Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Impaired dynein function preserves spinal interneuron survival and positioning in an ALS-like mouse model

Abstract

Impaired cytoplasmic dynein function has been implicated in amyotrophic lateral sclerosis (ALS) pathogenesis, yet the contributions of spinal interneurons to disease phenotypes remain unclear. We tested the hypothesis that hypomorphic dynein function in cholinergic neurons disrupts the development, survival, or positioning of inhibitory interneuron populations in the lumbar spinal cord. Using ChAT-Cre recombination, we generated four mouse genotypes with graded reductions in dynein activity in ChAT+ cells: Dync1h1+/+ (wildtype), Dync1h1−/+ (hemizygous wildtype), Dync1h1+/Loa (heterozygous Loa mutation), and Dync1h1−/Loa (hemizygous Loa). At 52 weeks of age, lumbar spinal cords (L3–L6) were harvested, cryosectioned, and immunostained for ChAT, GAD-67, Parvalbumin, and Calbindin. Cell counts were performed on confocal images from eight sections per mouse (N = 3 male mice/genotype), and radial distances from the central canal were normalised to gray matter width. Angular distributions were analysed via circular statistics. There were no significant genotype-dependent differences in the numbers of ChAT+, GAD-67+, Parvalbumin+, or Calbindin+ cells, nor in ChAT+ subpopulations (motor neurons versus interneurons) or double‐positive interneuron subsets (e.g., ChAT+–GAD-67+, Parvalbumin+–GAD-67+, Parvalbumin+–Calbindin+). Radial positioning relative to the central canal was similarly preserved across all markers and genotypes. Circular‐median tests revealed statistically significant shifts in mean angle for ChAT+, GAD-67+, and certain double‐positive cells, but these amounted to only 5–10° displacements, translating to lateral shifts of ~10–20 µm, well within single laminar bands, and are unlikely to impact circuit connectivity. Despite substantial motor deficits and hallmark TDP-43 pathology previously seen in these models, impaired dynein function does not precipitate interneuron loss or gross migratory defects in the lumbar spinal cord. Instead, our findings suggest that the primary contributions of dynein to ALS-like phenotypes likely arise from functional disruptions in axonal transport, synaptic maintenance, and neuronal physiology rather than from structural alterations or loss of interneuron populations.

Citation: Christoforidou E, Rowe JS, Simoes FA, Cassel R, Dupuis L, Leigh PN, et al. (2026) Impaired dynein function preserves spinal interneuron survival and positioning in an ALS-like mouse model. PLoS One 21(4): e0346246. https://doi.org/10.1371/journal.pone.0346246

Editor: Xiaona Wang, Zhengzhou University, CHINA

Received: October 15, 2025; Accepted: March 16, 2026; Published: April 2, 2026

Copyright: © 2026 Christoforidou et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the manuscript and its Supporting Information files.

Funding: This work was funded by the Motor Neurone Disease Association grants Christoforidou/Nov24/2494-793 (to EC, MH, and PNL) and Dupuis/Apr16/852-791(to LD and MH). The funders of this project had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Amyotrophic Lateral Sclerosis (ALS) is a fatal neurodegenerative disorder characterised by the progressive loss of motor neurons, leading to muscle weakness, paralysis, and eventual death [1]. Despite extensive research, the underlying mechanisms remain poorly understood, and therapeutic options are limited. Interneurons have historically been neglected in ALS research partly because they are less visible to pathologists using conventional methods [2]. Traditional pathological examinations often miss these cells due to their small size and less prominent morphological characteristics compared to motor neurons. However, interneurons play a crucial role in modulating motor neuron activity and overall neural circuitry [3].

Pathological changes in interneurons have been documented in ALS [2], but the importance of these changes has not yet been established. Specifically, interneurons might experience changes in their excitability, connectivity, or neurotransmitter release in response to ALS pathology [4–6]. Moreover, muscle twitches (fasciculations) observed in ALS have traditionally been attributed to cell-autonomous defects in motor neurons [7]. However, recent studies suggest that these muscle twitches may also be linked to abnormal interneuron function [5,8].

Previous research showed changes in the density of some interneuron subtypes in the cortex of people with ALS [9]. Furthermore, a mouse model carrying a point mutation (M323K) within the endogenous Tardbp gene which codes for the TDP-43 protein shows progressive cortical interneuron degeneration [10]. Although mutations in TARDBP are causative for only ~4% of familial and ~1% of sporadic cases of ALS, TDP-43 pathology is present in ~97% of all ALS cases [11,12]. Additionally, findings from Project MinE [13], containing whole genome sequencing (WGS) data from a large number of ALS cases (6,538) and controls (2,415), have identified ALS patients with rare variants in genes not typically associated with ALS, suggesting these may contribute to disease susceptibility. One such gene is dynein, cytoplasmic heavy chain 1 (DYNC1H1) which encodes the heavy chain subunit of cytoplasmic dynein 1 (henceforth ‘dynein’), a motor protein essential for neuronal cell motility and morphology, and the primary mediator of retrograde transport. Dynein has not consistently appeared in large-scale ALS genetic studies, but rare variants have been reported in motor neuron diseases, including certain ALS cases [14,15] and lower extremity–predominant spinal muscular atrophy (LED-SMA) [16,17]. While these variants are uncommon, they point to dynein dysfunction as a plausible contributor to disease mechanisms. As ALS is primarily sporadic, likely involving a combination of subthreshold genetic variants and environmental or lifestyle factors, gene variants in components of dynein may predispose individuals to ALS by perturbing interneuron networks within the motor neuron pools in the spinal cord. One possible cause of such perturbation is impaired dynein-mediated neuronal migration during neurodevelopment, which is a novel hypothesis in understanding ALS pathology. However, prior research has shown that mutations in dynein or its accessory proteins can lead to cortical neuron migration defects and thus brain malformations [18].

The F580Y Legs at odd angles (Loa) mutation in the mouse Dync1h1 disrupts dynein complex integrity and leads to defective cortical neuron migration, arborisation of neurons, and impaired retrograde axonal transport [19–24]. Importantly, our recent study has shown that the Loa mutation in mice can recapitulate cellular hallmarks of ALS by promoting aggregation of TDP-43 and p62 [25]. In another of our recent studies [26], we have shown that the combination of defective dynein with a TARDBP transgene (encoding TDP-43M337V) in mice triggered TDP-43 pathology, partially recapitulating human ALS. These results suggest a complex interplay between protein aggregation and impaired dynein function, and underscore the need to further investigate the role of dynein in ALS.

