Anchored FRET sensors detect local caspase activation prior to neuronal degeneration
© Figueroa et al; licensee BioMed Central Ltd. 2011
Received: 5 October 2010
Accepted: 23 May 2011
Published: 23 May 2011
Skip to main content
© Figueroa et al; licensee BioMed Central Ltd. 2011
Received: 5 October 2010
Accepted: 23 May 2011
Published: 23 May 2011
Recent studies indicate local caspase activation in dendrites or axons during development and in neurodegenerative disorders such as Alzheimer's disease (AD). Emerging evidences point to soluble oligomeric amyloid-β peptide as a causative agent in AD.
Here we describe the design of fluorescence resonance energy transfer (FRET)-based caspase sensors, fused to the microtubule associated protein tau. Specific caspase sensors preferentially cleaved by caspase-3, -6 or -9 were expressed in differentiated human neuroblastoma SH-SY5Y cells. The anchoring of the sensors resulted in high FRET signals both in extended neurites and soma and made analysis of spatiotemporal signal propagation possible. Caspase activation was detected as loss of FRET after exposure to different stimuli. Interestingly, after staurosporine treatment caspase-6 activation was significantly delayed in neurites compared to cell bodies. In addition, we show that exposure to oligomer-enriched amyloid-β peptide resulted in loss of FRET in cells expressing sensors for caspase-3 and -6, but not -9, in both soma and neurites before neurite degeneration was observed.
Taken together, the results show that by using anchored FRET sensors it is possible to detect stimuli-dependent differential activation of caspases and to distinguish local from global caspase activation in live neuronal cells. Furthermore, in these cells oligomer-enriched amyloid-β peptide induces a global, rather than local activation of caspase-3 and -6, which subsequently leads to neuronal cell death.
In comparison with current knowledge about axon growth and guidance, relatively little is known about the intracellular mechanisms of neurite (axon or dendrite) elimination and its relation to apoptosis is still not clear. It has been suggested that axonal degeneration may occur through a local self-destructive program, which is distinct from the proteolytic program that mediates apoptosis . On the other hand removal of dendrites from sensory neurons during pruning in Drosophila melanogaster is directed by local caspase-like activity . In addition, there is evidence for the involvement of the non-classical executioner caspase-6 in local axonal degeneration induced by trophic deprivation of cultured primary sensory neurons . In these neurons, localization of inactive procaspase-6 was observed in both the axons and cell bodies, whereas procaspase-3 was primarily detected in cell bodies . Caspase-6 activation, as detected by immunostaining of active caspase-6, has been reported to occur in hippocampus and cerebral cortex of mild, moderate, severe and very severe sporadic Alzheimer's disease (AD) . In human ischemia caspase-6 translocates to the nuclei  where it plays a role in nuclear lamin cleavage and subsequent chromatin condensation . In contrast, in AD brains immunostaining of active caspase-6 was concentrated in neurites , indicating a local effect in neurites.
In tubular cell structures, such as axons and dendrites, propagation of diffusion-mediated signaling is slower compared to in the cytosol of a spherical cell. Local caspase activation may not necessarily lead to a global activation of apoptotic processes in a cell (cf. ). It is also possible that different parts of a cell, such as nerve terminals, axons or dendrites, are more vulnerable to a particular type of toxic insult and that an apoptotic process is initiated in a specific location followed by spreading of the signal to the soma resulting in apoptotic nerve cell death. Although there are many reliable assays for apoptosis, they generally only distinguish between cells that are dead or alive. The incoherent picture of degenerative processes described in the literature calls for techniques that allow monitoring of apoptotic progression in greater detail.
To investigate early degenerative events in neurites and cell bodies we have developed the first model system for spatiotemporal monitoring of caspase activation in neuronal cells. Caspase-3, -6, and -9 have been implicated both during development of the nervous system and in neurodegenerative disorders, such as AD. Hence, fluorescence resonance energy transfer (FRET)-based sensors preferentially cleaved by caspase-3, -6 and -9 were designed. The sensors localize to microtubules and are therefore highly enriched in axons/neurites. Here, we show that our caspase sensors can in fact be used in live cell imaging of neurite degeneration and cellular apoptosis in an in vitro model using differentiated human neuroblastoma SH-SY5Y cells. Our experimental system enables monitoring of local activation of caspases and propagation of the signal in neurites and was used to study caspase activation induced by staurosporine, local oxidative stress and by the AD-associated amyloid β (Aβ) peptide.
