Region-specific features of early glial activation and Aquaporin-4 dysregulation in conditional mouse models of TDP-43 proteinopathies Authors: Gabriela Nieva, Florencia Vassallu, Amaicha Depino, Vanina Netti, Lionel Muller Igaz Published: 2026 ABSTRACT AbstractIntroductionAggregation and cytoplasmic mislocalization of TDP-43 are key features of several neurodegenerative diseases, including amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD). Neuroinflammatory processes mediated by glial cells play crucial roles in the pathophysiology of these and other diseases, defined as TDP-43 proteinopathies. Here, we characterized region-specific glial activation in two conditional mouse models: hTDP-43-WT (overexpressing nuclear wild-type human TDP-43) and hTDP-43-ΔNLS (expressing cytoplasmic TDP-43 with altered nuclear localization signal) following 1 month of transgene expression.Methods and ResultsImmunofluorescence analysis revealed distinct patterns of microglial activation across brain regions. hTDP-43-WT mice exhibited significant microgliosis in motor (MC) and somatosensory (SSC) cortices and hippocampal dentate gyrus (DG) with pronounced morphological alterations (i.e. increased soma size). Sholl analysis demonstrated reduced branching length and complexity in MC, SSC, and hippocampal subfields. hTDP-43-ΔNLS mice displayed more pronounced microglial activation in hippocampal regions (CA1, DG) compared to cortical areas, with significant increases in microglial density. Additionally, we observed region-specific cortical astrocytosis in both models, suggesting coordinated glial reactivity. hTDP-43-ΔNLS mice showed decreased polarization of astrocytic water channel Aquaporin-4 (AQP4) around vascular structures in SSC and hippocampal CA1/DG. The changes in AQP4 localization, which is critical for glymphatic function, support the hypothesis that this waste clearance system for the brain is altered in TDP-43 proteinopathies.ConclusionThese findings demonstrate that these different animal models of ALS/FTD induce distinct neuroinflammatory signatures, potentially contributing to the region-specific vulnerability observed in these diseases. Our data provide insights into early glial-mediated pathogenic mechanisms that could guide targeted therapeutic strategies for TDP-43 proteinopathies. Introduction Neurodegenerative diseases are a heterogeneous group of progressive disorders characterized by selective neuronal vulnerability, accumulation of misfolded proteins, and irreversible functional decline. Despite their clinical diversity, conditions such as Alzheimer’s disease (AD), Parkinson’s disease (PD), amyotrophic lateral sclerosis (ALS), and frontotemporal dementia (FTD) share convergent pathogenic features, most notably the formation of disease-specific protein aggregates that disrupt cellular homeostasis and neuronal integrity [1]. In ALS and a substantial proportion of FTD cases, mislocalization and aggregation of TAR DNA-binding protein 43 (TDP-43) represent a defining molecular hallmark [2], linking RNA dysregulation and proteostasis failure to neurodegeneration [3, 4]. Increasing evidence indicates that these proteinopathies are tightly coupled to innate immune activation within the central nervous system (CNS), where microglia and astrocytes undergo sustained phenotypic remodeling in response to chronic cellular stress [5]. Both constitutive and conditional mouse models expressing wild-type or mutant TDP-43 variants have demonstrated that mislocalization of TDP-43 from the nucleus to the cytoplasm disrupts RNA metabolism, induces protein aggregation, and provokes widespread neuroinflammation (i.e. Igaz et al. [6], see review by Armas et al. [7]). Neuroinflammation and reactive gliosis are now recognized as core components of neurodegenerative pathology across disease entities, detectable in patient biofluids, neuroimaging studies, postmortem tissue, and experimental models [1, 8]. Activated microglia and astrocytes can exert both neuroprotective and neurotoxic effects, influencing protein clearance, synaptic integrity, and neuronal survival in a context- and stage-dependent manner [5, 9]. Since the discovery of its involvement in neurodegenerative disease, the focus of research on TDP-43–driven pathology has been primarily on neuronal degeneration. However, more recent work has increasingly shifted to other cellular populations and processes, with growing evidence pointing to neuroinflammation and non-neuronal cells as active, early contributors to pathology, not mere consequences [10]. Glial cells, particularly microglia and astrocytes, are the primary orchestrators of this inflammatory response. Under pathological conditions, these cells undergo a transformation from homeostatic supporters to active participants in the disease process. However, their roles are complex, and their activation phenotypes are highly context-dependent [11, 12]. While the role of toxic astrocytes is well-established in SOD1-based animal models of ALS [10, 13], less is known about how TDP-43 pathology, which is relevant to most ALS cases, drives specific glial activation signatures. Pathological TDP-43 has been shown to engage inflammatory signaling pathways, including NF-κB, inflammasome activation, and cGAS–STING signaling [3, 4, 14], while genetic risk factors for ALS/FTD converge on immune regulatory pathways [4], supporting a bidirectional and self-reinforcing relationship between protein aggregation and neuroinflammation. Together, these observations identify gliosis not merely as a secondary response to neuronal loss but as an integral driver of disease progression and a potential therapeutic target in TDP-43–associated neurodegeneration [3]. To elucidate the role of TDP-43 