Abstract
Previous reports in various cells and species have shown that apoptotic cells are specifically and strongly labeled by certain c-Jun/N-terminal antibodies, such as c-Jun/sc45. This kind of immunoreactivity is confined to the cytoplasm. It is not due to c-Jun but appears to be related to c-Jun-like neoepitopes generated during apoptosis. This study was planned to gain further information about c-Jun-like immunostaining during apoptosis and to evaluate these antibodies as possible tools for characterizing cell death. Most of the experiments were performed in chick embryo spinal cord. When the apoptotic c-Jun-like immunoreactivity and caspase-3 immunostaining patterns were compared, we found that both antibodies immunostained the same dying cells in a similar pattern. In contrast to TUNEL staining, which reveals a positive reaction in both apoptotic and necrotic dying cells, active caspase-3 and c-Jun/sc45 antibodies are more selective because they stained only apoptotic cells. When cytosolic extracts from normal tissues were digested in vitro with caspase-3, c-Jun/sc45 immunoreactivity was strongly induced in several proteins, as demonstrated by Western blotting. Similar results were found when normal tissue sections were treated with caspase-3. Our results show that c-Jun/sc45 antibodies react with neoepitopes generated from cell proteins cleaved by activated caspases during apoptosis. We conclude that c-Jun/sc45 antibodies may be useful for detecting apoptosis. They can even be used in archival paraffin-embedded tissue samples.
Keywords
A key feature of apoptosis is a sequential activation of proteolytic system machinery. Most of the morphological characteristics of apoptotic cell death are a consequence of caspase activity (Nicholson and Thomberry 1997; Earnshaw et al. 1999). The morphological changes associated with apoptosis usually include nuclear retraction coupled with severe chromatin condensation and cell shrinkage. The structure of organelles is preserved until cell fragments or apoptotic bodies are engulfed by macrophages (Kerr et al. 1972; Savill et al. 1993). The main biochemical counterparts of apoptosis include DNA fragmentation at the nucleosome linkage and the proteolytic cleavage of several cellular proteins by caspases. These characteristics can be monitored and yield some of the biochemical hallmarks of apoptosis, such as DNA laddering and the detection of cleavage fragments resulting from caspase activity, such as fodrin, PARP, U1 small nuclear ribonucleoprotein particle (U1–170 kD) and lamins (Wyllie et al. 1980; Kaufmann 1989; Kaufmann et al. 1993; Lazebnik et al 1994; Greidinger et al. 1996; Cohen 1997; Earnshaw et al. 1999).
Histochemical techniques developed to visualize fragmented DNA are also commonly used for identifying apoptotic cells in situ but cannot be considered specific markers for apoptosis. The most commonly used method for this purpose is based on nick end-labeling by incubation with dUTP and deoxyribonucleotyl transferase (TdT) (TUNEL). It is now well recognized that the TUNEL assay often yields false-positives as a result of its ability to label non-apoptotic dying cells (Charriaut-Marlangue and Ben-Ari 1995; Portera-Caillau et al. 1997). Therefore, to validate this method for detecting apoptosis, TUNEL testing should be accompanied by other criteria, such as systematic ultrastructural evaluation (Olney and Ishimaru 1999). More recently, some immunohistochemical methods for detecting activated caspases have proved more valid for detecting apoptotic dying cells in tissues. Activation of caspases has been identified either by specific antibodies that recognize their cleaved form or by antibodies that react with neoantigenic sites emerging on target proteins as a consequence of the caspase activity. In this context, the following tests have been applied: antibodies against the C-terminal of the 32-kD actin fragment produced by ICE-like activity (Yang et al. 1998; Rossiter et al. 2000); antibodies against synthetic peptides mimicking the neo-N-terminal antigen SBDP120 generated by caspase cleavage of spectrin (Nath et al. 2000); and the antibody Ab127 against the peptide CKGDEVD situated immediately upstream of the caspase cleavage locus within PARP (Siman et al. 1999). Furthermore, some polyclonal anti-c-Jun antibodies are very efficient at detecting apoptotic cells by crossreacting with apoptosis-specific antigens other than those from c-Jun. This property is shared by antibodies in which the immunizing peptide TPTPQFLCPKNVTD corresponds to v-Jun 73–87 or mouse c-Jun 91–105 amino acids (Ayala et al. 1999; Ferrer et al. 2000; Terwell and van de Berg 2000). This property has been evidenced using commercially available antibodies against c-Jun, such as c-Jun/sc45 from Santa Cruz Technology (Santa Cruz, CA) or Ab-2 from Oncogene Science (Cambridge, MA). Both antibodies produce strong immunolabeling of apoptotic cell cytoplasm in a broad spectrum of cell types of several species, e.g., in rat brain during developmental and experimentally induced apoptosis (Ferrer et al. 1996a,1997a,b,c; Guégan et al. 1997; Pozas et al. 1997), in human brain tumors (Ferrer et al. 1996b) in neuronal death associated with Alzheimer disease (Anderson et al. 1996), in developing sympathetic neurons (Messina et al. 1996), in apoptotic facial motor neurons in neonatal rats induced by axotomy (Garrah et al. 1998; Casanovas et al. 2001), and in different types of chick embryo cells undergoing normal or experimentally induced apoptosis (Ayala et al. 1999).
