Astrocytes with TDP-43 inclusions exhibit reduced noradrenergic cAMP and Ca2+ signaling and dysregulated cell metabolism

Most cases of amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) have cytoplasmic inclusions of TAR DNA-binding protein 43 (TDP-43) in neurons and non-neuronal cells, including astrocytes, which metabolically support neurons with nutrients. Neuronal metabolism largely depends on the activation of the noradrenergic system releasing noradrenaline. Activation of astroglial adrenergic receptors with noradrenaline triggers cAMP and Ca2+ signaling and augments aerobic glycolysis with production of lactate, an important neuronal energy fuel. Astrocytes with cytoplasmic TDP-43 inclusions can cause motor neuron death, however, whether astroglial metabolism and metabolic support of neurons is altered in astrocytes with TDP-43 inclusions, is unclear. We measured lipid droplet and glucose metabolisms in astrocytes expressing the inclusion-forming C-terminal fragment of TDP-43 or the wild-type TDP-43 using fluorescent dyes or genetically encoded nanosensors. Astrocytes with TDP-43 inclusions exhibited a 3-fold increase in the accumulation of lipid droplets versus astrocytes expressing wild-type TDP-43, indicating altered lipid droplet metabolism. In these cells the noradrenaline-triggered increases in intracellular cAMP and Ca2+ levels were reduced by 35% and 31%, respectively, likely due to the downregulation of β2-adrenergic receptors. Although noradrenaline triggered a similar increase in intracellular lactate levels in astrocytes with and without TDP-43 inclusions, the probability of activating aerobic glycolysis was facilitated by 1.6-fold in astrocytes with TDP-43 inclusions and lactate MCT1 transporters were downregulated. Thus, while in astrocytes with TDP-43 inclusions noradrenergic signaling is reduced, aerobic glycolysis and lipid droplet accumulation are facilitated, suggesting dysregulated astroglial metabolism and metabolic support of neurons in TDP-43-associated ALS and FTD.

Taken together, these results demonstrate that in cultured rat astrocytes, expression of the ALS-and FTD-U-linked C-terminal fragment of TDP-43 causes the formation of TDP-43-containing cytoplasmic inclusions and partial sequestration of endogenous nuclear TDP-43 to the cytoplasm.
A small amount of TDP-43 inclusions in some cells can be seen inside the DAPI-labelled nuclear area, thus the role of nuclear TDP-43 inclusions on the accumulation of LDs cannot be excluded. However, since the thickness of the optical section of an individual confocal image was relatively large (1.2 μm) enabling the nuclear and cytosolic fluorescence signals to overlap, it is more likely, that the TDP-43 inclusions seen at the DAPI-labeled nuclear area are in fact located in the cytosol. 2+ ] i are reduced in TDP-43 208-414 -expressing astrocytes. Astrocytes appear to be the primary target of NA, an important neuromodulator in the CNS [28][29][30]40 . Through binding to ARs on the surface of astrocytes, NA activates cAMP and Ca 2+ signaling 37,39,56 , which may be dysregulated in various neurologic disorders 32,57-60 . To study whether ALS-and FTD-U-linked TDP-43 inclusions affect cAMP and Ca 2+ signaling, we monitored NA-induced changes in [cAMP] i and [Ca 2+ ] i using real-time confocal microscopy and genetically encoded FRET-based cAMP nanosensor Epac1-camps 53 or Fluo-4 AM dye, respectively, in astrocytes expressing WT (RFP-TDP-43 wt ) or mutant TDP-43 pDNA construct (RFP-TDP-43 208-414 ; Figs. 3 and 4 and Online resource 1, Fig. S1).
