Introduction

Huntington’s disease (HD) is an autosomal dominant neurodegenerative disorder resulting from an expanded CAG tandem repeat in the gene encoding huntingtin (HTT).1 The mutant form of HTT includes an expanded polyglutamine tract and elicits progressive psychiatric, cognitive and motor symptoms. A primary neuropathology associated with HD is the selective dysfunction, and eventual death, of medium spiny neurons within the striatum,2, 3, 4 although the cerebral cortex and other areas are also affected. The exact cause of such cellular dysfunction is still unclear, however, it is known that mitochondrial dysfunction, oxidative stress and excitotoxicity have a major role.5, 6, 7, 8 The amino acid, cysteine, has the potential to alleviate each of these features of HD pathology. The compound N-acetylcysteine (NAC) is able to cross the blood–brain barrier and directly donate cysteine to the central nervous system, acting as a potential therapy.9

Oxidative stress has been shown to occur in the brain of patients with HD,5, 6 while an increase in peripheral reactive oxygen species and markers of oxidative stress occur in those in the prodromal phase of the disorder.10 Furthermore, transgenic animal models expressing the amino (N)-terminal fragment of mutant form of Htt show oxidative stress occurring in the striatum before the onset of motor symptoms.11 Cysteine is the rate-limiting amino acid required for the synthesis of glutathione (GSH), the primary endogenous antioxidant in the brain.12 Administration of NAC increases levels of GSH, reducing oxidative stress.13 It has recently been shown that cysteine supplementation is able to alleviate some of the pathology observed in animal models of HD.14

NAC also has the capacity to modulate glutamate-induced toxicity. The elevated synaptic glutamate observed in excitotoxic states leads to dysregulated glutamate receptor activity. Evidence supporting an excitotoxic mechanism in HD includes the development of striatal medium spiny neuron degeneration following the administration of the N-methyl-D-aspartate antagonist, quinolinic acid.15 Furthermore, the R6/2 transgenic mouse model of HD was found to show increased N-methyl-D-aspartate activation before the onset of behavioral phenotypes,16 while the administration of N-methyl-D-aspartate antagonists were found to delay the progression of symptoms in HD patients.17, 18 However, the question remains as to why there is an increased glutamatergic signaling in HD. Astrocytic glutamate uptake may be reduced, leading to an increase in synaptic glutamate. Supporting this, RNA and protein levels of the astrocytic glutamate antiporter, GLT-1 are reduced in postmortem brains of HD patients and transgenic mice.19, 20 Furthermore, treatment with the GLT-1 agonist, ceftriaxone, was found to increase GLT-1 expression and decrease motor abnormalities in the R6/2 mouse model.21 NAC can alleviate excess synaptic glutamate by promoting astrocytic uptake.

Astrocytic glutamate transport can occur via several transporters, two of which are dependent on cysteine: Excitatory amino-acid transporter-1 (also known as GLT-1), which is directly dependent on cysteine and System xc−, which is dependent on cystine derived from oxidized cysteine.22, 23 Supplementation with NAC can both decrease synaptic glutamate and increase intracellular cysteine via these transporters.22, 24, 25 These actions of NAC are therapeutic in models of addiction.26, 27 Given the ability of NAC to increase GSH production and potentially decrease synaptic glutamate via GLT-1, there is the possibility of treating two of the primary neuropathological features of HD: excitotoxicity and oxidative stress. Therefore, the aim of the current study is to test the therapeutic potential of NAC in HD using the R6/1 transgenic mouse model.

Materials and methods

Mice

R6/1 transgenic hemizygote males28 were originally obtained from the Jackson Laboratory (Bar Harbor, ME, USA) and bred with CBB6 (CBA~C57/B6) F1 females to establish the R6/1 colony at the Florey Institute of Neuroscience and Mental Health. After weaning, animals were group housed (four mice per cage with two of each genotype) and maintained on a 12 h light/dark cycle with access to food and water ad libitum. To avoid any possible ‘litter effects’, we ensured that animals from each litter were represented in each experimental condition using appropriate randomization. Experimental animals were kept in open-top cages under standard housing conditions. All the experiments were performed on wild-type (WT) and R6/1 (HD) mice in accordance with the guidelines of the Florey Institute of Neuroscience and Mental Health Animal Ethics Committee and the National Health and Medical Research Council.

N-Acetylcysteine treatment

N-acetylcysteine (NAC) was obtained from Astral Scientific (#AM0108). NAC was prepared in clinical-grade saline (0.9%) at concentrations suitable for dosing at 500 mg kg−1 per day. Solutions were pH-matched to that of the vehicle (saline). NAC or vehicle were injected intraperitoneally (1 ml per 100 g body weight) daily, beginning at 8 weeks of age and continuing until the mice were culled at 17 weeks of age. NAC and saline treatments were allocated semi-randomly, stratified to ensure that there were no baseline differences in body weight or rotarod performance between treatment groups.