It has been suggested that these antibodies recognize distinct proteins of c-Jun (Pozas et al. 1997). Grand et al. (1995) found similar cytoplasmic immunoreactivity in different types of apoptotic cells and reported that the antibody recognized a novel 45-kD protein which they called “apoptosis specific protein” (ASP). However, more recently we have shown that c-Jun/sc45 antibodies recognize a neoantigenic site generated in apoptotic cells as a consequence of caspase proteolytic activity (Casas et al. 2001).
In the present study we examined the localization of the cytoplasmic immunoreactivity of c-Jun/sc45 in apoptotic cells to gain further information about c-Jun-like immunostaining in apoptosis. This was then compared with immunocytochemical results obtained with an antibody against the 17-kD fragment of cleaved caspase-3, which indicates the activation of caspase-3. We conclude that, in apoptotic cells, there is a co-localization of immunostaining by the two antibodies, with the c-Jun/sc45 antibody being more sensitive than the antibody against caspase-3 in recognizing apoptotic cells. We also show that the strong c-Jun/sc45 immunoreactivity characteristic of apoptosis may be reconstructed in situ in normal cells as a consequence of caspase-3 proteolytic attack.
Materials and Methods
Reagents
N-Methyl-d-aspartic acid (NMDA), β-bungarotoxin (β-Bgtx), normal goat serum (NGS), and protease inhibitors were purchased from Sigma (St Louis, MO). The caspase inhibitor Z-VAD-FMK was from Enzyme Systems Products (Livermore, CA). Active recombinant human caspase-3 (CPP32) was from PharMingen/Becton Dickinson (Le Pont de Claix, France). The c-Jun/sc45 polyclonal antibody was from Santa Cruz Biotechnology (Santa Cruz, CA). c-Jun/mAB monoclonal antibody was from Transduction Laboratories (Lexington, KY). The antibody against cleaved caspase-3 (17 kD) was provided by New England BioLabs (Beverly, MA). Secondary antibodies, goat anti-rabbit or goat anti-mouse and immunoglobulins, labeled with either Alexa Fluor 488 or 546, were purchased from Molecular Probes (Eugene, OR), as were 4′,6-diamidino-2-phenylidole and dihydrochloride (DAPI). Biotinylated goat anti-rabbit IgG and ABC complex were from Vector (Burlingame, CA). PVDF membranes were from Millipore (Bedford, MA). Peroxidase-conjugated secondary antibodies and the ECL detection system were manufactured by Amersham (Little Chalfont, UK). Durcupan ACM was from Fluka (Buchs, Switzerland). DABCO (1–4-diazabicyclot(2–2–2) octane) was from Aldrich (Milwaukee, WI).