While the expression of α 1 -and β 1 -ARs was not significantly different between the RFP-TDP-43 wt -and RFP-TDP-43 208-414 -expressing astrocytes (84.0% The expression of MCT1 transporters is reduced in TDP-43 208-414 -expressing astrocytes. Since NA-mediated lactate production was enhanced in astrocytes with TDP-43 inclusions (Fig. 5), which may affect the availability of extracellular lactate, we next investigated whether the presence of TDP-43 inclusions alters the expression level of astrocyte-specific monocarboxylate transporters (MCT), which are in astrocytes predominantly responsible for L-lactate transport across the plasma membrane. We immunostained astrocytes expressing RFP-TDP-43 wt and RFP-TDP-43 208-414 with antibodies against MCT1 and MCT4 transporters (Fig. 7). The expression level of MCT4 transporters did not significantly differ between the RFT-TDP-43 wt -and RFP-TDP-43 208-414 -expressing astrocytes (   www.nature.com/scientificreports www.nature.com/scientificreports/ Data are presented as means ± SEM and acquired from at least two different animals (one cell was recorded per coverslip). The Student's t-test, used to test significant differences between astrocytes expressing RFP-TDP-43 208-414 and RFP-TDP-43 wt , revealed the similarity of the responses, however, in Iso-treated astrocytes expressing RFP-TDP-43 208-414 there was a trend towards the reduction in the [lactate] i increase. (C,D) Pie graphs showing the responsiveness of astrocytes to (C) noradrenaline-and (D) isoprenaline-induced changes in intracellular lactate levels. Note that the probability of observing a response to NA with production of lactate in RFP-TDP-43 208-414 -expressing astrocytes was 1.6-fold higher compared to RFP-TDP-43 wt -expressing astrocytes, but not in Iso-treated cells (see also Table 2).

Discussion
Based on the new disease models, a number of recent studies have highlighted the involvement of non-neuronal cells in the pathogenesis of ALS and FTD-U, including astrocytes 2,7 , which provide metabolic and trophic support to motor neurons. However, the molecular mechanisms of astroglial-mediated neurotoxicity in ALS and FTD-U remain poorly understood. Here, we investigated whether astroglial expression of the C-terminal fragment of TDP-43 (TDP-43 208-414 ), a major component of ALS-and FTD-U-associated pathologic cytoplasmic inclusions, affects astroglial cell metabolism. TDP-43 inclusions may compromise astroglial metabolic support of neurons in ALS and FTD-U and contribute to CNS hypometabolism observed in patients with neurodegeneration.
Expression of C-terminal fragment of TDP-43 in isolated cortical rat astrocytes has led to the formation of cytoplasmic TDP-43 inclusions, which can mimic key biochemical features of TDP-43 proteinopathies 10,20 , consistent with studies in other cell types 20,64 . Astrocytes with cytoplasmic TDP-43 inclusions had a 3-fold lower amount of endogenous nuclear TDP-43 (which may cause a partial loss-of-function of nuclear TDP-43) compared with astrocytes expressing WT TDP-43, where most of the TDP-43 was present in the cell nucleus. This indicates that expression of the C-terminal fragment of TDP-43 in astrocytes affects trafficking of endogenous TDP-43 between the nucleus and the cytoplasm, consistent with reports on neurons 5 , glial cells from human tissue samples 55 , various cell lines (e.g. QBI-293 and tsBN2 cells 10 ), and muscle cells 65 . The results suggest that aggregated TDP-43 prevents newly synthesized endogenous TDP-43 from being imported into the nucleus and/ or inhibits the re-entry of existing cytoplasmic TDP-43 into the nucleus 10,13 . Interestingly, the transfection with inclusion-forming TDP-43, caused in neighboring non-transfected astrocytes (with no visible RFP signal) a small, www.nature.com/scientificreports www.nature.com/scientificreports/ albeit not significant, trend in reduced percentage of DAPI and TDP-43-colabeled nuclei, which suggests that the transfection with the C-terminal fragment of TDP-43 might affect the physiology of non-transfected cells. In pathological conditions, mimicked here by the expression of inclusion-forming TDP-43, astrocytes bear a significant functional plasticity, known as reactive astrogliosis 66 . These astrocytes secrete or distribute through gap junctions various cytotoxic factors, such as Lcn2 22,67,68 , interleukin-6, ciliary neurotrophic factor, etc. 69 , which can affect neighboring cells. Even though we did not specifically test whether astrocytes expressing inclusion-forming TDP-43 transform into a reactive phenotype, this outcome is possible. If this is the case, the observed reduction in the nuclear TDP-43 staining in some non-transfected astrocytes adjacent to inclusion-forming TDP-43 astrocytes may be a consequence of altered physiological state of these cells affecting the TDP-43 gene expression level or the synthesis/degradation/distribution of TDP-43.
