The amyloid precursor protein intracellular domain (AICD) disrupts actin dynamics and mitochondrial bioenergetics

Authors

  • Manus W. Ward,

    1. Department of Physiology and Medical Physics and RCSI Neuroscience Research Centre, Royal College of Surgeons in Ireland, Dublin 2, Ireland
    Search for more papers by this author
  • Caoimhín G. Concannon,

    1. Department of Physiology and Medical Physics and RCSI Neuroscience Research Centre, Royal College of Surgeons in Ireland, Dublin 2, Ireland
    Search for more papers by this author
  • Jacqueline Whyte,

    1. Department of Physiology and Medical Physics and RCSI Neuroscience Research Centre, Royal College of Surgeons in Ireland, Dublin 2, Ireland
    Search for more papers by this author
  • Ciara M. Walsh,

    1. Department of Physiology and Medical Physics and RCSI Neuroscience Research Centre, Royal College of Surgeons in Ireland, Dublin 2, Ireland
    Search for more papers by this author
  • Brian Corley,

    1. Department of Physiology and Medical Physics and RCSI Neuroscience Research Centre, Royal College of Surgeons in Ireland, Dublin 2, Ireland
    Search for more papers by this author
  • Jochen H. M. Prehn

    1. Department of Physiology and Medical Physics and RCSI Neuroscience Research Centre, Royal College of Surgeons in Ireland, Dublin 2, Ireland
    Search for more papers by this author

  • Manus W. Ward and Caoimhín G. Concannon contributed equally.

Address correspondence and reprint requests to Jochen H. M. Prehn, Ph.D., Department of Physiology and Medical Physics, Royal College of Surgeons in Ireland, 123 St. Stephen’s Green Dublin 2, Ireland. E-mail: prehn@rcsi.ie

Abstract

J. Neurochem. (2010) 10.1111/j.1471-4159.2010.06615.x

Abstract

The amyloid precursor protein (APP) is critically involved in the pathogenesis of Alzheimer’s disease, and is strongly up-regulated in response to traumatic, metabolic, or toxic insults to the nervous system. The processing of APP by γ/ε-secretase activity results in the generation of the APP intracellular domain (AICD). Previously, we have shown that AICD induces the expression of genes (transgelin, α2-actin) with functional roles in actin organization and dynamics and demonstrated that the induction of AICD and its co-activator Fe65 (AICD/Fe65) resulted in a loss of organized filamentous actin structures within the cell. As mitochondrial function is thought to be reliant on ordered actin dynamics, we examined mitochondrial function in human SHEP neuroblastoma cells inducibly expressing AICD/Fe65. Confocal analysis of the mitochondrial membrane potential (Δψm) identified a significant decrease in the Δψm in the AICD50/Fe65 over-expressing cells. This was paralleled by significantly reduced ATP levels and decreased basal superoxide production. Over-expression of the proposed AICD target gene transgelin in SHEP-SF parental cells and primary neurons was sufficient to destabilize actin filaments, depolarize Δψm, and significantly alter mitochondrial disrtibution and morphology. Our data demonstrate that the induction of AICD/Fe65 or transgelin significantly alters actin dynamics and mitochondrial function in neuronal cells.

Abbreviations used:
AD

Alzheimer’s disease

AICD

APP intracellular domain

APP

amyloid precursor protein

dox

doxycycline

iFA

inducible for Fe65 and AICD50

PBS

phosphate-buffered saline

ROI

region of interest

TRITC

tetramethylrhodamine B isothiocyanate

TMRM

tetramethyl rhodamine methyl ester

Alzheimer’s disease (AD) is a neurodegenerative disease of the CNS associated with progressive memory loss and the onset of dementia. One of the major pathological characteristics observed in AD patients is the formation of extracellular plaques which consist primarily of β-amyloid peptides derived from proteolytic processing of the amyloid precursor protein (APP), a ubiquitously expressed type I transmembrane protein (Selkoe 1991, 1994). The APP intracellular domain (AICD), is a small 6-kDa protein that originates from the cleavage of APP by γ-secretase activity at Val636 (AICD59), Ala638 (AICD57), or Leu645 (AICD50) corresponding to the γ- or ε-cleavage sites of the γ-secretase complex (Octave et al. 2000; Sastre et al. 2001; Yu et al. 2001; Muller et al. 2007). The role of AICD in the pathology of AD is still unclear, however, AICD has been shown to have transactivation potential (Cao and Sudhof 2001).