To study the contribution of impaired dynein function to the onset of ALS, we previously generated mice harbouring a spectrum of the wildtype or Loa alleles of Dync1h1 in cholinergic motor neurons using ChAT-Cre expressing mice. The ChAT-Cre line expresses Cre recombinase under the control of the endogenous choline acetyltransferase (Chat) promoter, with the inclusion of an internal ribosome entry site (IRES) allowing for the efficient expression of Cre in conjunction with the ChAT gene. The genotypes generated include: (1) Dync1h1+/+ (wildtype), (2) Dync1h1−/+ (hemizygous wildtype; one allele knocked out in ChAT+ cells), (3) Dync1h1+/Loa (heterozygous mutant; one allele carries the Loa mutation), and (4) Dync1h1−/Loa (hemizygous mutant; one Loa mutant allele, one allele knocked out in ChAT+ cells). This order was hypothesised to reflect a dose-dependent reduction in dynein function that leads to progressively worse ALS-like phenotypes.

In our previous characterisation of these mouse strains [25], both Dync1h1+/Loa and Dync1h1−/Loa mice displayed impairments in motor performance relative to wildtypes, including reduced grip strength detectable from approximately 4–8 weeks of age. These deficits persisted throughout adulthood and were accompanied by reduced body weight from around 8 weeks onwards. The more severe Dync1h1−/Loa genotype exhibited an earlier and steeper decline in grip strength and a progressively greater reduction in body weight compared with Dync1h1+/Loa mice, with the divergence between the two mutants becoming increasingly apparent with age and reaching statistical significance by ~40–52 weeks. In addition to these behavioural phenotypes, structural abnormalities at neuromuscular junctions (NMJs), including endplate fragmentation and reduced innervation, were already detectable at 8 weeks and persisted into later adulthood. Despite this, there was no loss of spinal motor neurons in our mouse models. It is possible that defects in the interneuron pools of the spinal cord may exacerbate the impact of aberrant NMJs, thereby contributing to the impaired motor function phenotype observed in these mice. This suggests that any aberrant interneuron migration could play a critical role in the overall motor deficits in these mice, highlighting the need for further investigation into the interplay between motor neurons and interneurons in this context.

Therefore, we hypothesised that disruption of dynein function may interfere with the development and/or survival of inhibitory interneurons in the spinal cord. To test this hypothesis, we explored whether there are changes in interneuron density and distribution in the spinal cord of our mouse models that exhibit other ALS-like pathology. Based on our previously observed phenotypic progression, we selected 52 weeks of age as a late adult timepoint at which motor deficits and associated neuromuscular pathology are well established. It is important to note that in these models the Loa mutation in Dync1h1 is present in all cells, whereas the additional reduction in dynein function occurs selectively in cholinergic neurons through Cre-mediated deletion of the floxed Dync1h1 allele. The inhibitory interneuron populations analysed in this study represent broader inhibitory classes that are not necessarily part of the ChAT lineage. Consequently, while these interneurons may carry the Loa mutation, they do not experience the same cholinergic-specific reduction in dynein function. As a result, the present study primarily examines whether dynein dysfunction in cholinergic neurons leads to indirect or circuit-level effects on the survival or spatial organisation of inhibitory interneuron populations within the lumbar spinal cord. By focusing on interneurons, we aimed to shed light on their potential contribution to ALS pathology, which has been underexplored, but which may be important in pathogenic mechanisms such as excitotoxicity [27–29].

Materials and methods

Animals

Procedures involving animals were approved by the University of Sussex Animal Welfare and Ethics Review Board (AWERB) and were conducted under a project licence issued by the UK Home Office in accordance with the UK Animals (Scientific Procedures) Act 1986. All experiments were performed in accordance with relevant institutional and national guidelines for the care and use of laboratory animals. Mice were on a C57Bl/6J background and were housed under a 12-hour light-dark cycle (lights on at 7 am) with free access to food and water and with up to five littermates. To generate mice with one Dync1h1 allele knocked out only in cholinergic (ChAT+) cells, a 3-step breeding program using Cre-Lox recombination was carried out as we previously described in [25]. This generated either wildtypes (Dync1h1+/+), heterozygous mutants (Dync1h1+/Loa), hemizygous wildtypes (Dync1h1−/+) where one Dync1h1 allele was absent only in ChAT+ cells, or hemizygous mutants (Dync1h1−/Loa) where again one Dync1h1 allele was absent only in ChAT+ cells and the Loa mutation was present in all cells. The genotyping procedure and the confirmation of successful recombination have also been previously described [25]. All procedures were conducted in accordance with the principles of the 3Rs. Animals were monitored regularly for health and welfare throughout the study by trained personnel in accordance with UK Home Office guidelines. Humane endpoints were predefined, and animals showing signs of pain, distress, or significant deterioration in condition (e.g., weight loss, impaired mobility, or abnormal behaviour) would have been humanely euthanised by perfusion fixation under terminal anaesthesia. No unexpected adverse events occurred during the course of the study.

Perfusion and tissue collection

Only males were used in this study. This was done to maintain consistency with our previous characterisation of this mouse model, in which sex-dependent differences in body weight, grip strength and neuromuscular junction phenotypes were observed, with males showing a more pronounced and consistent phenotype. At 52 weeks of age, male animals were anaesthetised with an overdose of pentobarbital and transcardially perfused using a peristaltic pump with oxygenated artificial cerebrospinal fluid (125 mM NaCl, 2.5 mM KCl, 2.5 mM CaCl2, 2 mM MgCl2.6H2O, 26 mM NaHCO3, 1.25 mM NaH2PO4.H2O, 25 mM C6H12O6). The vertebral column was dissected and cut cross-sectionally between vertebrae T9 and T10 and between vertebrae L6 and S1. The lumbar portion of the vertebral column (T10-L6) was drop-fixed in 4% formaldehyde for 24 hours, followed by removal of the spinal cord from the vertebrae which was then stored in 30% sucrose with 0.2% sodium azide at 4°C until cryosectioning.

