Cytoplasmic Expression of the ALS/FTD-Related Protein TDP-43 Decreases Global Translation Both in vitro and in vivo

TDP-43 is a major component of cytoplasmic inclusions observed in neurodegenerative diseases like frontotemporal dementia (FTD) and amyotrophic lateral sclerosis (ALS). To further understand the role of TDP-43 in mRNA/protein metabolism and proteostasis, we used a combined approach with cellular and animal models overexpressing a cytoplasmic form of human TDP-43 (TDP-43-ΔNLS), recapitulating ALS/FTD features. We applied in HEK293 cells a method for labeling de novo translation, surface sensing of translation (SUnSET), based on puromycin (PURO) incorporation. While control cells displayed robust puromycilation, TDP-43-ΔNLS transfected cells exhibited reduced ongoing protein synthesis. Next, by using a transgenic mouse overexpressing cytoplasmic TDP-43 in the forebrain (TDP-43-ΔNLS mice) we assessed whether cytoplasmic TDP-43 regulates global translation in vivo. Polysome profiling of brain cortices from transgenic mice showed a shift toward non-polysomal fractions as compared to wild-type littermates, indicating a decrease in global translation. Lastly, cellular level translational assessment by SUNSET was performed in TDP-43-ΔNLS mice brain slices. Control mice slices incubated with PURO exhibited robust cytoplasmic PURO signal in layer 5 neurons from motor cortex, and normal nuclear TDP-43 staining. Neurons in TDP-43-ΔNLS mice slices incubated with PURO exhibited high cytoplasmic expression of TDP-43 and reduced puromycilation respect to control mice. These in vitro and in vivo results indicate that cytoplasmic TDP-43 decreases global translation and potentially cause functional/cytotoxic effects as observed in ALS/FTD. Our study provide in vivo evidence (by two independent and complementary methods) for a role of mislocalized TDP-43 in the regulation of global mRNA translation, with implications for TDP-43 proteinopathies.

TDP-43 is a major component of cytoplasmic inclusions observed in neurodegenerative diseases like frontotemporal dementia (FTD) and amyotrophic lateral sclerosis (ALS). To further understand the role of TDP-43 in mRNA/protein metabolism and proteostasis, we used a combined approach with cellular and animal models overexpressing a cytoplasmic form of human TDP-43 (TDP-43-NLS), recapitulating ALS/FTD features. We applied in HEK293 cells a method for labeling de novo translation, surface sensing of translation (SUnSET), based on puromycin (PURO) incorporation. While control cells displayed robust puromycilation, TDP-43-NLS transfected cells exhibited reduced ongoing protein synthesis. Next, by using a transgenic mouse overexpressing cytoplasmic TDP-43 in the forebrain (TDP-43-NLS mice) we assessed whether cytoplasmic TDP-43 regulates global translation in vivo. Polysome profiling of brain cortices from transgenic mice showed a shift toward non-polysomal fractions as compared to wild-type littermates, indicating a decrease in global translation. Lastly, cellular level translational assessment by SUNSET was performed in TDP-43-NLS mice brain slices. Control mice slices incubated with PURO exhibited robust cytoplasmic PURO signal in layer 5 neurons from motor cortex, and normal nuclear TDP-43 staining. Neurons in TDP-43-NLS mice slices incubated with PURO exhibited high cytoplasmic expression of TDP-43 and reduced puromycilation respect to control mice. These in vitro and in vivo results indicate that cytoplasmic TDP-43 decreases global translation and potentially cause functional/cytotoxic effects as observed in ALS/FTD. Our study provide in vivo evidence (by two independent and complementary methods) for a role of mislocalized TDP-43 in the regulation of global mRNA translation, with implications for TDP-43 proteinopathies.
Although the most extensively described function of TDP-43 has been its role in mRNA splicing, there has been reports consistent with a role for this protein in translational regulation. A proteomic study in HEK293 cells showed that TDP-43 interact extensively with proteins of the translational machinery (Freibaum et al., 2010) and in vitro downregulation of TDP-43 alters global translational yield through a SKAR-dependent mechanism (Fiesel et al., 2012). Moreover, the ribosomal protein RACK1 seems to be a key mediator of the interaction of TDP-43 with the translational machinery (Russo et al., 2017).
Using inducible transgenic mice, we demonstrated that expression of a cytoplasmic form of TDP-43 (termed TDP-43-NLS) recapitulates multiple features of TDP-43 proteinopathies of the ALS/FTD spectrum, including progressive neurodegeneration, gliosis, global gene expression changes and behavioral abnormalities encompassing the motor, cognitive and social domains (Igaz et al., 2011;Alfieri et al., 2014). Other animal models with cytoplasmic TDP-43 expression, covering from Drosophila to mammalian systems, also support the idea that this mislocalization event might be pathogenically relevant (Miguel et al., 2011;Walker et al., 2015;Wang et al., 2015;Yin et al., 2019).
While most studies have focused on the regulation of global or specific translation using cell lines or primary neurons in culture, the consequences of in vivo TDP-43 manipulation (specifically, of its cytoplasmic form) are virtually unknown. In this work, we investigated both in vitro and in vivo if cytoplasmic expression of TDP-43 modifies global protein synthesis, using two different approaches, polysome profiling and surface sensing of translation (SUnSET) technique.