The function of AICD as a signalling protein is difficult to clearly define because of its rapid degradation by the proteasome (Nunan et al. 2001) or by the insulin degrading enzyme (Edbauer et al. 2002). Indeed the activity of AICD appears to be highly dependent on its stabilization with transcriptional co-activators including the cytosolic adaptor protein, Fe65. The AICD-Fe65 nuclear complex when formed is known to bind to the transcription factor CP2/LSF/LBP1 at the phosphotyrosine-binding 1 domain of Fe65 (Zambrano et al. 1998) and the histone acetyl-transferase tat-interactive protein (Cao and Sudhof 2001) initiating transcription. We have recently identified putative AICD target genes following the inducible expression of AICD alone or in combination with Fe65 in human SHEP-SF neuroblastoma cells using a DNA microarray approach (Muller et al. 2007). Of particular interest was our finding that the induction of AICD increased the expression of genes with a known function in the organization and dynamics of the actin cytoskeleton, including transgelin, resulting in pronounced changes in the organization of the actin cytoskeleton in neurons (Muller et al. 2007). Furthermore, an increased expression of transgelin and other genes associated with actin stability were also identified in brain samples from patients with AD (Muller et al. 2007).

Interestingly, a previous study has shown that in cells lacking AICD expression mitochondrial function appears to be significantly compromised (Hamid et al. 2007). Given that mitochondrial morphology, motility, and membrane potential (Δψm) are known to be altered when the actin cytoskeleton is distorted (Anesti and Scorrano 2006; Boldogh and Pon 2007), we here investigated how mitochondrial function is modulated in human neural cells over-expressing either AICD/Fe65 or its target gene, transgelin. In the course of the study, we have identified major changes in mitochondrial function in neural cells when actin dynamics is altered following the induction of AICD/Fe65 and the over-expression of transgelin.

Materials and methods

Plasmid construction

To create the pIRES2-EGFP-Transglein vector the entire open reading frame of transgelin was generated by PCR amplification with primers containing XhoI and BamHI restriction sites with human cDNA as template. PCR fragments were ligated into a XhoI–BamHI digested pIRES2-EGFP (Clontech, Mountain View, CA, USA) vector. All constructs were verified by sequencing.

Cell culture and transfection

SHEP-SF stably over-expressing AICD50/Fe65 under the control of a Tetracycline dependent promoter (Tet-Off system) were generated as previously described (Muller et al. 2007) and maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum, 2 mM glutamine, 100 U/mL penicillin, 100 mg/mL streptomycin, and 1 μg/mL of doxycycline (dox) in a humidified environment of 5% CO2 at 37°C. For expression of AICD50/Fe65, the dox was removed from medium for 72 h as previously described (Muller et al. 2007) prior to the initiation of experiments. In the experiments involving the over-expression of Trangelin, SHEP-SF cells were transfected with either the pIRES2-EGFP-transgelin or control empty vector (pIRES2-EGFP) on eight-well chamber slides or Willco dishes (Willco BV, Amsterdam, The Netherlands) using Metafectene (Biontex, Munich, Germany) as per manufacturers instructions 24 h prior to the initiation of experiments. Primary murine neocortical neurons were prepared, cultured, and transfected as previously described (Muller et al. 2007; Concannon et al. 2008).

Mitochondrial bioenergetics, distribution, and morphology

Cells were plated on Willco dishes and were loaded with tetramethyl rhodamine methyl ester (TMRM) (20 nM; Bio Sciences Ltd., Dun Laoghaire, Ireland) for 30 min at 37°C (in the dark) in experimental buffer (120 mM NaCl, 3.5 mM KCl, 0.4 mM KH2PO4, 20 mM HEPES, 5 mM NaHCO3, 1.2 mM Na2SO4, 1.2 mM CaCl2, 1.2 mM MgCl2, and 15 mM glucose; pH 7.4). The Willco dishes with cells were washed in fresh medium (containing TMRM 20 nM) before being mounted in sealed (mineral oil) non-perfusion (37°C) holder and placed on the stage of a LSM 510 Meta Zeiss confocal microscope (Carl Zeiss, Jena, Germany). TMRM was excited at 543 nm with a Helium Neon laser (LASOS Lasertechnik GmbH, Jena, Germany) (3% of output power) and the emission collected with through a 560 nm long pass filter. Z-stacks (six images per stack with 0.5 μm steps between images and the pinhole adjusted for the acquisition of 1-μm-thick optical slices) were collected at 30 s intervals and the resultant stacks deconvoluted and analyzed using Metamorph Software version 7.1 (Molecular devices, Berkshire, UK).

