Background
Mutations in transactivation response element DNA-binding protein 43 (TDP-43) cause rare forms of familial amyotrophic lateral sclerosis (fALS) and frontotemporal lobe dementia (FTLD) [
1,
2]. The incidence of TDP-43 mutations is estimated at approximately only 4% of all familial cases of ALS. However, abnormal TDP-43 cytosolic aggregates in motor neurons are a common pathological feature of the majority of sporadic and familial ALS autopsies [
3], suggesting that even in the absence of mutations, TDP-43 may be involved in disease pathogenesis.
TDP-43 is a nuclear riboprotein involved in regulation of RNA splicing [
4]. Under physiological conditions, TDP-43 localization is predominantly nuclear, but under stress TDP-43 translocates to the cytoplasm, where it localizes to stress granules. Mutant TDP-43, or even wild type protein when it is overexpressed, can form ubiquitin-positive aggregates in the extra-nuclear cell compartment.
In ALS-FTLD, mutant TDP-43 is depleted from the nucleus and accumulates in the cytosol, where it forms aggregates of phosphorylated protein [
3]. Therefore, the pathogenic mechanisms of mutant TDP-43 may include both loss of nuclear function and gain of extra-nuclear toxic functions. Indeed, numerous lines of evidence support a role for extra-nuclear TDP-43 in ALS-FTLD [
5,
6], and there is a strong possibility that TDP-43 oligomers spread from cell to cell in a prion-like fashion [
7].
Mitochondria have been suggested to be one of the multiple potential targets of TDP-43 extra-nuclear mislocalization and aggregation, since mitochondrial dynamics and distribution has been described in cellular and animal models of the disease. In particular, it was shown that both TDP-43 overexpression and suppression impair mitochondrial movement in cultured neurons and that mutant TDP-43 co-localizes with mitochondria [
8]. Furthermore, mitochondrial accumulation and aggregation was found in various TDP-43 animal models, including the transgenic mice expressing C-terminal fragments [
9] or full-length protein [
10]. In these models, mitochondrial aggregates in the cell bodies of motor neurons were accompanied by depletion of mitochondria in motor terminals and neuromuscular junctions, suggesting that mitochondrial axonal transport was impaired.
Mitochondrial axonal transport defects were identified in various TDP-43 animal models, from flies [
11] to mice [
12], and are likely to contribute to mitochondrial mislocalization and possibly to neuronal functional defects and degeneration. Interestingly, cytosolic localization of TDP-43 has also been linked to alterations in the interactions between mitochondria and endoplasmic reticulum [
13], leading to functional defects, predominantly in intracellular calcium handling, which could also have important implications for ALS-FTLD pathogenesis.
Recently, it was proposed that mutant cytosolic TDP-43 gains access to the mitochondrial matrix through a dysregulated import mechanism, and that TDP-43 in the matrix impairs mitochondrial respiratory chain activity by downregulation of complex I biosynthesis [
14].
Based on the growing body of evidence suggesting that mitochondria are targeted by mutant TDP-43, we decided to perform a thorough investigation of mitochondrial bioenergetics in multiple model systems, ranging from mutant TDP-43 transgenic mice to patient-derived cells harboring TDP-43 mutations, to asses the presence of mitochondrial dysfunction.
Methods
Chemical reagents
All reagents were from Sigma-Aldrich (St. Louis, MO) unless otherwise stated.
Animals
We used the strain B6;CB-Tg(Prnp-TARDBP*A315T)95Balo/J of TDP43A315T transgenic mice (from The Jackson Laboratory, Bar Harbor, ME). In all experiments, N-Tg littermates were used as controls.
The criterion for determining disease onset was the development of abnormal hind-limb extension (clasping). The criterion for survival was the inability of the mouse to right itself, when placed on its side (loss of righting reflex).