Immunocytochemistry
Cells on coverslips were fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) for 10 min, permeabilized with 0.2% Triton X-100 (Sigma) in PBS for 10 min, blocked with 5% powdered milk in PBS for 2 h, and incubated overnight with primary antibody at 4 • C. Primary antibodies used in this study (diluted in 5% powdered milk in PBS) were as follows: antitotal-TDP-43 [1:5000, rabbit polyclonal anti-TDP-43 1038C ] and anti-puromycin (1:1000 dilution, MABE343, clone 12D10, Millipore). Primary antibodies were visualized with secondary antibodies conjugated with Alexa Fluor 488 and Alexa Fluor 594 (1:500, donkey anti-mouse and donkey antirabbit IgG, respectively, Jackson Immunoresearch), and nuclei were detected using Hoescht 33258 (Sigma). Coverslips were mounted in gelatin-coated slides using Fluoro-gel (Thermo Fisher Scientific) and stored at 4 • C. Images were obtained with a Zeiss AxioImager 2 microscope equipped with APOTOME 0.2 structured illumination, using a Hamamatsu Orca Flash 4.0 camera. Four areas per image were analyzed using ImageJ software (version 1.52a). Mean puromycin and TDP-43 signals were calculated and normalized to empty vector (+PURO) group signal.
Animals hTDP-43-NLS line was generated as described in Igaz et al. (2011). Monogenic tetO-TDP-NLS mice were bred to Camk2a-tTA mice (Jackson Laboratory) generating non-transgenic, tTA monogenic, single tetO-TDP-43 transgenic mice, and bigenic mice expressing hTDP-43-NLS. Breeding mice and pups were treated with 0.2 mg/ml doxycycline hyclate (DOX; sc-204734A; Santa Cruz Biotechnology) in drinking water, to avoid pre-natal and post-natal developmental effects. Transgene expression was activated at weaning (post-natal day 28) by removing DOX from water and mice were analyzed after ∼1 month (at post-natal day 60). Tail DNA was used for genotyping, employing primers described in Igaz et al. (2011). The experimental protocol for this study was approved by the National Animal Care and Use Committee of the University of Buenos Aires (CICUAL). All animal procedures were performed according the NIH Guide for the Care and Use of Experimental Animals.