For the charaterization of Δψm, mitochondrial distrbution and mitochondrial length a series of protocols were derived. For Δψm measurements, a region of interest (ROI) was marked around the whole cell or neuronal axon and the integrated (total flouescence) TMRM fluorescence within each ROI determined using the Metamorph software. The values obtained for the control (EGFP transfected alone) cells were averaged and this value normalized to 100%. The same procedure was then followed for the induced (transgelin in combination with EGFP) cells line and the average data for these cells compared directly with the average value of the control cells and the data presented as a percentage of the control.

Mitochondrial distribution was measured as a function of distance from the nucleus within each cell. A ROI was drawn around the border of each cell and the integrated fluorescence for TMRM was measured (total accumulative fluorescence for the ROI). Sequential ROIs were then drawn at a distance of 5 and 10 μm around the nucleus (identified using the brightfield image) within each cell and the integrated TMRM fluorescence determined within each region and expressed as a percentage of the total TMRM fluorescence. The total mitochondrial mass for the region outside the 10 μm radius was determined by subtracting the total integrated fluorescence for the 0–5 μm and 5–10 μm regions from the total. The final values for each region within the cell were then normalized as a percentage of the total fluorescence to allow for a direct comparison between cells.

The characterization of mitochondrial morphology was based soley on a scoring mechanism of those mitochondria with a length greater or less than 5 μm and the resulting values were presented as a percentage of the total number of mitochondria for each cell. The deconvoluted z-stacks allowed most of the mitochondria within each cell to be clearly resolved and measured, however, often the number of mitochondria and the total fluoescence intensity within the peri-nuclear region of non-induced (eGFP transfected alone) cells was such that made it difficult to resolve individual mitochondria. In light of this, we were reserved in our scoring and aired more on the side caution and scored those mitochondria that could not be clearly resolved as having a length of less than 5 μm. The number of mitochondria within each cell typically varied in the region of 15–30, therefore, the total number of mitochondria for each cell was normalized to 100% and the associted scores (number of mitochondria with lengths greater or less than 5 μm) aligned with this to allow for a more accurate comparision between cells.

Assay of superoxide

A stock solution of 2 mM hydroethidine was prepared daily in 100% dimethyl sulphoxide and diluted to 100 mM in N2-purged water and was kept on ice until use. For single-cell imaging of superoxide production, Willco dishes containing cells (60–70% confluence) were mounted in a sealed non-perfusion (37°C) holder and placed on the stage of a LSM 510 Meta Zeiss confocal microscope with 2 mL of experimental buffer containing 1 μM hydroethidine. Hydroethidine when oxidized by superoxide produces the highly fluorescent ethidium which can be monitored over time. Ethidium fluorescence was monitored with excitation at 488 nM with an Argon laser (1% of output power) and the emission collected through a 560-nm-long pass filter. For the determination of changes in ethidium fluorescence over time, images were taken every 30 s. The resulting fluorescent images were processed using Metamorph Software version 7.1, with ROIs drawn around each cell and the integrated ethidium fluorecence measured over time. To compare changes in ethidium fluorescence between cells within each experiment and between experiments, the baseline value for each cell at the beginning of an experiment was normalized to 100% and the changes in fluorescence then expressed as a percentage increase from the baseline over time.

ATP assay

Following appropriate treatment in 24-well plates, the cells were lysed using a hypotonic lysis buffer (Tris-acetate buffer, pH 7.75) and 50 μL of the sample was reacted with 50 μL of the luciferin–luciferase reaction kit (ENLITEN ATP Assay System Bioluminescence Detection kit; Promega, Southampton, UK) to quantify ATP content. The amount of ATP was determined by a concentration standard curve, and ATP content values were normalized according to the protein concentration for each sample (μmol ATP/mg protein).