in triggering glial activation, we used two conditional transgenic mouse models with different patterns of TDP-43 expression in forebrain neurons: one overexpressing wild-type human TDP-43 with preserved nuclear localization (tTA/WT) and a second one expressing a human TDP-43 variant lacking the nuclear localization signal, leading to pathological cytoplasmic localization (tTA/ΔNLS) [6]. We hypothesized that these two TDP-43 species would induce different and region-specific early microglial and astroglial activation signatures, which might represent an early marker of disease progression. The 1-month post-transgene induction time window was selected to capture sustained glial changes prior to overt degeneration in these inducible TDP-43 models. Moreover, as ΔNLS mouse models have been shown to present more varied and severe phenotypes than TDP-43 WT expressing ones [6, 15–17], we also analyzed if Aquaporin-4 (AQP4) was dysregulated in the brain after neuronal cytoplasmic TDP-43 expression. AQP4, the most abundant water channel in the brain, is an integral component of the glymphatic system. Its distribution is strongly polarized in the endfeet of astrocytes, which ensheathe the blood vessels in the brain [18]. This specific localization allows AQP4 to be in close contact with the perivascular space, thus facilitating CSF influx into the brain parenchyma, which is crucial for the effective functioning of the glymphatic system [19]. In recent years, many studies have reported the pathophysiological role that AQP4 plays in several CNS neurodegenerative disorders such as PD and AD by affecting the clearance function of the glymphatic system [19, 20] and the formation of protein aggregates such as α-synuclein (α-Syn) in PD and β-amyloid (Aβ) in AD, thus contributing to the pathogenesis of neurodegeneration. The loss of AQP4 perivascular localization (and the impairment of glymphatic flow) was reported in AD patients as well as in animal models [19]. Here, we assessed the status of AQP4 polarization in the brains of tTA/ΔNLS mice to evaluate a potential contribution of glymphatic dysfunction to phenotype development in this model of TDP-43 proteinopathies. Our results demonstrate that 1 month of wild-type hTDP-43 overexpression or cytoplasmic hTDP-43 expression is sufficient to induce significant cortical and hippocampal gliosis. Collectively, these findings emphasize the relevance of studying the early phases of the pathological processes associated with neurodegenerative conditions and suggest that modulating these early events might eventually contribute to the development of therapeutic approaches for TDP-43 proteinopathies. Materials and methods Animals All animal experiments were approved by the National Animal Care and Use Committee (CICUAL) of the School of Medicine, University of Buenos Aires (UBA). All efforts were made to minimize animal suffering and to reduce the number of animals used. Conditional transgenic mice expressing either wild-type human TDP-43 (hTDP-43-WT) or TDP-43 with a mutated nuclear localization signal (hTDP-43-ΔNLS) were generated as previously described (see (Alfieri et al. [15]; Alfieri et al. [16]; Igaz et al. [6]; Silva et al. [21])). Both lines were backcrossed on a C57BL/6J background for at least 15 generations. Experimental groups consisted of animals positive for both the CaMKIIα-tTA and tetO-hTDP-43 transgenes (either WT or ΔNLS TDP-43 variants; from now on, tTA/WT and tTA/ΔNLS, respectively). The control group included nontransgenic littermates or those carrying only one of the transgenes (CaMKIIα-tTA or tetO-hTDP-43). Mice from both sexes were used, as in our previous studies [6, 15, 16, 21]. Mice were housed in a pathogen-free facility under controlled environmental conditions (23 ± 2°C; 40–60% relative humidity) with a 12-h light/dark cycle and ad libitum access to food and water. Environmental enrichment consisted of cardboard rolls and paper, replaced only when damaged or heavily soiled. Doxycycline treatment and transgene induction To prevent prenatal and postnatal transgene expression, all breeding pairs and offspring received drinking water containing doxycycline hyclate (0.2 mg/ml; sc-204734A, Santa Cruz Biotechnology, USA) from conception until weaning at postnatal Day 28 (P28), as previously described [15]. At P28, mice were switched to regular drinking water to induce transgene expression. Mice were analyzed at 2 months of age, corresponding to 1 month of continuous transgene induction. Genotyping Ear biopsies were collected between postnatal Days 15 and 20. DNA extraction and PCR amplification were performed in our Molecular Biology Laboratory using previously described primer sequences [15, 16, 21]. Perfusion and tissue processing Mice were deeply anesthetized with 5% chloral hydrate (1 ml/30 g) and transcardially perfused with phosphate-buffered saline (PBS, 0.01 M, pH 7.4) supplemented with 10 U/ml Heparin. The brains were immediately extracted and post-fixed overnight in 4% paraformaldehyde at 4°C, then cryoprotected in 10% and 30% sucrose in PBS. Coronal sections (50 µm) were obtained using a microtome (SM 2010R; Leica) and stored at −20°C in cryoprotecting solution (50% glycerol, 50% PBS). Immunofluorescence staining Double immunofluorescence was performed as follows: free-floating sections were rinsed in PBS (0.01 M) and permeabilized for 1 h at room temperature (RT) in PBS containing 1% Triton X-100. Sections were then blocked for 1 h in PBS containing 5% normal goat serum (NGS) and 0.3% Triton X-100. The primary antibodies (diluted in 0.3% Triton X-100 and 3% NGS in PBS) were incubated overnight at 4°C with the indicated dilutions: rabbit anti-Iba1 antibody (1:3000; Wako) for microglial labeling, rabbit