Low Concentrations of Dox Reduce Soluble TDP-43 ΔNLS Protein Expression and Pathology in TDP-43 rNLS8 Mice

The TDP-43 rNLS8 mice are normally maintained on 200 mg/kg Dox to suppress expression of the transgene, thus disease is initiated when Dox is removed. First, we examined if expression of hTDP-43 ΔNLS could be lowered by retaining Dox supplementation, but using 10–20-fold lower concentrations (10 mg/kg or 20 mg/kg in DietGel® respectively). DietGel® was added to the standard diet two days a week to address specific nutritional needs and to ensure proper hydration and feeding, particularly in this disease model where motor deficits can impair the ability to eat or drink normally. Hence, four groups of mice were created: Dox off (no Dox, no DietGel®), Diet gel only (no Dox + DietGel® ), 10 mg/kg (10 mg/kg Dox + DietGel®), and 20 mg/kg (20 mg/kg Dox + DietGel®). These groups were all monitored from the timepoint when 200 mg/kg Dox was either removed (Dox off and Diet gel only groups) or replaced with 10 mg/kg Dox + DietGel®(10 mg/kg Dox group), or 20 mg/kg Dox + DietGel® (20 mg/kg Dox group). Behavioural phenotypes (grip strength, neurological score, and rotarod performance) were assessed once per week until the Dox off and Diet gel mice reached late disease stage, then assessed weekly until the remaining groups reached end-stage. Body weight was measured once a week, and mice were monitored until disease end-stage. The brain and spinal cords were collected at disease end-stage to examine hTDP-43 ΔNLS expression and pathology. Five disease stages were categorised: pre-onset, disease onset, early disease, late disease stage, and humane end-stage. We applied a stringent humane endpoint for the study of 20% weight loss threshold to prioritise animal welfare in line with institutional guidelines, in contrast to previous studies that have defined 30% weight loss as the humane end-stage criterion (Walker et al., 2015). As detailed below, disease end-stage was reached at 5 and 6 weeks for Dox off and Diet gel mice, and 18 weeks for both 10 mg/kg and 20 mg/kg Dox groups.

Western blotting analyses of whole brain lysates from these animals, and mice where Dox was retained in the diet (Dox on, 200 mg/kg) for comparison, using an antibody specific for human, but not mouse, TDP-43, revealed a statistically significant reduction in soluble h-TDP-43 ΔNLS levels in mice supplemented with either 10 mg/kg Dox (4.8-fold, **p < 0.01) or 20 mg/kg Dox (2.71-fold, *p < 0.05), compared to the Diet gel alone group (Fig. 1A and B). No h-TDP-43 ΔNLS protein was present in the Dox on lysates, demonstrating that 200 mg/kg Dox effectively suppresses its expression at this concentration. No statistically significant differences were detected between the 10 mg/kg and 20 mg/kg Dox groups. Hence, low concentrations of Dox result in significant reductions in h-TDP-43 ΔNLS protein in the mouse brain. Similar Western blotting analyses of whole spinal cord lysates of these animals also revealed significantly less h-TDP-43ΔNLS in mice supplemented with 10 mg/kg Dox (2.64-fold, **p < 0.01) or 20 mg/kg Dox (2.23-fold, **p < 0.01) compared to the Diet gel alone group (Fig. 1C and D). No statistically significant differences were detected between both the 10 mg/kg and 20 mg/kg Dox groups. Hence, 10 mg/kg and 20 mg/kg Dox are enough to express significantly less h-TDP-43 ΔNLS in both the TDP-43 rNLS8 mouse brain and spinal cord compared to the Diet gel only group.