Tissue Processing for Light Microscopic Immunocytochemistry
Experiments were performed using tissues from chick embryos between embryonic days (E) 7.5 and 10. Samples for light microscopic immunocytochemistry were fixed by immersion in cold 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4 (PB), for 12 hr at 4C. Some samples were later cryoprotected in 30% (w/v) sucrose in the same buffer for 24 hr before cryostat sectioning, while other samples were processed for paraffin embedding. Sections were collected on SuperFrost Plus slides (Menzel-Glaser; Freiburg, Germany), air dried at room temperature (RT), and then processed for immunocytochemistry.
Light Microscopic Immunocytochemistry
Sections were first permeabilized for 45 min in PBS containing 0.1% Triton X-100. Unspecific staining was blocked by incubating sections with 10% NGS in PBS for 1 hr. The incubation with primary antibodies c-Jun/sc45 (diluted 1:250), caspase-3 (diluted 1:70), or c-Jun/MAb (diluted 1:200) was performed overnight at 4C. After washing in PBS, sections were incubated with the appropriate secondary fluorescent-labeled Alexa-Fluor antibody (1:250) containing DAPI (3 mM), for 1 hr at RT. After washing, sections were coverslipped with an anti-fading mounting medium (0.1 M Tris-HCl buffer, pH 8.5, 20% glycerol, 10% Mowiol, and 0.1% DABCO) and images were captured by a cooled CCD camera (Life Science Resources; Cambridge, UK) coupled to a Nikon Eclipse 600 fluorescence microscope fitted with different selective filters. Immunocytochemical controls included omission of the primary antibody and preincubation with immunogen peptides, which resulted in complete abolition of immunoreactivity.
A double immunostaining method was performed for simultaneous localization of both the active caspase-3 and the antigens revealed by c-Jun/sc45 antibodies. Since both antibodies were created in rabbit, we applied a technique of antibody elution described by Tramu et al. (1978). Briefly, samples were first processed for immunofluorescence using primary antibody anti-cleaved caspase-3 and Alexa-Fluor 488 goat anti-rabbit as a secondary antibody. After immunostaining, samples were mounted with glycerol (70%) in PBS and imaged. The coverslips were then detached by immersion in PBS and the immune complexes were removed by oxidation with a mixture of KMnO4 and H2SO4. Samples were subsequently processed to reveal the second-step immunofluorescence reaction resulting from c-Jun/sc45 antibody using Alexa-546 goat anti-rabbit as secondary antibody. The samples were finally imaged at selected areas matching those taken after the first immunofluorescence round.
Elution was carried out by immersing the slides in a mixture of 2.5% KMnO4 and 5% H2SO4 in distilled water for 10 min after 30 sec of destaining in 0.5% Na2S2O5. Sections were sequentially washed for 15 min in running tapwater, for 10 min in distilled water, and for 10 min in PBS before processing with the second immunofluorescence round. The efficiency of elution was checked before the second round of imunostaining by observing the complete elimination of the fluorescent signal.
Some immunofluorescence assays were carried out in 8-μm sections of paraffin-embedded samples. Formalin-fixed, paraffin-embedded sections of human tonsil were immunostained for c-Jun/sc45 using the avidin-biotin-peroxidase detection system (Elite ABC; Vector Laboratories). Sections were then counterstained with hematoxylin.
Electron Microscopic Immunocytochemistry
For pre-embedding, specimens were fixed by immersion in a mixture of cold 4% paraformaldehyde and in 0.1% or 0.25% glutaraldehyde in 0.1 M PB, pH 7.4, for 2–4 hr. Transverse 50–100-μm-thick sections from spinal cords were obtained using a vibratome. To block endogenous peroxidase activity, sections were incubated with 1% H2O2 in PBS for 45 min, washed in PBS, and then incubated with 10% NGS to reduce unspecific binding. Sections were then incubated overnight at 4C with c-Jun/sc45 antibody (diluted 1:200), washed in several changes of PBS, and incubated with biotinylated goat anti-rabbit IgG (diluted 1:100) for 60 min. After washing in several changes of PBS, sections were then incubated for 60 min with ABC complex, washed in PBS, and developed according to the DAB procedure. Samples were postfixed in PB 1% OsO4 for 2 hr, dehydrated in acetone, and flat-embedded in Durcupan ACM (Fluka). Ultrathin sections of selected areas containing labelled cell bodies were examined in a Zeiss EM910 electron microscope either with or without uranyl acetate and lead citrate counterstaining. Pre-embedding procedure was also performed using 1-nm gold-labeled goat anti-rabbit IgG (Auro-Probe One GAR; Amersham) as a secondary antibody with ultimate silver enhancement (IntenSE M silver enhancement kit; Amersham) according to the manufacturer's instructions. Labeled samples were flat-embedded and sectioned as described above.