Sucrose Gradient Preparation
Cortices of control (non-transgenic) or hTDP-43-NLS animals were homogenized in cold lysis buffer [15 mM Tris-HCl, pH 8.0, 300 mM NaCl, 15 mM MgCl 2 , 1% Triton X-100, 1 mg/ml heparin, protease inhibitors (Roche) and 0.1 mg/ml cycloheximide (Sigma-Aldrich Inc.)] at a ratio of 500 µl buffer /10 mg tissue. Homogenates were then placed on ice for 10 min and centrifuged at 5,000 rpm for 10 min at 4 • C. Supernatant, consisting in cytosolic extract containing polysomes, was collected and protein concentration was determined by bicinchoninic acid (BCA) method (Pierce). Sucrose gradient, ranging from 15 to 50%, were prepared in gradient buffer (15 mM Tris-HCl, pH 8.0, 300 mM NaCl, 5 mM MgCl 2 , 1 mg/ml heparin and 0.1 mg/ml cycloheximide). Briefly, 2.4 ml 50% sucrose solution was added to a tube and chilled for 30 min at −80 • C. The procedure was repeated with layers of sucrose solution at lower concentration (41.25, 32.5, 23.75, and 15%) and stored at 4 • C if used the next day. 0.5 to 1 mg of cytosolic protein extract were added to each gradient, and centrifuged in a SW40 rotor at 35,000 rpm for 2-4 h, at 4 • C. Five hundred microliters aliquots were separated and absorbance at 260 nm was measured (Eppendorf Biophotometer Plus, Germany). A peak area ratio between polysomal (P) and non-polysomal (NP, comprising RNPs and monosomes) fractions was determined for each group.

Western Blot
Five hundred microliters of sample from each fraction was subjected to methanol/chloroform protein extraction method and resuspended in sample buffer. Equal amounts of samples were resolved on 10% sodium dodecyl Sulfate polyacrilamyde gel electrophoresis (SDS-PAGE) and transferred onto PVDF membranes (Immobilon-P; Millipore). After transfer, membranes were stained with Ponceau S for total protein load assessment, then blocked in 5% powdered milk in 0.1% Tween-TBS, and incubated in anti-total TDP-43 primary antibody overnight (1038C, 1:20000). Detection of primary antibody was performed with horseradish peroxidaseconjugated secondary antibody (Jackson Immunoresearch), and ECL was used for blot development. Digital images were acquired using a G-BOX Imager (Syngene). Ponceau S staining was used as a control for equal protein loading in our immunoblot analysis of polysomal fractions, since typical housekeeping genes are barely detectable in polysome fractions (Reschke et al., 2013).

Tissue Processing
SUnSET labeling in brain slices was adapted from Cook et al. (2014), with some modifications. Brain coronal slices were prepared from 8-week old (1 month off-DOX) hTDP-43-NLS or control mice littermates. Slices were cohort and age matched. Mice were anesthetized with isoflurane and decapitated; brain was removed and chilled ice-cold in 95 O 2 -5% CO 2 -saturated, sucrose-based artificial cerebrospinal fluid (ACSF, 220 mM sucrose, 2.5 mM KCl, 1 mM CaCl 2 , 6 mM MgCl 2 , 1.25 mM NaH 2 PO 4 , 26 mM NaHCO 3 , and 10 mM dextrose). The brain was next glued onto a vibratome stage and acute sagittal slices (400 µm) were cut. Then, the slices were recovered at room temperature for an hour in ACSF (125 mM NaCl, 2.5 mM KCl, 2 mM CaCl 2 , 1.1 mM MgCl 2 , 1.25 mM NaH 2 PO 4 , 26 mM NaHCO 3 , and 10 mM glucose, pH 7.4). After recovery, slices were placed on a fine grid mesh in a petri dish and incubated in ACSF (C-group) or ACSF supplemented with 5 µg/µl puromycin (P7255, Sigma-Aldrich Inc.; C+ or NLS+ groups) at a final volume of 30 ml. Incubation time was 45 min under continuous aeration with carbogen (95% O 2 / 5% CO 2 ). Slices were then rinsed in ACSF and fixated in 4% paraformaldehyde (Sigma-Aldrich Inc.) in 0.1 M PBS for 24 h at 4 • C. Next day, slices were washed three times in 0.01 M PBS for 10 min each under continuous agitation in an orbital shaker, incubated with 10% sucrose in 0.01 M PBS for 2 h at 4 • C and stored in 30% sucrose-0.01 M PBS at −20 • C until use.

