DNA Damage and Activated Caspase-3 Expression in Neurons and Astrocytes: Evidence for Apoptosis in Frontotemporal Dementia

Experimental Neurology 163, 9 –19 (2000) doi:10.1006/exnr.2000.7340, available online at http://www.idealibrary.com on

DNA Damage and Activated Caspase-3 Expression in Neurons and Astrocytes: Evidence for Apoptosis in Frontotemporal Dementia Joseph H. Su,* Kathryn E. Nichol,* Tom Sitch,* Philip Sheu,† Charlie Chubb,† Bruce L. Miller,§ Kevin J. Tomaselli,㛳 Ronald C. Kim,‡ and Carl W. Cotman* *Institute for Brain Aging and Dementia, †Department of Electrical and Computer Engineering, and ‡Department of Pathology, University of California, Irvine, Irvine, California 92697; §Department of Neurology, UCSF School of Medicine, San Francisco, California 94143; and 㛳IDUN Pharmaceuticals, Inc., La Jolla, California 92037 Received February 18, 1999; accepted December 28, 1999

neuropathological characteristics of FTD consist mainly of frontal and anterior temporal cortical degeneration with neuronal loss, gliosis, and microvacuolation of lamina I to lamina III. No changes characteristic of Pick’s disease or Alzheimer’s disease were noted. Although patients with FTD due to Pick’s disease, Pick’s disease/FTD, or FTD/motor neuron disease share many clinical similarities and show a common topographical distribution of atrophy within frontal and anterior temporal lobes, the neuropathological changes underlying these different types of degeneration appear diverse (6, 18). Thus, the concept of FTD as a separate clinicopathological entity has gained increased recognition (9, 18, 30). Neuronal cell loss is a prominent feature of FTD (5, 6, 18). Cell death may be caused by either apoptosis or necrosis. Because DNA degradation occurs in both apoptosis and necrosis, both apoptotic and necrotic strand breaks can be detected with new histological techniques such as TdT-mediated in situ end-labeling (14) and in situ nick translation (16). We and others have shown that cells in the AD brain exhibit labeling for DNA strand breaks using TdT, whereas few such nuclei are detected in control brain (2, 25, 43). To understand the mechanisms of neuronal loss in FTD, it is necessary to determine whether DNA damage as well as other indices of apoptosis are present in the cells of FTD brain. Genetic studies in the roundworm identified the ced3 gene, whose function is required for normal apoptosis during development (52; for reviews, see 38). CED-3 is a cysteine protease. Numerous human CED-3 homologs have been identified and there are currently at least 10 human genes that exhibit homology to CED-3. To achieve consistency, “caspase” has been adopted as a root name for all members of this cysteine aspartate protease family (1). Each caspase is synthesized as an inactive proenzyme that is processed in cells undergoing apoptosis by self-proteolysis and/or cleavage by another protease. Caspase-3 (also known as CPP32,

Frontotemporal dementia (FTD) is a neurodegenerative disease which affects mainly the frontal and anterior temporal cortex. It is associated with neuronal loss, gliosis, and microvacuolation of lamina I to III in these brain regions. In previous studies we have described neurons with DNA damage in the absence of tangle formation and suggested this may result in tangle-independent mechanisms of neurodegeneration in the AD brain. In the present study, we sought to examine DNA fragmentation and activated caspase-3 expression in FTD brain where tangle formation is largely absent. The results demonstrate that numerous nuclei were TdT positive in all FTD brains examined. Activated caspase-3 immunoreactivity was detected in both neurons and astrocytes and was elevated in FTD cases as compared to control cases. A subset of activated caspase-3-positive cells were also TdT positive. In addition, the cell bodies of a subset of astrocytes showed enlarged, irregular shapes, and vacuolation and their processes appeared fragmented. These degenerating astrocytes were positive for activated caspase-3 and colocalized with robust TdT-labeled nuclei. These findings suggest that a subset of astrocytes exhibit degeneration and that DNA damage and activated caspase-3 may contribute to neuronal cell death and astrocyte degeneration in the FTD brain. Our results suggest that apoptosis may be a mechanism of neuronal cell death in FTD as well as in AD (228). © 2000 Academic Press Key Words: apoptosis; astrocyte; activated caspase-3; degeneration; DNA fragmentation; frontotemporal dementia; neuron; neuropathology.

INTRODUCTION

Frontotemporal dementia (FTD) or dementia of frontal lobe type (5, 17, 29, 33) is a neurodegenerative disease where patients typically present with behavioral disturbances that over time lead to dementia (17, 33). Brun and others (5, 29) have reported that the 9

0014-4886/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.

10

SU ET AL.

Yama, and Apopain) has been most intensively studied (for reviews, see 31). Caspase-3 normally exists in the cytosolic fraction of cells as an inactive precursor and is activated in early apoptosis. Activated caspase-3 is found only in cells undergoing apoptosis and consists of p18 (amino acids 29 –175) and p12 subunits (amino acids 176 –277) (34) that are derived from a 32-kDa proenzyme (pro-caspase-3) by cleavage at multiple aspartic acid sites. Several studies have implied that caspase-3 is both necessary and sufficient to trigger apoptosis. Multiple apoptotic signals activate caspase-3 (10, 11, 20). A caspase-3-specific tetrapeptide inhibitor, Ac-DEVD-CHO, can block the initiation of the cellular apoptotic program in response to apoptotic stimuli (20, 34). Addition of active caspase-3 to normal cytosol activates the apoptotic program (12). Thus, caspase-3 has properties of a cell death protease. Although mRNAs encoding CED proteases and caspase-3 have been detected in normal brain (13, 23, 24, 50), little is known about the expression of activated caspase-3 in diseased brain. Such information is important for understanding the molecular mechanisms of neuronal cell death in FTD and could be critical for developing strategies for managing these proteases with inhibitors. Astrogliosis is another prominent feature of FTD (5, 18, 29). Although the disparity between the number and distribution of astrocytes are noted (29), affected regions of FTD brain show an increased cerebral GFAP immunoreactivity within reactive astrocytes. Astrocyte activation is a universal response to brain injury. Elevation of GFAP is found wherever neuronal degeneration occurs (37, 43). Little is known, however, about astrocyte degeneration in FTD, even though fragmentation of the astroglial processes, termed clasmatodendrosis by Cajal, has been documented in brains with cerebral edema and ischemic cerebrovascular disease foci (7). We have observed that a small subset of astrocytes display TdT-positive labeling for DNA damage in AD brain (43). To evaluate pathogenesis in FTD, it is interesting to determine whether the astrocytes also exhibit degeneration in FTD brain. In the present study, we sought to evaluate activated caspase-3 expression, DNA fragmentation, and the relationship between these two variables in FTD brain. The results demonstrate that numerous nuclei were TdT positive in all FTD brains examined. Activated caspase-3 immunoreactivity was detected in neurons and was elevated in FTD cases. A subset of activated caspase-3 positive neurons were also TdT positive. In addition, the cell bodies of a subset of astrocytes showed enlargement, vacuolation, activated caspase-3 immunoreactivity, and TdT-positive staining. The processes of these cells were also disintegrated. These findings suggest that DNA damage and activated caspase-3 may contribute to neuronal cell death and