anti-GFAP antibody (1:2000; DAKO) for astrocytic labeling, and monoclonal mouse anti-hTDP-43 antibody (1:10 000; Proteintech). The next day, sections were washed 3 × 5 min in PBS containing 0.04% Triton X-100 and incubated for 4 h at RT in the dark with fluorophore-conjugated secondary antibodies diluted in 3% NGS, 0.3% Triton X-100 in PBS: rhodamine-conjugated goat anti-rabbit IgG (1:500; Jackson Laboratories) and Alexa Fluor 488–conjugated goat anti-mouse IgG (1:500; Invitrogen). After three washes in PBS, nuclei were counterstained for 30 min at RT in the dark with Hoechst 33342 (1 µg/ml; Sigma-Aldrich) and washed twice in PBS. Sections were mounted on glass slides and coverslipped with 30% glycerol in PBS. Slides were stored at 4°C in the dark until imaging. For AQP4/GFAP double immunofluorescence staining, brain tissue slices were permeabilized with 0.3% Triton X-100 in PBS for 15 minutes at RT and blocked with BSA 1 mg/ml (Sigma-Aldrich) in PBS with 0.03% Triton-X-100 for 1 h at RT. Samples were incubated with a rabbit anti-AQP4 primary antibody (1:/300; sc-20812 Santa Cruz) overnight at 4°C, followed by three washes in 0.1% Triton X-100/PBS and then incubated with secondary Alexa 488 donkey anti-rabbit antibody (1:300; 711-545-152, Jackson Laboratories) for 2 hours at RT. After washing, slices were incubated with a rat anti-GFAP antibody (1:500, 2.2B10, generously provided by Dr. Virginia Lee, University of Pennsylvania) overnight at 4°C and then with a secondary Alexa 647 goat anti-rat antibody (1:1000, Invitrogen) for 2 hours at RT. Finally, cell nuclei were identified using Hoechst 33342 (1 µg/ml; Sigma-Aldrich) for 30 minutes at RT. Coverslips were mounted with Vectashield mounting medium (VectorLabs). Image acquisition and analysis Fluorescent images were acquired using a Zeiss AxioImager 2 microscope equipped with the structured illumination system ApoTome 2 (Carl Zeiss, Germany), using a Hamamatsu Orca Flash 4.0 camera, under identical settings for all groups. For experiments with both tTA/WT and tTA/ΔNLS mice, two separate immunofluorescences were performed: one for Iba1 and one for GFAP, sometimes combined with hTDP-43 staining. The number of animals analyzed varied depending on the immunostaining and the brain region examined and are indicated in the corresponding figure legends. Microglial analysis from Iba1-stained sections included Iba1-positive (Iba+) area (normalized to the mean of the control group), microglial density (Iba1+ somas/mm2), and soma area of Iba1+ cells (µm2). Sholl analysis parameters included the number of intersections as a function of distance from the soma (µm), reflecting branching complexity; the total number of intersections; the total process length (µm), representing overall arbor size; and the maximum radial process extension (µm). Astrocytic reactivity in GFAP-stained sections was additionally quantified as GFAP-positive area (normalized to the mean of the control group). All measurements, image processing, and quantification were performed using Fiji [22]. Images for Iba1 and GFAP quantifications were captured at 20× magnification (N.A. 0.8 objective). For Sholl analyses, z-stacks of Iba1-stained microglia were acquired at 40× (N.A. 0.95 objective) using 1 µm step size, typically yielding 30–45 optical sections per sample. Maximum-intensity projections were generated in Fiji prior to binarization, skeletonization, and subsequent morphometric analysis. AQP4 polarization was quantified as previously described [23, 24] using 40× images acquired as described for GFAP and Iba1 staining. Briefly, 50 μm segments centered on blood vessels (as identified by vascular-shaped AQP4 localization) were analyzed using the line-plot tool in ImageJ. 5 vessels (range of 5–8 μm wide) were chosen per image (2 per region in each animal). The baseline fluorescence was defined as the average intensity over 10 μm, −20 μm to −10 μm from peak fluorescence. AQP4 polarization index (PI) was calculated as the ratio between the mean vascular endfeet AQP4 fluorescent signal and mean background signal. Average polarity was calculated, and the PI of each group was normalized to the control group. The higher the AQP4 PI, the greater proportion on immunoreactivity was restricted to perivascular regions, whereas the lower PI, the more evenly distributed immunoreactivity was between the perivascular endfeet and the cell body. Blood vessel diameter was determined from line plots by measuring the distance between the initial and terminal high-intensity AQP4 fluorescence signals, corresponding to the start and end of each vessel. Statistical analysis Graph Pad Prism 8 software (GraphPad Software, Inc., San Diego, CA, USA) was used for data visualization and statistical analysis. Comparisons between two groups for each brain region were made using unpaired two-tailed Student’s t-test. Sholl data were analyzed by two-way ANOVA with repeated measures followed by Sidak’s multiple comparisons test. Data are expressed as mean ± standard error mean (SEM), and significance was set at P < 0.05. Results Wild-type hTDP-43 overexpression induces widespread microglial activation and morphological remodeling Although microglial activation is a feature of several TDP-43 murine models of proteinopathies [6, 25] the specific regional contribution in early postinduction stages has not been investigated. We observed a significant increase in microglial activation across multiple cortical and hippocampal regions after 1 month of transgene expression in tTA/WT hTDP-43 mice (Fig. 1). Immunofluorescence analysis confirmed robust neuronal expression [6, 21] of human wild-type TDP-43 with predominant nuclear localization across cortical and hippocampal regions in tTA/WT