Cytoplasmic TDP-43 leads to early behavioral impairments without neurodegeneration in a serotonergic neuron-specific C. elegans model

Abstract

TDP-43 proteinopathies, such as amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD), are marked by the pathological cytoplasmic accumulation of TAR DNA-binding protein 43 (TDP-43), leading to progressive neuronal dysfunction and degeneration. To investigate the early functional consequences of TDP-43 mislocalization, we generated Caenorhabditis elegans models expressing either wild-type human TDP-43 or a variant with a mutated nuclear localization signal (ΔNLS), specifically in serotonergic neurons. These neurons were chosen because in C. elegans they regulate well-characterized behaviors, providing a straightforward readout of neuronal function. We found that expression of either TDP-43 variant impaired serotonin-dependent behaviors—including pharyngeal pumping, egg-laying, and locomotion slowing upon food encounter—with the cytoplasmic ΔNLS form causing more severe deficits. These behavioral impairments are evident even while the serotonergic neurons remain apparently normal in structure, suggesting that neuronal dysfunction precedes overt neurodegeneration. Moreover, the serotonergic HSN neurons that control egg-laying were also partially responsive to the selective serotonin reuptake inhibitor fluoxetine, suggesting that neurotransmitter release remains functional to some extent. Altogether, our findings demonstrate that TDP-43 expression causes neuronal dysfunction leading to behavioral deficits, even in the absence of detectable structural pathology and that its mislocalization to the cytoplasm results in more severe behavioral impairments. This C. elegans model provides a genetically tractable system to dissect early mechanisms of TDP-43-mediated neuronal dysfunction and to identify therapeutic strategies targeting predegenerative stages of ALS/FTD.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-36138-5.

Introduction

TAR DNA-binding protein 43 (TDP-43) is a ubiquitously expressed RNA-binding protein that plays critical roles in RNA metabolism, including transcription, splicing, transport, and stability¹. Under normal physiological conditions, TDP-43 is predominantly localized in the nucleus. However, in neurodegenerative disorders such as amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD), TDP-43 undergoes pathological mislocalization to the cytoplasm, where it forms insoluble aggregates - both hallmark features of these pathological conditions^2,3.

While ALS has been classically characterized by the degeneration of motor neurons⁴, emerging evidence suggests that other neuronal populations, including serotonergic neurons, may also be affected⁵. Studies in ALS mouse models have reported cell-autonomous pathological alterations in serotonergic neurons, and pharmacological modulation of serotonin signaling has demonstrated partial therapeutic benefits^6,7. These observations indicate that non-motor neuronal systems may contribute to disease pathogenesis, highlighting the importance of investigating TDP-43 pathology in different neuronal classes.

The nematode Caenorhabditis elegans has emerged as a powerful model organism for studying neurodegenerative processes, offering several unique advantages: a fully mapped nervous system consisting of exactly 302 neurons in the hermaphrodite, genetic tractability, and a short lifespan^8,9. Importantly, C. elegans exhibits well-characterized behaviors that are controlled by specific, identifiable neuronal circuits, allowing for precise assessment of neuronal function through behavioral assays^10–12.

Previous studies have successfully modeled various proteinopathies in C. elegans, including Parkinson’s disease, Alzheimer’s disease, and TDP-43 proteinopathies^13,14. These models commonly feature pan-neuronal or muscle-specific expression of human proteins (α-synuclein, β-amyloid, and TDP-43) linked to these pathologies, and frequently use locomotion as the primary behavioral readout^15,16. While valuable, these approaches present a key interpretive challenge: because expression is generally driven broadly across the nervous system or muscle, it is difficult to ascertain whether subsequent locomotion defects stem from the dysfunction of specific neuronal circuits or from systemic health issues¹⁴.

In this study, we focused on the serotonergic system of C. elegans, which consists of just three pairs of well-defined neurons that control quantifiable behaviors, including feeding (pharyngeal pumping), locomotion modulation, and egg-laying^17–22. The simplicity and well-characterized nature of this system make it particularly suitable for investigating how TDP-43 pathology affects specific neuronal circuits.