Immunofluorescence
Brain slices containing the motor cortex [approximately bregma 1.18 mm, (Paxinos and Franklin, 2008)] were selected and washed three times in 0.01 M PBS for 10 min. Permeabilization step was performed with 0.01 M PBS-1% Triton X-100 for 1 h at room temperature. Next, slices were incubated with blocking solution containing 5% normal goat serum in 0.01 M PBS-0.3% Triton X-100 for 1 h at room temperature. Sections were immunolabeled with rabbit polyclonal anti-TDP-43 (1038C, 1:15000) and mouse monoclonal anti-puromycin (MABE343 clone 12D10, Millipore, 1:4000) as primary antibodies; incubation was performed first at room temperature for 3 h followed by 48 h at 4 • C. Staining against the neuronal marker NeuN (mouse monoclonal MAB377, Millipore, 1:1000) or human TDP-43 protein (mouse monoclonal 60019-2, Proteintech, 1:10000) was performed in 50 µm slices as described in Alfieri et al. (2014). Then, slices were washed three times for 10 min each with 0.01 M PBS-0.04% Triton X-100 and incubated with Alexa Fluor 594 goat anti-rabbit IgG and Alexa Fluor 488 goat anti-mouse IgG (each at 1:500 dilution, Jackson laboratories) for 4 h at room temperature in darkness. Slices were washed three times for 10 min with 0.01 M PBS-0.04% Triton X-100 and subjected to Hoescht staining (1:2000 dilution in 0.01M PBS) for 30 min at room temperature in darkness. Finally, sections were washed in 0.01 M PBS two times (10 min each) and mounted in gelatin-coated glass slides with 30% glycerol in 0.01 M PBS, covered with glass coverslips, sealed and stored in darkness at 4 • C until use. All incubation and wash steps were performed in a 12-well plate under continuous agitation in an orbital shaker.

Image Detection and Analysis
Images were obtained with a Zeiss AxioImager 2 microscope equipped with APOTOME.2 structured illumination, using a Hamamatsu Orca Flash 4.0 camera. Two sections of layer 5 cells of motor cortex were analyzed using ImageJ software (version 1.52a). In non-transgenic (control, C) mice incubated with puromycin (C+), each cellular area was defined in the green channel corresponding to puromycin signal, as well Frontiers in Cellular Neuroscience | www.frontiersin.org FIGURE 1 | (B) Representative micrographs of total TDP-43 (red) and puromycin (green) IF staining. Control (empty vector) cells treated with vehicle (-PURO) served to assess non-specific staining and no puromycin signal was detected, while control (+ PURO) cells exhibited robust puromycilation. In both control groups, TDP-43 labeling was nuclear (as expected for endogenous expression), while TDP-43-NLS transfected cells showed intense, cytoplasmatic expression. A lower exposure for the TDP-43 channel is also shown. Puromycin staining is visibly decreased in TDP-43-NLS transfected cells (arrows). Scale bar: 10 µm. (C) Quantitative analysis of puromycilation signal revealed that cells with cytoplasmic TDP-43 (NLS+) exhibit a decreased ongoing translation. Staining intensity of TDP-43-NLS immunopositive cells show significant differences when compared to puromycin-labeled control cells and non-transfected (NLS-) cells. TDP-43 levels were significantly higher in NLS+ than in the control or non-transfected groups (as expected for transgene overexpression). Data represent mean ± SEM; n = 75 cells per condition (from three replicates). One-way ANOVA/Tukey's multiple comparison test: ****P < 0.0001, ***P < 0.001, **P < 0.01.
as background areas between cells; the same approach was employed in hTDP-43-NLS animals incubated with puromycin ( NLS+) but in the red channel (total TDP-43 signal). For non-transgenic mice incubated in ACSF (C-), a circular area centered in each red nucleus and background areas between cells were quantified. Background signal was normalized to individual cell areas, and a ratio between raw cell intensity signal and normalized background signal was calculated. The signal from individual cell populations and averaged signal per animal were compared between experimental groups.