mice, while no signal was detected in control littermates (Fig. 1a), confirming neuronal transgene expression and validating lack of colocalization with microglia, as expected. Quantitative analysis of immunofluorescence-stained sections revealed a significant elevation in the total immunopositive area for Iba1 in the motor cortex (MC, +21.0%), somatosensory cortex (SSC, +21.0%), and dentate gyrus (DG, +28.0%) compared to control littermates (Fig. 1b). This was accompanied by a significant increase in microglial density (Iba1+ somas/mm2) in the same regions (MC: +19.0%, SSC: +33.2%, DG: +33.5%) (Fig. 1c). Notably, when the CA1 region was analyzed as a whole, no significant differences in Iba1-immunopositive area were detected between groups. However, given the well-defined laminar organization of the hippocampal CA1, we performed a layer-specific analysis (Supplementary Figure S1). This approach revealed a significant increase in Iba1+ area in the stratum oriens (Or) and stratum radiatum (Rad) of tTA/WT hTDP-43 mice, whereas no changes were observed in the pyramidal layer (Py) (Supplementary Figure S1c). DG subregional analysis revealed significant increases for both the granular (Gr) and the polymorphic (Po) layers (Supplementary Figure S2d). A key indicator of microglial activation, soma hypertrophy, was evident through a significant enlargement of Iba1+ soma area in the prefrontal cortex (PFC: +15.2%), MC (+13.7%), SSC (+12.5%), and DG (+13.8%) (Fig. 1d). Widespread microglial activation across cortical and hippocampal regions following transgene expression in tTA/WT hTDP-43 mice. (a) Representative immunofluorescence images show Iba1+ microglia (red), human TDP-43 (green), and merged channels (plus nuclear marker Hoechst in blue) in the prefrontal cortex (PFC), motor cortex (MC), somatosensory cortex (SSC), CA1 hippocampal region, and dentate gyrus (DG) of control and tTA/WT mice after 1 month of transgene expression. Scale bar: 50 µm. (b–d) Quantitative analyses reveal a significant increase in Iba1+ area (normalized to control) (b) and microglial density (Iba1+ somas/mm2) (c) in the MC, SSC, and DG of tTA/WT mice compared with controls. Moreover, Iba1+ microglia exhibit enlarged soma area (µm2) (d) in the PFC, MC, SSC, and DG, a hallmark of microglial activation. Statistical significance was determined by Student’s t test (* P < 0.05; ** P < 0.01; *** P < 0.001; and **** P < 0.0001). Results are expressed as mean ± SEM. n = 8–12 animals per group (b-c); n = 144–175 cells per group (d) from the same number of animals as in B and C. In order to further evaluate morphological changes beyond cell size and density, we performed Sholl analysis on reconstructed microglial cells (Fig. 2). In tTA/WT hTDP-43 mice, microglia exhibited a significant reduction in process complexity. This was evidenced by a lower number of process intersections at increasing distances from the soma in the MC, SSC, CA1, and DG hippocampal regions (Fig. 2c-g, panel i). Consistent with this, we found a decrease in the total number of intersections (MC: −26.1%, SSC: −26.3%, CA1: −46.2% and DG: −39.3%) and total process length (MC: −28.1%; SSC: −25.5%, CA1:-48.2% and DG: −42.9%) in these areas (Fig. 2c-g, panels ii and iii). Maximum process extension was significantly lower in MC (−21.1%) and hippocampal CA1 (−35.9%) and DG (−41.1%) regions (Fig. 2c-g, panel iv). This shift towards a less ramified, more hypertrophied or bushy morphology, characterized by process retraction, is a well-established feature of microglial activation [26]. Sholl analysis reveals reduced microglial process complexity in multiple brain regions of tTA/WT hTDP-43 mice. (a) Z-stacks were acquired using a Zeiss AxioImager 2 ApoTome 2 system with a 40× objective and processed as maximum intensity projections. For each animal, five Iba1+ microglial cells were randomly selected, reconstructed in Fiji (ImageJ, NIH, USA), binarized, and skeletonized for process tracing. Sholl analysis quantified intersections of microglial processes with concentric circles radiating from the soma. Scale bar = 10 µm. (b) Representative maximum intensity Z-stack images of microglia from the MC regions of control and tTA/WT mice. Scale bar = 10 µm. (c-g) Quantitative Sholl analysis of Iba1+ microglia in the prefrontal cortex (PFC), motor cortex (MC), somatosensory cortex (SSC), CA1 hippocampal region (CA1), and dentate gyrus (DG) demonstrates significant morphological alterations after 1 month of transgene expression in tTA/WT mice compared with control littermates. Sholl profiles indicated a reduction in the number of intersections at increasing distances from the microglial soma (i), total number of intersections (ii), and total process length (iii) in MC, SSC, CA1, and DG hippocampal regions. Maximal process extension (iv) was reduced in tTA/WT mice, consistent with microglial activation and neuroinflammation. Statistical analysis of Sholl profiles (number of intersections vs. distance) was performed using two-way ANOVA followed by Sidak’s multiple comparisons test (* P < 0.05; ** P < 0.01; *** P < 0.001), whereas total number of intersections, total process length, and maximal process extension were analyzed using Student’s t test (* P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001); n = 15 cells/3 mice per group analyzed. Results are expressed as mean ± SEM. Cytoplasmic hTDP-43 expression elicits a more pronounced microglial activation phenotype We next investigated whether mice expressing the cytoplasmic TDP-43 variant would elicit a different microglial response. In tTA/ΔNLS hTDP-43 mice, we observed robust microglial activation and soma hypertrophy across several cortical and hippocampal areas analyzed (Fig. 3). Specifically, there