Phalloidin staining

Cells were grown to 60–70% confluence on glass coverslips. After induction, cells were fixed with 3.7% paraformaldehyde in phosphate-buffered saline (PBS) for 10 min at 20–25°C and permeabilized with 0.1% Triton X-100 in PBS for 5 min at 20–25°C. Cells were incubated with 10 μg/ml phalloidin-tetramethylrhodamine B isothiocyanate (TRITC; Sigma-Aldrich, Dublin, Ireland) in PBS for 20 min at 20–25°C. After two washes with PBS to remove excess stain, the coverslips were mounted in VectaShield (Vector Laboratories, Burlingame, CA, USA) and visualized using a Zeiss LSM 510 confocal microscope (Carl Zeiss, Jena, Germany). TRITC and GFP were excited using a series of different lasers and the emissions for each collected, GFP was excited at 488 nm with an argon laser (1% of output power), and the emission was collected through a 505- to 550-nm barrier filter. TRITC was excited at 543 nm with a helium neon laser (3% of output power), and the emission was collected through 560-nm-long pass barrier filter. Images were collected and were processed using MetaMorph software.

siRNA and transfection

Annealed siRNA duplexes targeting transgelin (5′-AGUCCAAAAUCGAGAAGAAdTdT-3′) or a control non-targeting sequence (5′-UUCUCCGAACGUGUCACGUdTdT-3′) were purchased from Sigma-Proligo (Paris, France). Cells were transfected with 100 nM of appropriate duplex using Metafectene (Biontex) as per manufacturer’s instructions.

Western blotting

Whole cell lysates and western blotting was performed as previously described (Concannon et al. 2008). Blots were subsequently probed with either a mouse monoclonal anti-cytochrome c-oxidase subunit IV antibody (Molecular Probes, Leiden, The Netherlands) diluted 1 : 1000, a rabbit polyclonal anti-transgelin (Abcam, Cambridge, UK) diluted 1 : 1000, or a mouse monoclonal anti-α-tubulin antibody (Sigma) diluted 1 : 5000. Horseradish peroxidase-conjugated secondary antibodies (Jackson ImmunoResearch, Plymouth, PA, USA) were diluted 1 : 10000 and detected using Supersignal West Pico Chemiluminescent Substrate (Pierce, Rockford, IL, USA) and imaged using a FujiFilm LAS-300 imaging system (Fuji, Sheffield, UK).

qPCR

Total RNA was extracted using the RNeasy mini Kit (Qiagen, Crawley, UK). First-strand cDNA synthesis was performed with 2 μg total RNA as template using Superscript II reverse transcriptase (Invitrogen, Paisley, UK) primed with 50 pmol random hexamers (New England Biolabs, Ipswich, MA, USA). qPCR was performed using the LightCycler 2.0 (Roche Diagnostics, Indianapolis, IN, USA) and the QuantiTech SYBR Green PCR kit (Qiagen) as per manufacturers’ protocols and 25 pmol of primer pair concentration (Sigma-Genosys). The data were analyzed using Lightcycler Software 4.0 with all samples normalized to GAPDH.

Statistics

Data are given as SEM. For statistical comparison, one-way analysis of variance followed by Tukey test were employed using SPSS software (SPSS GmbH Software, Munich, Germany). p-Values smaller than 0.05 were considered to be statistically significant.

Results

Overexpression of AICD and Fe65 alters actin dynamics and mitochondrial function

We have previously established that the induction of AICD/Fe65 resulted in a destabilization of actin filaments within SHEP-SF cells (Muller et al. 2007). As mitochondrial function and structure are highly sensitive to modifications in actin dynamics (Anesti and Scorrano 2006; Boldogh and Pon 2006), we aimed to investigate if mitochondrial function was altered following conditional induction of AICD50/Fe65 expression using a Tet-Off system. As previously demonstrated, dox removal for 72 h allowed for the expression of both AICD50 and Fe65 (Muller et al. 2007). Following induction, there was more than a 15-fold increase in the expression of AICD50, and 30-fold increase in expression of Fe65 (Fig. 1a).

Figure 1.