Western blot analysis, using anti-GFP antibodies, of extracts of SH-SY5Y cells expressing the caspase-3 (DEVD) or control (LEVA) FRET sensors displayed a specific band with an apparent molecular weight of 130 kDa (Figure 3B). After 3 hours of incubation with 10 μM staurosporine a specific band at approximately 100 kDa appeared in cells expressing the caspase-3 (DEVD) sensor. EYFP and ECFP both have an expected size of 27 kDa, whereas the 441 amino acid long tau isoform in human brain tissue has been reported to migrate with an apparent molecular weight of approximately 74 kDa . Thus, the ~100 kDa cleavage product is consistent with the expected migration of the EYFP-Tau fragment (~27 kDa + ~74 kDa) from caspase-dependent cleavage of the DEVD motif. In support of this is the fact that the ~100 kDa band was absent from staurosporine treated cells expressing the control (LEVA) sensor. In addition, the presence of caspase inhibitors prevented staurosporine-induced appearance of the ~100 kDa band (Additional file 1, Figure S1). The results demonstrate that decreased FRET intensity in apoptotic SH-SY5Y cells expressing the FRET sensor was correlated to caspase-dependent proteolysis of the sensor molecule.
Selective elimination (degeneration or retraction) of axons and dendrites, without death of the neuron is important during development of the nervous system in order to refine synaptic connections. However, axon elimination, as well as apoptosis [13, 14], also occurs in many neurological diseases or as a consequence of neuronal injury (reviewed by [1, 15]). Axonal degeneration involves a rapid breakdown of the axonal cytoskeleton caused by enzymatic degradation of neurofilaments. A process termed dying-back degeneration (or Wallerian-like degeneration) has been described in a variety of neurological disorders and is characterized by initial degeneration of distal regions of long axons followed by distal-to-proximal progression. In many neurodegenerative diseases, including AD, axonal and dendritic atrophies are observed, (e.g. ). Such neuritic atrophies will most likely contribute significantly to the clinical symptoms and occur prior to or in the absence of neuronal death. In a mouse model of AD in vivo imaging studies indicated that the Aβ deposits, typically observed in AD brain, can indeed induce breakage of axonal and dendritic branches .
Here, we describe the development and characterization of new FRET based sensors for spatiotemporal imaging and analysis of caspase activation in relation to neurite degeneration and neuronal cell death. A full-length tau moiety was added to the FRET sensors both to mediate association to microtubules in neurites and thus increase signal intensity as well as to restrict diffusion of the sensors to allow monitoring progression of the apoptotic signal in a spatiotemporal manner (cf. Figure 1). As expected our caspase sensors localize to the cytoskeleton when expressed in non-neuronal (HeLa) cells and also to the microtubule of the neurites when expressed in differentiated neuronal cells. FRET studies in differentiated SH-SY5Y cells exposed to staurosporine verify the sensors as early markers of apoptosis occurring prior to neurite retraction and any other morphological changes could be detected. Caspase activation was measured as loss of FRET and it was confirmed, by western blot analysis of sensor fragmentation, to correspond to cleavage between EYFP and ECFP where the caspase cleavage sites are localized.
In this study we focus on three different caspases; caspase-3, -6, and -9. Caspase-9 is the initiator caspase associated with the mitochondrial or intrinsic apoptotic pathway . Proapoptotic signals stimulating release of cytochrome c from mitochondria that promotes oligomerization of apoptotic peptidase activating factor 1 (Apaf-1), will lead to recruitment and activation of caspase-9. Caspase-9 null mice die perinatally and show a similar phenotype as mice with deletion of Apaf-1 or caspase-3 genes, including enlarged and deformed brains [19, 20]. This clearly indicates the importance of the intrinsic apoptotic pathway during development of the nervous system. Caspase-9 has also been proposed to be activated by an alternative pathway: It has been shown that endoplasmic reticulum (ER) stress can activate caspase-9 in the absence of cytochrome c and Apaf-1 [21, 22]. Interestingly, Aβ peptides have been shown to induce apoptosis mediated by ER stress . Furthermore, increased levels of active caspase-9 have been detected in AD brain [15, 24]. Also, in a recently published study  caspase-9 and the intrinsic apoptotic pathway was suggested to be involved in non-apoptotic caspase-3 activity in dendritic spines in Tg2576-APPswe mice. This activity correlated with the onset of memory deficits in these mice. Therefore, caspase-9 has been proposed to be involved not only in developmental apoptosis of neuronal cells but also in neurodegenerative disorders such as AD. Caspase-3 is the major executioner caspase and can be activated once an initiator caspase, like caspase-9, is activated. Caspase-6 is also often classified as an executioner caspase since it cleaves the nuclear lamin proteins during apoptosis . However, proteomics analysis showed that nearly half of identified neuronal caspase-6 targets were cytoskeleton-associated, indicating that caspase-6 most likely can function also before death signals reach the nucleus . In addition, recent studies have implicated caspase-6 in axonal degeneration not linked to apoptotic cell death .