were significant increases in microglial density in the MC (+30.9%), CA1 (+73.8%), and DG (+157.4%) regions (Fig. 3c) and immunopositive area for Iba1 in the same areas plus the SSC (MC: +46.9%, SSC: +25.6%, CA1: +61.7%, DG: +53.1%) (Fig. 3b) compared to controls. Of note, microglial soma size was significantly increased in all three cortical areas but not in hippocampal subregions (PFC: +102.5%, MC: +23.9%, SSC: +9.8%) (Fig. 3d). Microglial activation and soma hypertrophy across cortical and hippocampal regions in tTA/ΔNLS hTDP-43 mice. (a) Representative immunofluorescence images show Iba1+ microglia (red), human TDP-43 staining (green), and merged channels (plus nuclear marker Hoechst in blue) in the prefrontal cortex (PFC), motor cortex (MC), somatosensory cortex (SSC), CA1 hippocampal region, and dentate gyrus (DG) of control and tTA/ΔNLS mice after 1 month of transgene expression. Scale bar: 50 μm. Significant increases in Iba1+ area (b) and microglial density (c) are observed in the MC, CA1, and DG regions in the tTA/ΔNLS hTDP-43 animals compared to control animals after 1 month of transgene expression. Iba1+ microglia also demonstrate significantly increased soma size (d) in the PFC, MC, and SSC regions, which is a key morphological hallmark of neuroinflammation. The analysis was based on images acquired using a Zeiss AxioImager 2 ApoTome 2 system at a 20× objective. Statistical significance was determined by Student’s t test (* P < 0.05; ** P < 0.01; **** P < 0.0001). Error bars represent SEM. n = 7–11 animals per group (b-c); n = 122–263 cells per group (d) from the same number of animals as in B and C. To resolve whether microglial activation was uniformly distributed within hippocampal subfields, we conducted a layered analysis of Iba1 immunoreactivity to assess whether the increased Iba1 immunoreactivity observed in tTA/ΔNLS mice reflected uniform microglial changes across hippocampal layers (Supplementary Figure S2). Layer-specific analysis revealed a significant increase in Iba1-positive area in CA1 strata and in the granule cell layer of the DG, whereas no significant differences were detected in the polymorphic layer (Supplementary Figure S2c-d). These results indicate that microglial activation induced by cytoplasmic hTDP-43 is spatially different at the laminar level. This selectivity supports the notion of distinct neuroinflammatory signatures driven by nuclear and cytoplasmic TDP-43 overexpression. Sholl analysis in the tTA/ΔNLS hTDP-43 animals revealed an extensive and profound decrease in microglial ramification, indicating a transition to an activated, amoeboid-like state (Fig. 4b-f, panel i). Across MC, SSC, CA1, and DG, microglia showed a significant reduction in process complexity, total number of intersections (MC: −49.9%, SSC: −54.9%, CA1: −46.1%, and DG: −78.0%), and total process length compared to controls (MC: −52.9%, SSC: −55.0%, CA1: −44.8%, and DG: −77.1%) (Fig. 4b-f, panels ii, iii, and iv). This generalized process retraction suggests a more severe and widespread activated phenotype compared to that induced by the expression of WT-hTDP-43 transgene. Microglial ramification is significantly impaired across multiple brain regions in tTA/ΔNLS hTDP-43 mice after 1 month of gene expression, indicating an activated/amoeboid state. (a) Representative maximum intensity Z-stack images of microglia from the MC of control and tTA/ΔNLS mice. Scale bar = 20 µm. (a–f) Sholl analysis performed in prefrontal cortex (PFC), motor cortex (MC), somatosensory cortex (SSC), CA1 hippocampal region (CA1), and dentate gyrus (DG) demonstrates a significant reduction in microglial process complexity (i), total number of intersections (ii), and total length of microglial processes (iii) in the tTA/ΔNLS hTDP-43 animals compared to control littermates. This reduced ramification (microglial process retraction) (iv) is a key morphological hallmark of neuroinflammation and transformation to an activated/amoeboid phenotype. The analysis was performed after 1 month of transgene expression. Images used for quantification were acquired using a Zeiss AxioImager 2 ApoTome 2 system with a 40× objective. Statistical analysis of Sholl profiles (number of intersections vs. distance) was performed using two-way ANOVA followed by Sidak’s multiple comparisons test (** P < 0.01; *** P < 0.001, *** P < 0.0001), whereas total number of intersections, total process length, and maximal process extension were analyzed using Student’s t-test (* P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001). n = 15 cells/3 mice per group analyzed. Results are expressed as mean ± SEM. Differential astroglial activation highlights a widespread response in tTA/ΔNLS hTDP-43 mice Although reactive astrocytosis is a key hallmark of many neurodegenerative diseases [27], most studies in ALS/FTD mouse models focus on the glial activation in the spinal cord. To determine the early response of the astrocyte population in the brains of our inducible TDP-43 mouse models, we quantified the area covered by GFAP+ cells (Fig. 5). One month after induction, GFAP-based astrogliosis patterns differ between WT and ΔNLS models. Astroglial activation was selective and more regionally restricted in hTDP-43-WT mice, with GFAP upregulation confined primarily to MC (+363.0%) and SSC (+199.0%) (Fig. 5b). In contrast, tTA/ΔNLS mice displayed a significant and widespread astroglial activation across all cortical and hippocampal regions analyzed (PFC: +92.58%, MC: +259.0%, SSC: +215.5%, CA1: +50.4%, and DG: +51.3%) (Fig. 5c). This robust astroglial reactivity indicates that the presence of cytoplasmic TDP-43 triggers a more pervasive neuroinflammatory response from astrocytes compared to the wild-type