Bioenergetics and calcium uptake measurements in brain mitochondria
Forebrain mitochondrial fractions were freshly prepared from TDP43
A315T transgenic mice and age and sex matched N-Tg littermates by differential centrifugation of homogenates on a discontinuous Percoll™ gradient as previously described [
15,
16]. Mitochondria were obtained from the non-synaptosomal gradient layer and washed 3 × in buffer containing 75 mM sucrose, 225 mM mannitol, 10 mM HEPES; 2 mM EDTA pH 7.4.
ATP synthesis was measured in purified brain mitochondria using a luciferin-luciferase approach, as previously described [
17]. Glutamate (5 mM) and malate (2 mM) or succinate (5 mM) plus rotenone (1 μM) were used as oxidative substrates. Measurements were carried out by luminometry.
ROS emission was measured as Amplex Red (Thermo Fisher Scientific, Waltham, MA) fluorescence (555 nm excitation and 581 nm emission wavelengths) in the presence of exogenous horseradish peroxidase and mitochondrial H
2O
2 as described [
18,
19]. Briefly, 100 μg mitochondria were added to 1 mL incubation buffer (125 mM KCl, 20 mM HEPES, 0.2 mM EGTA, 2 mM KH
2PO
4, 200 μg/mL BSA, 1 μM Amplex Red, 4 U horseradish peroxidase, pH 7.2). Standard curves were used to calculate H
2O
2 emission rates after sequential addition of substrate (5 mM glutamate, 2 mM malate), 1 μM rotenone, and 1.8 μM antimycin A.
Mitochondrial Ca2+ uptake was estimated fluorimetrically with Fura 6 (340/380 nm excitation and 510 nm emission wavelengths) (Thermo Fisher Scientific) with sequential additions of 25 nmoles of Ca2+ to the incubation medium (125 mM KCl, 20 mM Hepes, 1 mM MgCl2, 2 mM KH2PO4, 0.2 mM ATP, 1 μM rotenone, 5 mM succinate, 0.3 μM Fura 6, pH 7.2).
Mitochondrial membrane potential was estimated using safranin O (excitation and emission wavelengths of 495 nm and 586 nm, respectively), as previously described [
15]. The incubation buffer contained 125 mM KCl, 20 mM HEPES, 1 mM MgCl
2, 2 mM KH
2PO
4, 0.2 mM ATP, 200 μg/mL BSA, 5 mM glutamate, 2 mM malate, 2 μM safranin O, pH 7.2). Mitochondrial membrane potential decay curves were obtained by repetitive additions of 25 nmol Ca
2+ or 2–16 nM of the respiratory chain uncoupler SF6847.
For mitochondrial respiration, 100 μg of brain mitochondria were resuspended in 0.5 ml of respiration buffer containing 125 mM KCl, 20 mM Hepes, 4 mM K
2HPO
4, 0.1 mg/ml BSA, pH 7.2 and 1 mM ADP. Glutamate (5 mM) and malate (2 mM) were used as oxidative substrates. Oxygen consumption was recorded with an oxygraph equipped with a Clark electrode (Hansatech, Norfolk, UK), as described [
20], before (state 4) and after the addition of ADP (state 3). SF6847 (0.1 μM) was used to fully uncouple mitochondrial respiration (state 3 uncoupled).
Respiratory chain complex I and complex IV enzymatic activities were measured spectrophotometrically, as previously described [
21].
Bioenergetics measurements in cultured cells
Skin fibroblasts (from the repository at the University of California, San Diego) were cultured in Dulbecco modified Eagle medium (DMEM) supplemented with 25 mM glucose, 4 mM glutamine, 1 mM pyruvate, and 10% fetal bovine serum. All fibroblast lines were coded to protect patients’ identity.
HEK293T (from American Type Culture Collection, ATCC, Manassas, VA), were grown in DMEM supplemented with 25 mM glucose, 4 mM glutamine, 1 mM pyruvate, and 5% fetal bovine serum.
All cells tested negative for mycoplasma contamination by PCR assays of the culture medium using previously described primer sets and amplification protocols [
22].