We evaluated the performance of the caspase sensors in differentiated neuroblastoma cells treated with staurosporine. We observed a significant decrease in FRET signal in cells expressing caspase-3 and -9 sensors, but not in cells expressing the non-cleavable control (Figure 3). Staurosporine is a non-specific inhibitor of protein kinases and is one of the best inducers of apoptosis in different types of cells, including neuroblastoma [27, 28]. Although the mechanism by which staurosporine induces apoptosis is not well known, the mitochondrial pathway has been shown to be involved [29, 30], consistent with our results. For caspase-3 and -9, we could not detect any time difference in the loss of FRET in neurites and soma. This indicates that staurosporine may induce a global activation of these two caspases in neuronal cells. However, when the loss of FRET in cells expressing the caspase-6 sensor was analyzed we observed a clear and significant delay in the neurites compared to the cell bodies (Figure 4). This indicates a difference in staurosporine-induced apoptotic signaling in axons/dendrites compared to the cell bodies. A previous study has indicated a spatiotemporal difference in caspase-3 compared to caspase-6 activation . In undifferentiated SH-SY5Y cells exposed to staurosporine caspase-6 activation, monitored by the appearance of a lamin B1 fragment in western blot analysis, was considerably delayed compared to cleavage of caspase-3 substrates . Thus, it may be speculated that staurosporine induces a delayed activation of caspase-6 that is initiated in the cell body and thereafter is spread to the neurites.
The ability to detect caspase activation at sub-cellular resolution was further established using a co-expressed mitochondrial photosensitizer (Killer red) , which enabled induction of apoptosis locally in the cell via the intrinsic pathway. We observed that if caspase-3 activation was initiated in one of the neurites the signal was quickly transferred to the cell body but not to other neurites (Figure 5). The other neurites degenerated as well after the cell had been exposed to the death stimuli in spite of the delayed caspase-3 activation.
Finally, we used differentiated neuroblastoma cells expressing caspase sensors to analyze the effect of the Aβ peptide on apoptotic signaling. Apoptotic signaling is believed to play an important role in a number of neurological disorders, including AD . If caspase activation is crucial for the progression of AD pathology it is important to learn how the activation occurs, where in the neuron this process takes place, and whether or not it will result in neuronal cell death. Aβ is a hydrophobic peptide generated by sequential proteolytic cleavage of amyloid precursor protein (APP) by β- and γ-secretase and is the leading candidate for activation of apoptotic mechanisms in AD [7, 33]. Aβ is highly prone to aggregate and in the brain it is found in different forms ranging from monomers and non-fibrillar aggregates, termed oligomers, to a highly fibrillar form. Recent evidences suggest that it is the diffusible Aβ oligomers that have the most toxic properties [34, 35]. Aβ oligomers have also been shown to bind to synaptic sites , which may be important for early synaptic failure observed in AD (cf. ). Furthermore, Aβ has been proposed to have a local effect on neurites causing degeneration and neuritic apoptosis using compartmented cultures of hippocampal neurons . In the present study we have exposed the cells to a preparation enriched in oligomer-enriched forms of Aβ with non-detectable levels of fibrils. Aβ exposure resulted in activation of caspase-3 and -6, but we could not detect any specific spatiotemporal effect of the peptide. Thus, our results suggest that in these neuronally differentiated cells there is no part of the cell that is more vulnerable to insult and that Aβ rather induces global caspase activation. Because the caspase-9 sensor resisted proteolysis, the intrinsic pathway may play a less important role in Aβ1-42 induced apoptosis of these cells. Most interestingly, it was recently shown in a mouse model that caspase activation precedes and leads to tangles (i.e., a characteristic neuropathological feature in AD). Evidences also indicated that caspase-3 and -6 were activated in the tangle positive cells .