Behavioral analysis

Accelerating rotarod

A rotarod (Ugo Basile model 7650; Sandown Scientific, Hampton, UK) was used as a measure of motor performance. A detailed protocol has been described previously.29 Experimenters were masked to both genotype and treatment. One day before the initial testing phase, mice were acclimatized to the rod; they were allowed to walk at the baseline speed for 2 min before allowing the rod to accelerate to 20 r.p.m. over 150 s. If the mice fell off during this time, they were placed back on the rod until the acclimatization period was complete.

Paw clasping

Mice were held by the tip of the tail against the edge of a table to prevent swinging motion. The mice were observed for a period of 30 s, with the highest clasping score occurring during this time being recorded. Mice were scored on a five-point scale (from 0 to 4) as described previously.30, 31 A score of 0 indicated that no clasping was present, whereas scores from 1 to 4 represented the number of paws clasping. Scorers were masked to both genotype and treatment.

Digigait

Assessment of gait was performed using the Digigait automated treadmill (Mouse Specifics, Boston, MA, USA). On the initial trial, mice were placed inside a plexiglass container on top of a transparent treadmill and allowed to habituate for 1 min. For all time points, the speed of the treadmill was increased incrementally from 0 to 15 cm s−1 over a period of 15 s. Video of the mouse from underneath was then recorded until the mouse ran without stopping for a period of 3 s. The contact of each paw on the treadmill was automatically detected via the Digigait system. Scorers masked to treatment and genotype manually corrected any errors made by the Digigait Software. The major parameters of gait to be analyzed relate to either the ‘duration’ or ‘consistency’ of various stride characteristics. The duration parameters include: total stride, stance, swing, propulsion and brake time. The consistency characteristics include: gait symmetry, paw overlap, paw positioning, stance width and the ataxia coefficient.

Forced swim test

The forced swim test was conducted at 12 weeks of age, as described previously.32 Total immobility time was scored manually by an experienced experimenter masked to both genotype and treatment.

Western blotting

Immediately after cervical dislocation, mice brains were dissected on ice and snap frozen in liquid nitrogen, before being stored at −80 °C. Tissue homogenates were prepared in Tris-lysis buffer containing EDTA, EGTA, glycerol, triton-X100, sodium fluoride, sodium pyrophosphate, Na3VO4, DTT and a protease inhibitor cocktail (Sigma, Castle Hill, NSW, Australia; #p8340). Ribolyser was used at mid-strength for ~10 s with one oscillation per second. Samples were allowed to rest for 10 min on ice before centrifugation at 180 000 g for 25 min at 4 °C. Supernatants were removed and protein concentrations were determined using BCA total protein assay (Thermo, Scoresby, VIC, Australia; #3225). Acrylamide (12%) bis/tris gels were precast and allowed to set at 4 °C overnight. Sample buffer was added to the samples before being heated at 95 °C for 3 min. Samples were then run in tris-glycine buffer using the Bio-Rad (Gladesville, NSW, Australia) MiniPROTEAN Tetra cell system. Gels were transferred to PVDF membrane before being blocked for 1 h in 5% BSA. Blots were washed three times in TBST before incubation in primary antibodies overnight at 4 °C. Primary antibodies were diluted in TBST containing 5% BSA: anti-GLT-1 (1:1000; Millipore, Bayswater, VIC, Australia; #ABN102); anti-xCT (1:500; Abcam, Waterloo, NSW, Australia; #AB37185); and anti-β-actin (1:1000; Cell Signaling Technology, Beverly, MA, USA; #8H10D10). The following day, blots were washed three times in TBST, before being incubated in IR secondary antibodies (1:5000; Li-Cor; Lincoln, NE, USA; #926-3211 and #926-68020) for 1 h at room temperature. Blots were once again washed three times before being imaged in the Li-Cor Odyssey IR detection system. Densitometry was carried out using the integrated intensity value for each band. Analyses of the results were carried out as ratio of protein-of-interest:β-actin.

HPLC

Immediately after cervical dislocation, mice brains were dissected on ice and snap frozen in liquid nitrogen, before being stored at −80 °C. Tissue homogenates were prepared in 0.1% formic acid using a motorized latex pestle for 10 s with one oscillation per second. Samples were then centrifuged at 8000 g for 15 min. The supernatants were then collected and analyzed for levels of reduced and oxidized glutathione (GSH and GSSG, respectively) using the method previously described.33

Protein carbonyl content

Protein carbonyl content was assayed using the Oxyblot kit (Millipore, #S7150). Mitochondrial and nuclear fractions were separated as described previously.34 Nuclear fractions were then prepared in lysis buffer, similarly to whole tissue (see above), while mitochondria were prepared in the mitochondrial isolation buffer. Total protein content was assayed before the preparation of samples. 2,4-Dinitrophenylhydrazine was added to samples to derivatize carbonyl groups from the protein side chains. Derivatized samples were then separated using electrophoresis, as described above. Western blot analysis was carried out, as described above, using the 2,4-dinitrophenylhydrazine antibody provided (1:150). Densitometry was carried out using the integrated intensity value for each band. Analyses of results were carried out as ratio of protein-of-interest:β-actin. Given the presence of multiple bands, the average value of all bands within each lane was used to give an overall measure of protein carbonyl content.