We generated transgenic C. elegans strains expressing either human wild-type TDP-43 (hTDP-43-WT) or a nuclear exclusion mutant (hTDP-43-ΔNLS), which lacks a functional nuclear localization signal and is retained in the cytoplasm)²³, specifically in serotonergic neurons. This targeted approach allowed us to address two key questions: (1) Does the expression of TDP-43 variants with different expected subcellular localizations lead to distinct degrees of behavioral impairment? (2) Are behavioral deficits associated with structural neurodegeneration or more subtle functional impairments?

Our results demonstrate that targeted expression of human TDP-43 in serotonergic neurons is sufficient to cause behavioral impairments, with the cytoplasmically-localized variant producing the most severe deficits, all while preserving neuronal integrity. These findings not only provide a novel animal model for investigating the effects of TDP-43 mislocalization on neuronal circuits but also link these cellular defects to quantifiable behavioral phenotypes. This model thus represents a valuable tool for initial screening of potential therapeutic interventions.

Materials and methods

C. elegans culture and maintenance

All C. elegans strains were grown at room temperature (22 °C) on Nematode Growth Media (NGM) agar plates with Escherichia coli OP50 as a food source. The wild-type reference strain used in this study is N2 Bristol. Some of the strains were obtained through the Caenorhabditis Genetics Center (CGC, University of Minnesota). Worm population density was maintained at a low level throughout their development and during the assays. All experiments were conducted on age-synchronized animals. This was achieved by placing gravid worms on NGM plates and removing them after two hours. The assays were performed on the animals hatched from the eggs laid in these two hours.

Transgenic strains were generated by microinjection of plasmid DNA containing either the construct Ptph-1::hTDP43 or Ptph-1::hTDP43(ΔNLS) at 10 ng/µL into the germ line of lin-15 (n765ts) mutants with the co-injection marker lin-15 rescuing plasmid pL15EK (80 ng/µl). At least three independent transgenic lines were obtained. Data are shown from a single representative line.

The strains used were:

N2 (wild-type).

MT15434 tph-1(mg280) II.

MT13471 [ptph-1::GFP].

MT1082 egl-1(n487) V.

OAR165 lin-15(n765ts); nbaEx20[Ptph-1::hTDP43(10) + lin-15 (80)].

OAR166 lin-15(n765ts); nbaEx21[Ptph-1::hTDP43(ΔNLS)) + lin-15 (80)].

OAR199 lin-15(n765ts); nbaEx20[Ptph-1::hTDP43(10) + lin-15 (80)]; ptph-1::GFP.

OAR200 lin-15(n765ts); nbaEx21[Ptph-1::hTDP43(ΔNLS)) + lin-15 (80)]; ptph-1::GFP.

Microscopy and image analysis

For microscopy, young adult worms (one day after L4 stage synchronization) were mounted in M9 with levamisole (10 mM) onto slides with 2% agarose pads. Images were acquired on confocal microscopy (LSCM; Leica DMIRE2) with 20X and 63X objectives. For tph-1::GFP expression levels analysis, animals containing the corresponding transcriptional GFP reporter were imaged using an epifluorescence microscope (Nikon Eclipse TE2000-5) coupled to a CCD camera (Nikon DS-Qi2) with 20X objective. Fluorescence intensity was quantified using a simple Macro that runs on Image J FIJI software. With this Macro, the background was initially subtracted, and then a region of interest, typically the neuron soma, was selected. The mean fluorescence intensity of this selected region was measured.

Pharyngeal pumping rate

Feeding rate was quantified essentially as previously described²⁴. Briefly, young adult worms (one day after L4 stage synchronization), were transferred to NGM plates seeded with E. coli OP50. Animals were filmed under a stereomicroscope at 75× magnification using an Allied Vision Alvium 1800 U-500 m camera, at 30 frames per second. Pharyngeal contractions in the terminal bulb were manually counted for 1 min by analyzing videos at 0.5× playback speed. All experiments were conducted blind to the experimental condition. At least three independent trials with ∼30 animals for each condition were performed.