 Over-expression of amyloid precursor protein intracellular domain (AICD)50/Fe65 alters actin dynamics resulting in a significant decrease in Δψm. (a) Cells were induced to express AICD50/Fe65 for 72 h at which point total RNA was extracted and the expression levels of AICD50 and Fe65 were established using qPCR. (b) iFA and iEGFP cells were maintained on Willco dishes for 72 h in the absence and presence of doxycycline (dox). Non-induced (+dox) and induced (–dox) cells were loaded with TMRM (20 nM) and fluorescence visualized by confocal microscopy. Induction of AICD50/Fe65 resulted in a decrease in TMRM fluorescence and a clustering of the mitochondria within the perinuclear region (iFA-dox). In contrast, induction of EGFP alone (-dox) did not significantly alter mitochondrial morphology or TMRM florescence. Bar, 20 μm and images are representative of those from three independent cultures. (c) Quantification of raw TMRM fluorescence for iFA and iEFGP in the presence and absence of doxycycline (n = 20 cells for each condition). *p < 0.01 difference between induced (-dox) and not induced (+dox). (d) Western blot anlaysis of the expression levels of Cox IV in non-induced (+dox) and induced (-dox) cells expressing either EGFP alone (iEGFP) or AICD50/Fe65. Tubulin served as loading control.

We have previously demonstrated that induction of AICD50/Fe65 is associated with dramatic changes in actin dynamics (Muller et al. 2007). To investigate how AICD50/Fe65 induction and the subsequent changes in actin dynamics impacted mitochondrial function and morphology, we employed the potentiometric cationic fluorescent probe TMRM (Nicholls and Ward 2000; Ward et al. 2000) and high resolution confocal microscopy at a single-cell level to monitor mitochondrial dynamics and membrane potential (Δψm). The clones inducibly expressing either AICD50/Fe65 (iFA) or enhanced green fluorescent protein were induced for 72 h and then loaded with TMRM (20 nM non-quenched mode) to monitor mitochondrial function. In the non-induced and induced iEGFP cells (Fig. 1b, upper left hand panels), and the non-induced iFA50 cells (Fig. 1b, third left hand panel), the mitochondria retained a predominantly threadlike morphology, were mobile and maintained a high Δψm (Fig. 1c; see also Movie S1, iFA Non-induced.avi). In contrast to this, the induction of AICD50/Fe65 expression (Fig. 1b, lower left hand panel) was associated with a clustering of sedentary mitochondria at the nucleus, with the mitochondria having a significantly decreased Δψm (Fig. 1c; see also Movie S2, iFA Induced.avi). This data would suggest that the induction of AICD50/Fe65 exerted a significant impact on mitochondrial function and localization. Furthermore, the changes in TMRM fluorescence did not result from alterations in mitochondrial biomass as western blotting revealed similar levels of Cox IV expression following AICD50/Fe65 expression (Fig. 1d).

The mitochondrial production of reactive oxygen species has been shown to highly dependent on the maintenance of a high Δψm (Skulachev 1996). Utilizing the fact that hydroethidine when oxidized by superoxide produces the highly fluorescent ethidium (Bindokas et al. 1996; Castilho et al. 1999), we set out to establish if a decrease in Δψm following the induction of the iFA50 clone also resulted in a concomitant decrease in superoxide production. Cells were loaded with hydroethidine (1 μM) and the basal rate of superoxide production was monitored over time (Fig. 2a). A significant decrease (Fig. 2b) in superoxide production (hydroethidine oxidation) was observed following induction of AICD/Fe65 (iFA; 1.17% ± 0.15% per 1 min) when compared with non-induced cells (0.49 ± 0.14% per 1 min).

Figure 2.

 Over-expression of amyloid precursor protein intracellular domain (AICD)50/Fe65 decreases superoxide production and cellular ATP levels. (a) The iFA cells were loaded with hydroethidine and ethidium fluorescence monitored over time in the presence or absence of doxycycline (dox). Induction of iFA (-dox) resulted in a marked decrease in the basal rate of superoxide production within the cell. (b) Ethidium fluorescence in induced and non-induced iFA cells expressed as a % increase in fluorescence over time (n = 8 cells for each condition). *p < 0.01 difference between non-induced and induced iFA cells. (c) ATP levels were measured in iFA cells in the presence or absence of AICD50/Fe65 over-expression for 72 h. *p < 0.01 difference between ATP levels in non-induced and induced iFA cells. (d) ATP levels were measured in iEGFP cells in the presence or absence of doxycycline for 72 h.