protein. Region-specific astroglial activation in tTA/WT and widespread activation in tTA/ΔNLS hTDP-43 mice. (a) Representative immunofluorescence images show GFAP+ astrocytes (red) and nuclear marker Hoechst (blue) in control and tTA/WT mice (top two rows), and in control and tTA/ΔNLS hTDP-43 mice (two bottom rows), across the prefrontal cortex (PFC), motor cortex (MC), somatosensory cortex (SSC), CA1 hippocampal region (CA1), and dentate gyrus (DG). Scale bar: 50 µm. (b, c) Quantification of GFAP+ area (normalized to control) revealed selective astroglial activation in MC and SSC of tTA/WT mice (b), whereas tTA/ΔNLS hTDP-43 animals exhibited significant increases in GFAP+ area across all analyzed regions (c), consistent with robust astroglial reactivity and neuroinflammation. Images were acquired using a Zeiss AxioImager 2 ApoTome 2 system. Statistical analysis was performed using Student’s t-test (** P < 0.01; *** P < 0.001; **** P < 0.0001). Results are expressed as mean ± SEM. n = 6–7 animals per group (b); n = 8–11 animals per group (c). Decreased AQP4 PI in SCC and hippocampus of hTDP-43-ΔNLS animals To investigate whether glymphatic system function, which relies on astrocytic AQP4 expression, is altered in hTDP-43-ΔNLS mice, we evaluated AQP4 expression and polarization, as well as the diameter of blood vessels in selected brain regions (SSC, CA1, and DG). We focused on this animal model given its more widespread astroglial activation, which provided an appropriate context to explore early alterations in astrocytic AQP4 polarization. Figure 6 shows AQP4 (green channel) and GFAP (red channel) expression and the merged channel. AQP4 is highly polarized to perivascular astrocytic end feet in control animals, but the AQP4 PI was reduced in SSC (−18.5%), CA1 (−32.6%), and DG (−46.0%) of hTDP-43-ΔNLS mice. Blood vessel diameter was reduced only in DG of these animals, likely due to the higher rate of neurodegeneration reported in this area [6, 15]. Remarkably, while the spatial distribution of the water channel was altered, no changes in total AQP4 immunofluorescence intensity were observed in the CA1, DG, or SSC of hTDP-43-ΔNLS mice compared to controls (not shown). This early decrease in AQP4 polarization suggests that mice expressing cytoplasmic TDP-43 likely present a loss of perivascular organization rather than a quantitative depletion of the protein itself. Perivascular AQP4 polarization is decreased in the cortex and hippocampus of tTA/ΔNLS hTDP-43 mice. hTDP-43-ΔNLS mice show decreased AQP4 PI around vascular structures in SSC and hippocampal CA1/DG in comparison to control animals. Representative images of AQP4 (green channel), GFAP (red channel), and merged channel (plus nuclear marker Hoechst in blue) in SSC (a), CA1 (b), and DG (c). Images are representative of 4 animals (10 vessels/animal) in 2 independent experiments (Scale bar: 100 μm). Inserts show enlarged boxed areas from the green channel image (inset: 28 μm). Quantification of AQP4 polarization index (PI) and vessel diameter (scale bar: 28 μm). PI was calculated as the ratio between the mean vascular end foot AQP4 fluorescent signal and mean background signal. Results are expressed as mean ± SEM. * P < 0.05, *** P < 0.0001 control vs. hTDP-43-ΔNLS mice (Student’s t test). Discussion In this study, we provide a detailed characterization of early, region-specific glial responses to neuron-specific expression of different TDP-43 variants. Our results demonstrate that 1 month of either overexpression of wild-type hTDP-43 or cytoplasmic hTDP-43 expression is sufficient to induce significant cortical and hippocampal gliosis. Although both microglia and astrocytes exhibited morphological signatures of activation, the nature, spatial distribution, magnitude of this response, and cellular profile of glial activation are contingent upon both the specific brain region and the subcellular location of hTDP-43. Our findings challenge a simplistic, binary ‘resting versus activated’ view of microglia, supporting a model of a complex activation spectrum [28–30]. Microglia and astrocytes exhibit region and context-dependent responses across the CNS. Single-cell transcriptomics studies have demonstrated that microglial states exist along a continuum shaped by anatomical location and pathological context [31]. Understanding how early glial responses are spatially organized in different models of TDP-43 proteinopathies, characterized by nuclear or cytoplasmic TDP-43 localization, may enable a better definition of region-specific patterns of vulnerability and their relationship to functional outcomes. A recent study in human ALS tissue and the rNLS8 TDP-43 mouse model reported that microglia display a stage-dependent transition from hypertrophic to a later dystrophic morphology, and these changes correlate with increased load of pTDP-43 [32]. In the tTA/WT mice described here, PFC microglia exhibited significant soma enlargement, yet Sholl analysis suggested that process complexity might be preserved. This dissociation between soma hypertrophy and process retraction may represent an early or primed activation state, where microglia increase their metabolic activity while preserving their surveillance territory. This phenotype aligns with findings in early-stage models of Alzheimer's disease and chronic stress, where initial changes in soma size precede dramatic alterations in process morphology [33, 34]. In contrast, the more profound process retraction observed in motor and sensory cortices in tTA/WT mice, and globally in tTA/ΔNLS mice, indicates a more advanced shift toward a reactive, potentially phagocytic state. The widespread and severe morphological changes in the tTA/ΔNLS model underscore the potent pro-inflammatory