Egg laying assay

Egg laying behavior was assessed using standard procedures²⁵. Briefly, worms were synchronized at the L4 stage. One day after synchronization (young adult stage), individual animals were transferred to NGM plates seeded with E. coli OP50 as a food source. After one hour, gravid adults were removed, and the number of deposited eggs was manually counted. At least three independent biological replicates were performed, with 22 worms assessed per experiment.

To evaluate the effect of fluoxetine on egg-laying, we followed an adapted methodology based on previously reported protocols²⁶. Briefly, individual young adult worms were transferred into separate wells of a 96-well microtiter plate, each containing 100 µL of M9 buffer supplemented with fluoxetine at a concentration of 0.5 mg/mL, which has been previously established for this assay^26,27. As previously reported, worms in M9 buffer alone (without food) under these basal conditions lay a negligible number of eggs²⁶. Control groups (M9 buffer alone) were included for each strain in every experiment. For each strain and treatment condition, a minimum of eight animals were tested per replicate.

Enhanced slowing response

Locomotion velocity was quantified using the Multi-Worm Tracker (MWT) (Rex Kerr, https://sourceforge.net/projects/mwt/). Raw tracking data were processed with Choreography (MWT’s feature extraction software) and analyzed using custom MATLAB scripts (The MathWorks, Inc.). To ensure robust tracking, we excluded objects that were not tracked for at least 20 s or that moved less than 5 body lengths during the recording.

Animals were placed on bacterial (E. coli OP50) lawn-seeded plates (arranged in a donut-shaped pattern to confine food to the edges). A droplet (5 µl) of M9 buffer containing the worms was deposited at the center of the plate and allowed to dry before recording. Locomotion was captured at 30 frames per second using an Allied Vision Technology Guppy Pro GPF 125 C IRF Camera.

To assess the food-induced slowing response, we computed the ratio between the speed upon encountering the bacterial lawn and the initial speed (movement in the absence of food). A lower ratio indicates a stronger slowing response.

Gene expression analysis by RT-qPCR

To assess the expression levels of the hTDP-43 transgene, total RNA was extracted from approximately 100 adult worms using Bio-Zol RNA isolation reagent (PBL-Productos Bio-Logicos). Before extraction, worms were lysed with five cycles of cold-warm shock, alternating between liquid nitrogen and a 35 °C water bath. The extracted RNA was reverse-transcribed into complementary DNA (cDNA) using EasyScript^® Reverse Transcriptase (TransGen Biotech).

Gene expression analysis was performed by quantitative real-time PCR (qPCR) on a Rotor-Gene 6000 ThermoCycler (Corbett Research) using a SYBR Green Master Mix. The cycling conditions were as follows: initial incubation at 50 °C for 2 min, denaturation at 95 °C for 2 min, and 45 cycles of 15 s at 95 °C, 30 s at 60 °C, and 30 s at 72 °C. A melting curve analysis was generated after amplification to verify the specificity of the PCR products. All experiments included at least three independent biological replicates, with each replicate consisting of three technical replicates.

The act-1 (actin) gene was used as an endogenous reference for normalization. The cycle threshold (Ct) was determined automatically by the instrument’s software. For each sample, the ΔCt was calculated as the difference between the Ct of the gene of interest (hTDP-43) and the Ct of the reference gene:

The ΔΔCt method was then applied to calculate the relative change in gene expression. The hTDP-43-WT strain was used as the control (calibrator) condition. The ΔΔCt was calculated for each experimental condition (e.g., hTDP-43-ΔNLS) as follows: ΔΔCt = ΔCt (experimental condition) – ΔCt (control condition). The relative expression (fold change) was finally determined using the formula: Relative Expression (Fold Change) = 2^(–ΔΔCt). Ct values above 35 were considered below the reliable detection limit. Primers used have been previously designed and validated²⁸.