astrocyte degeneration and that astrocytes exhibit degeneration in FTD brain. MATERIALS AND METHODS

Human Tissue Postmortem tissue was obtained from the Institute for Brain Aging and Dementia Tissue Repository. The superior frontal gyrus and anterior pole of temporal lobe from six FTD and six control cases were examined. Fresh brain tissue was fixed in 10% formalin in 0.1 M Sorensen’s buffer, pH 7.3, for 48 –72 h and stored in 0.1 M PBS (0.02% sodium azide) or fixed in 4% paraformaldehyde in 0.1 M Sorensen’s buffer, pH 7.3, for 24 h at 4°C. All experiments were performed using freefloating 40- to 50-␮m sections cut on a Vibratome and collected in PBS, pH 7.4. Frontotemporal dementia and control cases were matched for age and PMD as closely as possible (Table 1) (mean ages: FTD ⫽ 69 years, control ⫽ 71 years; mean PMD AD ⫽ 4.5 h, control ⫽ 6.3 h). TdT 3⬘-OH DNA Strand-Break Labeling ApopTag peroxidase kits (Oncor, MD) were used to detect digoxigenin-nucleotide residues added by TdT to the 3⬘-OH termini of DNA strand breaks generated during DNA fragmentation in neurons. ApopTag staining was performed according to protocol as described previously (44, 47, 48). Tissue sections were incubated in TdT enzyme and digoxigenin– dUTP reaction buffer. Sections were then incubated in anti-digoxigenin-antibody conjugated with peroxidase. The digoxigenin– dUTP–peroxidase complex was visualized by reacting with DAB (Vector Labs, CA) to generate a brown reaction product. Negative controls were performed by substituting distilled water for TdT enzyme in the preparation of working solution and were negative. Analysis of control tissue incubated with distilled water substituted for TdT enzyme has been previously published, demonstrating the high specificity of TdT staining (2). Quantitative Evaluation of TdT-Labeled Cells Three ⫻ 20 fields, each field corresponding to an approximate area of 2.5 mm 2 of tissue, were chosen from four FTD and four control cases which reflected the overall TdT labeling on the entire slide. All images were photographed on 35-mm slides and digitized with a SprintScan 35-mm Slide Scanner. Following digital capture a “magic wand” tool was used to designate pixel density range for primary, secondary, and background populations (Figs. 1a and 1b). In this case, dark cells assumed as cells of irreversible DNA damage were selected for the primary population, and light cells assumed as DNA strand breaks in response to injury as well as part of the normal aging processes

11

APOPTOSIS IN FRONTOTEMPORAL DEMENTIA

TABLE 1 FTD and Control Case Information and Relative Caspase-3 Labeling in Superior Frontal Cortex and Anterior Pole of Temporal Lobe

Case no.

Age

Sex

PMD

MMSE

1 2 3 4 5 6

62 77 62 72 68 74

M F M F M M

5h 5h 5h 5h 3.5 h 3.5 h

22/30 n/a n/a n/a 0/30 18/30

1 2 3 4 5 6

74 70 81 64 64 76

M M F F M F

8 4.5 6 8 7 4.5

n/a n/a n/a n/a n/a n/a

Pathological diagnosis

Pick bodies

Motor involvement

Activated caspase-3 labeling

FTD FTD FTD FTD FTD FTD/right basal ganglia infarct* Control Control Control Control Control Control

— — — — — —

— — — — — —

⫹⫹⫹⫹ ⫹⫹⫹ ⫹⫹⫹⫹ ⫹⫹⫹ ⫹⫹⫹⫹ ⫹⫹⫹⫹

— — — — — —

— — — — — —

⫹ ⫾ ⫹ ⫹⫹ ⫹⫹ ⫹⫹

Abbreviations: PMD, postmortem delay; MMSE, Mini-Mental State Examination, maximum possible score ⫽ 30, decreasing scores represent more severe dementia; n/a, not available. * Rare punctate foci of cystic degeneration was identified elsewhere.

were selected for the secondary population. The green background was selected as the background population. Cell counts were then analyzed by an independent samples t test. Given this pixel color information, the application generated a Gaussian-space distribution for each of the three populations. A value for the probability of its membership in each of the three color populations was

FIG. 1. Photomicrographic demonstration of the pertinent steps of computer-assisted color analysis. (a) TdT-labeled nuclei. (b) Same picture showing dark- (arrowhead) and light- (arrow) labeling nuclei. This demonstrates the thresholding task, which assigns labels such as “light” and “dark” according to the criteria of the experimenter. (c) Dark-labeled nuclei, inside blue boxes, and light-labeled nuclei, inside red boxes; the computer then tabulates cell-count information according to the criteria selected in b. (d) Dark- and light-labeled nuclei shown by blue and red, respectively. At stages shown in c and d, the experimenter can refine selection criteria, if necessary, for greater accuracy.

then generated for each pixel in the image. This probability was then modified by a Bayesian classifier that examined the pixels based on prior probabilities gathered and maintained from previous evaluations. The evaluation entered into its final stage when the probability values were used to create a deterministic output (Fig. 1c). The pixels, once assigned to light cells, dark cells, or background, were grouped together into true objects that were then screened based on highlevel criteria. In this case, the objects were only screened for size. The application finished by displaying those objects recognized as cells by boxes and outputting a final count of both the primary and secondary populations selected (Fig. 1d). For the data in this experiment, an evaluation was run initially for all the images in each case. All evaluations built their criteria off the criteria from the previous evaluation. Once all of the images had been run, the application contained a set of criteria trained to effectively count cells in all of the samples within a reasonable margin of error (estimated at ⬍5%). Finally, this set of criteria was applied to each of the images and a cell count for light and dark cells was taken for each image. Immunocytochemistry Tissue was processed as described previously (45, 46). Briefly, tissue sections were treated for 20 min with 1.0% H 2O 2 to inactivate endogenous peroxidases. Before incubation with antibodies to caspase-3, some sections were immersed in boiling water for 5 min. Sections were incubated overnight at room temperature in primary antibody, rinsed, and incubated in

12

SU ET AL.