Cryosectioning

Fixed spinal cords were mounted in optimal cutting temperature (OCT) compound (Agar Scientific, AGR1180) and sliced cross-sectionally in 30 μm-thick sections using a cryostat at −17°C. Each section was kept in antifreeze buffer (30% ethyleneglycol, 30% glycerol, 192.5 mM NaOH, 244 mM NaH2PO4) at 4°C until immunostaining.

Immunohistochemistry of spinal cord sections

Two sections per lumbar L3-L6 segment were immunostained for ChAT, Calbindin, Parvalbumin, and GAD-67 at the same time per animal. This approach ensured systematic sampling across the lumbar enlargement, with an equal number of sections analysed from each anatomical level in every animal. Sampling based on anatomical levels was used instead of fixed section intervals because the mouse genotypes exhibit significant differences in body weight at this age, which could plausibly result in differences in spinal cord length. All steps were performed in Coplin jars with ~40 ml of solution, except for the antibody incubations and the antigen retrieval step. Briefly, the tissue was mounted onto Superfrost Plus glass microscope slides (Fisher, 10149870) then washed three times (15 min each) in phosphate-buffered saline (PBS). Antigen retrieval was performed with sodium citrate buffer (10 mM Sodium citrate, 0.05% Tween 20, pH 6.0) prewarmed to 80–95°C for 30 minutes on a rocker at 30 osc/min. Permeabilization was performed with 0.05% Tween-20 (in PBS) for 10 minutes at 100 rpm (three times). Blocking was performed with PBS containing 0.1% Tween-20, 0.1M glycine, and 5% horse serum (Fisher, 11520516) for 90 minutes at 100 rpm. The sections were then probed overnight in a humidity chamber on an orbital shaker at 100 rpm and at 4°C in PBS containing the primary antibody solution (0.05% Tween-20, 1% horse serum) for polyclonal goat anit-ChAT (1:50 dilution, Merck/Sigma, AB144P), polyclonal guinea pig anti-calbindin D28k (1:200 dilution, Synaptic Systems, 214 004), polyclonal rabbit anti-parvalbumin (1:500 dilution, Synaptic Systems, 195 002), and polyclonal chicken anti-GAD67 (1:50 dilution, Bio-Techne, NBP1−02161). Following three 10-min PBS washes at 100 rpm, secondary probing was performed for 3 hours in the dark in PBS containing 0.05% Tween-20 and the relevant secondary antibodies for ChAT (1:250 dilution, donkey anti-goat IgG H + L Alexa Fluor® 405, Abcam, ab175664), Calbindin (1:500 dilution, Rhodamine TRITC-AffiniPure donkey anti-guinea pig Ig G H + L, Jackson Immuno Research, 706-025-148-JIR), Parvalbumin (1:500 dilution, Alexa Fluor 647-AffiniPure donkey anti-rabbit IgG H + L, Jackson Immuno Research, 711-605-152), and GAD-67 (1:500 dilution, donkey anti-chicken IgY H + L Highly Cross Adsorbed Alexa Fluor® 488, Life Technologies, A78948). Finally, four washes in PBS (15 min each) were performed at 100 rpm, and the slides were sealed with Invitrogen™ ProLong™ Glass Antifade Mountant (Fisher, 15820993).

Image acquisition

Imaging was performed blinded to genotype. Images were acquired using a Leica SP8 confocal microscope with LAS X version 3.4 software. A 20x objective was used; however, since the entire spinal cord cross-section could not be captured in a single field of view, a tile scan acquisition was performed. Each tile of the spinal cord cross-section was imaged as a z-stack in four separate fluorescence channels, with each channel corresponding to a specific immunohistochemical marker: ChAT, Calbindin, Parvalbumin, or GAD-67. The individual tiles were then merged into a single mosaic image using the mosaic merge function of LAS X with statistical overlap blending.

Image processing

Image analysis was performed blinded to genotype using custom-written ImageJ macros and MATLAB functions which we made publicly available [30]. The complete dataset analysed is also accompanying this publication (S1 File). First, maximum intensity projections were generated from z‐stacks of each colour channel image to produce 2D representations of spinal cord cross‐sections, allowing us to distinguish the different cell populations present in the tissue. Next, a region of interest (ROI) delineating the gray matter was manually drawn and saved for subsequent masking. The gray matter ROI was used to mask out non-gray matter areas. Then, a background subtraction step was implemented by employing a rolling ball algorithm with a radius of 20 pixels, which estimates and subtracts the smoothly varying background intensity from the image, enhancing the contrast of cellular structures. Manual thresholding was then applied to each individual channel image using the ‘MaxEntropy’ method which maximises the entropy between the dark (background) and bright (foreground) regions, to ensure accurate segmentation of the cells. The thresholded images were converted to binaries, and noise reduction followed by morphological closing were performed. Particle analysis was then used with cell type-specific size limits to extract cell centroids and boundaries. Specifically, ChAT-positive cells are expected to have a 10–40 µm diameter (area of 78.5–1256.6 µm2) [31,32], Parvalbumin-positive cells a 10–20 µm diameter (area of 78.5–314 µm2) [33], GAD-67-positive cells a 8–20 µm diameter (area of 50–314 µm2) [34], and Calbindin-positive cells a 10–25 µm diameter (area of 78.5–490.9 µm2) [35]. Additionally, when investigating ChAT-positive subpopulations, we divided them into interneurons of 10–25 μm diameter and motor neurons of 25–40 μm diameter.

Colocalisation analysis

To quantify colocalisation, we loaded each cell’s ROI—defined on marker A (e.g., ChAT)—and measured how much of that ROI overlapped thresholded pixels in the binary mask for marker B (e.g., GAD-67). We then computed the fraction of the ROI’s area that was positive for marker B and only retained cells for which this overlap was ≥ 20%. This 20% cutoff was chosen to exclude spurious edge‐touches or small contacts while still capturing genuine co-expression events.

Cell distribution analysis

The central canal was manually delineated, and its centroid was computed. Subsequently, we computed the straight-line distance from each cell’s centre to the central canal’s centre, and we determined the orientation angle by using the inverse tangent of the differences in their horizontal and vertical positions. Finally, to adjust for differences in tissue size between samples, we normalised the cell distances relative to the gray matter. We did this by creating a binary mask from the saved gray matter ROI, computing a distance map that assigns each pixel a value based on its distance from the tissue edge, and then dividing each cell’s distance (sampled from this map) by the maximum distance value to obtain normalised distances.