For mitochondrial membrane potential and mitochondrial content measurements skin fibroblasts were seeded at the density of 1.5 × 104 cells/well in replicates of eights in 96-well tissue culture plates in growth medium incubated at 37 °C in 5% CO2. The following day, all cells were washed with cultured medium and loaded with 50 nM tetramethylrhodamine methyl ester (TMRM; 544ex, 590em; Thermo Fisher Scientific) and 450 nM Mito Tracker Green (MTG; 490ex, 516em; Thermo Fisher Scientific) for 30 min at 37 °C in phenol-free DMEM containing 5 mM glucose, 4 mM glutamine, and 1 mM pyruvate, half of the wells additionally contained the protonophore carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP; 2 μM) to completely depolarize mitochondria and obtain background TMRM and MTG fluorescence, which were subtracted from total fluorescence levels. After washing with DMEM, MTG and TMRM fluorescence were simultaneously recorded in a plate reader equipped with a polychromator (Spectramax 5; Hitachi, Tokyo, Japan). MTG and TMRM fluorescence values were expressed as relative fluorescence units per milligram of total cellular proteins (DC Protein Assay; Bio-Rad, Hercules, CA).
Total ATP content in fibroblasts was measured by luciferase reactions in a luminometry plate reader, according to the manufacturer’s guidelines (Promega, Madison, WI). Cells were seeded at the density of 1.5 × 104 cells/well in replicates of nines in 96-well tissue culture plates in growth medium incubated at 37 °C in 5% CO2. On the following day, triplicates of wells were incubated with either DMEM containing 5 mM glucose, 4 mM glutamine, and 1 mM pyruvate (ATP baseline), DMEM containing 5 mM 2-Deoxy-D-glucose (2DG), 4 mM glutamine, and 1 mM pyruvate to block glycolysis (ATP 2DG) or DMEM containing 5 mM glucose, 4 mM glutamine, 1 mM pyruvate, and 1 μM oligomycin to block the mitochondrial ATPase (ATP Oligo) for 90 min. Cells were washed with PBS and lysed in 30 μl of trichloroacetic acid (2.5% W/V) on ice for 30 min. Following lysis, 20 μl aliquots were pipetted into a separate plate for protein determination for data normalization, and 45 μl of Tris-acetate (400 mM, pH 8.0) buffer was added to each well of the remaining lysates. Bioluminescence was measured promptly after adding and mixing 20 μl of luciferase reagent. Luminescence values were normalized against an ATP standard and normalized by protein content.
Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were measured in human fibroblasts or TDP-43 transfected HEK293T cells with a Seahorse XF96 Flux Analyzer (Agilent, Santa Clara, CA). Cell lines were seeded in 12 wells of a XF 96-well cell culture microplate at a density of 1 × 104 cells/well (fibroblasts) and 2 × 104 cells/well (HEK293T cells) in 200 μL of growth medium and incubated for 24 h at 37 °C in 5% CO2. After replacing the growth medium with 200 μL of XF Assay Medium supplemented with 5 mM glucose, 1 mM pyruvate and 4 mM glutamine, pre-warmed at 37 °C, cells were degassed for 1 h before starting the assay procedure, in a non-CO2 incubator. OCR and ECAR were recorded at baseline followed by sequential additions of 1 μM oligomycin, 2 μM FCCP, and 0.5 μM Antimycin A (AA) plus 0.5 μM Rotenone (Rot). Non-mitochondrial oxygen consumption (in the presence of AA + Rot) was subtracted from all OCR values, and outliers technical replicates outside the 2 standard deviation of the mean were discarded for both ECAR and OCR. In fibroblasts, values were normalized by the mean protein value of each line measured by modified Lowry protein assay. In HEK293T cells, values were normalized by total DNA content measured by fluorescence of DAPI stain, using DNA standards for quantification.
Cultured cells transfection and immunostaining
HEK293T or HeLa cells were transiently transfected with plasmids expressing recombinant WT or mutant human TDP-43-myc or empty vector, using FuGENE-HD (Promega), according to the manufacturer’s protocol. For ER and mitochondrial contact analyses, HeLa cells were co-transfected with ddGFP plasmids targeted to mitochondria and ER (Tom20-ddGFP and calN-ddGFP, respectively) together with DSred2-Mito. Cells were imaged 48 h later on the Leica TCS SP5 confocal microscope on a live imaging stage, controlled at 37 °C.