Statistical Analysis
Data are expressed as means ± SEM. Statistical analysis was performed using a standard unpaired two-tailed student's t-test when comparing only two groups on one measure. A oneway ANOVA followed by Tukey's multiple-comparison post hoc test was used when comparing three groups. Prism Version 8 (GraphPad Software) was used for statistical analysis and differences were considered to be statistically significant when the probability value was < 0.05.

RESULTS
As a first approach to study at the cellular level the effect of cytoplasmic TDP-43 expression on protein synthesis, we adapted the SUnSET methodology (Schmidt et al., 2009) to measure single-cell puromycilation by immunocytochemistry (Figure 1). Puromycin incorporation into nascent polypeptides enables direct visualization of actively translating ribosomes. We transfected either empty vector or TDP-43-NLS in HEK293 cells and, during the last 20 min of the experiment, applied puromycin (+ PURO) or vehicle (-PURO) to the cultures ( Figure 1A). Representative micrographs show immunofluorescence staining using antibodies against total TDP-43 or puromycin ( Figure 1B). Control (empty vector) cells treated with vehicle (-PURO) were used to assess non-specific staining and showed virtually no puromycin signal, while control (+PURO) cells displayed robust puromycilation. As expected, TDP-43 staining showed nuclear, endogenous TDP-43 for both control groups, while TDP-43-NLS transfected cells demonstrate abundant, cytoplasmic expression in a proportion of HEK293 cells. Remarkably, analysis of puromycilation signal in TDP-43-NLS transfected cells demonstrates that cells with cytoplasmic TDP-43 (NLS + ) show a decrease in ongoing protein synthesis (arrows, Figure 1B). Quantification of puromycin staining intensity in TDP-43-NLS positive cells show significant differences against both the puromycin-labeled control cells and TDP-43-NLS negative (non-transfected) cells ( Figure 1C). As expected, TDP-43 levels are significantly higher in TDP-43-NLS cells than in the control or non-transfected groups ( Figure 1C). Overall, these results demonstrate at cellular resolution that cytoplasmic TDP-43 decreases ongoing protein synthesis in cultured cells. Next, we wanted to extrapolate these results in vivo, taking advantage of our inducible human TDP-43-NLS mouse model with CamKIIα-driven transgene expression enriched in forebrain neurons (Igaz et al., 2011). We first aimed to study global protein synthesis by polysome profiling (Figure 2), a method that allows for translational state assessment by differential separation of mRNA-protein complexes in a sucrose gradient (Chasse et al., 2017). TDP-43-NLS or control mice were generated (Figure 2A) and transgene induction was started at P28 to avoid developmental deficits. Animals were sacrificed at P60 (after 1 month of transgene expression), a time point when these mice already develop ALS/FTD-like phenotypes (Igaz et al., 2011;Alfieri et al., 2014;Walker et al., 2015) ( Figure 2B). Immunofluorescence analysis of TDP-43-NLS mice demonstrate robust cytoplasmic transgene expression in cortical neurons, as revealed by an antibody that specifically stains human TDP-43 (i.e., the transgene product) but not endogenous mouse TDP-43 ( Figure 2C). Polysome profiles from control (non-transgenic) or TDP-43-NLS mice cortical homogenates show an increase in the free ribonucleoprotein (RNP) fraction (low density) and a decrease in translationally active fractions of high density (polysomes) in transgenic mice ( Figure 2D). Interestingly, immunoblots using total TDP-43 antibody show that overexpression of TDP-43-NLS leads to an increased presence of TDP-43 in most gradient fractions, but especially in those with higher density. Quantification of the ratio between areas under the curve for polysomal and nonpolysomal (i.e., non-translationally active) fractions demonstrate a significant reduction in transgenic mice (Figure 2E), indicating that cortical global protein synthesis is reduced in vivo due to cytoplasmic TDP-43 expression.