Angle transformation for laterality correction

To ensure that cell distribution analyses were not confounded by arbitrary left-right orientations of free-floating spinal cord sections, we applied a transformation to the angle data. Angles were originally computed in ImageJ relative to the central canal, using the mathematical convention where 0° corresponds to the right, 90° to the dorsal side, 180° to the left, and 270° to the ventral side. Because the true anatomical left and right sides cannot be determined for each section (given the inherent symmetry of the tissue and the possibility of different mounting orientations) we transformed the angle data to eliminate any bias. Specifically, for each cell, if the angle was between 90° and 270° (indicating a location on the left side), it was transformed using: θ′ = 180° − θ. Angles on the right side (0°–90° and 270°–360°) remained unchanged. This transformation effectively mirrored left-side angles onto the right side while maintaining dorsoventral positioning (Fig 1). This transformation ensured that left-right differences were removed, preventing the detection of artificial laterality due to arbitrary assignment of left and right.

thumbnail
Download:
Fig 1. Angle transformation for laterality correction.

Cells in the gray matter between 90° and 270° (left side; white cells) were mirrored onto the right side of the gray matter. The angle of cells originally on the right side (black cells) was not altered.

https://doi.org/10.1371/journal.pone.0346246.g001

Statistical analysis

Statistical analysis of linear data was performed using GraphPad Prism version 10. The Shapiro-Wilk test was used to confirm that all data were normally distributed and then an ordinary one-way analysis of variance (ANOVA) or a nested one-way ANOVA was used for the analysis of cell numbers and cell distances, respectively. No post hoc tests were used as none of the main effects were statistically significant at the 5% alpha level. Analysis of circular data was performed using the CircStat toolbox [36] in MATLAB R2024b. A circular median test (equivalent to a Kruskal-Wallis test for linear data) was used to compare cell angles between genotypes, followed by the Kuiper post hoc test, and adjusted for multiple comparisons using the False Discovery Rate (FDR) approach, specifically via the Benjamini–Hochberg procedure and through custom-written code which we made publicly available [30].

Results

We immunostained spinal cord cross-sections from lumbar levels L3–L6 of mice with increasingly reduced dynein function and then performed confocal microscopy to identify different cell populations. An example of an immunostained spinal cord cross-section from each genotype is shown in Fig 2. First, we examined whether there were any changes in the numbers of different cell populations in the lumbar spinal cord of mice, by looking at each marker independently. We found no statistically significant differences in the number of ChAT+ cells between any of the genotypes (Fig 3A). This suggests there is no loss of cholinergic motor neurons in the lumbar spinal cord of mice with increasingly reduced dynein function. Previously, in a different set of 12 mice (three per genotype), we observed a non-significant trend towards cell loss with increasing dynein dysfunction, when counting Nissl-stained motor neurons (assumed but not confirmed to be ChAT+) specifically in the sciatic motor pool of the lumbar (L3–L6) spinal cord [25]. Here, the same non-significant trend was observed with the ChAT+ cells even without restricting the counting to the sciatic motor pool. This corroborates our prior conclusions that there is no significant motor neuron loss in the lumbar spinal cord of these mice, however, this could also be due to both studies being underpowered (i.e., only three biological replicates per genotype). It should be noted, however, that our analysis of ChAT+ cells was done globally on the entire gray matter, therefore, a small proportion (5–10%) of ChAT+ cells in Fig 3A are likely to be interneurons residing in the intermediate zone of the gray matter (lamina VII) and in regions bordering the central canal (lamina X). To investigate this, we divided the ChAT+ cells into two subpopulations based on their size, with interneurons represented by cells of 10–25 µm diameter (~78.5–491 µm2 area) and motor neurons represented by cells of 25–40 µm diameter (491–1257 µm2 area) [37]. As with the overall ChAT+ cells, this subpopulation analysis indicated no significant differences in the number of ChAT+ interneurons or motor neurons between genotypes, despite a trend towards a reduction in cell numbers with increasing dynein dysfunction (Fig 3B-C). Additionally, we also did not observe a significant loss of GAD-67+, Calbindin+, or Parvalbumin+ cells (Fig 3D-F), although, GAD-67+ cells showed the same non-significant trend towards a reduction of cell numbers with increasing dynein dysfunction as ChAT+ cells (Fig 3D).

thumbnail
Download:
Fig 2. Representative confocal microscopy images of immunostained spinal cord cross-sections from each genotype.

(A) Full mosaic-merge images. Images were adjusted for brightness and contrast for illustrative purposes only. (B) Corresponding insets from A to show examples of positively stained cells more clearly. Scale bars = 500 μm (A) or 200 μm (B).

https://doi.org/10.1371/journal.pone.0346246.g002

thumbnail
Download:
Fig 3. Number of ChAT+ (A-C), GAD-67+ (D), Calbindin+ (E), Parvalbumin+ (F), Parvalbumin+−Calbindin+ (G), Parvalbumin+−GAD-67+ (H), and ChAT+−GAD-67+ (I) cells per biological replicate. The number of cells analysed is the sum of immunostained positive cells from eight spinal cord cross-sections (two sections per lumbar level L3–L6) per biological replicate. Data are shown as mean ± standard deviation and were analysed by an ordinary one-way ANOVA. N = 3 biological replicates per genotype.

https://doi.org/10.1371/journal.pone.0346246.g003

The above single‐marker analyses label broad classes, such as all cholinergic cells (ChAT+) or all GABAergic cells (GAD-67+), but many functionally distinct interneurons are defined by co-expression of two proteins. Therefore, we next investigated each double‐positive population across genotypes to determine whether dynein impairment selectively compromises particular spinal cord microcircuits, even if overall single‐marker counts remain unchanged. Firstly, a subset of cholinergic neurons (ChAT+) around the central canal (lamina X) co-express GAD-67, and these are thought to be cerebrospinal-fluid–contacting interneurons, which are sometimes called central autonomic or central cluster cells [38]. Secondly, almost all Parvalbumin+ neurons in the spinal cord are GABAergic and thus co-express GAD-67. These are typically called fast-spiking GABAergic interneurons. The two main populations include dorsal horn presynaptic inhibitory interneurons, which gate afferent input, and ventral Ia‐inhibitory interneurons [39]. Finally, Parvalbumin and Calbindin are co-expressed in Renshaw cells, the classic recurrent‐inhibitory interneurons in laminae VII/VIII, which is expected to make up only a small subset of all Parvalbumin+ cells [39]. In our tissues, we did not observe a significant difference in the number of any of these double-positive cell populations between genotypes (Fig 3G-I). This suggests that impairing dynein function does not produce overt loss of any of the major spinal interneuron populations or their more restricted double‐positive subsets.