Fig. 1

Expression of less phosphorylated, insoluble hTDP-43∆NLS in the brain and spinal cord of TDP-43 rNLS8 mice supplemented with 10 mg/kg and 20 mg/kg Dox. A Western blotting using a human-specific TDP-43 antibody (with GAPDH as a loading control) of RIPA-soluble brain lysate fractions prepared from mice analysed in this study: 200 mg/kg Dox on (control), Diet gel (control), 10 mg/kg Dox, and 20 mg/kg Dox. B Quantification of Western blots shown in (A). Densitometry analysis of relative intensity of h-TDP-43 ΔNLS to GAPDH in brain lysates of Diet gel, 10 mg/kg Dox, 20 mg/kg Dox mice, normalised to the Diet gel group. Data are represented as mean ± SEM, mixed effects analysis, one-way Anova, Tukey’s multiple test comparison, *p < 0.05, **p < 0.01, NS = non-significant (p > 0.05), n= 3 (Diet gel group), n = 4 (10 mg/kg and 20 mg/kg Dox groups). C Western blotting using a human-specific TDP-43 antibody (with anti-hTDP-43 and anti-GAPDH as a loading control)  of RIPA-soluble spinal cord lysate fractions prepared from mice analysed in this study: 200 mg/kg Dox on (control), Diet gel (control), 10 mg/kg Dox, and 20 mg/kg Dox. D Densitometry analysis of the relative intensity of h-TDP-43 ΔNLS to GAPDH in spinal cord lysates of Diet gel, 10 mg/kg Dox, 20 mg/kg Dox mice, normalised to the Diet gel group. Data are represented as mean ± SEM, mixed effects analysis, one-way Anova, Tukey’s multiple test comparison, **p < 0.01, n = 3, NS = non-significant (p > 0.05), (Diet gel group), n = 4 (10 mg/kg and 20 mg/kg Dox groups). E, F Representative confocal microscopy maximum projection images following immunohistochemistry using a pan-TDP-43 (detecting both human and mouse TDP-43) antibody (green) and DAPI staining (blue) in the brain (E) and spinal cord (F) of 200mg/kg Dox on (control), Diet gel alone supplemented mice, 10 mg/kg and 20 mg/kg Dox groups. Scale bar = 20 μM. Cytoplasmic TDP-43 expression is indicated by the white arrows, and inset TDP-43 inclusions are indicated by the yellow arrows (G) Western blotting using an anti-phospho-TDP-43 (Ser409/410) antibody, and total protein staining in Urea soluble and insoluble brain fractions prepared from Diet gel, 10 mg/kg and 20 mg/kg Dox mice. H Quantification of Western blots shown in (G). Densitometry analysis of the relative intensity of phospho-TDP-43 (Ser409/410) relative to total protein staining in brain lysates in (G). Data are represented as mean ± SEM, mixed effects analysis, one-way Anova, Tukey’s multiple test comparison,***p < 0.001,****p < 0.0001, NS = non-significant (p > 0.05), n = 3 (Diet gel), n = 4 (10 mg/kg and 20 mg/kg Dox groups). I Western blotting using an anti-phospho-TDP-43 (Ser409/410) antibody, and total protein staining in Urea soluble and insoluble spinal cord fractions prepared from Diet gel, 10 mg/kg Dox and 20 mg/kg Dox mice. J Quantification of Western blots shown in (I). Densitometry analysis of the relative intensity of phospho-TDP-43 to total protein staining in spinal cord lysates of mice in (I), normalised to the Diet gel group. Data are represented as mean ± SEM, mixed effects analysis, one-way Anova, Tukey’s multiple test comparison, *p < 0.05, **p < 0.01, NS = non-significant (p > 0.05), n = 3 (Diet gel group), n = 4 (10 mg/kg and 20 mg/kg Dox groups). Graphs shown in (B), (D), (H), and (J) represent the average of at least two technical replicates of each lysate

We next examined the presence of TDP-43 pathology in brain and spinal cord neurons of the mice at disease end-stage using immunohistochemistry and confocal imaging. Cytoplasmic h-TDP-43 ΔNLS and cytoplasmic inclusions were detected in both spinal cord and brain tissues in the Diet gel, 10 mg/kg and 20 mg/kg Dox groups (Fig. 1E and F). The insoluble cell pellet fractions (solubilised by urea) were also probed by Western blotting using a phospho-TDP-43 (Ser409/410) specific antibody, representing insoluble, aggregated phosphorylated TDP-43. Significantly less insoluble, phosphorylated h-TDP-43 ΔNLS was present in brains of mice supplemented with either 10 mg/kg or 20 mg/kg Dox (4.65-fold, ***p < 0.001 and 18.8-fold,****p < 0.0001 less) compared to the Diet gel only group (Fig. 1G and H). Similarly, significantly less insoluble phosphorylated h-TDP-43ΔNLS was present in spinal cords of mice supplemented with 10 mg/kg or 20 mg/kg Dox (1.78-fold and 2.69 -fold less respectively, **p<0.01 and *p < 0.05) compared to the Diet gel only group (Fig. 1I and J). No statistically significant differences were detected between the 10 mg/kg and 20 mg/kg Dox groups. Together, these results reveal  that both 10 mg/kg and 20 mg/kg Dox supplementation induce expression of TDP-43 with pathological features; cytoplasmic, aggregated, insoluble phosphorylated h-TDP-43 in the mouse brain and spinal cord.