FIG. 2. TdT-labeled nuclei in control and FTD. (a) TdT-labeled nuclei in control brain. Positive nuclei are usually light (arrow). (b) TdT-labeled in FTD brain. Positive nuclei are robustly labeled (arrow). (Left inset) Higher magnification shows chromatin aggregation. (Right inset) Higher magnification shows apoptotic bodies (arrow). Scale bars in a and b, 50 ␮m.

biotinylated secondary antibody and avidin– biotin complex for 1 h, respectively (Vector Labs, CA). The final color product for single labeling was visualized by diaminobenzidine (DAB) for a brown reaction product or an SG kit (Vector Labs) for a blue-gray reaction product. Immunostaining intensity of caspase-3 was rated as follows: ⫹, weak; ⫹⫹, moderate; ⫹⫹⫹, intense; ⫹⫹⫹⫹ very intense. In double-labeling experiments, bound antibodies were detected using a DAB kit for the first antigen and an SG kit for the second antigen. Sections incubated in parallel without primary antibody failed to develop specific staining. Antibodies An antibody (CM1) was used to identify activated caspase-3 (IDUN Pharmaceuticals, Inc., La Jolla, CA). The protocols employed for the production of CM1 antibody and its characterization have been described elsewhere (41). Briefly, this antibody was raised against a 13-amino-acid peptide sequence from the carboxyl terminus of the p18 subunit of cleaved caspase-3. On Western analysis, this antibody recognized the large (p18) subunit of processed caspase-3 and did not recognize the unprocessed p32 zymogen form or the processed p12 subunit of caspase-3. CM1 immunolabeled apoptotic but not normal or necrotic neurons in vitro. CM1 immunoreactivity was absent in the nervous system of caspase-3-deficient mouse embryos and in neurons cultured from caspase-3-deficient mice. One commercially available polyclonal antibody, PharMingen caspase-3 antibody (65906E), was also used to recognize both the 32-kDa unprocessed pro-caspase-3 and the 17-/18-kDa subunit of the active caspase-3. This antibody was raised in rabbits using recombinant human CPP32-His6 as immunogen and was characterized by Western blot analysis and immunohistochemistry of paraffin-embedded tissue sections (23). Two well-characterized monoclonal antibodies, AT8 and PHF-1, were used to detect neurofibrillary degeneration. The specificity of these antibodies have been de-

scribed in detail previously (3, 4). In addition, an antibody (Dako, CA) against glial fibrillary acidic protein (GFAP) or an antibody against S-100 (␤-subunit) (Sigma, St. Louis, MO) was used as an astrocyte marker. Western Blot Analysis Western blot analysis was performed to confirm the presence of the active subunit of caspase-3 in FTD human brain tissue. Frozen brain samples were homogenized in 10 volumes extraction buffer (100 mM Tris, 10 mM EDTA, and 1% SDS) with a protease inhibitor cocktail (1:1000 ZVAD, 1:200 PMSF, and 1:200 pepstatin ⫹ leupeptin). Protein content was assessed using a BCA assay. Samples were boiled for 5 min then diluted with sample buffer (0.5 M Tris, 37% glycerol, 1% SDS, 5% ␤-mercaptoethanol, and 0.04% bromphenol blue). Samples from FTD and control-tissue homogenates were run on a 12% SDS–PAGE gel (100 ␮g/lane) and transferred to PVDF membrane. Membranes were blocked 2 h in TTBS with 2% horse aserum and 2% BSA. Membranes were then incubated 1 h in the IDUN CM1 antibody for the 17-/18-kDa active caspase-3 subunit (1:3000 in TTBS) or the Pharmingen caspase-3 antibody (65906E) (1:1000), which recognizes the 32-kDa pro-enzyme, as well as the 17-/18-kDa active subunit. Membranes were washed and incubated in the appropriate secondary antibody. Following another washing series, membrane proteins were visualized using a commercially available ECL kit (Amersham Pharmaceuticals, Inc.). RESULTS

TdT-Labeled Nuclei in Age-Matched Control and FTD Brain ApopTag kits were used to detect DNA strand breaks generated during DNA fragmentation. Low levels of TdT labeling were observed in six control brains exam-

APOPTOSIS IN FRONTOTEMPORAL DEMENTIA

FIG. 3. DNA fragmentation, assessed by TdT labeling, was quantified by computer. Number of TdT-positive cells in four aged control cases and four FTD cases is shown above. FTD cases had vastly more light- and dark-labeled nuclei showing DNA fragmentation than control cases. Abbreviations: L, lightly stained cell; D, darkly stained cell; Ave, average.

ined (Fig. 2a). Positive nuclei were usually light. Robust-labeled nuclei were rare. On average, there were 10 times more light-labeling cells than dark-labeling cells in control brain (Fig. 3). In contrast, numerous nuclei with evidence of DNA damage were present in all FTD brains examined (Fig. 2b). In general, TdT-positive nuclei were numerous in gray matter; however, a few strong TdT-positive nuclei were also detected in white matter. TdT-positive nuclei were present in almost all layers of cortex in gray matter, although robustly labeled cells were predominantly seen in layer III and in layer V. On average, there were 10 times more dark-labeling cells in FTD compared to those in controls (Fig. 3). For both lightand dark-labeled cells, the number of cells in FTD was significantly greater than the number in control brain (P ⬍ 0.05), though we caution that the sample size was small (n ⫽ 4). A subset of these labeled nuclei exhibited the classic, distinct morphological characteristics of apoptosis (chromatin aggregation, nuclear shrinkage, and formation of apoptotic bodies) (Fig. 2b, insets). The morphological characteristics and distinct distribution of the TdT-labeled cells suggest that many of them are neuronal in type, consistent with our previous observations (42, 45, 46). Activated Caspase-3 Expression in Age-Matched Control and FTD Brain CM1, which detects activated caspase-3, was examined in six normal control cases with an age range similar to the FTD cases. Activated caspase-3 immu-