Our polysome profiling data strongly suggest that global translation was decreased in the brains of TDP-43 mice, but this technique does not allow for examination of cellspecific populations. To elucidate if TDP-43-NLS mice display altered ongoing protein synthesis in vivo at the neuronal level, we applied the SUnSET method in slices from control or transgenic mice and studied puromycilation levels in neurons from layer 5 of the motor cortex (Figure 3). We evaluated protein synthesis after 1 month of TDP-43-NLS expression, using acute thick slices (400 µm) from coronal vibratome FIGURE 2 | Transgenic TDP-43-NLS mice have an altered polysome profile in brain cortex. (A) Generation of hTDP-43-expressing mice. CamKIIa/tTA mice are crossed with tetO/hTDP-43-NLS mice. In bigenic mice, doxicycline (DOX) inhibits tTA binding to the tetracycline-responsive element (tetO), thus repressing hTDP-43 expression. Non-transgenic littermates were used as controls. (B) Timeline and experimental workflow. Mating and breeding of animals was carried in the presence of DOX to avoid developmental effects. Transgene expression was induced at weaning (P28) by removing DOX from water (OFF DOX), and mice were analyzed at ∼1 month post-induction (P60). Then, brain cortices were subjected to polysome profiling. (C) Representative micrographs of human-specific TDP-43 antibody IF in control (Ctrol) and TDP-43-NLS mice at 1 month off DOX. Transgenic animals show robust cytoplasmic expression of the transgene in cortical neurons. Hoescht 33258 was used as a nuclear marker. Scale bar: 10 µm. (D) Representative cortical homogenate polysome profiles in non-transgenic (Ctrol) and transgenic (TDP-43-NLS) mice. In TDP-43-NLS animals, the free ribonucleoprotein (RNP) fraction is increased while the translationally active fractions (polysomes, P) are decreased. Immunoblots against total TDP-43 using protein isolated from the individual fractions revealed that overexpression of TDP-43-NLS increases the presence of total TDP-43, especially in the polysomal fractions. Ponceau S was used as a loading control. (E) Quantification of the ratio between the areas under the curve for polysomal (P) and non-polysomal (NP) fractions. Transgenic mice display a significant reduction, indicating that global translation in brain cortex is decreased due to cytoplasmic TDP-43 expression. Data represent mean ± SEM; n = 5 for Ctrol, n = 3 for TDP-43-NLS. *P < 0.05, significantly different from Ctrl mice (Student's t-test).
quantification. Low magnification immunofluorescence images form Layer 5 of the motor cortex show that non-transgenic, control mice (Ctrol) incubated with vehicle (ACSF) instead of puromycin (-PURO) showed normal nuclear staining of endogenous TDP-43 ( Figure 3B). As expected, anti-PURO signal was negligible, although some non-specific staining of blood vessels can be observed (this is likely due to remaining IgG in the slices reacting with the anti-PURO Ab raised in mice). Slices from control mice incubated with puromycin display robust cytoplasmic signal in most cells, indicating proper labeling of ongoing protein synthesis. As expected, TDP-43 endogenous levels are also nuclear and with intensity comparable to control (-PURO) group ( Figure 3B). TDP-43-NLS mice slices incubated with PURO show the expected pattern of increased cytoplasmic expression of TDP-43, while these cells display a reduction in puromycilation ( Figure 3B). A high-magnification view of cortical motor neurons from layer 5 clearly show that cells overexpressing cytoplasmic TDP-43 display reduced levels of anti-PURO signal ( Figure 3C). Neuronal identity of cells with cytoplasmic TDP-43 was corroborated by NeuN and TDP-43 double immunofluorescence on thin cortical slices from TDP-43-NLS mice ( Figure 3D). Quantification at the cellular level demonstrates that TDP-43-NLS mice display a significant increase in TDP-43 signal intensity compared to both control groups, and that PURO treatment does not affect TDP-43 levels ( Figure 3E). Cellular analysis of puromycilation in those same neurons exhibits the expected significant increase in control groups treated with (+ PURO) vs. without (-PURO) puromycin. Importantly, neurons in TDP-43-NLS mice display a drastic reduction in protein synthesis as revealed by anti-PURO signal ( Figure 3E). These results are also evident when we pool data from each individual animal, with a significant reduction of puromycin signal in neurons from transgenic mice ( Figure 3F).