Because TDP-43 is involved in multiple aspects of RNA processing, any changes in TDP-43 nuclear level may have detrimental effects on astroglial physiology 3 , including on cell metabolism. LD accumulation has been observed in glial cells, including astrocytes, in early stages of neurodegeneration 46 . Recently, alterations in lipid metabolism (accumulation of cholesteryl esters, determined with lipidomic analysis) in spinal cords from SOD1 G93A transgenic mouse model have been reported that might be linked to astrogliosis and LD formation in astrocytes 48 , since a population of astrocytes isolated from the degenerating spinal cords of the same animal model exhibited significant abundance of LDs as well as autophagic and secretory vesicles, all characteristic features of cellular stress and inflammatory activation 70 . However, the mechanisms leading to increased LD accumulation in astrocytes are poorly understood and may, among others, rely on altered neuronal mitochondrial function and ROS as well as on astrocyte-neuron lactate shuttle 47 , in particular on astroglial lactate-derived lipid production in neurons and transfer of excess lipids in lipoprotein-like particles (ApoE) from neurons to astrocytes 47,71 . We show here that in astrocytes with TDP-43 inclusions the LD presence is enhanced (the size and the number of LDs) in the absence of neighboring neurons, suggesting the existence of an alternative astroglial-mediated mechanism triggering accumulation of LDs through altering the balance between biogenesis and degradation of LDs. Accumulation of LDs in astrocytes with cytoplasmic TDP-43 inclusions may be a response to cellular inflammation, which is typically found in the pathology of various neurologic disorders, including ALS 66,72 . Here, LDs are hypothesized to be an important source of energy for proliferation and may serve a protective role by gathering free fatty acids to protect cells against lipotoxicity 73 .
Besides alterations in LD metabolism, changes in noradrenergic regulation of glucose metabolism were observed in astrocytes with cytoplasmic TDP-43 inclusions. Astrocytes with TDP-43 inclusions exhibited downregulation of β 2 -ARs and a 35% reduction in NA-mediated increase in [cAMP] i . Consistent with our results, dysregulation of astrocytic β 2 -AR/cAMP signalling has been suspected to contribute to the pathology of a number of other neurologic disorders, including multiple sclerosis, Alzheimer's disease, human immunodeficiency virus encephalitis, and others 74 . Reduction of β 2 -AR expression was reported in human white matter astrocytes obtained from post mortem brain tissue of patients with multiple sclerosis 57,75 . Moreover, it was demonstrated both in vitro and in vivo that the presence of Alzheimer's disease associated amyloid beta peptide (Aβ) in prefrontal cortical neurons leads to internalization and degradation of β 2 -ARs, which leads to subsequent attenuation of cAMP signalling 59 . Since aerobic glycolysis in astrocytes is upregulated with β-AR/cAMP signaling, one would expect that astrocytes with TDP-43 inclusions and a reduced expression of β 2 -ARs will exhibit reduced β-adrenergic mediated aerobic glycolysis and lactate production. When we stimulated astrocytes with Iso, a selective β-AR agonist, although the responsiveness of cells to Iso was unchanged, there was a trend in reduction of [lactate] i increase, since the amplitude in [lactate] i increase was ~2-fold lower (P = 0.08) in astrocytes with TDP-43 inclusions, consistent with downregulation of β 2 -AR and cAMP signaling in these cells. In contrast to Iso stimulation, the amplitude and the rate of [lactate] i increase upon NA stimulation was unaltered in astrocytes with TDP-43 inclusions, but the probability of activating aerobic glycolysis, measured as increased responsiveness to NA, was increased by 1.6-fold in astrocytes with TDP-43 inclusions. When viewing the astroglial population as a whole, this means that glycolytic lactate production upon NA stimulation is facilitated in astrocytes with TDP-43 inclusions.
