, the plasma compartment was diluted, containing organelles clustered around the nucleus (thick arrow). A mixed cell population was noted at 48 –72 hr, and some cells contained dispersed nucleoli. The nucleoli (i–l; thin arrows) seemed to divide in two. The chromatin was marginated as a thin rim to the inner nuclear membrane (m; thin arrow). The cytoplasm compartments were easily identified and well organized. However, large cytoplasm vacuoles (thick arrows) were observed in i and m. Scale bars ⫽ 5 ␮m.

911

extracellular space, which was also confirmed by DWMRI (Vorisek et al., 2002). Starting at 6 hr, the corpus callosum and external capsulae were observed as bright areas in ADCW and T2 maps, thus showing increased water movement between both hemispheres. However, the structural changes on the contralateral side were not detected by DW-MRI (Vorisek et al., 2002), whereas EM showed less severe morphological changes. The threshold at which some cells undergo lethal changes and others recover depends on the individual metabolic cell status at each time point (Collins, 1989). The ATP levels within some of the cells are probably restored at 6 hr after lesioning. ATP-dependent membrane pumps, such as the sodium/potassium and calcium/ hydrogen exchange pumps, function optimally when supplied with enough ATP. Subsequently, the membrane potential and cell homeostasis are maintained, and the cell survives. Therefore, there is a possibility for neuronal reorganization after 6 hr. The beam-walking test showed functional recovery at 48 hr and 72 hr. Also, at 48 and 72 hr, dispersed nucleoli were found, indicating cell division (Alberts, 1989) and reorganization. However, only little is known about the metabolic alterations in the end stages of ischemic cell changes, and the exact mechanism of endogenous neuromodulation cannot yet be explained. We conclude that the photochemical lesion is a useful model for studying injured cells. By using EM, subtle morphological changes can be studied and the criteria of irreversible cell death detected. Our study showed complete cell death in the lesioned area and the ultrastructural criteria of irreversible cell death on the contralateral side at 3 hr after lesioning, so we suggest 3 hr to be a breakpoint at which cells either progress toward lethal changes or recover. Thereafter, neuronal reorganization could be possible, in that metabolically active cells are still present in the contralateral hemisphere. REFERENCES Abo M, Chen Z, Lai LJ, Reese T, Bjelke B. 2001. Functional recovery after brain lesion— contralateral neuromodulation: a fMRI study. Neuroreport 25:1543–1547. Alberts B. 1989. Molecular biology of the cell. New York: Garland. Auer RN, Kalimo H, Olsson Y, Siesjo BK. 1985a. The temporal evolution of hypoglycemic brain damage. I. Light- and electron-microscopic findings in the rat cerebral cortex. Acta Neuropathol 67:13–24. Auer RN, Kalimo H, Olsson Y, Siesjo BK. 1985b. The temporal evolution of hypoglycemic brain damage. II. Light- and electron-microscopic findings in the hippocampal gyrus and subiculum of the rat. Acta Neuropathol 67:25–36. Brown AW. 1977. Structural abnormalities in neurones. J Clin Pathol 30(Suppl 11):155–169. Brown AW, Brierly JB. 1972. Anoxic-iscemic cell change in rat brain: light microscopic and fine-structural observations. J Neurol Sci 16:59 – 84. Chesler M. 1990. The regulation and modulation of pH in the nervous system. Prog Neurobiol 34:401– 427. Collins RC. 1989. Selective vulnerability of the brain: new insights into the pathophysiology of stroke. Ann Intern Med 110:992–1000. Cragg BG. 1976. Ultrastructural features of human cerebral cortex. J Anat 121:331–362. Feeney DM, Boyeson MG, Linn RT, Murray HM, Dail WG. 1981. Responses to cortical injury: I. Methodology and local effects of contusions in the rat. Brain Res 211:67–77.

912

Andersson et al.