Together, these findings indicate that reductions in dynein activity, even when sufficient to drive clear motor deficits in these mice, are not accompanied by statistically significant cell death of either motor neurons or key inhibitory interneuron classes in the lumbar spinal cord. Functionally, this preservation of cell number suggests that the motor dysfunction seen in dynein‐deficient animals is more likely to arise from impairments in axonal transport, synaptic connectivity, or neuronal physiology, rather than from degeneration of spinal neurons themselves. Future studies employing electrophysiological recordings and synaptic ultrastructure analyses will be needed to determine whether hypomorphic dynein perturbs circuit activity or synaptic maintenance in the absence of cell death.

Next, we assessed the distribution of different cell types within the gray matter. The central canal is a well-defined, anatomically conserved structure in the spinal cord, and therefore, we used it as a reference point to standardise measurements across samples. We first calculated the distance of each immunopositive cell from the central canal and normalised it to the gray matter boundaries to ensure that any observed differences were not simply due to variability in overall tissue size. The distance from the central canal can reveal gradients in cell positioning. Certain interneuron populations might cluster close to the central canal, while others are preferentially located in more peripheral regions of the gray matter. Such gradients can suggest differences in developmental origins or migratory cues [40]. In our tissues, we did not observe any statistically significant differences in the distances of cells from the central canal between genotypes, for any of the cell markers we used, either independently (Fig 4A-F) or double-positive (Fig 4G-I). This suggests that any dynein dysfunction does not appear to disrupt the spatial organisation of spinal neurons in the lumbar enlargement. In other words, both single‐marker and co‐labelled cell populations occupy their normal positions relative to the central canal regardless of genotype. This preservation of positional gradients implies that impaired dynein does not affect the ability of the cells to follow the migratory cues and microenvironmental signals guiding interneuron and motor neuron placement during development.

thumbnail
Download:
Fig 4. (A-I) Normalised distances of cells relative to the central canal.

The distances were normalised to the size of the gray matter to account for size differences between spinal cord cross-sections. The normalised distance is a unitless fraction between 0 (closest to the central canal) and 1 (farthest from the central canal). Data are from eight spinal cord cross-sections (two sections per lumbar level L3–L6) per biological replicate. Boxes represent the interquartile range with a line in the middle representing the median. The whiskers represent the minimum and maximum values. Data were analysed by a nested one-way ANOVA. N = 3 biological replicates per genotype.

https://doi.org/10.1371/journal.pone.0346246.g004

In addition to measuring the radial distance of cells from the central canal, we analysed the angular distribution of each cell relative to this central landmark. The angle provides directional information, delineating which areas of the spinal cord the cells occupy. This analysis allowed us to determine whether cells exhibit a preferential orientation, such as a bias toward dorsal or ventral regions, independent of their distance from the central canal. For example, this is already known to be the case for cholinergic motor neurons, which should predominantly be in the ventral horns [41]. Due to the inherent circularity of the angle data, where 0° is equivalent to 360°, standard linear plots like bar graphs and violin plots are not suitable for accurately representing angle measurements. Thus, we employed circular-specific visualisations (i.e., rose diagrams) to display the directional distribution of cell orientations across all lumbar levels together. As expected, we observed most ChAT+ cells (likely motor neurons) to be in the ventral horns (between 270° and 360°), whereas a small proportion (likely interneurons) was located around the central canal (Fig 5A). The angular distribution of Calbindin+, GAD-67+, and Parvalbumin+ cells were also visualised (Fig 5B-D). Similar visualisations were created for double-positive cells (Fig 6A-C).

thumbnail
Download:
Fig 5. (A-D) Spatial distribution of immunopositive cells displayed as rose diagrams (i.e., circular histograms) of cell angles relative to the central canal. All left side angles were mirrored to the right side to avoid displaying artificial laterality due to the inability to distinguish the true left and right sides of free-floating spinal cord cross-sections. The length of each bar represents the percentage of cells found at a particular angle (not the distance from the central canal; distance information is ignored in these plots). Data are from eight spinal cord cross-sections (two sections per lumbar level L3–L6) per biological replicate. N = 3 biological replicates per genotype.

https://doi.org/10.1371/journal.pone.0346246.g005

thumbnail
Download:
Fig 6. (A-C) Spatial distribution of double-positive cells displayed as rose diagrams (i.e., circular histograms) of cell angles relative to the central canal. All left side angles were mirrored to the right side to avoid displaying artificial laterality due to the inability to distinguish the true left and right sides of free-floating spinal cord cross-sections. The length of each bar represents the percentage of cells found at a particular angle (not the distance from the central canal; distance information is ignored in these plots). Data are from eight spinal cord cross-sections (two sections per lumbar level L3–L6) per biological replicate. N = 3 biological replicates per genotype.

https://doi.org/10.1371/journal.pone.0346246.g006

Statistical analysis using a circular median test (equivalent to a Kruskal-Wallis test for linear data) to compare cell angles between genotypes (ignoring lumbar level) revealed a significant effect of genotype for ChAT+ cells (p < 0.0001). Further subdivision of the ChAT+ cells into interneuron and motor neuron subpopulations based on cell size as before, revealed that the significance remained in the interneuron but not the motor neuron subpopulation. Post hoc tests for ChAT+ cells (total) and ChAT+ interneurons were done using the circular-data specific Kuiper test and adjusted for multiple comparisons using the False Discovery Rate (FDR) approach, specifically via the Benjamini–Hochberg procedure. This revealed that the significant differences in the angular distribution of ChAT+ cells/interneurons were between all possible pairwise genotype comparisons (all p < 0.0001) except for Dync1h1+/Loa versus Dync1h1‒/Loa (not significant; Fig 7A-B). The angles of GAD-67+ cells similarly displayed a significant effect of genotype (p < 0.0001), and the post hoc pairwise comparisons revealed this to be between all possible pairwise genotype comparisons (p = 0.033 for Dync1h1+/Loa versus Dync1h1‒/Loa and p < 0.0001 for all others; Fig 7C). Furthermore, the circular median test for ChAT+−GAD-67+ and Parvalbumin+−GAD-67+ double-positive cells also displayed a statistically significant effect of genotype (p < 0.0001 and p < 0.05, respectively). Post hoc tests revealed that the significant differences were again between all possible pairwise genotype comparisons (Fig 7D-E). The angles of Calbindin+, Parvalbumin+, and Parvalbumin+−Calbindin+ double-positive cells did not display a statistically significant effect of genotype (p = 0.2963, p = 0.1588, and p = 0.1453, respectively) and thus no post hoc tests were conducted for these.