Low Concentrations of Dox Delay Disease Onset in TDP-43 rNLS8 Mice

To determine whether 10 mg/kg and 20 mg/kg Dox supplementation delay disease onset in TDP-43 rNLS8 mice, analysis of neurological scores was performed weekly. A previously used 4-point scale was used by an observer blinded to the treatment groups (Wright et al., 2021): Score 0: normal, with no observable impairment. Score 1: mild impairment, characterised by abnormal hindlimb splay, slightly slower gait, and a normal righting reflex. Score 2: moderate impairment, indicated by a partially or fully collapsed hindlimb, with at least one foot dragging along the cage floor, and delayed righting reflex. Score 3: severe impairment, with rigid paralysis or minimal movement in the hindlimbs, which can no longer be used for forward motion, and a markedly slow righting reflex (Wright et al., 2021). Disease onset was indicated by the date each animal first reached a neurological score of 1.

Disease onset was significantly delayed in TDP-43 rNLS8 mice supplemented with 10 mg/kg and 20 mg/kg Dox (median onset = 6 and 7 weeks after removing 200 mg/kg Dox, respectively) compared to both Dox off mice and the Diet gel only group (median onset = 2 and 3 weeks, respectively, all ***p < 0.001) (Fig. 2A). Similarly, the average date of disease onset was significantly delayed in TDP-43 rNLS8 mice supplemented with 10 mg/kg and 20 mg/kg Dox (mean = 6.25, 6.75 weeks after replacing 200 mg/kg Dox, respectively) compared to both Dox off and Diet gel only groups (mean = 2 and 3 weeks, respectively, *p < 0.05 **p < 0.01) (Fig. 2B). A statistically significant difference was detected between the Dox off and Diet gel only supplemented groups (****p < 0.0001), suggesting that the extra nutrition provided by DietGel® delays disease onset by one week. No statistically significant differences in disease onset were obtained between the 10 mg/kg and 20 mg/kg Dox groups (Fig. 2B), despite an apparent visual distinction between the two groups (Fig. 2A). Hence, these data indicate that supplementation with either 10 mg/kg or 20 mg/kg Dox delays disease onset in TDP-43 rNLS8 mice.