13

noreactivity was present in some neurons. Those neurons positive for activated caspase-3 showed weak to moderate staining within the soma and proximal neurites (Fig. 4a) in five of six cases. In three control cases, light to moderate immunoreactivity for activated caspase-3 was also found within glial cells with morphology resembling that of astrocytes (not shown). Activated caspase-3 immunoreactivity was significantly more intense in six FTD brains examined compared to control cases (Table 1). There was no apparent relationship of activated caspase-3 immunoreactivity with PMD or different fixation protocols. Neurons immunoreactive for activated caspase-3 in FTD cases were clearly detected in pyramidal cell bodies and neurites (dendrites and axons) (Fig. 4b). Activated caspase-3 immunoreactivity was also detected within nuclei (Fig. 4b, inset). Robustly labeled cells were usually pyramidal neurons. Cortical granule neurons also showed moderate to intense activated caspase-3 immunoreactivity. Activated caspase-3 immunoreactivity was localized in the soma, nucleus, or neurites in a fine punctate pattern

FIG. 4. Activated caspase-3 expression in age-matched control and FTD brain. (a) Activated caspase-3 expression is faint in control brain (arrows indicate neurons). (b) Activated caspase-3 expression is more robust in FTD brain (arrows indicate neurons). (Left inset) Higher magnification shows activated caspase-3 immunoreactivity in neurons of FTD brain. Note that activated caspase-3 immunoreactivity is localized in soma, neurites, and nucleus. This immunoreactivity demonstrates the active caspase-3 (17/18 kDa) form only. (Right inset) Activated caspase-3 expression is also detected in a subset of AD neurons. (c) Caspase-3 expression in neurons in the FTD brain labeled with PharMingen caspase-3 antibody (65906E). Note that the nuclei appear almost devoid of 65906E labeling (arrows). This antibody demonstrates largely caspase-3 (32 kDa) proenzyme immunoreactivity, with limited immunoreactivity to the active 17-/18-kDa caspase-3 form. Note that these neurons appear more healthy than those labeled in b. Scale bars in a– c, 50 ␮m. (d) Activated caspase-3 expression in astrocytes (arrows) of FTD brain suggests possible apoptotic mechanisms at work in glia as well. Scale bar, 25 ␮m.

14

SU ET AL.

FIG. 5. Relationship of neurons positive for activated caspase-3-labeled to TdT-labeled nuclei in FTD. (a) There is a colocalization between many neurons displaying activated caspase-3 (17/18 kDa) immunoreactivity and TdT labeling (arrows). Arrowhead indicates astrocyte. Scale bar, 50 ␮m. (Inset) Higher magnification shows colocalization of neurons positive for activated caspase-3 to TdT-positive nuclei. (b) A subset of neurons with strong activated caspase-3 expression lack TdT-positive nuclei (arrows). A subset of nuclei with robust TdT labeling lack positive staining for CM1 (arrowheads). Scale bar, 18 ␮m. FIG. 6. Relationship of degenerating astrocytes labeled with GFAP or S-100␤ to TdT-positive nuclei or activated caspase-3 expression. (a) Few GFAP-positive astrocytes (arrow) can be found in white matter of aged control brain. (b) Numerous GFAP-positive reactive astrocytes (arrows) are detected in white matter of FTD brain. (c) Some soma of GFAP-positive astrocytes (blue) have an irregular, enlarged appearance (arrows) and contain robustly labeled TdT-positive nuclei (brown). The processes are rough and irregular and appear disintegrated. (Inset) Higher magnification showing degenerating astrocyte. Note that the soma of the astrocyte (blue) shows enlargement with a compartment lacking GFAP immunoreactivity with a robustly labeled TdT-positive nucleus (brown). (d) Many reactive astrocytes labeled with S-100␤ (blue) are positive for activated caspase-3 (arrows). (Insets) Higher magnification showing that degenerating astrocytes (blue) with vacuolations (arrows) are positive for activated caspase-3 (brown). Scale bars in a– d, 25 ␮m.

in these cells. The pattern of activated caspase-3 expression displayed a laminar distribution profile. Neurons in layers II, III, and V of frontal cortex or anterior pole of temporal lobe showed stronger activated caspase-3 immunoreactivity. Layer I appeared almost devoid of robust staining. Activated caspase-3 immunoreactivity was also detected within a subset of astrocytes (Fig. 4d). Immunoreactivity for activated caspase-3 in astrocytes in FTD brain appear to be stronger than that in control brains examined. 65906E antibody (PharMingen, CA) labeled neurons and glial cells and the number of 65906E-labeled cells was usually greater than that of CM1-labeled cells.

Interestingly, most nuclei appeared devoid of 65906E labeling, though it was faintly present in occasional neurons (Fig. 4c). Colocalization of Activated Caspase-3 Immunoreactivity and DNA Damage within Neurons in FTD Brain CM1 immunostaining combined with TdT end-labeling was conducted to evaluate activated caspase-3 expression in relationship to DNA fragmentation. Examination of neurons double-labeled for CM1 and TdT revealed that a subset of neurons was double labeled in