Primer sequences for RT-qPCR

Gene	Direction	Primer sequences (5´−3´)
hTDP-43	Forward	CCTAATTCTAAGCAAAGCCAAGATG
hTDP-43	Reverse	ACAGCGCCCCACAAACA

Thrashing assay

Locomotion in liquid was assessed using a standard thrashing assay. L4 animals were individually transferred to the wells of a multi-well plate (1 animal per well), each containing 150 µL of M9 buffer without food. After a 5-minute habituation period, thrashing rates were manually quantified under a dissecting microscope as previously described^29,30. A thrash was defined as a change in the direction of bending at the midbody^29,31. Specifically, the number of times the worm’s body concavity was directed to the same lateral side was counted. This count was then multiplied by two to yield the total number of thrashes per minute, thereby accounting for a full cycle of left and right bends. All assays were performed by two independent investigators who were blinded to the genotype of the samples throughout the counting process.

Statistical analysis

All data are presented as the mean ± standard deviation (SD) from at least three independent biological replicates. All experiments were scored blindly. GraphPad Prism version 8.01 and SigmaPlot version 12.0 were used for graph generation and statistical analyses, respectively.

The specific statistical tests applied in each experiment are detailed in the corresponding figure legends. For comparisons among three or more independent groups (e.g., different genetic strains), we used the Kruskal-Wallis test with Dunn’s post-hoc test for non-parametric data (Fig. 1b and c and S2) or one-way ANOVA with Holm-Sidak´s post-hoc test for parametric data (Figs. 1a, 2a, 3 and 4). For direct comparisons between two conditions within the same strain (e.g., control vs. fluoxetine), an unpaired two-tailed Student’s t-test was used (Fig. 2a). All raw data are available in the Open Science Framework repository (https://osf.io/zpw43/?view_only=9dc9afe51f6a4e3a959d7f17c6f8a8f7).

Fig. 1.

Behavioral effects of nuclear and cytoplasmic hTDP-43 expression on serotonergic neurons in C. elegans. To assess the effect of nuclear, wild-type human TDP-43 (hTDP-43-WT) or a cytoplasmic form with a mutated nuclear localization signal (hTDP-43-ΔNLS) on serotonergic neuron function in C. elegans, we expressed both forms specifically in serotonergic neurons and evaluated their impact on serotonin-dependent behaviors, including feeding, locomotion modulation, and egg-laying. (a). Pharyngeal pumping rate (pumps per minute in the presence of food), a measure of food intake. Upper: Schematic of C. elegans pharynx. Lower: Quantification of pumping rate. n = 30 animals per condition from 3 independent experiments (One-Way ANOVA with Holm-Sidak’s post-hoc test). (b) Food-induced slowing response. After 1 h of food deprivation, speed was measured upon food re-encounter using Worm-Tracker software for video acquisition and MATLAB for analysis. Slowing response was calculated as a percentage of initial speed. n = 26–34 animals per condition from 4 independent experiments (Kruskal-Wallis with Dunn’s post-hoc test). (c) Egg-laying frequency. Age-synchronized young adults were transferred to food-seeded plates for 1 h, followed by egg quantification. n = 22 animals per condition from 3 independent experiments (Kruskal-Wallis test with Dunn’s post-hoc test). All data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001 (comparisons between strains); #p < 0.05 (comparisons versus wild-type).

Fig. 2.

Fluoxetine-induced egg laying in animals expressing human TDP-43 variants in serotonergic neurons. Egg-laying response to fluoxetine (1.6 mM) was evaluated. Fluoxetine induces egg-laying in C. elegans²⁶. (a) Egg-laying was quantified over one hour in M9 buffer for wild-type, tph-1 null mutants (lacking 5-HT synthesis), and strains expressing hTDP-43-WT or hTDP-43-ΔNLS (n = 23 animals per condition). It is important to note that this assay was conducted in M9 buffer without food, resulting in minimal basal egg-laying, in contrast to the assay on solid medium containing bacterial lawns presented in Fig. 1. Within-strain comparisons (fluoxetine vs. control): Unpaired two-tailed Student’s t-test (*p < 0.05, **p < 0.01, ***p < 0.001). Between-strain comparisons: One-way ANOVA with Holm-Sidak’s post-hoc test (###p < 0.001, ##p < 0.01, #p < 0.05). (b) The egl-1 mutant (HSN neuron-deficient) showed no increase in egg-laying upon exposure to fluoxetine, confirming the essential role of HSNs in this response.