Pharmacological Treatment of Embryos
Drug applications were preformed in fertilized chicken eggs from COPAGA (Lleida; Catalonia, Spain) and incubated in the laboratory. A group of E8 embryos was treated with a single dose of 1 mg NMDA (Sigma). This neurotoxin was dissolved in saline and dropped directly onto the chorioallantoic membrane (CAM) through a window in the shell. Embryos were sacrificed 12 hr later (E8.5).
In another set of E7 embryos, a 1-μl dose of β-Bgtx (100 ng) dissolved in saline solution was injected into the right leg.
After treatment, the shell window was sealed with adhesive tape and the eggs were returned to the incubator until the time of sampling.
Western and Dot-blot Assays
Fresh spinal cords from E10 embryos were homogenized using a Polytron in 2 volumes of ice-cold 50 mM Tris-HCI buffer (pH 7.4) containing 1 mM EDTA, 5 mM mercaptoethanol, 10 μg/ml pepstatin, 10 μg/ml aprotinin, 10 μg/ml leupeptin, and 100 μg/ml phenylmethylsulfonyl fluoride. After ultracentrifugation at 150,000 X g for 1 hr, the supernatants corresponding to the cytosolic fraction were transferred into fresh tubes, snap-frozen in liquid nitrogen, and stored at −80C until further analysis. Before use, samples were boiled for 5 min and 20 μl of protein was loaded in SDS-polyacrylamide gels for Western blots or dotted directly onto an Immobilon PVDF transfer membrane (Millipore). For Western blotting, proteins were first separated in 7.5% SDS-PAGE gels using the Bio-Rad mini Protean II electrophoresis system and then blotted onto Immobilon PVDF transfer membranes using a semi-dry system (Pharmacia; Uppsala, Sweden). The membranes were blocked with 5% fat-free dried milk and incubated with sc45 antiserum (diluted 1:1000). Immuoreactive bands or dot-blots were detected using a horseradish peroxidase-conjugated anti-rabbit secondary antibody with a bioluminescent ECL detection system (Amersham).
In Vitro Assay for Induction of Caspase-3-mediated Proteolysis
Extracts of cytosolic fraction from normal E10 spinal cord were treated with active caspase-3 to examine whether the appearance of c-Jun7/sc45 antibody-reacting antigenic epitopes in the apoptotic neuron cytoplasm emerged as a result of caspase-3 activity. Histological sections from normal E10 spinal cord were also treated with active caspase-3. Extracts were treated according a modification of the method described by Stennicke and Salvensen (1997). Digestion was performed by incubation with active caspase-3 in a medium containing 50 mM Tris-HCl buffer (pH 7.4), 25 mM Hepes, pH 7.4, 1 mM EDTA, 5 mM DTT, 0.1% CHAPS, 10% sucrose, and a cocktail of protease inhibitors. The enzymatic reaction was carried out at 37C for 1 hr in a 20-μl volume containing 60 μg of cytoplasmic protein extracts and 3 ng/μl of active caspase-3. The specificity of caspase-3 proteolysis was tested by adding its inhibitor Z-VAD (150 nM). Samples were later used for Western and dot-blot analysis.
For capase-3 in situ proteolysis assay on cryostat sections, active caspase-3 was dissolved in a buffer containing 25 mM Hepes, pH 7.4, 1 mM EDTA, 5 mM DTT, 0.1% CHAPS, and 10% sucrose (Buffer A). Ten-μm-thick sections that were collected on SuperFrost Plus slides and air-dried overnight, were washed three times for 10 min at 37C with Buffer A and then incubated with the same buffer, which contained 2 ng/μl active caspase-3, at 37C for 4 hr. Samples were subsequently washed in PBS and processed for immunofluorescence. The specificity of caspase proteolytic activity was tested by adding Z-VAD (150 nM final).