When TMRM is employed in the non-quenched mode (as utilized here), the fluorescence obtained is related to the membrane potential across the inner mitochondrial membrane. Therefore, major changes in the TMRM fluorescence either at a whole cell or at a single mitochondrial level can be directly attributable to changes in the H+ gradient across the inner membrane and as such the net driving force available across the mitochondrial membrane to drive the synthesis of ATP through complex V (ATP synthase). Indeed, ATP levels were found to be significantly reduced following the induction of AICD50/Fe65 (Fig. 2c), an event which correlated with a significant reduction in TMRM fluorescence (Fig. 1c and d). From these results, it was apparent that the induction of AICD50/Fe65 significantly modified cellular bioenergetics.

Overexpression of transgelin alters actin dynamics, mitochondrial morphology, and Δψm

Transgelin has been shown to play a role in the organization and stability of the actin cytoskeleton (Goodman et al. 2003) and was found to be differentially regulated following the over-expression of AICD50/Fe65 in human neural cells (Muller et al. 2007). We next set out to determine if an increased expression of transgelin was sufficient to mimick the effect of AICD50/Fe65 induction. To this end, a pIRES2-EGFP-transgelin vector was generated and transiently transfected into SHEP-SF parental cells and primary cortical neurons, and changes in actin dynamics and mitochondrial function were monitored. Over-expression of transgelin resulted in a destabilization of the actin fibers within the cell (Fig. 3a, right panels) resulting in mitochondria with a significantly reduced Δψm localized to the perinuclear region of the cell (Fig. 3b left bottom panel GFP positive cell, Fig. 3c).

Figure 3.

 Over-expression of transgelin alters actin dynamic resulting in major changes in mitochondrial energetics and localization. (a) SHEP-SF parental cells were transfected with plasmids for the expression of either EGFP alone (pIRES2-EGFP) or transgelin in combination with EGFP (pIRES2-EGFP-transgelin) and stained with rhodamine-phalloidin (red) to visualize F-actin. Expression of transgelin resulted in the disorganization of actin fibers within the cell body and clumping of actin fibers at the periphery, suggesting altered actin dynamics. (b) SHEP-SF parental cells were transfected as described above. After 24 h, cells were loaded with TMRM (20 nM) and fluorescence visualized by confocal microscopy. Over-expression of transgelin resulted in a decrease in TMRM fluorescence and a clustering of the mitochondria within the perinuclear region. In contrast, transfection with EGFP alone did not significantly alter mitochondrial morphology or TMRM florescence. Bar, 20 μm. Images are representative of those from three independent cultures. (c) Quantification of TMRM fluorescence for SHEP-SF cells transfected with EGFP or transgelin and EGFP relative to non-transfected cells (n = 18 cells for each condition). *p < 0.001 difference between cells transfected with EGFP alone and transgelin.

To investigate where the changes in actin dynamics associated with AICD50/Fe65 expression were mediated in a trangelin-dependent manner we utilized siRNA to inhibit the expression of transgelin in the presence and absence of AICD50/Fe65 expression. As demonstrated in Fig. 4(a) the AICD50/Fe65 mediated induction of transgelin was significantly attenauted by the siRNA duplex targeting transgelin (Fig. 4a). Subsequent phalloidin staining of AICD50/Fe65 expressing cells demonstrated the loss of organized actin filaments in control siRNA transfected cells, a phenotype which was reversed by the transgelin siRNA (Fig. 4b).

Figure 4.

 siRNA-targeting transgelin reverses the actin phenotype following amyloid precursor protein intracellular domain (AICD)50/Fe65 expression. (a) The iFA cells were transfected with either a siRNA duplex targeting transgelin or a control sequence. Cells were subsequently cultured in the presence and absence of doxycycline to induce AICD50/Fe65 expression. Expression levels of transgelin were assessed by western blotting. Tubulin served as loading control. (b) Cells were treated as described above and rhodamine-labeled actin filaments were visualized by confocal microscopy. Induction of both AICD50 and Fe65 (-dox) revealed disorganization of actin fibers within the cell body and clumping of actin fibers at the periphery, suggesting a destabilization of actin dynamics. In contrast, induction of AICD50/Fe65 in the presence of transgelin siRNA prevented the disorganization of actin filaments.