Fig. 3.

ADF and NSM serotonergic neuron integrity in TDP-43-expressing strains. (a) Schematic of head serotonergic neurons (ADF and NSM) in adult C. elegans. (b) Quantification of left and right side neuronal occurrence (worms with detectable neurons/total worms). (c) Representative confocal images showing animals expressing tph-1::GFP reporter constructs in wild-type, hTDP-43-WT and hTDP-43-ΔNLS expressing animals. Neuronal positions are indicated by arrowheads (ADF: white; NSM: orange). (d) Fluorescence intensity quantification in the neuronal soma (ImageJ). n = 17–18 animals/condition. One-way ANOVA with Holm-Sidak’s test (ns = not significant). All data are presented as mean ± SD.

Fig. 4.

HSN serotonergic neuron integrity in TDP-43-expressing strains. (a) Schematic of the HSN serotonergic neuron position near the vulva in adult C. elegans. (b) Quantification of left and right side HSN neuron occurrence (worms with detectable HSN neurons/total worms). (c) Representative confocal images showing HSN neurons (orange arrowheads) in wild-type, hTDP-43-WT and hTDP-43-ΔNLS expressing animals. (d) Fluorescence intensity quantification of HSN neuronal soma (ImageJ). n = 27–29 animals/condition. One-way ANOVA with Holm-Sidak’s test (ns = not significant). All data are presented as mean ± SD.

Results

To investigate the neuronal consequences of cytoplasmic TDP-43 accumulation in vivo, we generated transgenic C. elegans strains expressing human wild-type TDP-43 (hTDP-43-WT) or a nuclear localization-deficient variant (hTDP-43-ΔNLS) specifically in serotonergic neurons. To direct expression specifically to serotonergic neurons, we used the well-characterized tph-1 promoter. This exact promoter fragment is derived from the gene encoding tryptophan hydroxylase, the rate-limiting enzyme for serotonin synthesis, and has been rigorously demonstrated to drive gene expression exclusively in the ADF, NSM, and HSN serotonergic neurons in hermaphrodites³². The serotonergic system in C. elegans serves as an ideal platform for investigating neuronal dysfunction, as its activity states are tightly coupled to food availability and directly modulate three core, quantifiable behaviors: (1) feeding rate, (2) locomotion speed, and (3) egg-laying frequency. These stereotypic responses provide a behavioral triad that is highly reproducible across experiments, mechanistically well-characterized, and easily quantifiable through standardized assays^17–22.

In C. elegans, feeding rate can be quantified by measuring pharyngeal pumping - the rhythmic contraction of the muscular pharynx, which functions as the worm’s feeding organ³³. This tubular structure connects the mouth to the intestine and exhibits stereotyped pumping motions that are easily countable under microscopy. The pumping frequency directly correlates with food intake and is modulated by serotonergic signaling²⁰. We, therefore, quantified the pharyngeal pumping of young animals on food. We observed striking differences between strains. Wild-type animals displayed robust pharyngeal pumping (333.4 ± 13.8 pumps/min) when exposed to bacterial food (Fig. 1a). As previously reported³², tph-1 (mutants deficient in serotonin synthesis) showed severely impaired pumping (158.2 ± 14.1 pumps/min, p < 0.001 versus wild-type) (Fig. 1a). Animals expressing hTDP-43-WT exhibited an intermediate phenotype (283.3 ± 14.1 pumps/min), suggesting partial disruption of serotonergic signaling. Most notably, hTDP-43-ΔNLS animals showed significantly reduced pumping rates (221.6 ± 12.8 pumps/min, p < 0.001 versus hTDP-43-WT) that approached but did not reach the severe deficit observed in tph-1 null mutants (Fig. 1a).