Results
c-Jun/sc45 and Caspase-3 Antibodies Displayed an Overlapping Pattern of Immunoreactivity in Dying Cells
Assays were made on developing spinal cord motor neurons (MNs). The natural programmed cell death of this neuron population has been extensively characterized in chick embryo. In this system, the MN number is adjusted employing an apoptotic programmed cell death (PCD) process that follows a well-defined temporal and spatial pattern. For lumbar segments this occurs from E6 to E10 and affects about 50% of initially differentiated MNs (Hamburger 1975; Chu-Wang and Oppenheim 1978; Oppenheim 1991). In chick LMC, MNs are large cells, clustered at the ventral horn, which are easily identifiable in transverse sections of spinal cord. At the peak PCD period (on E7–8), we found at least three or four dying cells per section in each ventral horn (not shown). Histochemical assays were therefore performed at E7.5. All sections examined displayed large cells scattered among LMC revealing strong cytoplasmic immunoreactivity to c-Jun/sc45. Similar cellular morphology was found using antibodies against active caspase-3. After c-Jun/sc45 immunostaining cells showed a typical neuronal morphology, with a Golgi-like pattern that extensively delineated their somata and neurites, although nuclei were excluded from immunolabeling (Figure 1A). Active caspase-3-immunoreactive MNs showed a strong cytoplasmic labeling that was only slightly extended to proximal neurites (Figure 1B). When c-Jun/sc45 and active caspase-3 immunoreactivity were combined with the TUNEL reaction and DAPI nuclear staining in a multifluorescent assay, it was clear that most cells immunostained with these antibodies possessed pyknotic nuclei (Figures 1A′ and 1B′) and also exhibited a nuclear TUNEL-positive reaction indicative of DNA fragmentation (Figures 1A′ and 1B′). Pyknotic cells were readily identifiable after DAPI staining because their nuclei displayed brightly fluorescent chromatin clumps.
The final common pathway in apoptotic cell death involves caspase-mediated proteolysis of several substrates whose cleavage products accumulate in dying cells. Some, like those from PARP or spectrin, can be specifically detected by antibodies and have often been used as markers for apoptosis. More recently, antibodies against the active form of caspase-3 that work on tissue sections have made it possible to carry out specific in situ labeling of apoptotic cells.
Cryostat sections of E7.5 chick embryo spinal cord were double-immunostained using both rabbit polyclonal antibodies to demonstrate the relationship between active caspase-3 labeling and c-Ju/sc45 antigens in dying cells (see Materials and Methods). This method made it possible to demonstrate that dying cells displaying cytoplasmic immunoreactivity for active caspase-3 were also immunoreactive to c-Jun/sc45 (Figures 1C1 and 1C2 ). Although elution by oxidation produced some tissue waste, the microscopic structure was sufficiently preserved to demonstrate the co-localization of both antibodies in dying cells.
Ultrastructural immunolabeling of c-Jun/sc45 antigens in apoptotic cells was achieved by applying preembedding procedures involving peroxidase or 1-nm colloidal gold particles (Figures 1D and 1E). Accumulated immunoreactive deposits generated by peroxidase were mainly observed in the cytoplasm of dying cells (Figure 1D). Similar results were obtained using immunogold followed by a silver enhancement preembedding procedure (Figure 1E). Neurons exhibiting immunolabeling for c-Jun/sc45 showed a nuclear morphology of chromatin condensation characteristic of apoptotic cells.
c-Jun/sc45 Immunolabeling, but not TUNEL, Can Distinguish Between Apoptotic and Necrotic Cell Death
Apoptosis in chick embryo ventral horn can be dramatically stimulated by injecting β-Bgtx. At 12–24 hr after β-Bgtx administration there was a dramatic increase in the number of apoptotic cells showing positive c-Jun/sc45 immunoreactivity, pyknotic nuclei, and TUNEL reaction. These neurons also displayed an ultrastructural apoptotic morphology (Ayala et al. 1999). Similar results were obtained through active caspase-3 immunocytochemistry (data not shown). Double staining again revealed that active caspase-3 co-localized with c-Jun/sc45 immunoreactivity.