Fig. 2

Disease onset is delayed in TDP-43 rNLS8 mice supplemented with 10 mg/kg and 20 mg/kg Dox compared to Diet gel and Dox off groups. A Kaplan–Meier curves representing age of disease onset in weeks, defined as an individual mouse achieving a neurological score = 1 (NS = 1). Log-rank (Mantel-Cox) test, ***p < 0.001. Median disease onset; Dox off group = 2 weeks, Diet gel group = 3 weeks, 10 mg/kg Dox group = 6 weeks and 20 mg/kg Dox group = 7 weeks. B Disease onset is delayed in 10 mg/kg and 20 mg/kg Dox mice, mean age of onset = 6.25 and 6.75 weeks, respectively, compared to Dox off and Diet gel mice (mean = 2 and 3 weeks, respectively). Mixed effect analysis mean ± SD, one-way Anova, Tukey’s multiple test comparison, ****p < 0.0001, **p < 0.01, *p < 0.05, n = 3 (Dox off and Diet gel groups), n = 4 (10 mg/kg and 20 mg/kg Dox groups). C Neurological scores are significantly higher in both Dox off and Diet gel mice at 2, 3, 4, and 5 weeks compared to 10 mg/kg and 20 mg/kg Dox groups. Mixed effect analysis, two-way Anova, Tukey’s multiple test comparison, mean ± SD, ****p < 0.0001, n = 3 (Dox off and Diet gel group), n = 4 (10 mg/kg and 20 mg/kg Dox group). D Neurological score analysis throughout life until disease end-stage reveals 10 mg/kg and 20 mg/kg Dox mice have delayed disease onset compared to Dox off and Diet gel groups, mean ± SD. E Body weights of 10 mg/kg and 20 mg/kg Dox mice are significantly higher than Dox off and Diet gel groups at 2, 3, 4, and 5 weeks, mean ± SD, mixed effect analysis, two-way Anova, Tukey’s multiple test comparison, ***p < 0.001, **p < 0.01, *p < 0.05, n = 3 (Dox off and Diet gel group), n = 4. Black (*) and red (*) asterisks refer to comparisons to Dox off or Diet gel mice, respectively. F Body weight analyses throughout life until disease end-stage reveal 10 mg/kg and 20 mg/kg Dox mice have higher body weight compared to Dox off and Diet gel groups, mean ± SD. G Kaplan–Meier curves representing age of disease onset in weeks, defined as an individual mouse achieving a weight loss of > 10% from peak body weight. Log-rank (Mantel-Cox) test, **p < 0.01. Median disease onset; Dox off group = 2 weeks, Diet gel group = 3 weeks, 10 mg/kg Dox group and 20 mg/kg Dox group = 7 weeks. H Body weight analyses reveal 10 mg/kg and 20 mg/kg Dox mice (mean = 7 and 6.75 weeks, respectively) have delayed disease onset compared to Dox off and Diet gel groups (mean = 2 and 3 weeks, respectively). Mixed effect analysis of age of disease onset, indicated by age each mouse reaches weight loss of > 10% from peak body weight, mean ± SD, Tukey’s multiple test comparison, *p < 0.05, n = 3 (Dox off and Diet gel group), n = 4 (10 mg/kg and 20 mg/kg Dox group)

We also assessed neurological scores throughout the study until end stage to monitor disease progression. Statistically significant improvements in neurological scores were detected in mice supplemented with 10 mg/kg and 20 mg/kg Dox compared to controls throughout the disease course. Neurological scores in the Dox off group worsened progressively and reached a score of 3 by week 5. In contrast, the Diet gel only group displayed scores of 0 to 2 over the same timeframe, whilst the Dox-supplemented groups (both 10 mg/kg and 20 mg/kg) displayed neurological scores of 0 over the same period (****p < 0.0001, Fig. 2C). Significant improvements in neurological scores in mice supplemented with either 10 mg/kg or 20 mg/kg Dox compared to Dox off and Diet gel only groups throughout the disease course were evident (Fig. 2D).

To confirm these findings, body weights were also analysed weekly to identify disease onset, as previous (Wright et al., 2021). This was determined to be the mean number of weeks after replacing 200 mg/kg Dox until the loss of 10% weight was reached from peak body weight (Wright et al., 2021). A statistically significant difference in body weights was detected from week 2 onwards in TDP-43 rNLS8 mice supplemented with either 10 mg/kg or 20 mg/kg Dox compared to both Dox off and Diet gel supplemented groups (***p < 0.001,**p < 0.01, *p < 0.05 weeks) (Fig. 2E). The 10 mg/kg and 20 mg/kg Dox mice initially gained weight but then began to lose weight at approximately 6 weeks (Fig. 2F).

Disease onset was significantly delayed in TDP-43 rNLS8 mice supplemented with 10 mg/kg and 20 mg/kg Dox (median = 7 weeks for both groups) compared to both Dox off and Diet gel only groups (median = 2 and 3 weeks, respectively, all **p < 0.01) (Fig. 2G). Similarly, the average date of disease onset was significantly delayed in TDP-43 rNLS8 mice supplemented with 10 mg/kg or 20 mg/kg Dox (mean = 7 and 6.75 weeks, respectively) compared to both Dox off and Diet gel only groups (mean = 2 and 3 weeks, respectively, both *p < 0.05 (Fig. 2H). No statistically significant differences were detected between the Dox off and Diet gel groups, nor between the 10 mg/kg and 20 mg/kg Dox groups, the latter implying that 10 mg/kg is sufficient to delay disease onset in TDP-43 rNLS8 mice. These findings suggest that 10 mg/kg and 20 mg/kg Dox supplementation delays disease onset in TDP-43 rNLS8 mice by at least 4 weeks.