Disruption of mitochondrial localization and size following transgelin expression

To further examine the effects of transgelin expression we performed more detailed quantitative analysis of the distribution and morphology of mitochondria following transgelin expression in the SHEP-SF neural cells. As demonstrated in Fig. 5(a) and (b), the vast majority of the mitochondria within transgelin expressing cells were primarily localized within a 0–5 μm radius of the nucleus. In subsequent experiments we examined the effects of transglein expression on mitochondrial energetics, morphology and distribution in mouse neocortical neurons. As noted with the SHEP-SF cells, expression of transgelin resulted in a dramatic reduction in TMRM fluorescence compared to non-transfected neurons or neurons expressing the empty vector (Fig. 5c and d). We noted an almost complete loss of polarized mitochondria within the axon and dendritic processes, and an accumulation of perinuclear mitochondria with reduced Δψm within the cell body (Fig. 5c; bottom left panel). This was in stark contrast to the mitochondrial morphology and localization identified in neurons that expressed EGFP alone (Fig. 5c upper left panel). Furthermore, quantification of mitochondrial size revealed a significant increase in the number of smaller (< 5 μm) mitochondria in the transgelin expressing neural cells (Fig. 5e). Taken together, our data suggest that increased transgelin expression is associated with marked changes in mitochondrial size and distribution.

Figure 5.

 Over-expression of transgelin in murine neocortical neurons alters mitochondrial energetics and localization. (a, b) SHEP-SF parental cells were transfected with plasmids for the expression of either EGFP alone (pIRES2-EGFP) or transgelin in combination with EGFP (pIRES2-EGFP-transgelin) and loaded with TMRM. TMRM fluorescence was measured in sequential regions of interest (0–5 μm, 5–10 μm, and > 10 μm) around the nucleus and mitochondrial distribution measured as a function of total cellular fluorescence. *p < 0.01 difference between control (EGFP alone n = 16) and transgelin induced cells (n = 16). (c) Cortical neurons transfected with EGFP or transgelin and EGFP for 24 h before incubation with TMRM (20 nM). Over-expression of transgelin resulted in a clustering of the mitochondria within the cell body and a loss of mitochondria from the processes when compared with control transfection neurons (EGFP alone). Bar, 10 μm and images are representative of those from three independent cultures. (d) Quantification of TMRM fluorescence in murine cortical neurons transfected to express EGFP alone (n = 8) or EGFP and transgelin (n = 7). *p < 0.01 difference between neurons transfected with EGFP alone and transgelin and EGFP. (e) Changes in mitochondrial morphology were measured in SHEP-SF parental cells transfected with plasmids for the expression of either EGFP or transgelin in combination with EGFP. Mitochondria were scored as having a length greater or lower than 5 μm. *p < 0.01 difference between control (EGFP alone n = 16) and transgelin expressing cells (n = 16).

Discussion

The function of AICD as a signaling protein and its potential role in AD has until recently been difficult to clearly define. Previously, we have established a key role for AICD and its co-activator Fe65 in the regulation of genes involved in the organization and dynamics of the actin cytoskeleton, including transgelin (Muller et al. 2007). Here, we have further established that the induction of AICD/Fe65 and the resultant alteration in actin dynamics can significantly modulate mitochondrial function and cellular bioenergetics. Furthermore, over-expression of the AICD target gene transgelin induced a similar phenotype to that of AICD50/Fe65 induction, with clumping of actin filaments and a significantly decreased Δψm and altered mitochodrial localization and morphology evident in SHEP neural cells and murine neocortical neurons.

Previous studies have noted that the deposition of neurofilaments within axons are predominant features in the development of the neurodegenerative disorders such as amyotrophic lateral sclerosis (Delisle and Carpenter 1984; Collard et al. 1995; Tudor et al. 2005), and AD (Galloway et al. 1990; Heredia et al. 2006). Furthermore, a recent quantitative proteome analysis of a presymptomatic A53T α-synuclein Drosophila model of Parkinson’s disease identified a number of dysregulated proteins including those associated with membranes, the endoplasmic reticulum, the actin cytoskeleton, and mitochondria (Xun et al. 2008). Alteration in actin dynamics and their relationship to cellular function have almost exclusively been modeled in yeast cells to date. Gourlay et al. (2004) utilizing the versatility of yeast to examine the role of the actin cytoskeleton in their model of aging have identified that the depletion of the gene scp1, a yeast homolog of the mammalian transgelin, increased actin dynamics, and the longevity of yeast. In addition, over-expression of scp1 in yeast was found to decrease actin dynamics, with a resultant depolarization of Δψm (Gourlay et al. 2004), an outcome similar to that identified here where a marked depolarization of Δψm was evident following the induction of AICD50/Fe65 (Fig. 1) or the over-expression of transgelin (Fig. 2). In addition to a decrease in Δψm, we also identified a marked decrease in cellular ATP level and reactive oxygen species production after 72 h. Interestingly, a number of studies have identified that the mitochondrial respiratory capacity is significantly altered when their ability to interact with the actin cytoskeleton is impeded (Koch et al. 2003; Boldogh and Pon 2006, 2007). Indeed, one of the earliest hallmarks of AD is that of a reduced respiratory capacity with a significant decrease in cytochrome oxidase activity (Chagnon et al. 1995; Hirai et al. 2001).