thumbnail
Download:
Fig 7. (A-E) Summary of post hoc pairwise comparisons for cell types that displayed significant main effects. The left column shows heatmaps of false discovery rate (FDR) values to indicate statistical significance. The right column shows the mean angle of the cells (circles) and the 95% confidence interval (error bars) for each genotype.

https://doi.org/10.1371/journal.pone.0346246.g007

It is important to interpret these statistically significant differences in context. Although the circular statistics revealed genotype differences for ChAT+, GAD-67+, ChAT+–GAD-67+, and Parvalbumin+–GAD-67+ cells, the actual shifts in mean angle remain modest (only 5–10°), and the variability within groups is low, as indicated by the small circular variance values in Table 1. Such tiny angular displacements translate to minimal lateral shifts, on the order of a few tens of microns along a roughly 300 µm–radius gray‐matter arc, which are well within the width of a single laminar band. Moreover, both ChAT+–GAD-67+ cerebrospinal-fluid–contacting interneurons and Parvalbumin+–GAD-67+ fast‐spiking interneurons occupy highly discrete niches, and thus a shift of only a few degrees (10–20 µm) would not meaningfully alter their synaptic partners or circuit roles. Therefore, despite the mathematically significant differences, their magnitude is not large enough to be of clear biological consequence. Overall, while the statistical analysis identifies significant differences, careful consideration of effect size and the biological context is required to determine if these differences have meaningful implications for spinal cord organisation.

thumbnail
Download:
Table 1. Circular means and variances from cell angle data for cell types displaying statistically significant differences between genotypes.

https://doi.org/10.1371/journal.pone.0346246.t001

Discussion

The goal of this study was to determine whether impaired dynein function alters the survival or positioning of spinal interneurons in a mouse model of ALS-like pathology. Contrary to our original hypothesis, we observed no significant loss of any major interneuron population, including cholinergic, GABAergic, parvalbumin- and calbindin-positive cells, across genotypes with progressively reduced dynein activity. Subdividing ChAT+ cells into presumed motor neurons versus smaller‐diameter interneurons likewise revealed no genotype‐dependent differences in cell number. Furthermore, the radial distribution of every marker, quantified as the normalised distance from the central canal, was preserved, indicating that gross migratory defects or laminar misplacement do not accompany reduced dynein function in these mice.

Functionally, maintaining the correct laminar and radial distributions of the cells we investigated is critical for proper circuit assembly and synaptic connectivity. Therefore, our data suggest that the motor impairments observed in dynein‐deficient mice are unlikely to stem from gross mislocalisation of neurons within the gray matter. Instead, they reinforce the view that the primary contributions of dynein to motor performance hinge on its roles in intracellular transport and synaptic maintenance rather than on the initial positioning or survival of spinal cord neurons. However, several studies have demonstrated that compromising the motor activity of dynein (Loa mutation), expression (RNAi), cargo–linkage via its accessory proteins (BICD2), or its coding sequence in humans (multiple DYNC1H1 variants) consistently produces radial migration delay, abnormal cortical lamination, or overt malformations of cortical development [18,22,42–46]. The mechanistic consensus is that dynein-driven nuclear movement, both inter‑kinetic in progenitors and somal translocation in neurons, is indispensable for building a correctly layered cerebral cortex, therefore, there is a mismatch between the brain and spinal cord effects of impaired dynein function, based on our current findings.

We also tested whether dynein impairment subtly shifts the angular distribution of cells. While circular‐statistical analyses uncovered statistically significant genotype effects for ChAT+, GAD-67+, and certain double-positive populations, the measured shifts were extremely modest (5–10°) along an arc of ~300 µm, which are distances well within a single laminar band, and are thus unlikely to alter circuit connectivity in any meaningful way.

Functionally, the preservation of both cell numbers and macro‐anatomical positioning argues against interneuron death or gross mislocalisation as drivers of the motor deficits in dynein‐deficient animals. Instead, our data reinforce a model in which the primary contributions of dynein to neuromuscular dysfunction likely arise from disrupted intracellular transport, synaptic maintenance, and protein homeostasis, phenomena we have previously linked to TDP-43 aggregation and p62 upregulation in these same mouse lines [25]. Indeed, interneuron hyperexcitability and synaptic disconnection, rather than cell loss per se, have been implicated in ALS pathogenesis by prior studies of both motor and non‐motor circuits [5,8,27–29].

There are important limitations to note. First, our analysis sampled three biological replicates per genotype with eight sections each; the resulting cell counts, while substantial, may still lack power to detect very small population changes. Therefore, the present findings should be interpreted as indicating that no gross interneuron loss or major spatial disorganisation is present in these models at 52 weeks of age. Detecting smaller effects would require larger cohorts and additional analyses, which were beyond the scope of the present study. Second, we focused exclusively on classical inhibitory markers (GAD-67, parvalbumin, calbindin) and cholinergic cells; other interneuron subtypes (e.g., glycinergic, somatostatin‐expressing) or excitatory interneurons were not assessed. Third, all measurements were made at 52 weeks of age, potentially missing dynamic developmental or early‐disease alterations in migration or survival. Finally, our image‐processing pipeline relied on thresholding and size‐based segmentation, which may misclassify cells near the chosen parameter boundaries.

Future work should therefore extend these structural analyses to functional assays such as patch‐clamp recordings or calcium imaging of identified interneurons, in vivo transport assays to directly measure dynein‐dependent cargo movement, and connectomic studies to assess synapse integrity. Examining earlier developmental time points, incorporating lineage‐tracing of migrating precursors, and applying single‐cell transcriptomics could reveal transient or subtype‐specific vulnerabilities. Finally, conditional ablation of dynein subunits in interneuron populations would help dissect cell‐autonomous from non‐cell‐autonomous contributions to ALS‐like phenotypes.