Live cell calcium imaging was performed with HeLa cells co-transfected with WT or mutant TDP-43 and mitochondrially targeted GCamp6. 48 h after transfection, cells were perfused with 20 μM ATP in imaging medium containing (156 mM NaCl, 3 mM KCl, 2 mM MgSO
4, 1.25 mM KH
2PO
4, 10 mM glucose, 1 mM EGTA and 10 mM HEPES, pH 7.35), followed by replenishment of ER calcium with 2 mM calcium to induce store operated calcium entry, as described previously [
23].
Brain mitochondria isolation for TDP-43 western blots
Brains from TDP43A315T transgenic mice and littermate N-Tg controls were cryopreserved by cutting the whole brain into 30–40 mg sections, placing them into solution containing 300 mM sucrose and 20% (by volume) DMSO, freezing in liquid nitrogen, and storing them at −80 °C until the experiment. For mitochondria isolation, ~100 mg of frozen brain tissue was slowly (1 h) thawed on ice, then the cryopreservation solution was replaced with homogenization medium (HM) comprising 225 mM sucrose, 75 mM mannitol, 1 mM EGTA, 20 mM HEPES (pH 7.4), and 1 mg/ml bovine serum albumin essentially fatty acid free (BSA). Brain tissue was homogenized in 10 ml of HM with a Dounce tissue grinder (15 ml volume, glass tube-glass pestle) manually by 40 strokes, in ice (step 1). All further procedures were performed at 4 °C. The homogenate was centrifuged at 2000 g × 5 min; the supernatant was further centrifuged at 12000 g ×10 min (step 2). The pellet from this step was collected, resuspended in 1 ml of HM, loaded on top of 9 ml of 23% Percoll™ separation medium, and centrifuged at 31,000 g × 10 min. The Percoll™ separation medium was prepared by dissolving 225 mM sucrose, 75 mM mannitol, 1 mM EGTA, 20 mM HEPES in 100% Percoll™ and adjusting pH to 7.4; this medium was diluted with HM to 23% Percoll™. The pellet from this step was collected and washed 2 times by centrifuging at 12000 g × 10 min in HM. The final pellet was dissolved in 200 μl of HM devoid of BSA (HM-BSA). For digitonin treatment, this pellet was thoroughly resuspended in 10 ml of HM-BSA and 20 μl of 10% digitonin DMSO solution (0.02%, final digitonin concentration) was added to the suspension. After incubating on ice for 5 min, the suspension was centrifuged at 12000 g × 10 min, the pellet was resuspended in digitonin-free HM-BSA and centrifuged again at 12000 g × 10 min. The final solid pellet was resuspended in 100 μl of HM-BSA and stored frozen at −20 °C until the experiment.
For assaying combined synaptic + nonsynaptic mitochondria, the isolation procedure was modified as follows. The supernatant from step 2 above was collected, and treated with 0.02% digitonin for 5 min in ice, then centrifuged at 12000 g × 10 min. The pellet was resuspended in 10 ml of HM and centrifuged 12,000 g × 10 min. To remove proteins electrostatically attached to mitochondria membranes, the pellet from this step (contains synaptic + non-synaptic mitochondria) was resuspended in 2 ml of medium comprising 6 M KCl, 20 mM HEPES (pH 7.4) and incubated in ice for 5 min with occasional gentle agitation. The suspension was diluted with HM-BSA to 10 ml and centrifuged 12,000 g × 10 min, and the pellet was collected. To prepare digitonin -treated mitochondria from this fraction, mitochondria pellet was diluted to 10 ml with HM-BSA and treated with 0.01% digitonin for 7 min in ice, with occasional gentle agitation. The suspension was centrifuged 12,000 g × 10 min, the pellet was collected and washed in 10 ml of HM-BSA and centrifuging 12,000 g × 10 min; the procedure was repeated once. The final solid mitochondria pellet was resuspended in 100 μl of HM-BSA and stored frozen at −20 oC until the experiment.