Low Concentration of Dox Delays Disease Progression in TDP-43 rNLS8 Mice

To determine whether 10 mg/kg and 20 mg/kg Dox supplementation delays onset of motor dysfunction and disease progression in TDP-43 rNLS8 mice, we examined behavioural phenotypes previously described in this model throughout the disease course. First, the accelerating rotarod test was used to monitor motor coordination, balance, and endurance. Disease progression, from the early disease stage to the late disease stage, by measuring the latency to fall from the rotarod with a threshold of over 50 s was also assessed.

Consistent with previous observations, a dramatic decline in motor performance was evident in the Dox off and Diet gel alone groups, at 3, 3.5, and 4 weeks after Dox removal, respectively. Rotarod performance was significantly improved in mice supplemented with 10 mg/kg and 20 mg/kg Dox from week 3 onwards (weeks 3, 3.5, and 4 weeks (***p < 0.001, **p < 0.01, *p < 0.05)), until the Dox off group and Diet gel alone groups reached late stage (Fig. 3A and B). There were no significant differences between 10 mg/kg and 20 mg/kg for the full duration of the experiment, demonstrating there was no effect on increasing the concentration of Dox from 10 to 20 mg/kg on rotarod performance (Fig. 3A and B). Disease progression from onset to early and late disease stage was assessed by the latency to fall from the rotarod below 50 s, which was significantly delayed in TDP-43 rNLS8 mice supplemented with 10 mg/kg and 20 mg/kg Dox (median = 9 weeks after replacement of 200 mg/kg Dox, respectively) compared to both Dox off mice and Diet gel only group (median 3 and 3.75 weeks, respectively, all ***p < 0.001) (Fig. 3C).

Fig. 3

Motor performance, grip strength, and survival are improved in TDP-43 rNLS8 mice supplemented with 10 mg/kg and 20 mg/kg Dox compared to Diet gel and Dox off groups. A Rotarod performance (latency to fall in seconds) is significantly improved at 3, 3.5, and 4 weeks in 10 mg/kg and 20 mg/kg Dox mice compared to Dox off and Diet gel  groups, indicating improvement in motor function. Mixed effects analysis, mean ± SD, two-way Anova, Tukey’s multiple test comparison, ***p < 0.001, **p < 0.01, *p < 0.05, n = 3 (Dox off and Diet gel groups), n = 4. (10 mg/kg and 20 mg/kg Dox groups). Black (*) and red (*) asterisks refer to comparisons to Dox off or Diet gel mice respectively. B Rotarod performance (latency to fall in seconds) throughout life until disease end-stage reveals the improved rotarod performance of 10 mg/kg and 20 mg/kg Dox mice compared to Dox off and Diet gel mice, mean ± SD. C Kaplan–Meier curves representing rotarod latency to fall, defined as an individual mouse < 50s. Severe motor deficits were detected by poor performance on the rotarod test, with latencies to fall dropping below 50 s. Log-rank (Mantel-Cox) test, ***p < 0.001. 10 mg/kg and 20 mg/kg Dox mice display improved performance compared to Dox off and Diet gel groups, median = 9 weeks for both, in contrast to 3 and 3.5 weeks for Dox off and Diet gel groups, respectively. D Muscle function, indicated by grip strength, is significantly improved at 3, 3.5, and 4 weeks in 10 mg/kg and 20 mg/kg Dox mice compared to Dox off and Diet gel mice. Mixed effects analysis, mean ± SD, two-way Anova, Tukey’s multiple test comparison, ****p < 0.0001, ***p < 0.001, *p < 0.05, n = 3 (Dox off and Diet gel group), n = 4 (10 mg/kg and 20 mg/kg Dox group). Black (*) and red (*) asterisks refer to comparisons to Dox off or Diet gel mice respectively. E Muscle function, indicated by grip strength, throughout life until disease end-stage reveals improved grip strength of mice supplemented with 10 mg/kg and 20 mg/kg Dox compared to both Dox off and Diet gel groups, mean ± SD. F Kaplan–Meier curves representing grip strength, defined as an individual mouse achieving < 100(g) force. Log-rank (Mantel-Cox) test, ***p < 0.001. 10 mg/kg and 20 mg/kg Dox display improved muscle function compared to Dox off and Diet gel groups, median = 13 weeks for both compared to 3 for Dox off and Diet gel groups. Severe motor deficits were  detected by poor muscle function, with grip strength dropping below 100 g force. G Kaplan–Meier curves representing probability of survival, defined as an individual mouse achieving the humane endpoint criterion. Log-rank (Mantel-Cox) test, ***p < 0.001. 10 mg/kg and 20 mg/kg Dox mice reach a neurological score of 3 at a median of 18 weeks compared to 5 or 6 weeks for the Dox off or Diet gel groups, respectively. H Table summarising disease progression in all groups, as assessed by grip strength, rotarod performance, and survival