In summary, impaired dynein function in cholinergic motor pools does not precipitate cell death or gross mislocalisation of key spinal interneurons. Rather, our findings shift the focus toward the essential roles of dynein in maintaining synaptic connectivity and proteostasis, underscoring the importance of functional rather than structural disruptions in ALS pathogenesis.

Supporting information

S1 File. The complete dataset that was statistically analysed.

It includes the transformed angles and normalised distances of each immunopositive cell, relative to the central canal.

https://doi.org/10.1371/journal.pone.0346246.s001

(XLSX)

References

  1. 1. Al-Khayri JM, et al. Amyotrophic lateral sclerosis: insights and new prospects in disease pathophysiology, biomarkers and therapies. Pharmaceutics. 2024;17(10).
  2. 2. Lorenc F, Dupuis L, Cassel R. Impairments of inhibitory neurons in amyotrophic lateral sclerosis and frontotemporal dementia. Neurobiol Dis. 2024;203:106748. pmid:39592063
  3. 3. Goffin L, Lemoine D, Clotman F. Potential contribution of spinal interneurons to the etiopathogenesis of amyotrophic lateral sclerosis. Front Neurosci. 2024;18:1434404. pmid:39091344
  4. 4. Van Den Bosch L, Van Damme P, Bogaert E, Robberecht W. The role of excitotoxicity in the pathogenesis of amyotrophic lateral sclerosis. Biochim Biophys Acta. 2006;1762(11–12):1068–82. pmid:16806844
  5. 5. King AE, et al. Excitotoxicity in ALS: overstimulation, or overreaction?. Exp Neurol. 2016;275(1):162–71.
  6. 6. Odierna GL, Vucic S, Dyer M, Dickson T, Woodhouse A, Blizzard C. How do we get from hyperexcitability to excitotoxicity in amyotrophic lateral sclerosis?. Brain. 2024;147(5):1610–21. pmid:38408864
  7. 7. de Carvalho M, Kiernan MC, Swash M. Fasciculation in amyotrophic lateral sclerosis: origin and pathophysiological relevance. J Neurol Neurosurg Psychiatry. 2017;88(9):773–9. pmid:28490504
  8. 8. Fogarty MJ. Driven to decay: Excitability and synaptic abnormalities in amyotrophic lateral sclerosis. Brain Res Bull. 2018;140:318–33. pmid:29870780
  9. 9. Maekawa S, Al-Sarraj S, Kibble M, Landau S, Parnavelas J, Cotter D, et al. Cortical selective vulnerability in motor neuron disease: a morphometric study. Brain. 2004;127(Pt 6):1237–51. pmid:15130949
  10. 10. Godoy-Corchuelo JM, Ali Z, Brito Armas JM, Martins-Bach AB, García-Toledo I, Fernández-Beltrán LC, et al. TDP-43-M323K causes abnormal brain development and progressive cognitive and motor deficits associated with mislocalised and increased levels of TDP-43. Neurobiol Dis. 2024;193:106437. pmid:38367882
  11. 11. Scotter EL, Chen HJ, Shaw CE. TDP-43 Proteinopathy and ALS: Insights into Disease Mechanisms and Therapeutic Targets. Neurotherapeutics. 2015;12(2):352–63.
  12. 12. Chen H-J, Topp SD, Hui HS, Zacco E, Katarya M, McLoughlin C, et al. RRM adjacent TARDBP mutations disrupt RNA binding and enhance TDP-43 proteinopathy. Brain. 2019;142(12):3753–70. pmid:31605140
  13. 13. Project MinE ALS Sequencing Consortium. Project MinE: study design and pilot analyses of a large-scale whole-genome sequencing study in amyotrophic lateral sclerosis. Eur J Hum Genet. 2018;26(10):1537–46. pmid:29955173
  14. 14. Mentis A-FA, Vlachakis D, Papakonstantinou E, Zaganas I, Patrinos GP, Chrousos GP, et al. A novel variant in DYNC1H1 could contribute to human amyotrophic lateral sclerosis-frontotemporal dementia spectrum. Cold Spring Harb Mol Case Stud. 2022;8(2):a006096. pmid:34535505
  15. 15. Scarlino S, Domi T, Pozzi L, Romano A, Pipitone GB, Falzone YM, et al. Burden of Rare Variants in ALS and Axonal Hereditary Neuropathy Genes Influence Survival in ALS: Insights from a Next Generation Sequencing Study of an Italian ALS Cohort. Int J Mol Sci. 2020;21(9):3346. pmid:32397312
  16. 16. Harms MB, Ori-McKenney KM, Scoto M, Tuck EP, Bell S, Ma D, et al. Mutations in the tail domain of DYNC1H1 cause dominant spinal muscular atrophy. Neurology. 2012;78(22):1714–20. pmid:22459677
  17. 17. Li JT, et al. Expanding the phenotypic and genetic spectrum of neuromuscular diseases caused by DYNC1H1 mutations. Frontiers in Neurology. 2022;13.
  18. 18. Tsai M-H, Cheng H-Y, Nian F-S, Liu C, Chao N-H, Chiang K-L, et al. Impairment in dynein-mediated nuclear translocation by BICD2 C-terminal truncation leads to neuronal migration defect and human brain malformation. Acta Neuropathol Commun. 2020;8(1):106. pmid:32665036
  19. 19. Grabham PW, Seale GE, Bennecib M, Goldberg DJ, Vallee RB. Cytoplasmic dynein and LIS1 are required for microtubule advance during growth cone remodeling and fast axonal outgrowth. J Neurosci. 2007;27(21):5823–34. pmid:17522326
  20. 20. Hafezparast M, Klocke R, Ruhrberg C, Marquardt A, Ahmad-Annuar A, Bowen S, et al. Mutations in dynein link motor neuron degeneration to defects in retrograde transport. Science. 2003;300(5620):808–12. pmid:12730604
  21. 21. Ori-McKenney KM, Xu J, Gross SP, Vallee RB. A cytoplasmic dynein tail mutation impairs motor processivity. Nat Cell Biol. 2010;12(12):1228–34. pmid:21102439