c-Jun/sc45 and active caspase-3 double fluorescent immunolabeling demonstrates that signals from both antibodies co-localize in apoptotic motor neurons. c-Jun/sc45 and active caspase-3 immunofluorescence were visualized using Alexa-Fluor 488 or Alexa-Fluor 546. Immunostaining was combined with the TUNEL technique to detect DNA fragmentation and DAPI was used to visualize nuclear morphology. Some motor neurons showed strong cytoplasmic immunoreactivity to c-Jun/sc45 with a Golgi-like immunostaining pattern (
TUNEL, c-Jun/sc45 and active caspase-3 immunolabeling of chick embryo spinal cord after an excitotoxic lesion induced by NMDA. A large number of TUNEL-positive cells appeared scattered throughout the spinal cord gray matter 2 hr after stimulation with NMDA (
A single application of NMDA in chick embryos older than E8 induced a massive excitotoxic lesion in spinal cord (Calderó et al. 1997). Excitotoxic MNs showed an ultrastructural morphology indicative of necrotic cell death, with extensive cytoplasmic vacuolization and disruption of cytoplasmic organization that included fragmentation of the nuclear envelope (Figures 2E and 2F). When spinal cord tissue from NMDA-treated embryos was processed for TUNEL assay, it was found that large areas of spinal cord were full of TUNEL-positive cells (Figures 2A and 2C). These were especially numerous in the medial and dorsal regions of the spinal cord (including white matter), whereas their density was lower in the ventral horn. Despite this, all these cells were TUNEL-positive and pyknotic, but only a very small proportion of them displayed c-Jun/sc45 and active caspase-3 immunoreactivity (Figures 2B and 2D). Such cells were presumably apoptotic neurons that normally exist in LMCs at these ages regardless of NMDA stimulation (Figures 2A′–2D′). These cells were usually smaller than motor neurons and their morphology might correspond to small interneurons or non-neuronal dying cells rather than to dying motor neurons (Figures 2B′ and 2C′).
Neoepitopes Emerging as a Consequence of Caspase-mediated Proteolysis Were Responsible for c-Jun/sc45 Immunoreactivity in Apoptosis: In Situ and In Vitro Generation on Healthy Cells Through Caspase-3 Digestion
Caspase-3 is a cysteine protease that plays a central role in apoptosis (Cohen 1997). caspase-3 activity has also been detected in chick motor neurons undergoing programmed cell death in vivo and in vitro (Li et al. 1998).
Antisera capable of specifically recognizing the active form of caspase-3 or certain neoantigenic epitopes deriving from its cleavage products have been used as histological markers for either in vivo or in vitro apoptotic cells (Kouroku et al. 1998; Tanaka et al. 2000). We have recently shown that the so-called c-Jun-like specific immunoreactivity for apoptosis (Ayala et al. 1999) is in fact due to neoantigenic sites recognized by c-Jun/sc45 antibodies. After cleavage by caspase-3, one of them, the seryl-tRNA synthetase, exposes amino acid residues that have sequence homology with the c-Jun immunizing peptide (Casas et al. 2001). Because these experiments were carried out in human neuroblastoma cell lines, we would like to validate this mechanism with other models and species determine the reason for the widespread specificity of c-Jun-like immunoreactivity in apoptosis.
Incubation of cytosolic extracts from E10 spinal cord tissues with active caspase-3 resulted in the emergence of neoantigens that reacted with c-Jun/sc45 antibody, as demonstrated by dot-blot analysis (Figure 3A–2). This effect was inhibited by the presence of Z-VAD, a caspase-3 inhibitor, in the incubation medium (Figure 3A–3). When these samples were analyzed by Western blotting, a large number of new immunoreactive bands appeared as a consequence of caspase-3 digestion (Figure 3A′-2). This effect was inhibited by the presence of Z-VAD (Figure 3A′-3).