The enhanced slowing response, where fasted animals abruptly decelerate upon encountering food, is a well-established serotonin-mediated behavior in C.elegans^17,18,34,35. To test whether this response was affected by TDP-43 expression, we analyzed locomotion of 1 h-fasted animals as they approached and encountered bacterial lawns (Fig. 1b). Consistent with previous reports, wild-type animals exhibited a sharp deceleration when approaching the food edge (40.1 ± 30.6% of the initial speed) (Fig. 1b). In contrast, serotonin-deficient tph-1 mutants exhibited a markedly reduced slowing response (71.1 ± 31.9% initial speed). While we observed no significant difference between wild-type and hTDP-43-WT animals, those expressing hTDP-43-ΔNLS in serotonergic neurons exhibited severely diminished slowing, comparable to that of tph-1 mutants (72.1 ± 26.1% of the initial speed).

To complete our functional characterization of serotonergic dysfunction in TDP-43-expressing animals, we quantified egg-laying frequency, a behavior primarily controlled by the HSN serotonergic neurons^36,37. As expected, tph-1 null mutants exhibited dramatically reduced egg-laying rates (6.2 ± 1.6 eggs/hour) compared to wild-type animals (12.9 ± 2.4 eggs/hour; p < 0.001), confirming the essential role of serotonin in this behavior (Fig. 1c). Animals expressing hTDP-43-WT showed an intermediate phenotype (9.4 ± 1.4 eggs/hour; p < 0.05 versus wild-type), while those expressing hTDP-43-ΔNLS displayed more severe impairment in egg laying (7.3 ± 1.3 eggs/hour; p < 0.05 versus hTDP-43-WT) (Fig. 1c).

Collectively, our behavioral analyses reveal a consistent pattern across three distinct serotonin-dependent processes: feeding, locomotion modulation, and egg-laying. In general, we observed a graded phenotypic severity, with hTDP-43-ΔNLS causing more pronounced dysfunction than hTDP-43-WT. Since qPCR analysis detected no significant differences in transgene mRNA levels between the transgenic strains (Fig. S1), the observed behavioral differences are unlikely to stem from differential expression levels. To confirm that the observed behavioral deficits were not due to a generalized neurological impairment, we assessed locomotion under conditions independent of serotonergic signaling. We measured swimming behavior in a liquid medium in the absence of food. In this context, the serotonergic system is inactive^18,35, and locomotion depends almost exclusively on the core locomotor circuit, which requires the coordinated function of cholinergic and GABAergic motor neurons³⁸. We did not detect significant alterations on swimming speed in hTDP-43 transgenic strains under these specific conditions (Fig. S2), indicating that the fundamental neuromuscular machinery required for movement remains intact. This suggests that the impairments identified in our primary assays are highly specific to serotonin-dependent behaviors.

Taken together, our results demonstrate that cytoplasmic accumulation of hTDP-43 severely disrupts serotonergic neuron function. The phenotypic severity of the hTDP-43-ΔNLS strain closely resembled the complete serotonin deficiency of tph-1 mutants in most behavioral assays, while expression of wild-type hTDP-43 produced intermediate defects. This pattern indicates that cytoplasmic mislocalization of hTDP-43 profoundly impairs serotonergic signaling mechanisms.