Besides cAMP, Ca 2+ signals through activation of α 1 -AR/G q -protein signaling pathway have important role in regulation of NA-mediated glucose metabolism in astrocytes 33 . Abnormal Ca 2+ homeostasis has been observed in astrocytes isolated from SOD1 G93A animals. In particular, excess Ca 2+ release from ER stores upon purinergic/ G q -protein signaling pathway activation has been reported due to abnormal ER Ca 2+ accumulation 76 . If such a mechanism exists in astrocytes with TDP-43 inclusions, binding of NA to α 1 -AR may lead to excess Ca 2+ -release from ER and enhanced aerobic glycolysis despite downregulation of β 2 -AR/cAMP signaling pathway. However, contrary to the results obtained on SOD1 G93A animal model, we observed a 31% reduction in Ca 2+ signaling in astrocytes with TDP-43 inclusions, even though the level of α 1 -AR expression was unchanged. It has been reported that β-AR/cAMP and α 1 -AR/Ca 2+ signaling pathways interact in astrocytes, enhancing each other 39 , therefore downregulation of β 2 -ARs may reduce the noradrenergic Ca 2+ response via α 1 -ARs, explaining the reduction of both noradrenergic cAMP and Ca 2+ signals despite an unaltered expression of α 1 -ARs.
It has been reported that morphologic changes in astrocytes exhibit a bell-shaped dependency on [cAMP] i 37 . If aerobic glycolysis in astrocytes displays a similar bell-shaped dependency on [cAMP] i , astrocytes with reduced β 2 -AR/cAMP signalling may actually attain enhanced metabolic responsiveness to noradrenergic stimulation. Moreover, perturbances in TDP-43 may change the transcriptome and proteome of a cell affecting the expression level of enzymes involved in LD metabolism and aerobic glycolysis, as was observed in other cell types with a TDP-43 knock down 16 . Whether expression level of metabolic enzymes is altered in astrocytes with TDP-43 inclusions and whether this contributes to the observed enhanced astroglial metabolism needs to be investigated in the future.
Increased lactate production upon NA stimulation in astrocytes with TDP-43 inclusions may lead to a better metabolic support of neurons with lactate due to lactate flux generated between astrocytes and neurons 77 . This (2020) 10:6003 | https://doi.org/10.1038/s41598-020-62864-5 www.nature.com/scientificreports www.nature.com/scientificreports/ is, however, contradictory to the CNS hypometabolism observed in patients with neurodegenerative diseases, including ALS 43,45 . Recently, decreased expression of lactate MCT1 and MCT4 transporters has been reported post mortem in the motor cortex of ALS patients compared to non-ALS patients 78 . Moreover, downregulation of MCT1 mRNA in the spinal cords of early symptomatic and endstage SOD1 G93A transgenic mice model of ALS has been observed, presumably in glia (oligodendrocytes and astrocytes) 78 . Down-regulation of the lactate MCT4 transporter has been also demonstrated in spinal cord astrocytes from patients with ALS with SOD1 mutations and in pre-symptomatic SOD1 G93A transgenic mice 43 . We show here that expression of C-terminal fragment of TDP-43 (TDP-43 208-414 ) per se in astrocytes causes a reduction in the expression of astroglial lactate MCT1 transporters. Lower expression of astroglial MCTs may decrease the lactate release capacity in astrocytes. Thus, despite facilitated NA-mediated aerobic glycolysis in astrocytes with TDP-43 inclusions, lactate may accumulate inside cells, and the metabolic support of neurons in patients may be decreased significantly contributing to neuron degeneration in ALS and FTD-U. Consistent with this hypothesis, cAMP signaling and lactate homeostasis in the brain were noted to be reduced in patients with some forms of ALS 43,45 .