APOPTOSIS IN FRONTOTEMPORAL DEMENTIA

15

all FTD cases examined in this study (Fig. 5). Another subset of neurons, most of them pyramidal neurons with intensely CM1 staining, did not associate with TdT-labeled nuclei. However, both populations showed an increase in immunoreactivity for activated caspase-3. Furthermore, a subset of cells with intensely TdT-labeled nuclei did not exhibit strong or moderate CM1 staining. Consistent with classic descriptions of FTD neuropathology, few/no neurofibrillary tangles were detected in adjacent sections stained with PHF-1 or AT8 antibody (not shown). Astrocyte Degeneration in FTD Cases GFAP staining revealed numerous astrocytes in gray and white matter of all FTD brains examined, though the number and distribution varied from case to case. Furthermore, the cell bodies of these astrocytes were enlarged, processes were thickened, and the intensity of GFAP immunoreactivity within them was increased, suggesting that these are reactive astrocytes (Figs. 6a and 6b). Furthermore, astrocytosis was most severe in immediately subcortical white matter. Interestingly, some soma of astrocytes showed enlarged, irregular shapes. GFAP immunoreactivity in these cells was nonuniform with apparently some somal compartments lacking GFAP immunoreactivity. This gave the appearance of vacuolation. In the neuropil, there were numerous GFAP-positive punctate structures, which were 2– 6 ␮m in diameter and frequently did not connect to cell bodies. The diameter of proximal processes was rough and irregular, and their GFAP immunoreactivity displayed a similar uneven appearance (Fig. 6c). These changes in morphology, in addition to positive caspase-3 and TdT immunoreactivity, suggest that these are degenerating astrocytes. Such degenerating astrocytes were abundant in the white matter and also observed in the gray matter, particularly prominent in three of the six FTD cases examined. GFAP immunostaining combined with TdT end-labeling was used to evaluate the relationship of TdTpositive nuclei or activated caspase-3 expression to degenerating astrocytes. Examination of astrocytes double-labeled for GFAP and TdT revealed that the majority of abnormally appearing GFAP-positive astrocytes colocalized with robust TdT-labeled nuclei (Fig. 6c). A subset of TdT-labeled nuclei in these cells appeared shrunken and had been dislocated towards the periphery of the cell body (Fig. 6c, inset). Examination of astrocytes double-labeled for S-100␤ and CM1 revealed a prominent colocalization between active or degenerating astrocytes and CM1 immunoreactivity (Fig. 6d).

FIG. 7. Active caspase-3 is visible in FTD brain homogenates as an 18-kDa band. This band can be seen with either the IDUN antibody (top) or the Pharmingen antibody (bottom). Controls do not exhibit the 18-kDa active caspase-3 with either antibody. The Pharmingen antibody also recognizes the larger 32-kDa pro-caspase, also present in one of four controls. Abbreviations: F1-2, FTD brain homogenates; F2*, second homogenized sample of FTD brain; C1-4, control brain homogenates.

Active Caspase-3 in FTD Western blot analysis confirmed the presence of the active subunit (17/18 kDa) of caspase-3 in FTD brain tissue. As shown in Fig. 7, this active subunit was largely absent in control tissue. This was true when using either the IDUN or Pharmingen antibody as the probe. The caspase-3 proenzyme was observed in one control case using the Pharmingen antibody. Nonetheless, this case still lacked the active form, while the FTDs expressed the active form, even with the PharMingen antibody (Fig. 7). This confirms the immunocytochemical findings. Case Report A case report illustrates the nature of pathology in a well-characterized or “classic” FTD patient. A 59-year-old right-handed man made errors in calculations and over 2 years was demoted from an estimator to a handyman and was forced to retire. He wore unmatched shoes or socks, tucked his jacket into his pants, buttoned shirts inside out and put deodorant or shaving lotion in his hair. He waved to pictures on walls. From being easygoing he became stubborn and irritable. A religious awakening led him to spend hours in church; he argued with his wife and friends regarding his new religious ideas. He became emotional, cried when people left him, and refused to attend his father’s funeral. His eating habits changed; he nibbled con-

16

SU ET AL.

stantly, repetitively spat, and ate coffee and banana peels. He was alternatively placid and irritable with a remote, bizarre, and robotlike affect. MMSE was 22. Digit span was 7 and verbal output fluent. He showed a severe anomia, named only two animals in 1 min but had only mild difficulty with reading and memory. His calculations were poor but he copied drawings well. Head CT showed mild atrophy, most marked in the right temporal lobe. SPECT showed decreased rCBF of the anterior temporal regions, greater on the right than the left. He died 4 years after examination. Autopsy showed asymmetric, primarily right-sided temporal ⬎ frontal atrophy with no plaques, neurofibrillary tangles, Pick bodies, or Lewy bodies. Neuronal loss and gliosis were noted in the temporal ⬎ frontal areas and were worse on the right than the left side. Numerous TdT-positive nuclei were found in the anterior pole of temporal lobe and superior frontal gyrus. Furthermore, caspase-3 immunoreactivity in neurons was upregulated and exhibited a prominent colocalization with TdT-labeling nuclei. Degenerating astrocytes were also detected in this FTD case. DISCUSSION

One goal of this study was to examine whether neurons in FTD brains exhibit DNA damage and possibly degenerate via apoptosis. TdT labeling has recently been extensively used for studying DNA damage in development, aging, and pathology; however, to our knowledge, there have been no reports regarding TdT labeling indicating DNA damage in FTD cases. In this study, we observed robust TdT-labeled nuclei in FTD brains, whereas few such nuclei were detected in control brains. Many of these robust TdT-labeled nuclei exhibited the classic, distinct morphological characteristics of apoptosis. Furthermore, quantitative data clearly showed that the number of dark TdT-labeled nuclei in FTD brains was significantly greater than that in control brain. These results are not likely to be due to postmortem DNA degradation or fixation time because the postmortem delay of the cases used in this study is short (average 4.5 h), well under the established threshold for PMD effects (2); it has also been shown that there is no obvious relationship between TdT labeling and storage time of the tissue (28). Thus, the DNA damage observed represents antemortem DNA damage. Another goal of this study was to examine the expression of activated caspase-3 and the relationship of TdT-labeled nuclei to activated caspase-3 positive neurons. In this study, we demonstrated that activated caspase-3 immunoreactivity was detected in neurons and astrocytes in both control and FTD brain. We also demonstrated that activated caspase-3 immunoreactivity was increased within neurons and astrocytes in FTD brain. Furthermore, a subset of neurons was dou-