To determine whether neurons expressing TDP-43 variants retain functionality, we pharmacologically challenged the animals with fluoxetine and assessed the egg-laying behavior. In wild-type animals, the HSN serotonergic neurons control egg-laying through the coordinated release of serotonin, acetylcholine, and neuropeptides^36,37. Fluoxetine is a selective serotonin reuptake inhibitor that robustly stimulates egg-laying^26,39,40. Strikingly, this response to fluoxetine only partially depends on serotonin, as tph-1 mutants (lacking serotonin synthesis) retain significant fluoxetine sensitivity⁴⁰. In contrast, it has been demonstrated that fluoxetine resistance occurred in egl-1 mutants where HSN neurons are absent, indicating that these neurons mediate both serotonin-dependent and independent components of the response^26,36,41. Unlike egl-1 mutants, which completely lacked fluoxetine responsiveness, we found that all TDP-43-expressing strains retained some capacity to increase egg-laying following fluoxetine treatment (Fig. 2), indicating that HSN neurons expressing TDP-43 variants maintain basic viability and function. However, quantitative analysis revealed significant differences in the magnitude of this response. While wild-type animals showed robust induction of egg-laying by fluoxetine (11.6 ± 2.0 eggs/hour), hTDP-43-WT expressing worms exhibited a markedly reduced response (5.8 ± 1.5 eggs/hour, p < 0.01 vs. wild-type). Importantly, hTDP-43-ΔNLS animals showed even weaker induction (4.5 ± 1.2 eggs/hour, p < 0.01 vs. hTDP-43-WT; p < 0.001 vs. wild-type), indicating more severe functional impairment. This progressive deficit indicates that cytoplasmic TDP-43 accumulation is a key driver of neuronal dysfunction, even in the absence of complete loss of function.

Finally, to determine whether the observed behavioral impairments were associated with structural alterations in serotonergic neurons, we examined neuronal morphology in young adult animals (the same age at which behavioral deficits were detected) using a transcriptional reporter (Ptph-1::GFP). This marker selectively labels the serotonergic neurons ADF, NSM, and HSN, enabling detailed visualization of their soma positioning, neurite projections, and overall integrity. The use of GFP reporters to assess neuronal integrity is an extensively validated method in C. elegans for multiple neuronal classes, including serotonergic neurons^42–45. Strikingly, we detected no significant morphological abnormalities in any of the examined neurons across strains (Figs. 3 and 4). The somas of ADF, NSM, and HSN neurons were present in 100% of animals, with no differences in their positioning or survival compared to wild-type controls. Furthermore, quantitative analysis of GFP fluorescence intensity—a proxy for neuronal health—revealed no differences between control and TDP-43-expressing strains (p > 0.05 for all comparisons) (Figs. 3 and 4).

These results demonstrate that the deficits in serotonin-dependent behaviors caused by hTDP-43-WT and hTDP-43-ΔNLS expression (e.g., impaired feeding, locomotion modulation, and egg-laying) occur even without neuronal loss or gross morphological disruption. The dissociation between behavioral dysfunction and preserved neuronal structure suggests that cytoplasmic TDP-43 accumulation impairs subtle, functional aspects of serotonergic signaling—such as neurotransmitter synthesis, vesicular release, or postsynaptic responsiveness—even before morphological evidence of neurodegeneration becomes apparent.

Discussion

TDP-43 proteinopathies are a hallmark of several neurodegenerative disorders, including ALS and FTD. While traditionally associated with motor neuron degeneration, recent studies have highlighted that other neuronal populations—such as serotonergic neurons—may also be affected in these conditions. For example, reduced serotonergic innervation has been reported in the spinal cord of ALS patients and the brainstem of SOD1(G86R) mouse models⁴⁶. Furthermore, increased expression of 5-HT2B serotonin receptors has been documented in the spinal cord of ALS mice at symptom onset, and Htr2b gene knockout exacerbated disease severity in SOD1(G86R) mice^47,48.

In this study, we generated transgenic C. elegans expressing human TDP-43 (hTDP-43) in serotonergic neurons, using either the wild-type (hTDP-43-WT) or a nuclear localization-deficient mutant (hTDP-43-ΔNLS) that accumulates in the cytoplasm. In these strains, we assessed how targeted mislocalization of TDP-43 affects neuronal function and serotonergic-related behavior. While this work does not seek to fully recapitulate the complex etiology of human ALS, it provides a powerful and simplified in vivo platform to isolate and study the early, circuit-specific functional deficits caused by TDP-43 pathology. The key advantage of this approach is that it provides a set of robust and easily quantifiable behavioral readouts—pharyngeal pumping, egg-laying, and locomotor modulation—to directly assess the functional integrity of these specific neurons in a living animal.