In conclusion, the results of the present study show that expression of the C-terminal fragment of TDP-43 in isolated rat cortical astrocytes leads to the formation of TDP-43-positive cytoplasmic inclusions, a hallmark of ALS. These astrocytes exhibit reduced NA-mediated cAMP and Ca 2+ signaling, whereas both aerobic glycolysis and LD presence appear facilitated. Although NA-mediated L-lactate production was increased in astrocytes with TDP-43 inclusions, the expression of lactate MCT1 transporters was reduced, suggesting decreased astroglial lactate release capacity (Fig. 8). Thus, these findings reflect an astroglial stressed state that may fail to adequately metabolically support neurons in ALS and FTD-U, leading to neurotoxicity.

Methods
Cell culture and transfection. Unless noted otherwise, all chemicals were of the highest quality available and purchased from Sigma-Aldrich (Merck KGaA, Darmstadt, Germany).
Primary astrocyte cultures were prepared from the cerebral cortices of 2-to 3-day-old Wistar rats, as described previously 79 , and grown in high-glucose Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 1 mM sodium pyruvate, 2 mM l-glutamine and 25 µg/ml penicillin-streptomycin in a vaporized atmosphere containing 95% air and 5% CO 2 at 37 °C until reaching 70%-80% confluence. Confluent cultures were shaken overnight at 225 rpm and the medium was changed the next morning; this was repeated three times. After the third overnight shaking, the cells were trypsinized and put in flat tissue culture tubes with 10 cm 2 growth area. After reaching confluence again, the cells were subcultured onto 22-mm diameter poly-l-lysine-coated glass coverslips. This procedure yielded astrocytes with > 95% purity, determined by anti-glial fibrillary acidic protein (GFAP) antibody staining 80 .
Colocalization analysis of Alexa Fluor 488 (immunolabelled TDP-43), TagRFP 584 (RFP-TDP-43 pDNA constructs), and DAPI fluorescence signals was performed on exported TIFF files using ColocAna software (Celica Biomedical, Ljubljana, Slovenia) 81 that counts all red, green, blue and colocalized (green and red or green and blue) pixels within the image. The threshold for the colocalized pixel count was set at 20% of the maximal green, red or blue fluorescence intensity, respectively. Colocalization was expressed as the ratio of colocalized to green or blue pixels (as percentages). Kruskal-Wallis one-way analysis of variance (ANOVA) on ranks (followed by Dunn's test) or Mann-Whitney U test were performed to determine statistical significance between the experimental groups. P < 0.05 was considered significant.
To determine the expression level of different ARs and MCTs, the number of α 1 -AR-, β 1 -AR-, β 2 -AR-, MCT1and MCT4-positive green fluorescence pixels with fluorescence intensity above the threshold of 10% (α 1 -/β 1 -/ β 2 -AR-positive cell area) or 20% (MCT1 and MCT4-positive cell area) of maximal fluorescence and the number of all pixels (total cell area) per cell cross-sectional area were determined for each cell separately in RFP-TDP-43 wt -and RFP-TDP-43 208-414 -expressing astrocytes using ZEN software (Zeiss, Oberkochen, Germany). The percentage of the AR-and MCT-positive cell area relative to the total cell area was calculated for RFP-TDP-43 wt -and RFP-TDP-43 208-414 -expressing astrocytes for individual receptor and transporter type. The Mann-Whitney U test was performed to determine statistical significance between the experimental groups. P < 0.05 was considered significant.
To determine the extent of DAPI and TDP-43-colabeled nuclei in non-transfected astrocytes (with no visible RFP signal) adjacent to the RFP-TDP-43 wt -and RFP-TDP-43 208-414 -expressing astrocytes (with visible RFP signal) we measured the percentage of DAPI and TDP-43-colabeled nuclei per all DAPI nuclei in neighboring non-transfected cells (with no visible RFP signal) in the RFP-TDP-43 wt and RFP-TDP-43 208-414 experimental groups. In the control non-transfected group, we analyzed the nuclei of all cells. Some partially visible DAPI-stained nuclei at the edges of the confocal images were also taken into account in the analysis, however, only if no RFP signal in or around the visible part of the nuclei was observed. Kruskal-Wallis one-way ANOVA on ranks, followed by Dunn's test was performed to determine statistical significance between the experimental groups. P < 0.05 was considered significant.