effect of cytoplasmic TDP-43, which is a key pathological hallmark in the majority of ALS/FTD cases [2]. A study on TREM2, a transmembrane protein expressed by microglia, demonstrated its involvement in the phagocytic clearance of pathological TDP-43 [35]. Moreover, TREM2 deficiency resulted in reduced microglial activation in response to TDP-43 accumulation, leading to neuronal damage and motor impairments. The expression levels of TREM2 in the spinal cord and cortex from ALS patients were significantly greater than in control patients. Co-immunoprecipitation essays revealed that TREM2 interacts with hTDP-43 in both mice and humans with ALS [35]. Recent work from the same group in rNLS8 mice, an ALS model with pan-neuronal cytoplasmic TDP-43 overexpression, showed that TREM2-deficiency decreased the occurrence of rod-shaped microglia, a unique subpopulation involved in modulating excitatory inputs through the engulfment of excitatory synapses, impacting on both neuronal activity and behavioral outputs [36]. In addition to phagocytic pathways, cytokine signaling also links TDP-43 pathology to microglial activation. IL-6 trans-signalling has been shown to mediate TDP-43–induced inflammatory responses, and its selective blockade reduces pro-inflammatory cytokine release, highlighting its role as an amplifier of microglial reactivity [37]. A region-specific pattern emerges when comparing microglial activation with behavioral alteration across models. We previously demonstrated that wild-type hTDP-43 overexpression induces early deficits in social interaction and working memory (Y-maze) after 1 month of transgene expression, while motor function remains preserved until later stages (6–12 months post-induction) [16]. In our current study, we found that microgliosis and astrogliosis at 1 month are significantly present in the MC and the DG. This early activation in hippocampal and cortical regions provides a neuropathological basis for the initial cognitive and social impairments observed before overt motor decline. Furthermore, our previous work [21] showed that social deficits in this model are irreversible after transgene suppression, suggesting that the early glial remodeling (such as the soma hypertrophy observed in the PFC) marks a stage of metabolic activation that precedes or coincides with permanent circuit dysfunction. In contrast, tTA/NLS mice develop a much more aggressive phenotype, including hyperlocomotion, impaired motor coordination (rotarod), and severe social/cognitive deficits within 1 month [15]. These findings align closely with the exhibited widespread microglial activation (MC, SSC, CA1, and DG) reported in the current study. While this does not imply causality, it supports the notion that the regional extent of microglial activation is associated with the severity and features of the behavioral impairments. The significant microglial process retraction (transitioning to an amoeboid state) and widespread astrogliosis in these mice reflect a severe neuroinflammatory signature that corresponds to the rapid functional deterioration and long-term memory deficits previously reported by our group. The early microglial activation in both MC and SSC regions aligns with some features of the clinical presentation of ALS/FTD, where patients can present with both motor and cognitive symptoms [38, 39]. Our findings suggest that microglial responses might be an early event in disease progression, potentially contributing to the initial phases of neurodegeneration. Consistent with an early role of microglia, innate immune-sensing pathways such as cGAS have been identified as upstream regulators of TDP-43-associated pathology. cGAS activation in microglia drives functional and transcriptional alterations, while its inhibition ameliorates these changes and improves disease outcomes in vivo [40]. Moreover, recent evidence supports that microglia can actively drive TDP-43 pathology. In a progranulin-deficient model, disease-associated microglia induced TDP-43 aggregation and neuronal death via complement signaling, indicating that microglia can act upstream of neurodegeneration rather than merely responding to it [41]. Among all glial outcomes assessed, the astroglial response emerged as the most striking discriminator between WT and ΔNLS TDP-43 models. The restricted astrogliosis in tTA/WT mice versus the widespread, robust reaction in tTA/ΔNLS mice suggests that astrocytes are highly sensitive to the presence of cytoplasmic TDP-43. Another mouse model of TDP-43 proteinopathies, prpTDP-43A315T, showed that both the loss and gain of function of TDP-43 are sufficient to induce glial reactivity characterized by astrogliosis and microglial activation [25]. This strong astroglial response could have deep consequences for neuronal health [42]. Reactive astrocytes can lose critical homeostatic functions and release pro-inflammatory cytokines, contributing to a toxic environment [43]. Our data indicate that astrogliosis is not just a general marker of neuroinflammation but might be a key indicator of the severity and nature of the underlying proteinopathy. Differences in glial responses found between WT-hTDP-43 and ΔNLS-hTDP-43 models align with recent single-cell transcriptomic studies showing that microglia and astrocytes adopt regionally distinct reactive states in ALS, FTD, and AD [44, 45]. Such heterogeneity may explain why certain cortical and hippocampal regions are particularly vulnerable in early disease stages. Our results suggest that glymphatic dysfunction may contribute to ALS pathogenesis, as evidenced by alterations in diffusion tensor imaging studies in patients in comparison to healthy individuals [46, 47]. Elevated AQP4 with loss of astrocytic