DISCUSSION
Protein synthesis is an essential process for all cellular types, but neurons are particularly sensitive to proper control of this metabolic process due to their high polarization and the demands of plasticity mechanisms (Kapur et al., 2017). Several studies have suggested a role for translational dysfunction in a wide range of neurodegenerative diseases, including Alzheimer's, tauopathies, Huntington's, Prion disease, and the ALS/FTD spectrum (Ding et al., 2005;Halliday et al., 2015;Sun et al., 2015;Meier et al., 2016;Russo et al., 2017;Kamelgarn et al., 2018;Zhang et al., 2018;Joag et al., 2019;Koren et al., 2019).
In this report, we show both in cultured cells and, more importantly, in cortical neurons in vivo that cytoplasmic TDP-43 leads to altered protein synthesis, expanding our understanding of the possible role of this RNA-binding protein (RBP) in ALS/FTD pathogenesis.
There are other examples of RBPs involved in neurological disease (Thelen and Kye, 2019), including FMRP (related to Fragile X syndrome) and FUS (associated with ALS/FTD) (Donlin-Asp et al., 2017;Zhou et al., 2018). FMRP regulates activity-dependent translation of its mRNA cargoes (Davis and Broadie, 2017), and some of its functions are performed in cooperation with other RBPs, including TDP-43 and Staufen 1 (Chu et al., 2019). FUS is similar in several aspects to TDP-43, and recent reports demonstrated that diseaserelated mutations in FUS can lead to suppression of global (Kamelgarn et al., 2018) or axonal translation (Lopez-Erauskin et al., 2018). However, most of these studies exploring the relationship between neurodegenerative disease-related proteins and translation were performed either in cultured cell lines or in primary cultures derived from different transgenic mouse models. A remarkable in vivo study recently showed that TDP-43 and DISC1 co-aggregate in frontotemporal lobar degeneration human samples and animal models, leading to disruption of dendritic local translation and aberrant behavior (Endo et al., 2018).
We show here that TDP-43-NLS expression reduces the amount of puromycin incorporated into translating peptides, and found that although TDP-43-NLS is abundantly localized to polysome fractions, there are fewer polysome-associated mRNAs. This raises the possibility that mislocalized TDP-43 may disrupt mRNA recruitment to polysomes. In addition, the direction of change in translation opposes that seen with TDP-43 knockdown (Fiesel et al., 2012), suggesting that rather than loss of function, there is a gain of toxic function that might be mediated by TDP-43 abnormal expression/localization. Further studies investigating these possibilities are warranted.
What could be the potential players involved in this translational effect mediated by TDP-43-NLS? A report using SH-SY5Y cells has shown that cytoplasmic TDP-43 interacts with both RACK1 and ribosomal protein S6 (rpS6) as demonstrated by immunoprecipitation, and that TDP-43-NLS co-localizes in the cytoplasm with RACK1 and rpS6, suggesting that TDP-43 associates to translational machinery (Russo et al., 2017). RACK1 is an associated ribosomal protein and functions as a docking site for several proteins on the translational machinery; in addition, it has been shown to accumulate in TDP-43 positive inclusions in motor neurons of ALS spinal cord cases (Russo et al., 2017). Other potential candidates are those identified in a proteomic study using HEK293T cells in which a main cluster of TDP-43interacting proteins include a cytoplasmic "translational" cluster, which contains multiple translation initiation and elongation factors, and ribosomal subunits (Freibaum et al., 2010). Although it is a predominantly nuclear protein, it has been previously shown that TDP-43 shuttles between the nucleus and cytoplasm (Winton et al., 2008), and alteration of cytoplasmic TDP-43 levels might lead to abnormal formation of complexes involved in translation, due to an altered stoichiometric ratio between TDP-43 and its protein and RNA partners. Lastly, TDP-43 has been shown to regulate the alternative splicing