Lipid droplet staining and data analysis. Control (non-transfected) astrocytes and astrocytes transfected with RFP-TDP-43 wt or RFP-TDP-43 208-414 plasmid were incubated in fresh growth medium for 24 h. Cells were then fixed in 4% formaldehyde in PBS for 5 min and stained with 1 µg/ml BODIPY 493/503 (Molecular Probes by Life Technologies, Thermo Fisher Scientific, Massachusetts, USA), a fluorescent LD marker, for 6 min at room temperature. Excess dye was washed off and the coverslips were mounted onto glass slides using SlowFade antifade reagent with DAPI (Molecular Probes by Life Technologies, Thermo Fisher Scientific, Massachusetts, USA) and carefully sealed. Stained cells were imaged with an inverted Zeiss LSM780 confocal microscope with a Plan apochromatic 40×/1.4 oil immersion objective (Carl Zeiss, Jena, Germany) using 488-nm Ar-ion, 543-nm He-Ne and 405-nm diode laser excitation. Emission spectra were acquired sequentially with 505-to 530-nm bandpass (BODIPY 493/503 ), 560-nm long pass (TagRFP 584 ) and 445-to 450-nm bandpass (DAPI) emission filters.
To determine the LD content in individual astrocytes, the number of BODIPY 493/503 -positive green fluorescence pixels with fluorescence intensity above the threshold of 20% of maximal fluorescence (Lipid droplet S) and the number of all pixels (Cell S) were determined for each cell separately in control, RFP-TDP-43 wt -and RFP-TDP-43 208-414 -expressing astrocytes in ZEN software (Zeiss, Oberkochen, Germany). The values obtained were normalized to the average LD content in control non-transfected cells. The mean number of BODIPY 493/503 -labeled LDs per cell and the mean LD perimeter were determined in cross-sections of individual cells using ImageJ, Analyze Particles function after applying 20% threshold and signal intensity (watershed) segmentation on individual images. Then, assuming that all LDs are spherical, the mean LD diameter was estimated from the perimeter values with the equation d = C/π, where d is the diameter and C is the LD perimeter. Kruskal-Wallis one-way ANOVA on ranks, followed by Dunn's test was performed to determine statistical significance between the groups. P < 0.05 was considered significant.

FRET measurements of intracellular cAMP and lactate levels and data analysis. Cells
co-expressing the FRET-based nanosensor Epac1-camps or Laconic and RFP-tagged TDP-43 pDNA construct (RFP-TDP-43 wt or RFP-TDP-43 208-414 ) were examined 24-30 h after transfection with a fluorescence microscope Zeiss Axio Obsever.A1 (Zeiss, Oberkochen, Germany), with a CCD camera and monochromator Polychrome V (Till Photonics, Graefelfing, Germany) as a monochromatic source of light with a wavelength 436 nm/10 nm. Dual emission intensity ratios were recorded using an image splitter (Optical Insights, Tucson, AZ, USA) and two emission filters (465/30 nm for CFP [cyan fluorescent protein] or mTFP [monomeric teal fluorescent protein] and 535/30 nm for YFP [yellow fluorescent protein] or Venus). Images were acquired every 3.5 s for Epac1-camps and 10 s for Laconic. Exposure time was 100 ms.
In some experiments astrocytes co-expressing Laconic and RFP-TDP-43 wt or RFP-TDP-43 208-414 were examined 24 h after transfection with a fluorescence microscope (Zeiss Axio Observer.A1 (Zeiss, Oberkochen, Germany)) with a Axiocam 702 camera and Colibri.2 Lamp Module (Zeiss, Oberkochen, Germany) as a source of light with a wavelength 433 nm. Dual emission intensity ratios were recorded using an image splitter (Optical Insights, Tucson, AZ, USA) and two emission filters (469-491 nm for ECFP and 530-4095 nm for EYFP). Images were acquired at intervals of 10 s with exposure time of 100 ms.