Western blot analyses
Samples were lysed in RIPA buffer (Thermo Fisher Scientific) and centrifuged at 14,000 rpm at 4 °C for 20 min. Supernatants were diluted in 2X Laemmli sample buffer (Bio-Rad). Proteins were separated on 10 or 4–20% gradient Mini-PROTEAN TGX gels (Bio-Rad) and transferred to nitrocellulose blotting membranes (Bio-Rad). Membranes were probed with primary antibodies against FLAG tag (M2, Sigma), MCU (Sigma), MICU1 (Abcam, Cambridge, MA), cytochrome C (Cell Signaling, Danvers, MA), Cyclophilin D (Millipore, Billerica, MA), Tim23 (BD Biosciences, San Jose, CA), VDAC1 (Abcam), COX1 (Abcam), GAPDH (Thermo Fisher Scientific), citrate synthase (Abcam), and an oxidative phosphorylation (OXPHOS) cocktail (Abcam) overnight at 4 °C. Blots were then probed with horseradish peroxidase-conjugated anti-mouse (Jackson ImmunoResearch, West Grove, PA) or anti-rabbit (Thermo Fisher Scientific) secondary antibodies and detected using enhanced chemiluminescence (Bio-Rad).
Statistical analyses
All data are presented as mean ± standard deviation in all experiments, except for the results of Figs.
4 and
5, which are presented as mean ± standard error of the mean. The results were compared using Student’s t-test or, when more than one condition was examined, one-way ANOVA with Bonferroni correction. In either case, a
p value <0.05 was considered significant.
Discussion
Mitochondrial dysfunction is one of the known pathogenic events associated with ALS. In particular, bioenergetic impairment has been amply described in cellular and mouse models of fALS caused by SOD1 mutations [
30] and, more recently, by other genetic forms of the disease, such as mutations in C9Orf72 [
31,
32], VCP [
33], and CHCHD10 [
34]. TDP-43 mutations cause rare cases of fALS, but cytosolic aggregation of wild type TDP-43 in motor neurons is a prominent pathological feature of ALS, including the most prevalent sporadic form of the disease [
3]. Mislocalization and aggregation of TDP-43 have been shown to cause abnormalities of mitochondrial morphology and dynamics in cultured neuron systems [
35] and in vivo, in peripheral neurons of animal models of TDP-43 ALS [
11,
12]. The mechanisms of these abnormalities and their impact on disease pathogenesis remain to be fully elucidated, but mutant TDP-43 was suggested to impair complex I in cultured cells [
36]. Moreover, mutant TDP-43 was recently proposed to impair mitochondrial function directly from the matrix compartment, where it causes respiratory chain dysfunction by inhibiting complex I translation [
14]. Such mechanism of mitochondrial damage by intra-mitochondrial TDP-43 was proposed in TDP-43
A315T transgenic mice and patient-derived cells.
In this study, we sought to validate the findings of bioenergetic dysfunction in TDP-43A315T transgenic mice, patient fibroblasts, and transfected cell expressing TDP-43. Overall, we did not confirm previous findings of mitochondrial bioenergetics defects in any of these models. In particular, mitochondria isolated from the brain of TDP-43A315T transgenic mice did not show impairment of mitochondrial respiration, ATP generation, and calcium handling. Similarly, there was no impairment of bioenergetic functions in skin fibroblasts harboring a pathogenic TDP-43 mutation or in HEK293T cells overexpressing various mutant forms of TDP-43.
We found that a portion of TDP-43 was peripherally associated with the surface of intact mouse brain mitochondria. This association, despite not interfering with mitochondrial bioenergetics, could affect inter-organellar communication. Indeed, TDP-43 was shown to interfere with the structures that link mitochondria and ER, also known as mitochondria associated membranes or MAMs [
37]. Our ddGFP assay, did not detect an overall change in the amount of ER-mitochondrial contacts, but this method does not specifically detect MAMs. Therefore, more work, using a variety of experimental approaches, which are beyond the scope of this study on mitochondrial function, will have to be done to further explore the effects of the peripheral association of TDP-43 with mitochondria.