of SKAR (Fiesel et al., 2012), a protein that positively affects cap-dependent protein translation (Ma et al., 2008). TDP-43 downregulation leads in vitro to a shift in SKAR α/β isoform ratio, which in turn increases global translation (Fiesel et al., 2012). Interestingly, Yang et al. (2014) used a mouse model to show that partial TDP-43 knockdown increased SKAR isoforms with exon 3 excluded, leading to a lower α/β isoform ratio in brain and spinal cord due to altered alternative splicing. Different mechanisms induced by ALS/FTD-related proteins may lead to translational repression. Pathological forms of these proteins can trigger stress granule formation and/or can activate ER stress-induced pathways (i.e., PERK) leading to translational modulation events such as eIF2α phosphorylation (Cestra et al., 2017). High levels of misfolded proteins stimulate the Unfolded Protein Response (UPR), which tend to stabilize proteostasis and may lead to protein synthesis inhibition. Chronic, sustained activation of these pathways can lead to cell death (Li et al., 2006;Sano and Reed, 2013). ER stress markers have been detected in samples from ALS patients and animal models (Matus et al., 2013); remarkably, ER-stress induced HIPK2 activation and neurodegeneration has been described in TDP-43-NLS mice (Lee et al., 2016). Alleviation of disease-related phenotypes can be achieved in neurodegenerative disease models by pharmacological and genetic interventions which lead to protein translation restoration (Moreno et al., 2013;Halliday et al., 2017). In this line, targeting UPR signaling is an attractive goal, given the variety of proof-of-concept drugs that act at different steps, such as PERK activation (i.e., GSK2606414) or eIF2α signaling (ISRIB), potentially enabling translational recovery (Costa-Mattioli et al., 2007;Kim et al., 2014;Halliday et al., 2015;Radford et al., 2015).
Given that our TDP-43-NLS mouse model recapitulates several FTD/ALS-related behavioral phenotypes (Alfieri et al., 2014), our present results showing global translational inhibition suggest that this alteration can be part of a pathogenic mechanism affecting not only neuronal survival but also multiple behavioral domains. An important step in future studies will be to apply strategies that tend to restore protein synthesis levels and evaluate if the neurodegenerative and behavioral phenotypes of TDP-43 animal models can be ameliorated.
In conclusion, we have shown here by two different approaches that expression of cytoplasmic TDP-43 leads to a decrease in global translation. Although the role of TDP-43 in protein synthesis regulation has been discussed and explored recently, our in vivo results underscore its importance as a mechanistic link between the possible pathogenic molecular changes and the clinical and pathological manifestations of TDP-43 proteinopathies.

DATA AVAILABILITY STATEMENT
The datasets generated for this study are available on request to the corresponding author.

ETHICS STATEMENT
The animal study was reviewed and approved by National Animal Care and Use Committee of the University of Buenos Aires (CICUAL).

AUTHOR CONTRIBUTIONS
SC, LL, and LI designed the study. SC and AV performed the cell culture experiments. LL performed the polysome profiling and immunoblot experiments. SC performed in vitro and in vivo SUnSET studies, image processing, and data analysis. MB provided key reagents, funding, and helped design and analyze the in vitro experiments. SC and LI wrote the manuscript. LI conceived and supervised all aspects of the project and obtained study funding. All authors read and approved the final manuscript.

FUNDING
This work was supported by research grants to LI from University of Buenos Aires (UBACyT) and the Agencia Nacional de Promoción Científica y Tecnológica (ANPCyT) under grants PICT 2015-0975 and PICT 2017-2140 and to MB under grant PICT 2016-0130. LI and MB were members of Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET). SC and LL were supported by post-doctoral fellowships from CONICET, and AV by a doctoral fellowship from CONICET.