Next, grip strength of the mice was analyzed to monitor changes in muscle strength, both in the forelimbs and hindlimbs. Consistent with previous observations (Walker et al., 2015; Wright et al., 2021), dramatic muscle weakening was detected from 3 weeks onwards in the Dox off group, and in the Diet gel alone group. Grip strength was significantly greater in mice supplemented with 10 mg/kg and 20 mg/kg Dox from week 3 onwards (weeks 3, 3.5 and 4 weeks, ****p < 0.0001, ***p < 0.001, *p < 0.05) (Fig. 3D and E) until the disease late stage was reached by the Dox off and Diet gel groups. There were no differences between 10 mg/kg and 20 mg/kg for the full duration of the experiment, demonstrating no effect of increasing the concentration of Dox from 10 to 20 mg/kg (Fig. 3D and E). Disease progression from onset to early and late disease stage, measured by muscle strength of an individual mouse reaching a force below 100 g, was significantly delayed in TDP-43 rNLS8 mice supplemented with 10 mg/kg or 20 mg/kg Dox (median = 13 weeks, respectively), compared to both Dox off mice and Diet gel only groups (median = 3 and 4 weeks, respectively, all ***p < 0.001) (Fig. 3F). Hence, low Dox supplementation delays disease progression and the onset of motor deficits.

Low Concentrations of Dox Extend Survival in TDP-43 rNLS8 Mice

Next, we examined whether low Dox supplementation extends lifespan in the TDP-43 rNLS8 mouse model. The humane endpoint was identified by the age at which the mice reached a neurological score =3 (complete paralysis of both limbs, slow righting reflex and > 20% body weight loss). Kaplan–Meier analysis revealed a statistically significant increase in survival in TDP-43 rNLS8 mice supplemented with 10 mg/kg or 20 mg/kg Dox (median = 18 weeks each) compared to both Dox off mice and Diet gel only groups (median = 5 and 6 weeks, respectively) (Fig. 3G and H; χ2 = 18.46, df = 3 ***p < 0.001). Previously, survival in the TDP-43 rNLS8 mouse was reported to be 10 weeks following Dox removal using 30% weight loss as the end-stage criterion (Walker et al., 2015). However, our findings using the more stringent 20% weight loss threshold-showing that the Dox off mice survive only to 5–6 weeks following Dox removal-are consistent with other studies within our laboratory (data not shown). Hence overall, these results indicate that low concentrations of Dox effectively slow the impairment of motor deficits and grip strength, and extend survival in the TDP-43 rNLS8 mouse model.