ble-labeled for CM1 and TdT. An interesting finding of the present study was that activated caspase-3 immunoreactivity was detected within the nuclei of neurons. In contrast, the PharMingen caspase-3 antibody (65906E), which detects pro-caspase, did not typically label neuronal nuclei, suggesting that activated caspase-3 is translocated from cytoplasm to nucleus and may contribute to nuclear DNA fragmentation when cells are engaged in the apoptotic program. As shown in Figs. 4b and 4c, the neurons expressing the pro-caspase still appear relatively healthy, while those in which active caspase is detected do not. Presumably, it is the activated form which is responsible for cleavage of nuclear substrates, such as DNA-dependent protein kinases and poly (ADP-ribose) polymerase (PARP) after its translocation to the nucleus. These results are consistent with recent studies in vitro suggesting that caspase-3 is activated during cell degeneration and that overexpression of caspase-3 accelerates cell death (for review, see 31). It has been reported that in cells triggered to undergo apoptosis, the translocation of cytochrome C from mitochondria to the cytoplasm is one of the earliest events (21, 51). Cytochrome C is involved in the cleavage and activation of caspase-3 (26). Additionally, Bcl-2 can regulate cell death mediated by caspase-3 induction (8, 19); overexpression of Bcl-2 prevents the release of signaling molecules, such as cytochrome C, aborting the activation of caspase-3 (21, 51). We have also demonstrated that a subset of CM1positive pyramidal neurons were negative for TdT; conversely, another subset of cells were CM1 negative but showed robust TdT-labeled nuclei. These data may reflect a sequence of apoptosis or alternative mechanisms. It has been shown that caspase-3 is activated in early apoptosis and caspase activation precedes the appearance of TdT-positive expression (32). Alternatively, these observations could explain that not only activation of caspase-3 but also other molecular mechanism play a role in neuronal cell death in FTD. These observations are supported by a recent study showing that TdT is positive in both activated caspase-positive and -negative cells in the naturally occurring dorsal root ganglia apoptosis of mouse embryos and DNA fragmentation can occur in some dorsal root ganglia neurons independent of the activation of caspase-3 (22). Interestingly, recent studies have shown that DNA fragmentation factor (DFF) and caspase-activated deoxyribonuclease are an important link between caspase activation and DNA fragmentation (27, 39). It is also possible that DNA damage itself can exist at some level in the absence of active caspase-3 expression. The role that these mechanisms may play in the extensive neuronal loss observed in FTD must be further elucidated. In addition to FTD, caspase-3 appears involved in neurodegeneration in AD (Fig. 4b, left inset). This is consistent with other reports of possible

17

APOPTOSIS IN FRONTOTEMPORAL DEMENTIA

active caspase-3 in select neurons in AD brain (15), although our data suggest the affected neurons are not restricted to those with granulovacular degeneration (GVD) (42) since staining was seen in the entorhinal cortex, a region not affected by GVD. As is apparent from Fig. 6, a subset of astrocytes appeared to be degenerating in FTD cases in this study. Their cell bodies are enlarged and irregular in shape with compartments lacking immunoreactivity for GFAP (vacuolation) and their processes are disintegrated. Furthermore, the majority of these degenerating astrocytes colocalized with robust TdT-labeled nuclei and activated caspase-3. Such morphological alterations resemble “clasmotodendritic” astrocytes documented in brains with cerebral edema (7). This result is unlikely to be due to postmortem artifact because three of the cases with prominently degenerating astrocytes had short postmortem delay, whereas in the remaining three cases with similar PMD, astrocyte degeneration was less abundant. This result is also supported by an FTD case with long PMD (13 h) in which less abundant astrocyte degeneration was observed (unpublished observations). Similar astrocytic changes have occasionally been observed in brains with cerebrovascular disease and Alzheimer’s disease (49). Caspase-3 has been implicated in the cleavage of nuclear proteins including PARP and DNA-PK, which are involved in DNA repair and protection of DNA (for review, see 40). Thus, it is possible that activated caspase-3 may play an important role in astrocyte degeneration. The role of degenerated astrocytes in the pathogenesis of FTD in not clear. One possibility is that these cells may be impaired in their ability to take up and inactivate excess excitatory amino acids, such as glutamate, causing excitotoxic stress, neuronal cell death, oligodendrocyte death, and further gliosis. They may also fail to regulate extracellular K ⫹ and Ca ⫹ concentrations and pH. Additionally, degenerating astrocytes may fail to produce lactate, a neuronal and oligodendroglial substrate. Degeneration of white matter astrocytes may lead to a depletion of oligodendroglial factors. It is also possible that degenerating astrocytes might compromise the local blood– brain barrier (for review, see 35). Last, the degeneration of astrocytes could have effects on synaptic efficiency, leading to late-stage neuronal degeneration similar to that observed in development (during which synaptic efficiency is often compromised) (35). In conclusion, our findings suggest that DNA damage is present in FTD brain. Activated caspase-3 is elevated in FTD cases and activated caspase-3 expression and TdT labeling are colocalized in a subset of cells of FTD brain. Degenerating astrocytes are activated caspase-3 positive and colocalized with robust TdT-labeled nuclei. These data suggest that DNA damage and activated caspase-3 may have an important

role in neuronal cell death and astrocyte degeneration in FTD. ACKNOWLEDGMENTS The authors thank Dr. A. Anderson and Ms. J. Martin for helpful comments on this manuscript.

REFERENCES 1.

2.

3.

4.

5. 6. 7.

8.

9.

10.

11.

12.

13.

14.

15.