endfeet polarization has been reported in the brainstem and cortex of the SOD1G93A rat model of ALS [48]. Even though an increase in AQP4 mRNA expression in the spinal cord of SOD1G93A mice was reported as the disease progressed [49, 50], AQP4 polarized localization at the endfeet of astrocytes decreased at the disease onset and end stages [49]. In this study, we observed that the severity of AQP4 depolarization followed a regional pattern, with the most significant reduction occurring in the DG, where we also detected a significant reduction in blood vessel diameter. This finding provides structural evidence of neurovascular unit disruption in the area most affected by AQP4 mislocalization. The generalized and robust astroglial activation observed across all cortical and hippocampal regions in hTDP-43-ΔNLS mice likely induced AQP4 mislocalization without affecting its expression levels. As astrocytes transition from homeostatic to a reactive state, the fundamental remodeling of their processes may lead to the redistribution of this water channel from the astrocytic endfeet towards the soma or other astrocytic processes and the subsequent loss of AQP4 polarity observed in hTDP-43-ΔNLS mice. Our results align with previous observations where TDP-43 depletion in astrocytes led to AQP4 becoming dispersed throughout the cytoplasm rather than anchored at the plasma membrane [51], pointing out that TDP-43 may modulate AQP4 expression and relocalization. Specifically, cytoplasmic TDP-43 mislocalization may interfere with the mRNA metabolism or stability of AQP4 anchoring proteins required to tether AQP4 to the endfeet, such as dystroglycans and/or syntrophin, thereby facilitating its redistribution. These alterations could contribute to glymphatic failure, leading to inefficient clearance of misfolded TDP-43 and other neurotoxic aggregates. For instance, in the hippocampus of AD patients, the presence of phosphorylated TDP-43 inclusions was associated with alterations in blood–brain barrier integrity and reduced astrocytic AQP4 organization, which correlated with disease severity [12, 52]. The reduction in AQP4 area fraction in blood vessels observed in AD patients led these authors to speculate that AQP4 relocates from the astrocytic endfeet to other compartments of the cell, leaving the glymphatic system affected. While direct functional assays of glymphatic flow were not performed in this study, the decrease in AQP4 PI in the context of intense astrogliosis is highly predictive of impaired clearance. In agreement with our results, similar early-stage findings in the rNLS8 TDP-43 model demonstrated that glymphatic dysfunction precedes neurodegeneration, even when total AQP4 protein levels remain normal [53]. Recent studies by the same group showed that cortical AQP4 mRNA expression was decreased after 7 days of transgene induction [54]. These findings highlight early glymphatic dysfunction in ALS, suggesting its potential as a therapeutic target. Even though altered glymphatic function was detected at 3 days post cytoplasmic TDP-43 induction, its function normalized to control levels by Day 7, but cortical AQP4 mRNA expression was significantly altered. By 21 days post-induction, glymphatic dysfunction was evident across all brain regions assessed and coincided with widespread degeneration, significant motor impairments, and astrogliosis, but AQP4 total expression showed no differences with the control group [54]. Although AQP4 polarization was not assessed in that study, we hypothesize that it might be likely altered, as we observed 4 weeks after CamK2 promoter-driven cytoplasmic TDP-43 overexpression. The generalized astrogliosis observed in hTDP-43-ΔNLS mice might be both a cause and a consequence of AQP4 dysregulation, reinforcing a cycle of neuroinflammation and glymphatic dysfunction. Differences between models highlight the complexity of TDP-43 pathology, with nuclear vs. cytoplasmic localization shaping the neuroinflammatory and clearance mechanisms. Although it is unclear whether and how microglial and astrocytic responses might be causally linked, or whether they represent parallel downstream effects of TDP-43 levels/mislocalization, these findings emphasize the importance of glial dysregulation as an early driver of pathology in TDP-43 proteinopathies, with implications for therapeutic strategies aimed at modulating microglial phenotypes. Some limitations of the study include: (1) our data provide insights into region-specific glial activation present at 1 month after induction; however, temporal relationships to neuronal injury and clearance function remain to be established in future studies. (2) Sex-specific differences in glial activation and glymphatic function have been reported; the current design did not include sex-stratified analyses, limiting generalization to both sexes. (3) Because stainings were performed separately for each line, direct magnitude comparisons between tTA/WT and tTA/ΔNLS were not performed. (4) Regional hTDP-43 neuronal expression and penetrance were not quantified in both lines at this time point; thus, differences in Iba1 and GFAP staining may partly reflect expression load or anatomical targeting, which is not identical in tTA/WT and tTA/ΔNLS lines [6]. Our study reveals that gliosis plays critical but distinct roles in hTDP-43-WT and hTDP-43-ΔNLS mice. While WT-TDP-43 overexpression induces moderate glial activation, cytoplasmic TDP-43 mislocalization leads to widespread gliosis, altered AQP4 distribution, and potential glymphatic impairment. These results highlight the need for therapeutic strategies targeting both neuroinflammation and protein clearance to mitigate TDP-43-related neurodegeneration. Supplementary Material