Moreover, in a recent study, Hamid et al. (2007) created H4 human neurglioma cells stably expressing an inducible allele of the AICD with a vesicular stomatitis virus tag under the control of a Tet-Off system. When AICD expression was turned off, the cells were found to have a significantly reduced endoplasmic reticulum Ca2+ and higher cytosolic basal Ca2+ and most significantly the cells contained mitochondria with a hyperpolarized Δψm (Hamid et al. 2007). Here, we have induced the expression of AICD (Figs 1 and 2) and its target gene transgelin (Figs 3–5) in neural cells, a subsequent depolarization of the Δψm and a change in mitochondrial distribution and morphology. Therefore, it would appear that a close association exists between mitochondrial energetics and the relative expression of AICD and its target genes in neural cells.

Mitochondrial transport and function are also known to be highly dependent on the maintenance of an intact cytoskeleton, with alterations in actin dynamics resulting in altered mitochondrial localization, morphology, motility, and Δψm in yeast (Drubin et al. 1993; Lazzarino et al. 1994; Simon et al. 1995; Smith et al. 1995; Boldogh et al. 1998; Gourlay et al. 2004). Indeed, the maintenance of an intact cytoskeleton has also been shown to be a key factor in the transport of mitochondria within neuronal axons with impaired cytoskeletal function linked to mitochondrial dysfunction in disease states such as amyotrophic lateral sclerosis (Collard et al. 1995). Treatment with cytochalsin D which results in actin aggregation has been demonstarted to result in impaired mitochondrial movement in hippocampal neurons (Ligon and Steward 2000). For neurons to function properly, it is imperative that mitochondria can be transported freely to (anterograde) and from (retrograde) synaptic sites, where mitochondria are known to be highly active (Miller and Sheetz 2004; Lee and Peng 2006) and play a functional role in the regulation of synaptic Ca2+ (Yang et al. 2003; Brown et al. 2006; Ly and Verstreken 2006; Mironov and Symonchuk 2006). We here have demonstrated that mitochondria of AICD50/Fe65 induced neural cells have a reduced proton motive force across the membrane and a reduced capacity to produce ATP, but more detailed studies on mitochondrial dynamics are required to investigate if this is because of reduced or altered mitochondrial motility. Interestingly, a number of reports suggest that the functionality of specific mitochondrial channels (Voltage dependent Anion channel and the KATP Channel) are dependent on their association with a dynamic actin cytoskeleton (Xu et al. 2001). Indeed, Gourlay and Ayscough (2005) have presented a model in which they suggest an intriguing possibility that actin has a conserved role in cell death by regulating the mitochondrial voltage dependent anion channel and following the stabilization of actin this channel opens with a resultant dissipation of the Δψm. Our findings may also be important in the context of neuronal vulnerability to other stressors in AD. Indeed alterations in Δψm are known to alter neurotransmission resulting in an extensive release of glutamate (Ward et al. 2007) with reduced energy availability in neurons having a negative impact on neuronal outcome, e.g. during Ca2+-mediated injury (Ward et al. 2007; Weisova et al. 2009).

In this study, we have highlighted a potential role for the APP cleavage product AICD in the pathology of AD, identifying that an induction of AICD and/or its target gene transgelin resulted in significant alterations in actin dynamics that in turn lead to significant impairments of mitochondrial function. As neurons require a constant transport of mitochondria to regions of high energy demand, an inhibition of this process through alterations in actin stability may contribute to their degeneration and loss.

Acknowledgements

This work was supported by grants from the Health Research Board, Higher Education Authority PRTLI (National Biophotonics and Imaging Platform Ireland) and Science Foundation Ireland (08/IN.1/B1949) to JHMP.

Ancillary