c-Jun/sc45-immunoreactive neoepitopes emerge in cytosolic extracts after treatment with caspase-3. Blots were stained with c-Jun/sc45 antibodies (
Sections from paraformaldehyde-fixed E10 chick embryo spinal cord were digested with active caspase-3 for 4 hr and processed for c-Jun/sc45 immunostaining to establish a closer correlation between caspase-3 activity and c-Jun/sc45 immunohistochemistry. Compared with undigested controls, it was clear that overall immunoreactivity increased in digested specimens (Figure 4). Enhancement of the immunoreactivity signal was especially manifest in all motor neuron somata (Figure 4B). As expected, motor neurons that became immunopositive to c-Jun/sc45 displayed normal morphology without pyknotic chromatin and also showed an extensive cytoplasmic immunostaining (Figures 4B′ and 4B″). In other words, the pattern of c-Jun/sc45 immunostaining induced in normal MNs by caspase-3 digestion closely mimicked that found in apoptotic MNs from undigested tissue. When samples were simultaneously incubated with capase-3 and Z-VAD, the induction of positive c-Jun/sc45 immunoreactivity was prevented in normal cells (Figure 4C). These samples exhibited a similar immunostaining pattern to the controls. (Figures 4C′ and 4C′).
c-Jun/sc45 Antibodies Were Capable of Detecting Apoptosis in Paraffin-embedded Specimens
In pathology laboratories, tissues are regularly embedded, processed, and stored for long periods in paraffin blocks. As a result, for retrospective studies the application of immunohistochemical procedures that can be used on paraffin sections is highly desirable. Because the c-Jun/sc45 antibody is an excellent tool for detecting apoptosis, we tested its applicability for routinely paraffin-processed sections of E7.5 embryo spinal cord samples stored for 3 years. We also compared the immunoreaction with c-Jun/sc45 in apoptotic neurons with that obtained with anti-active caspase-3 antibodies (Figure 5). Similar results were obtained with the two antibodies and specific staining of pyknotic MNs in a Golgi-like pattern was observed, as in the cryostat sections (Figures 5A and 5B). However, the intensity of the immunostaining displayed by the c-Jun/sc45 antibodies was much stronger than that obtained by caspase-3 active antibodies (Figures 5A and 5B).
Normal tissues and cells became strongly immunoreactive to c-Jun/sc45 when digested by caspase-3. (
c-Jun/sc45 and active caspase-3 immunoreactivity in apoptosis remains present in paraffin-embedded samples. c-Jun/sc45 and active caspase-3 fluorescent immunolabeling after paraffin processing of samples from E-7 chick embryo spinal cord. c-Jun/sc45 (
c-Jun/sc45 can also detect apoptotic cells in human lymphoid formalin-fixed, paraffin-embedded tissue samples, as demonstrated in sections from human tonsil (Figure 5D).
Discussion
In several works, polyclonal antibodies (c-Jun/sc45 or AB-2 from Oncogene) against the peptide TPTPQ-FLCPKNVTD, corresponding to 91–105 mouse c-Jun amino acids, have been used to study the involvement of this transcription factor in cell death and specific cytoplasmic immunostaining of apoptotic cells was also described (Grand et al. 1995; Ferrer et al. 1996a,b, 1997a,b,c; Messina et al. 1996; Soriano et al. 1996; Pozas et al. 1997; Garrah et al. 1998). However, it was later seen that this kind of immunolabeling was not the result of c-Jun, but of a crossreaction with cytoplasmic neoantigens that emerge in apoptotic cells (Ayala et al. 1999; Casas et al. 2001). Monoclonal antibodies directed to other c-Jun peptides otherwise exhibited the expected nuclear immunostaining corresponding to the true localization of c-Jun and were incapable of staining the cytoplasm of apoptotic cells (Casanovas et al. 2001).
Here we confirm that c-Jun/sc45 cytoplasmic immunolabeling is associated with dying pyknotic cells, and we also show by electron microscopy that these cells display ultrastructural characteristics of apoptosis. Cells immunostained with c-Jun/sc45 also exhibited positive cytoplasmic immunoreactivity for active caspase-3, which is considered a general hallmark of apoptosis (Fernandes-Alnemri et al. 1994; Kuida et al. 1996; Martins et al. 1997). The presence of activated caspase-3 in dying chick embryo developing MNs also confirms the studies of Li et al. (1998), who showed that DEVD-CHO is capable of rescuing these MNs from natural death.