In conclusion, we show that the anchored caspase sensors localize to the cytoskeleton, that they are differentially cleaved depending on stimuli and that they can be used to detect local caspase activation and to follow the spreading of caspase activation from one part of the cell to another. Under the conditions used in this study we could not detect any specific spatiotemporal effect of Aβ on caspase activation. Caspase-3 and -6, but not -9, were activated both in soma and in neurites. It will be interesting to use the anchored FRET sensors in experiments designed to cause retraction and loss of synaptic contacts without causing cell death. This is indeed a promising new tool in research on AD and other neurodegenerative conditions.
To construct the localized caspase sensors, EYFP was amplified from the EYFP-C1 plasmid (BD Biosciences) using the following primers: forward, 5'-TATTAGATCTCATGGTGAGCAAGGGCGA-G-3', reverse, 5'-ACGTGAATTCCTTGTACAGCTCGTCCATGC-3'. The resulting PCR product was digested with BglII-EcoRI, gel-purified and ligated into the BglII-EcoRI of the ECFP-C1 vector (BD Biosciences). The longest isoform of human tau (441 amino acids) was amplified from the pBThtau40 plasmid (kindly provided by Dr. Jacek Biernat, Max-Planck-Gesellschaft) using the following primers: forward, 5'-ATTAGAATTCATGGCTGAGCCCCGCCAG-3', reverse, 5'-ATGCGGATCCGTGATCACAAAC-3'. The resulting PCR product was digested with EcoRI-BamHI, gel-purified and ligated into the EcoRI-BamHI of the ECFP-EYFP vector. To create cleavage sites for caspase-3, -6 and -9 plus an un-cleavable control sequence single stranded sense 32-mers; 5'-CCGGAGCAGATGAGGTCGATGCCGATGAGGTC-3' (caspase-3, DEVDx2), 5'-CCGGAGCAGTCGAGATAGATGCCGTAGAGATC-3' (caspase-6, VEIDx2), 5'-CCGGAGCACTGGAGCACGATGCCCTGGAGCAC-3' (caspase-9, LEHDx2) and 5'-CCGGAGCACTGGAGGTCGCCCTGGAGGTCGCC-3' (control, LEVAx2) were annealed to complementary anti-sense 32-mers; 5'-GATCGACCTCATCGGCATCGACCTCATCTGCT-3' (caspase-3), 5'-GATCGATCTCTACGGCATCTATCTCGACTGCT-3' (caspase-6), 5'-GATCGTGCTCCAGGGCATCGTGCTCCAGTGCT-3' (caspase-9) and 5'-GATCGGCGACCTCCAGGGCGACCTCCAGTGCT-3' (control) generating double stranded DNA with caspase cleavage sites in tandem and BspEI-BglII sticky ends that were ligated into the BspEI-BglII of the ECFP-EYPP-Tau vector. One base was changed in the BglII site in the linker region to facilitate restriction analysis.
HeLa (CCL-2, American Type culture Collection) and U2OS cells (ATCC-HTB-96, LGC standards) were cultured in Dulbecco's Modified Eagle Medium (DMEM, Invitrogen) supplemented with 10% fetal bovine serum (FBS, Invitrogen), 100 U/ml penicillin and 100 μg/ml streptomycin (PEST, Invitrogen). Human neuroblastoma SH-SY5Y cells (CRL-2266, ATCC) were cultured in DMEM/F12 supplemented with 10% FBS, 100 U/ml PEST. The cells were maintained in a humidified 5% CO2 atmosphere at 37°C. Cells were transfected using FuGENE HD (Roche). In live cell imaging experiments cells were seeded on to glass cover slips prior to differentiation. The neuroblastoma cells were differentiated with 10 μM retinoic acid (RA) for three days before transfection after witch cells were incubated for one day prior to the experiment. For western blot analysis cells were seeded at a density of 25.000 cells/cm2 into Petri dishes (Nunc, Denmark). The following day the cells were treated with 10 μM RA for 72 h. After 24 h, the cells were transfected with plasmids encoding different caspase sensors and incubated for 48 h prior to the experiment.