Coverslips with transfected astrocytes were mounted in a superfusion recording chamber on the microscope stage. Imaging was performed at room temperature (22-24 °C). One cell per experiment was recorded. The FRET signals (CFP/YFP (Epac1-camps) and mTFP/Venus (Laconic) signals) were obtained from the integration of the ratio signal over the entire cell using Life Acquisition software (Till Photonics, Graefelfing, Germany) or ZEN (Carl Zeiss, Jena, Germany). In the graphs, the FRET signal was reported as the ratio of the CFP/YFP (Epac1-camps) and mTFP/Venus (Laconic) fluorescence signals after subtracting the background fluorescence from the individual fluorescence signals using Excel (Microsoft, Seattle, WA, USA). The values of the FRET ratio signals were normalized to 1.0. An increase in the FRET ratio signal reflects an increase in [cAMP] i or [lactate] i .
Before the experiments, astrocytes were kept in extracellular solution containing sodium bicarbonate for 10 min (10 mM Hepes/NaOH [pH 7.2], 3 mM d-glucose, 135.3 mM NaCl, 1.8 mM CaCl 2 , 2 mM MgCl 2 , and 5 mM KCl, 0.5 mM NaH 2 PO 4 ·H 2 O, 5 mM NaHCO 3 ), and then treated with 100 µM NA (non-selective AR agonist) or 100 µM Iso (selective β-AR agonist) following a 100-to 200-s baseline. Experiments were conducted with the addition of a bolus solution; 200 µl of extracellular solution containing NA was added by pipette to 200 µl of bath solution in the recording chamber. The application of a control bolus solution without reagents had no significant effect on the FRET signal, as reported previously 82,83 . Extracellular solution osmolality was 295-305 mOsm, measured with the Osmomat 030 freezing point osmometer (Gonotech GmbH, Germany).
In experiments with Epac1-camps, single-exponential increases to maximum functions (F = F 0 + c × (1 − exp(−t/τ))) were fitted to the FRET ratio signals using SigmaPlot. The time constant (τ) and amplitude changes in the FRET ratio signal (ΔFRET [%]) were determined from the fitted curves. F is the FRET ratio signal at time t, F 0 is the baseline FRET ratio signal, c is the FRET ratio signal amplitude of F − F 0 , and τ is the time constant of the individual exponential component. In experiments with Laconic, the maximal (initial) rates of the FRET ratio signal increase (ΔFRET/Δt) were calculated as the slope of the linear regression function (ΔFRET [%] = slope [%/min] × Δt [min]) fitting the initial FRET ratio signal increase. In these experiments, changes in the FRET ratio signal (ΔFRET [%]) were calculated by subtracting the mean maximal FRET ratio signals from the mean baseline FRET ratio signals.
Unless stated otherwise, the Student's t-test was performed to determine statistical significance between the experimental groups; P < 0.05 was considered significant.

Fluo-4 AM measurements of cytosolic Ca 2+ and data analysis.
Astrocytes expressing RFP-TDP-43 wt or RFP-TDP-43 208-414 were incubated for 30 min at room temperature in medium containing 2 µM Fluo-4 AM dye (Molecular Probes, Invitrogen) and then transferred to dye-free medium for at least 30 min before the experiments to allow for cleavage of the AM ester group. The cells were excited with an Argon-ion laser at 488 nm, and time-lapse images were obtained every 3 s with an inverted Zeiss LSM780 confocal microscope and an 20x objective. Emission light was acquired with a 505-530-nm band-pass emission filter. Fluo-4 AMlabeled astrocytes were after 100 s baseline stimulated with 100 μM NA. Total recording time was ~400 s (135 frames). In individual cells, Fluo-4 AM intensity was quantified within a region of interest and expressed as the relative change in fluorescence: ΔF/F 0 = (F − F 0 )/F 0 , where F 0 denotes the baseline fluorescence level after subtraction of background fluorescence. The peak and cumulative increase in Fluo-4 AM ΔF/F 0 were determined using Microsoft Excel.
Mann-Whitney U test was performed to determine statistical significance between the experimental groups. P < 0.05 was considered significant.

Data availability
All data generated or analyzed during this study are included in this published article and its Supplementary Information files or are available from the corresponding author on reasonable request.