Fuxe K, Agnati LF. 1991. Volume transmission in the brain: novel mechanisms for neuronal transmission. New York: Raven Press. Garcia JH, Lossinsky AS, Kauffman FC, Conger KA. 1978. Neuronal ischemic injury: light microscopy, ultrastructure and biochemistry. Acta Neuropathol 43:85–95. Garcia JH, Yoshida Y, Chen H, Li Y, Zhang ZG, Lian J, Chen S, Chopp M. 1993. Progression from ischemic injury to infarct following middle cerebral artery occlusion in the rat. Am J Pathol 142:623– 635. Garcia JH, Liu KF, Ho KL. 1995. Neuronal necrosis after middle cerebral artery occlusion in Wistar rats progresses at different time intervals in the caudoputamen and the cortex. Stroke 26:636 – 643. Ghadially FN. 1975. Ultrastructural pathology of the cell. London: Butterworth & Co. Ltd. Ginsberg MD. 1990. Local metabolic responses to cerebral ischemia. Cerebrovasc Brain Metab Rev 2:58 –93. Kalimo H, Garcia JH, Kamijyo Y, Tanaka J, Trump BF. 1977. The ultrastructure of “brain death”. II. Electron microscopy of feline cortex after complete ischemia. Virchous Archiv B Cell Pathol 25:207–220. Kalimo H, Olsson Y, Palajarvi L, Soderfeldt B. 1982. Structural changes in brain tissue under hypoxic-ischemic conditions. J Cereb Blood Flow Metab 2(Suppl 1):19 –22. Kerr JF, Wyllie AH, Currie AR. 1972. Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. Br J Cancer 26:239 –257. Kraig RP, Chesler M. 1990. Astrocytic acidosis in hyperglycemic and complete ischemia. J Cereb Blood Flow Metab 10:104 –114. Lee JM, Zipfel GJ, Choi DW. 1999. The changing landscape of ischaemic brain injury mechanisms. Nature 399(Suppl):A7–A14. Lipton P. 1999. Ischemic cell death in brain neurons. Physiol Rev 79:1431– 1568. Lloreta-Trull J. 2000. Electron microscopy in pathology articles: a retrospective appraisal. Ultrastrct Pathol 24:105–108. Mac Manus JP, Linnik MD. 1997. Gene expression indused by cerebral

ischemia: an apoptotic perspective. J Cereb Blood Flow Metab 17:815– 832. Mazel T, Richter F, Vargova´ L, Sykova´ E. 2002. Changes in extracellular space volume and geometry induced by cortical spreading depression in immature and adult rats. Physiol Res 51(Suppl 1):85–93. Nicholson C, Sykova´ E. 1998. Extracellular space structure revealed by diffusion analysis. Trends Neurosci 21:207–215. Paxinos G, Watson C. 1998. The rat brain in stereotaxic coordinates. Sydney: Academic Press. Petito CK, Pulsinelli WA. 1984. Sequential development of reversible and irreversible neuronal damage following cerebral ischemia. J Neuropathol Exp Neorol 43:141–153. Pierpaoli C, Righini A, Linfante I, Tao-Cheng JH, Alger JR, Di Chiro G. 1993. Histopathologic correlates of abnormal water diffusion in cerebral ischemia: diffusion-weighted MR imaging and light and electron microscopic study. Radiology 189:439 – 448. Reese T, Porszasz R, Baumann D, Bochelen D, Boumezbeur F, McAllister KH, Sauter A, Bjelke B, Rudin M. 2000. Cytoprotection does not preserve brain functionality in rats during the acute poststroke phase despite evidence of non-infarction provided by MRI. NMR Biomed 13:361–370. Studzinski GP. 1999. In: Hames BD, editor. Apoptosis, a practical approach. New York: Oxford University Press. Sykova´ E. 1997. The extracellular space in the CNS: its regulation, volume and geometry in normal and pathological neuronal function. Neuroscientist 3:28 – 41. Vorˇ´ıˇsek I, Ha´jek M, Tintera J, Nicolay K, Sykova´ E. 2002. Water ADC, extracellular space volume, and turtuosity in the rat cortex after traumatic injury. Magn Reson Med 48:994 –1003. Watson BD, Dietrich WD, Busto R, Wachtel MS, Ginsberg MD. 1985. Induction of reproducible brain infarction by photochemically initiated thrombosis. Ann Neurol 17:497–504.