Aβ1-42 (American peptide) was dissolved in hexafluoroisopropanol (HFP) to remove any preexisting structures in the lyophilized stocks and was then stored at -70°C. The peptide was prepared to obtain a solution enriched in oligomeric, but devoid of fibrillar, forms of the peptide, cf. . The oligomer-enriched Aβ1-42 was prepared freshly before each experiment. HFP was allowed to evaporate from the tube using speedvac system for 5 min at room temperature. Thereafter, Aβ1-42 was dissolved in dimethyl sulfoxide (DMSO) to a concentration of 5 mM. Subsequently, phosphate bovine saline (PBS) was added to give a final Aβ1-42 concentration of 100 μM. Finally, the preparation was vigorously vortexed at room temperature for 10 min, whereafter Aβ1-42 was allowed to aggregate by incubation for 24 h at 4°C at 200 rpm. Analysis of the Aβ preparation was made by western blot and Thioflavin T assay (Additional file 2, Figure S2).
Imaging was performed on a Leica TCS-SP laser scanning confocal microscope (Leica, Heidelberg, Germany) with a 63.0 × 1.32 oil immersion objective using a 442 nm 10 mW HeCd laser, 488 nm laser line of a 40 mW Ar-laser and 561 nm laser line of a yellow diode pumped solid-state 50 mW laser. Emission was collected between 455-505 nm (CFP), 515-575 nm (YFP and FRET), and 580-650 nm (KillerRed) using a long pass dicroic mirror (LP) 450 for CFP and FRET, LP500 for YFP and triple dicroic 488/561/633 for KillerRed. The laser lines were scanned sequentially to minimize bleed through. Induction of KillerRed phototoxic effect was achieved by scanning a region of the cells using the 561 nm laser line. Live cell imaging was performed by mounting cells in sealed chambers with serum free media buffered with HEPES. In experiments with staurosporine and Aβ, the agent was added to the media when mounting the cells. Temperature was maintained at 37°C using the Cube and Box microscope incubator system (Live Imaging Services). Projections of image stacks were used for quantification. Images were processed and quantified using ImageJ (Rasband, W.S., Image J, U. S. National Institutes of Health, Bethesda, Maryland, USA, http://rsb.info.nih.gov/ij/, 1997-2009). To acquire the ratio-metric images the intensity in the FRET channel was divided by the intensity in the CFP channel. A pseudo-color was applied and adjusted to visualize the changes in FRET. For graphic representation, images were segmented in a semi-automatic process rendering soma and neurite regions of interest. Values were normalized in relation to the mean value prior to changes in FRET. The last frame before decreased FRET could be detected in each individual experiment was chosen as zero time point to enable alignment of graphs from different experiments. At least three individual experiments were performed. Spatiotemporal studies of FRET signals in neurites and soma were statistically analyzed using unpaired two-tailed Student' s T-test.
For western blot analysis of caspase sensor fragmentation human neuroblastoma SH-SY5Y cells treated with Aβ1-42 or staurosporine were harvested as follows: First, the cells were washed twice with PBS containing 0.7 μg/ml aprotinin (from bovine) and 0.8 mM PMSF. Thereafter, the cells were lysed in lysis buffer (150 mM NaCl, 50 mM TrisCl, 1 mM EDTA, 1% NP-40), containing complete protease inhibitor cocktail, for 20 min at 4°C. In order to remove cell debris the lysate was centrifuged at 10 000 g for 5 min at 4°C. The supernatant was then used for western blot analysis. 30 μg protein from each separate sample was separated by electrophoresis on 8% SDS-polyacrylamide gels and subsequently transferred to polyvinyldine difluoride membranes at 350 mA for 3 h. The membrane was blocked in PBS, containing 5% dry milk (w/v) and 0.1% Tween, for 1.5 h, followed by over night incubation at 4°C with primary rabbit polyclonal or mouse monoclonal anti-GFP antibody (1:1000) (R&D systems). Finally, secondary horseradish peroxidise-coupled donkey anti-rabbit or sheep anti-mouse antibodies (1:5000) were incubated for 1 h at room temperature and the detected proteins were visualised by chemiluminescence using Western LightingTM-PLUS-ECL kit (Perkin Elmer, Netherlands). All western blot reagents were from GE Healthcare or Bio-Rad Laboratories.
We thank Konstantin Lukyanov for the kind gift of KillerRed plasmid, Dr. Jacek Biernat for the kind gift of pBThtau40 plasmid, and Dr. Marie Beckman for valuable discussion. This work was supported by grants from the Swedish Research Council to EH (#2006-5977, #2006-3468 and #2010-4481) and to KI (#2007-2798 and #2010-4505).
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.