Loss of Motor Neuron and Neuroinflammation are Present in Low Dox TDP-43 rNLS8 Mice at Disease end-Stage

We then examined tissues from disease end-stage animals to determine if the same pathological characteristics are present in mice treated with 10 mg/kg or 20 mg/kg Dox . Neuroinflammation and motor neuron degeneration are known features present in the TDP-43 rNLS8 model (Spiller et al., 2016; Walker et al., 2015). Motor neuron loss begins two weeks after Dox removal, with significant loss by the sixth week, particularly in the spinal cord, when approximately 30% of motor neurons are lost. We first examined motor neuron loss in mice supplemented with 20 mg/kg Dox at disease end-stage. The total number of all motor neurons (both α and γ), and α-motor neurons only, remaining in  the ventral horn regions of the spinal cord following Nissl staining was quantified as previously described (Huang et al., 2012; Mitchell et al., 2015). Motor neurons were identified by the presence of cells with large soma and intense blue cytoplasmic Nissl staining, whereas α-motor neurons were specifically detected by the presence of Nissl-stained cells with diameters of > 20 μm (C. Huang et al., 2012; Mitchell et al., 2015). However, no significant differences in the numbers of either total or α-motor neurons were detected between the Diet gel and the 20 mg/kg Dox groups at end-stage (Fig. 4A–C). Neuroinflammation was then assessed by GFAP immunohistochemistry of  brain tissues of TDP-43 rNLS8 mice. However, no significant differences in the total number of GFAP-positive cells were detected between the Diet gel, 10 mg/kg or 20 mg/kg Dox groups (Fig. 4D–E) at disease end-stage. Hence, despite the slowed disease course and extension of lifespan by low Dox supplementation, two pathological features—motor neuron loss and neuroinflammation—are retained in the TDP-43 rNLS8 mice once animals have reached disease end-stage.

Fig. 4

Loss of motor neurons and neuroinflammation are present in low Dox TDP-43 rNLS8 mice at disease end-stage. A Bright-field microscopy following Nissl staining using Toluidine blue of both ventral horn regions of the lumbar spinal cord (section thickness: 10 µm) from Diet gel and 20 mg/kg Dox mice, scale bar = 50 μm. B Quantification of images in (A) reveals there were no differences in the total number of α-motor neurons, identified by Nissl-stained cells with a diameter > 20 μM, present in Diet gel and 20 mg/kg Dox mice at disease end-stage (six weeks and 18 weeks, respectively)   two-tailed, unpaired t test, NS = non-significant (p > 0.05). C Quantification of images in (A) reveals there were no differences in total number of motor neurons (both α and γ), identified by the presence of all cells with large soma and intense Nissl-staining, present in Diet gel and 20 mg/kg Dox mice at disease end-stage (six weeks and 18 weeks respectively), two-tailed, unpaired t test, NS = non-significant (p > 0.05). D Representative confocal microscopy maximum projection images following immunohistochemistry of lumbar spinal cord sections (10 µm thickness) using anti-GFAP antibodies (green), co-stained with nuclear marker DAPI (blue) of Diet gel, 10 mg/kg Dox, and 20 mg/kg Dox mice, scale bar = 20 μm. E Quantification of images in (D) reveals there were no differences in the total number of GFAP-positive cells/mm2 between groups. Mixed effects analysis mean ± SD, one-way Anova, Tukey’s multiple test comparison, NS = non-significant (p > 0.05), n= 3 (Diet gel group), n = 4 (10 mg/kg and 20 mg/kg Dox group)

As no significant differences were detected between the 10 mg/kg and 20 mg/kg groups for any of the parameters assessed, the data from these two treatment groups were combined to increase the statistical power of the analysis. Again, a statistically significant delay in disease onset was present compared to Dox off and Diet gel only groups, detected by either neurological score (Supplementary Fig. 2A, ****p < 0.0001) or body weight loss (Supplementary Fig. 2B, ***p < 0.001). Similarly, a statistically significant delay in motor performance, grip strength or lifespan was present compared to Dox off and Diet gel only groups, (Supplementary Fig. 3A, *p < 0.05, ***p < 0.001, Supplementary Fig. 3B,,****p < 0.0001, Supplementary Fig. 3C, ****p < 0.0001, *p < 0.05, Supplementary Fig. 3D ***p < 0.001, Supplementary Fig. 3E, ****p < 0.0001). We also performed post-hoc power calculations of the separate 10 mg/kg and 20 mg/kg groups using the observed effect sizes and variability, which revealed the study was sufficiently powered (power > 95%) to detect statistically significant differences at α<0.05 when the 10 mg/kg and 20 mg/kg groups were analysed separately. This suggests that the magnitude of the treatment effect was large enough to be reliably detected with the initial sample size.