Alnemri, E. S., D. J. Livingston, D. W. Nicholson, G. Salvesen, N. A. Thornberry, W. W. Wong, and J. Yuan. 1996. Human Ice/Ced-3 protease nomenclature. Cell 87: 171. Anderson, A. J., J. H. Su, and C. W. Cotman. 1996. DNA damage and apoptosis in Alzheimer’s disease: Colocalization with c-Jun immunoreactivity, relationship to brain area, and effect of postmortem delay. J. Neurosci. 16: 1710 –1719. Biernat, J., E.-M. Mandelkow, C. Schroter, B. LichtenbergKraag, B. Steiner, B. Berling, H. Meyer, M. Merchen, A. Vandermeeren, M. Goedert, and E. Mandelkow. 1992. The switch of tau protein to an Alzheimer-like state includes the phosphorylation of two serine-proline motif upstream of the microtubule binding region. EMBO J. 11: 1593–1597. Bramblett, G. T., M. Goedert, R. Jakes, S. E. Merrick, J. Q. Trojanowski, and V. M.-Y. Lee. 1993. Abnormal tau phosphorylation at Ser 396 in Alzheimer’s disease recapitulates development and contributes to reduced microtubule binding. Neuron 10: 1089 –1099. Brun, A. 1987. Frontal lobe degeneration of non-Alzheimer type. I. Neuropathology. Arch. Gerontol. Geriat. 6: 193–208. Brun, A. 1993. Frontal lobe degeneration of non-Alzheimer type revisited. Dementia 4: 126 –131. Cajal, S. R. Y. 1913. Contribucio´n al conocimiento de la neuroglia del cerebro humano. Trab. Lab. Inv. Biol. Univ. Madrid 11: 255–315. Chinnaiyan, A. M., K. O’Rourke, B. R. Lane, and V. M. Dixit. 1997. Interaction of CED-4 with CED-3 and CED-9: A molecular framework for cell death. Science 275: 1122–1126. Clinton, J., D. M. A. Mann, and G. W. Roberts. 1993. Frontal lobe dementia is not a variant of prion disease. Neurosci. Lett. 164: 1– 4. Darmon, A. J., T. J. Ley, D. W. Nicholson, and R. C. Bleackley. 1996. Cleavage of cpp32 by granzyme b represents a critical role for granzyme B in the induction of target cell DNA fragmentation. J. Biol. Chem. 271: 21709 –21712. Datta, R., H. Kojima, D. Banach, N. J. Bump, R. V. Talanian, E. S. Alnemri, R. R. Weichselbaum, W. W. Wong, and D. W. Kufe. 1997. Activation of a CrmA-insensitive, p35-sensitive pathway in ionizing radiation-induced apoptosis. J. Biol. Chem. 272: 1965–1969. Enari, M., R. V. Talanian, W. W. Wong, and S. Nagata. 1996. Sequential activation of ICE-like and CPP32-like proteases during Fas-mediated apoptosis. Nature 380: 723–726. Fernandes-Alnemri, T., A. Takahashi, R. Armstrong, J. Krebs, L. Fritz, K. J. Tomaselli, L. Wang, Z. Yu, C. M. Croce, G. Salveson, W. C. Earnshaw, G. Litwack, and E. S. Alnemri. 1995. Mch3, a novel human apoptotic cysteine protease highly related to CPP32. Cancer Res. 55: 6045– 6052. Gavrieli, Y., Y. Sherman, and S. A. Ben-Saason. 1992. Identification of programmed cell death In situ via specific labeling of nuclear DNA fragmentation. J. Cell Biol. 119: 493–501. Gervais, F. G., D. Xu, G. S. Robertson, J. P. Vaillancourt, et al. 1999. Involvement of Caspases in proteolytic cleavage of Alz-

18

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

SU ET AL. heimer’s amyloid-␤ precursor protein and amyloidogenic A␤ peptide formation. Cell 97: 395– 406. Gold, R., M. Schmied, G. Rothe, H. Zischier, H. Breitschope, H. Wekerle, and H. Lassmann. 1993. Detection of DNA fragmentation in apoptosis: Application of in situ nick translation to cell culture systems and tissue sections. J. Histochem. Cytochem. 41: 1023–1030. Gustafson, L. 1987. Frontal lobe degeneration of non-Alzheimer type. II. Clinical picture and differential diagnosis. Arch. Gerontol. Geriat. 6: 209 –223. Hauw, J. J., C. Duyckaerts, D. Seilhean, S. Camilleri, V. Sazdovitch, and G. Rancurel. 1996. The neuropathologic diagnostic criteria of frontal lobe dementia revisited—A study of ten consecutive cases. J. Neural Trans. 47(Suppl.): 47–59. Irmler, M., K. Hofmann, D. Vaux, and J. Tschopp. 1997. Direct physical interaction between the Caenorhabditis elegans ‘death proteins’ CED-3 and CED-4. FEBS Lett. 406: 189 –190. Jacobson, M. D., M. Weil, and M. C. Raff. 1996. Role of Ced3/ Ice-family proteases in staurosporine-induced programmed cell death. J. Cell Biol. 133: 1041–1051. Kluck, R. M., E. BossyWetzel, D. R. Green, and D. D. Newmeyer. 1997. The release of cytochrome c from mitochondria: A primary site for Bcl-2 regulation of apoptosis. Science 275: 1132–1136. Kouroku, Y., K. Urase, E. Fujita, K. Isahara, Y. Ohsawa, Y. Uchiyama, M. Y. Momoi, and T. Momoi. 1998. Detection of activated Caspase-3 by a cleavage site-directed antiserum during naturally occurring DRG neurons apoptosis. Biochem. Biophys. Res. Commun. 247: 780 –784. Krajewska, M., H. G. Wang, S. Krajewski, J. M. Zapata, A. Shabaik, R. Gascoyne, and J. Reed. 1997. Immunohistochemical analysis of in vivo patterns of expression of CPP32 (Caspase-3), a cell death protease. Cancer Res. 57: 1605– 1613. Kumar, S., and M. F. Lavin. 1996. The Ice family of cysteine proteases as effectors of cell death. Cell Death Differ. 3: 255– 267. Lassmann, H., C. Bancher, H. Breitschopf, J. Wegiel, M. Bobinski, K. Jellinger, and H. M. Wisniewski. 1995. Cell death in Alzheimer’s disease evaluated by DNA fragmentation in situ. Acta Neuropathol. 89: 35– 41. Liu, X. S., C. N. Kim, J. Yang, R. Jemmerson, and JHS. Others. 1996. Induction of apoptotic program in cell-free extracts—Requirement for dATP and cytochrome C. Cell 86: 147–157. Liu, X. S., H. Zou, C. Slaughter, and X. D. Wang. 1997. DFF, a heterodimeric protein that functions downstream of caspase-3 to trigger DNA fragmentation during apoptosis. Cell 89: 175– 184. Lucassen, P. J., W. C. J. Chung, J. P. Vermeulen, M. V. L. Campagne, J. H. V. Dierendonck, and D. F. Swaab. 1995. Microwave-enhanced in situ end-labeling of fragmented DNA: Parametric studies in relation to postmortem delay and fixation of rat and human brain. J. Histochem. Cytochem. 43: 1163– 1171. Mann, D. M. A., P. W. South, J. S. Snowden, and D. Neary. 1993. Dementia of frontal lobe type—Neuropathology and immunohistochemistry. J. Neurol. Neurosurg. Psychiat. 56: 605– 614. Miller, B. L., J. L. Cummings, J. Villanueva-Meyer, K. Boone, C. M. Mehringer, I. M. Lesser, and I. Mena. 1991. Frontal lobe degeneration: Clinical, neuropsychological, and SPECT characteristics. Neurology 41: 1374 –1382.