Our behavioral analyses revealed that both hTDP-43 variants impaired serotonin-dependent behaviors such as pharyngeal pumping, egg-laying, and food-induced locomotion modulation. Notably, expression of cytoplasmic hTDP-43-ΔNLS caused more pronounced deficits than hTDP-43-WT, with phenotypes that mimicked those of tph-1 mutants, which are unable to synthesize serotonin³². This graded severity suggests that cytoplasmic mislocalization of TDP-43 compromises serotonergic function, potentially by partially disrupting neurotransmitter synthesis, release, or downstream signaling^49,50, rather than by inducing complete cell loss.

Consistent with this hypothesis, we found that TDP-43-expressing neurons remained structurally intact. Fluorescence microscopy using a tph-1::GFP reporter revealed no overt abnormalities in cell body positioning, axonal projections, or GFP intensity. Moreover, pharmacological exposure to fluoxetine, a selective serotonin reuptake inhibitor, induced egg-laying in TDP-43-expressing animals, albeit to a lesser extent than in wild-type controls. This partial responsiveness implies that serotonergic neurons retain the capacity to release neurotransmitters, further supporting the notion of an early functional, rather than structural, impairment. Interestingly, although fluoxetine can activate egg-laying independently of serotonin in tph-1 mutants, its effects are abolished in egl-1 animals, which lack the serotonergic HSN neurons altogether^26,36. The preserved fluoxetine responsiveness in our TDP-43 models suggests that HSN neurons remain viable and responsive, even in the presence of cytoplasmic TDP-43 accumulation.

The functional deficits resulting from TDP-43-ΔNLS expression are consistent with the expected pathological cytoplasmic mislocalization driven by this mutation, a phenomenon long associated with the severity of neuronal damage in TDP-43 proteinopathies^2,51; the stark phenotypic differences we observed strongly affirm that disrupting TDP-43’s nucleocytoplasmic transport is a key determinant of neuronal dysfunction. This is supported by the well-established role of the ΔNLS mutation in driving cytoplasmic accumulation^23,28. In addition, while our data establish a clear link between TDP-43 expression and serotonergic dysfunction, the precise subcellular mechanism remains to be fully elucidated. Future studies employing techniques such as synaptic vesicle labeling, optogenetic stimulation, and calcium imaging in defined neurons will be essential to directly assess functional parameters like neurotransmitter synthesis, synaptic transmission efficiency, and postsynaptic responsiveness.

Altogether, our results support a model in which cytoplasmic TDP-43 accumulation disrupts serotonergic function without inducing overt neurodegeneration. This dissociation between behavioral dysfunction and structural preservation mirrors clinical findings in early-stage neurodegenerative diseases, where subtle behavioral impairments often precede detectable neuronal loss⁵². Our model thus provides a tractable platform for dissecting the early, circuit-specific effects of TDP-43 pathology.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

DR and LMI designed the research; AL, SR and DR performed the experiments; AL, DR, FV, SR and LMI analyzed data; AL prepared all the figures; DR, FV, and LMI wrote the manuscript. All authors edited, revised, and approved the manuscript before submission.

Funding

This work was supported by Grants from: (1) Agencia Nacional de Promoción de la Ciencia y la Tecnología ANPCYT Argentina to LMI (PICT 2019 − 1585) and DR (PICT 2019 − 0480 and PICT-2021-I-A-00052); (2) Consejo Nacional de Investigaciones Científicas y Técnicas, Argentina to DR (PIP No. 11220200101606CO); and (3) Universidad Nacional Del Sur to DR (PGI: 24/B291). The funders had no role in the study design, data collection, and analysis, decision to publish, or preparation of the manuscript.

Data availability

The datasets generated and/or analysed during the current study are available in the OSF repository (https://osf.io/zpw43/?view\_only=9dc9afe51f6a4e3a959d7f17c6f8a8f7).

Declarations

Competing interests

The authors declare no competing interests.