  22. 22. Ori-McKenney KM, Vallee RB. Neuronal migration defects in the Loa dynein mutant mouse. Neural Dev. 2011;6:26. pmid:21612657
  23. 23. Deng W, Garrett C, Dombert B, Soura V, Banks G, Fisher EMC, et al. Neurodegenerative mutation in cytoplasmic dynein alters its organization and dynein-dynactin and dynein-kinesin interactions. J Biol Chem. 2010;285(51):39922–34. pmid:20889981
  24. 24. Garrett CA, Barri M, Kuta A, Soura V, Deng W, Fisher EMC, et al. DYNC1H1 mutation alters transport kinetics and ERK1/2-cFos signalling in a mouse model of distal spinal muscular atrophy. Brain. 2014;137(Pt 7):1883–93. pmid:24755273
  25. 25. Simoes FA, Christoforidou E, Cassel R, Dupuis L, Hafezparast M. Severe dynein dysfunction in cholinergic neurons exacerbates ALS-like phenotypes in a new mouse model. Biochim Biophys Acta Mol Basis Dis. 2025;1871(1):167540. pmid:39428001
  26. 26. Christoforidou E, Simoes FA, Gordon D, Talbot K, Hafezparast M. Aberrant dynein function promotes TDP-43 aggregation and upregulation of p62 in male mice harboring transgenic human TDP-43. Amyotrophic Lateral Sclerosis and Frontotemporal Degeneration. 2023;24(7–8):746–55.
  27. 27. Turner MR, Kiernan MC. Does interneuronal dysfunction contribute to neurodegeneration in amyotrophic lateral sclerosis?. Amyotroph Lateral Scler. 2012;13(3):245–50. pmid:22424125
  28. 28. Do-Ha D, Buskila Y, Ooi L. Impairments in motor neurons, interneurons and astrocytes contribute to hyperexcitability in ALS: underlying mechanisms and paths to therapy. Mol Neurobiol. 2018;55(2):1410–8. pmid:28160214
  29. 29. Crabé R, Aimond F, Gosset P, Scamps F, Raoul C. How degeneration of cells surrounding motoneurons contributes to amyotrophic lateral sclerosis. Cells. 2020;9(12):2550. pmid:33260927
  30. 30. Christoforidou E. Interneuron Microscopy Analysis. Zenodo. 2025.
    • 31. Hughes DI, Boyle KA, Kinnon CM, Bilsland C, Quayle JA, Callister RJ, et al. HCN4 subunit expression in fast-spiking interneurons of the rat spinal cord and hippocampus. Neuroscience. 2013;237:7–18. pmid:23357121
    • 32. Swett JE, Wikholm RP, Blanks RH, Swett AL, Conley LC. Motoneurons of the rat sciatic nerve. Exp Neurol. 1986;93(1):227–52. pmid:3732460
    • 33. Medlock L, Sekiguchi K, Hong S, Dura-Bernal S, Lytton WW, Prescott SA. Multiscale computer model of the spinal dorsal horn reveals changes in network processing associated with chronic pain. J Neurosci. 2022;42(15):3133–49. pmid:35232767
    • 34. Wang Y, Pillai S, Wolpaw JR, Chen XY. H-reflex down-conditioning greatly increases the number of identifiable GABAergic interneurons in rat ventral horn. Neurosci Lett. 2009;452(2):124–9. pmid:19383426
    • 35. Merkulyeva N, et al. Distribution of 28 kDa calbindin-immunopositive neurons in the cat spinal cord. Front Neuroanat. 2015;9:166.
    • 36. Berens P. CircStat: AMATLABToolbox for Circular Statistics. J Stat Soft. 2009;31(10).
    • 37. Krosnowski K, Ashby S, Sathyanesan A, Luo W, Ogura T, Lin W. Diverse populations of intrinsic cholinergic interneurons in the mouse olfactory bulb. Neuroscience. 2012;213:161–78. pmid:22525133
    • 38. Gotts J, Atkinson L, Yanagawa Y, Deuchars J, Deuchars SA. Co-expression of GAD67 and choline acetyltransferase in neurons in the mouse spinal cord: A focus on lamina X. Brain Res. 2016;1646:570–9. pmid:27378584
    • 39. Alvarez FJ, Jonas PC, Sapir T, Hartley R, Berrocal MC, Geiman EJ, et al. Postnatal phenotype and localization of spinal cord V1 derived interneurons. J Comp Neurol. 2005;493(2):177–92. pmid:16255029
    • 40. Laumonnerie C, Tong YG, Alstermark H, Wilson SI. Commissural axonal corridors instruct neuronal migration in the mouse spinal cord. Nat Commun. 2015;6:7028. pmid:25960414
    • 41. Alkaslasi MR, Piccus ZE, Hareendran S, Silberberg H, Chen L, Zhang Y, et al. Single nucleus RNA-sequencing defines unexpected diversity of cholinergic neuron types in the adult mouse spinal cord. Nat Commun. 2021;12(1):2471. pmid:33931636
    • 42. Tsai J-W, Lian W-N, Kemal S, Kriegstein AR, Vallee RB. Kinesin 3 and cytoplasmic dynein mediate interkinetic nuclear migration in neural stem cells. Nat Neurosci. 2010;13(12):1463–71. pmid:21037580
    • 43. Romero DM, Zaidi D, Cifuentes-Diaz C, Maillard C, Grannec G, Selloum M, et al. A human dynein heavy chain mutation impacts cortical progenitor cells causing developmental defects, reduced brain size and altered brain architecture. Neurobiol Dis. 2023;180:106085. pmid:36933672
    • 44. Fiorillo C, Moro F, Yi J, Weil S, Brisca G, Astrea G, et al. Novel dynein DYNC1H1 neck and motor domain mutations link distal spinal muscular atrophy and abnormal cortical development. Hum Mutat. 2014;35(3):298–302. pmid:24307404
    • 45. Möller B, Becker L-L, Saffari A, Afenjar A, Coci EG, Williamson R, et al. The expanding clinical and genetic spectrum of DYNC1H1-related disorders. Brain. 2025;148(2):597–612. pmid:38848546
    • 46. Ge WR, et al. Genotype and phenotype of DYNC1H1-related malformations of cortical development: a case report and literature review. Front Neurol. 2023;14.