Detection of apoptotic cells on tissue samples by light microscopy is currently achieved by TUNEL assay. However, the validity of TUNEL for identifying apoptotic cells has been seriously questioned since the observation that, in many cases, cells dying by necrosis show DNA laddering and TUNEL-positive staining (Charriaut-Marlangue and Ben-Ari 1995; van Lookeren Campagne et al. 1995; Dong et al. 1997; Gwag et al. 1997; Sohn et al. 1998; Ishimaru et al. 1999; Fujikawa et al. 2000). In our case, the experimental induction of excitotoxic necrosis by excitatory amino acids resulted in the appearance of a large number of TUNEL-positive cells in spinal cord. However, according to their ultrastructural morphology these cells were not apoptotic and were labeled neither by c-Jun/sc45 nor by active caspase-3 antibodies. In contrast, the induction of a massive wave of apoptotic MN death following a leg injection of β-Bgtx (Pittman et al. 1978) results in a parallel increase in the number of lateral motor column cells that are simultaneously positive for active caspase-3, c-Jun/sc45, and TUNEL. Thus, positive cytoplasmic c-Jun/sc45 immunoreactivity overlaps with caspase-3 activation and is a specific marker for apoptosis. This indicates that the c-Jun/sc45 antibody is a more efficient tool than TUNEL for detecting apoptosis in histological sections.
Comparable results were obtained using a specific antibody for caspase-3-polyADP-ribose-polymerase proteolyzed product (Siman et al. 1999). This antibody showed an immunoreactivity pattern on apoptotic cells similar to that displayed with c-Jun/sc45 antibodies.
The ability of the sc45 antibody to detect neoepitopes emerging as a consequence of caspase-mediated proteolysis was further demonstrated by the fact that normal cells in fixed and sectioned tissue samples may acquire strong cytoplasmatic c-Jun/sc45 immunoreactivity if they are incubated with caspase-3. This staining property did not emerge when caspase-3 activity was blocked by Z-VAD. Furthermore, when cytosolic extracts from chicken spinal cord were treated with caspase-3, a marked increase in c-Jun/sc45-immunoreactive signal was observed in dot-blots. Western blotting analysis indicated that the increase in immunoreactivity was due to a broad spectrum of proteins that migrated to different molecular weights. These bands must have emerged as a consequence of caspase-3 activity because they failed to appear in the presence of Z-VAD. This confirmed our previous results in neuroblastoma cells showing that one of the neoantigens recognized by c-Jun/sc45 antibodies was the caspase-3-cleaved seryl-tRNA-synthetase (Casas et al. 2001).
It is not surprising that c-Jun/sc45 antibodies recognize a variety of proteins after caspase proteolysis because it is known that caspase-3 cleaves many different substrates including poly-(ADP-ribose) polymerase (PARP), fodrin, retinoblastoma protein, huntingtin, DNA-dependent protein kinase, gelsolin, ribonucleoprotein A1, vimentin and others (Fernandes-Alnemri et al. 1995; Martin et al. 1995; An and Dou 1996; Goldberg et al. 1996; Song et al. 1996; Kothakota et al. 1997; Brockstedt et al. 1998; Hashimoto et al. 1998) that may share common epitopes at their cleavage sites.
Given the fact that c-Jun/sc45 antibodies are much more sensitive and specific than other immunohistochemical tools for detecting apoptosis, we suggest that they could be used as excellent and reliable markers for qualitative and quantitative studies on tissue sections from a broad diversity of tissues and species. In addition, c-Jun/sc45 antibodies may also be used for determining proteolytic activity in apoptosis and identifying new target proteins for caspase-3, which could be cleaved by caspase-3 during apoptosis.
Footnotes
Acknowledgements
Supported by Ministerio de Educación y Ciencia (SAF97–0083), Ajuntament de Lleida and grants from the Fundació La Marató de TV3 and Fundación “La Caixa.” The technical assistance of Ester Vazquez and Carmen Guerris is greatly appreciated.