Earlier studies had shown that TDP-43 overexpression alters mitochondrial dynamics in neurons, resulting in decreased motility accompanied by abnormal clustering and fragmentation [
10,
12,
38]. Based on our results showing that mitochondrial bioenergetics is essentially unaffected by mutant TDP-43, we conclude that the mitochondrial dynamics abnormalities are not the result of intrinsic mitochondrial bioenergetics defects or intramitochondrial localization of mutant TDP-43. It is logical to speculate that abnormal mitochondrial movement and morphology could be caused by alterations in components of the mitochondrial dynamics machinery, which are localized in extra-mitochondrial compartments or on the outer mitochondrial membrane. This could lead to impaired motility and mitochondrial clustering along the neuronal processes. Dysregulation of mitochondrial transport by mutant TDP-43 could impair the correct positioning of mitochondria relative to major sites of energy utilization, such as synapses. For example, the levels of the outer membrane cargo adaptor Miro1, a protein necessary for the docking of mitochondria on microtubule tracks, were decreased in spinal cord of mice expressing mutant TDP-43 as well as ALS patients [
35]. We propose that TDP-43 causes alterations of proteins involved in mitochondrial transport, such as Miro1, or components of the fusion and fission apparatus, without TDP-43 internalization in mitochondria or bioenergetic defects.
Interestingly, brain mitochondrial calcium uptake capacity was increased in TDP-43A315T transgenic mice relative to controls. This result suggests that TDP-43 A315T mitochondria were less prone to undergo calcium induced permeability transition than controls, a further indication that they did not suffer from bioenergetic impairment.
Like the mouse model, cultured cells expressing mutant TDP-43 did not show bioenergetic defects, both when endogenous levels of the protein were present, in patient-derived fibroblasts, and when the mutant proteins were overexpressed, in transfected cells. The results in fibroblasts were in line with a recent report, which compared C9Orf72 mutant and TDP-43 mutant fibroblasts, and detected respiratory chain defects in the former, but not the latter [
31]. Furthermore, similar to the transgenic mouse brain, cells expressing A315T TDP-43 had increase mitochondrial calcium uptake upon ER calcium release.
The reasons for the discrepancy between the present study and a recent report in which mitochondrial respiratory chain defects were associated with mitochondrial import of mutant TDP-43 [
14] is not immediately clear, especially in regards to the TDP-43
A315T mouse model, since we utilized mice from the same transgenic line and at similar ages as in the aforementioned report. It is possible that differences resulted from the protocols of mitochondrial purification, which were partly dissimilar in the two studies. It could be hypothesized that mitochondria from mutant TDP-43 brains are more fragile and could be damaged by the isolation procedure, which was not the case in our study, because soluble proteins of the intermembrane space were retained in the mitochondrial preparation. On the other hand, membrane fragility could not explain a decline in the activity of individual respiratory chain complexes, such as complex I, whose activities are measured on fractionated mitochondrial membranes and do not depend on mitochondrial integrity. Specifically, we did not find evidence of a decrease in either complex I-driven respiration and ATP synthesis or complex I oxidoreductase activity in brain mitochondria from TDP-43
A315T transgenic mice, suggesting that OXPHOS is preserved in these mitochondria. Furthermore, the levels of respiratory chain subunits that we analyzed were unchanged in TDP-43
A315T transgenic mouse brain mitochondria, suggesting that their synthesis was not impaired.
Unlike previous reports [
14,
36], we did not detect respiratory defects in mutant TDP-43 human fibroblasts. A possible explanation for this discrepancy in findings could be ascribed to the different TDP-43 mutations investigated in previous studies and the present one, since we analyzed lines from three patients from one family with the N352S mutation and their healthy relatives. Nevertheless, we also tested the effects of recombinant mutant TDP-43 expression in HEK293T cells, an approach similar to the one utilized previously [
14], but did not detect respiratory defects in these mutant cells either.