31. 32.

33.

34.

35.

36.

37.

38. 39.

40. 41.

42.

43.

44.

45.

46.

47.

48.

49.

Nagata, S. 1997. Apoptosis by death factor. Cell 88: 355–365. Namura, S., J. Zhu, K. Fink, M. Endres, A. Srinivasan, K. J. Tomaselli, J. Yuan, and M. A. Moskowitz. 1998. Activation and cleavage of caspase-3 in apoptosis induced by experimental cerebral ischemia. J. Neurosci. 18: 3659 –3668. Neary, D., J. S. Snowden, B. Northen, and P. Goulding. 1988. Dementia of frontal lobe type. J. Neurol. Neurosurg. Psychiat. 51: 353–361. Nicholson, D. W., A. Ali, N. A. Thornberry, J. P. Vaillancourt, C. K. Ding, M. Gallan, Y. Gareau, P. R. Griffin, M. Labelle, Y. A. Lazebnik, N. A. Munday, S. M. Raja, M. E. Smulson, T.-T. Yamin, V. L. Yu, and D. K. Miller. 1995. Identification and inhibition of the ICE/CED-3 protease necessary for mammalian apoptosis. Nature 376: 37– 43. McCall, M. A., R. G. Gregg, R. R. Behringer, M. Brenner, C. L. Delaney, E. J. Galbreath, C. L. Zhang, R. A. Pearce, S. Y. Chiu, and A. Messing. 1996. Targeted deletion in astrocyte intermediate filament (Gfap) alters neuronal physiology. Proc. Nat. Acad. Sci. USA 93: 6361– 6366. Nieto-Sampedro, M., and F. Mora. 1994. Active microglia, sick astroglia and Alzheimer type dementias. NeuroRep. 5: 375–380. Norton, W. T., D. A. Aquino, I. Hozumi, F. C. Chiu, and C. F. Brosnan. 1992. Quantitative aspects of reactive gliosis—A review. Neurochem. Res. 17: 877– 885. Patel, T., G. J. Gores, and S. H. Kaufmann. 1996. The role of proteases during apoptosis. FASEB J. 10: 587–597. Sakahira, H., M. Enari, and S. Nagata. 1998. Cleavage of CAD inhibitor in CAD activation and DNA degradation during apoptosis. Nature 391: 96 –99. Satoh, M. S., and T. Lindahl. 1994. Enzymatic repair of oxidative DNA damage. Cancer Res. 54: 1899s–1901s. Srinivasan, A., K. A. Roth, R. O. Sayers, K. S. Shindler, A. M. Wong, L. C. Fritez, and K. J. Tomaselli. 1998. In situ immunodetection of activated caspase-3 in apoptotic neurons in the developing nervous system. Cell Death Differ. 5: 1004 – 1016. Stadelmann, C., T. L. Deckwerth, A. Srinivasan, C. Bancher, B. Wolfgang, K. Jellinger, and H. Lassmann. 1999. Activation of caspase-3 in single neurons and autophagic granules of granulovacuolar degeneration in Alzheimer’s disease: Evidence for apoptotic cell death. AJP 155: 1459 –1466. Steward, O., E. R. Torre, L. L. Phillips, and P. A. Trimmer. 1990. The process of reinnervation in the dentate gyrus of adult rats: time course of increases in mRNA for glial fibrillary acidic protein. J. Neurosci. 10: 2373–2384. Su, J. H., A. J. Anderson, B. J. Cummings, and C. W. Cotman. 1994. Immunohistochemical evidence for apoptosis in Alzheimer’s disease. NeuroRep. 5: 2529 –2533. Su, J. H., B. J. Cummings, and C. W. Cotman. 1993. Identification and distribution of axonal dystrophic neurites in Alzheimer’s disease. Brain Res. 625: 228 –237. Su, J. H., B. J. Cummings, and C. W. Cotman. 1994. Subpopulations of dystrophic neurites in Alzheimer’s brain with distinct immunocytochemical and argentophilic characteristics. Brain Res. 637: 37– 44. Su, J. H., G. Deng, and C. W. Cotman. 1997. Bax protein expression is increased in Alzheimer’s brain: Correlation with DNA damage, Bcl-2 expression, and brain pathology. J. Neuropathol. Exp. Neurol. 56: 86 –93. Su, J. H., T. Satou, A. J. Anderson, and C. W. Cotman. 1996. Upregulation of Bcl-2 is associated with neuronal DNA damage in Alzheimer’s disease. NeuroRep. 7: 437– 440. Tomimoto, H., I. Akiguchi, H. Wakita, T. Suenaga, S. Nakamura, and J. Kimura. 1997. Regressive changes of as-

APOPTOSIS IN FRONTOTEMPORAL DEMENTIA troglia in white matter lesions in cerebrovascular disease and Alzheimer’s disease patients. Acta Neuropathol. 94: 146 –152. 50. Wang, L., M. Miura, L. Bergerson, H. Zhu, and J. Yuan. 1994. Ich-1, an ice/ced-3-related gene, encodes both positive and negative regulators of programmed cell death. Cell 78: 739 – 750.

51.

19

Yang, J., X. S. Liu, K. Bhalla, C. N. Kim, and JHS. Others. 1997. Prevention of apoptosis by Bcl-2: Release of cytochrome C from mitochondria blocked. Science 275: 1129 –1132. 52. Yuan, J. Y., S. Shaham, S. Ledoux, M. H. Ellis, and R. H. Horvitz. 1993. The c-elegans cell death gene ced-3 encodes a protein similar to mammalian interleukin-1-beta-converting enzyme. Cell 75: 641– 652.