Mitochondrial stress test

All bioenergetic and mitochondrial function analyses were performed using the Seahorse XF24 Extracellular Flux Analyzer (Seahorse Bioscience, Billerica, MA, USA). Immediately after cervical dislocation, the brain was removed and dissected on ice. Samples were then finely chopped in buffer containing mannitol (200 mM), sucrose (50 mM), KH2PO4 (5 mM), EGTA (1 mM), MOPS (5 mM) and BSA (0.1%) to aid in preserving mitochondria. The tissue was then resuspended in the same buffer containing dimethyl sulfoxide (20%), before being slowly frozen on dry ice and stored at −80 °C. Tissue was allowed to defrost on ice before mitochondria were isolated as described previously.34 Mitochondria were then seeded on to 24-well Seahorse V7 plates. Mitochondrial function was determined as previously described35 and basal respiration (state II), adenine diphosphate (ADP)-supplemented respiration with saturated substrate succinate (state III), ADP-deprived respiration (state IV) and respiratory control rate (state III:state IV ratio) were determined from these analyses. Each sample was measured in triplicate, with average values taken from successful wells. Successful wells were those that acted functionally by displaying normal responses to oligomycin (decreased oxygen consumption rate), FCCP (increased oxygen consumption rate) and antimycin A (complete shutdown of oxygen consumption rate). This frozen mitochondria protocol was validated by directly comparing freeze–thaw samples to fresh mitochondrial samples on the same plate. No differences in mitochondrial respiration were found (data not shown).

Statistics

Statistical analyses were performed using IBM SPSS statistics Version 21.0 (IBM, Armonk, NY, USA) and GraphPad Prism 6 (GraphPad Software, La Jolla, CA, USA). Rotarod, clasping, Digigait and body weight data were analyzed using three-way repeated measures analysis of variance. If sphericity was violated, Greenhouse–Geisser corrections were used. If significant interactions were found, alpha-levels for post hoc comparisons were corrected using Bonferroni’s method. All other data were analyzed using two-way analyses of variance. Alpha levels for post hoc comparisons were corrected using Tukey’s honestly significant difference. In all the cases, significance was set at P<0.05.

Results

NAC delays the onset and slows the progression of motor deficits

Paw clasping is a primary motor endophenotype that is observed in the R6 transgenic models of HD.28 A significant interaction of age, treatment and genotype was found on the development of the clasping phenotype (F(9,35)=4.268, P=0.001; Figure 1a). Post hoc analyses revealed that NAC significantly delayed the onset of the clasping phenotype in HD mice. Specifically, saline-treated HD mice differed from control mice from 10 weeks (P=0.017), whereas NAC-treated HD mice did not differ from controls until 14 weeks (P=0.030). Furthermore, NAC-treated HD mice performed significantly better than saline-treated HD mice from 11 weeks (P=0.003) onward, showing reduced clasping scores until the final data point of 17 weeks (P<0.001).

Figure 1
figure 1

Treatment with NAC delays the onset and progression of motor deficits in HD mice and has antidepressant-like effects in both HD and WT mice. (a) Analysis of the clasping phenotype showed a significant age × treatment × genotype interaction. Post hoc analyses revealed that NAC significantly delayed the onset of the clasping phenotype in HD mice, with saline-HD mice showing clasping behavior from 10 weeks onward, while NAC-HD mice did not show clasping behavior until 14 weeks. NAC-treated HD mice also showed significantly lower levels of clasping behavior relative to saline-treated HD mice from 11 weeks onward. (b) A significant age × treatment interaction was found in Rotarod. Post hoc analyses revealed that the treatment effect began at 12 weeks of age, with NAC-HD mice performing better than saline-HD mice from 11 weeks of age. There was also a significant age × genotype interaction, with post hoc analyses revealing that genotype deficits were present from 8 weeks of age, progressively worsening over the lifespan of the mice. (c) Representation of the different phases of stride that can be measured using the Digigait, including total stride, stance, swing, brake and propel duration. (d) A significant treatment × genotype interaction was found in the propel/brake ratio. Post hoc tests show that saline-HD mice have a higher ratio than both NAC-HD and saline-WT mice. (e) A significant treatment effect was seen in the time spent immobile during the FST, with NAC-treated mice spending less time immobile than saline-treated mice. Error bars represent mean±s.e.m.; n=10–14; treatment effects: *P<0.05, **P<0.01, ***P<0.001. FST, forced swim test; HD, Huntington’s disease; NAC, N-acetylcysteine; WT, wild type.

Full size image

The Rotarod was also used to assess motor coordination, and revealed a significant interaction of age and genotype (F(9,35)=7.993, P<0.001). Post hoc analyses revealed that HD mice displayed rotarod deficits from baseline (8 weeks; P=0.001) and progressively worsened until 17 weeks (P<0.001). However, a significant interaction of treatment and age was also observed (F(9,35)=2.746, P=0.015; Figure 1b). Post hoc analyses show that NAC-treated HD mice performed better than saline-treated HD mice from 11 weeks (P=0.038) of age onward.

In a separate analysis of the lower dose of 250 mg kg−1 per day (data not shown), a significant age × genotype interaction was also found (F(7,33)=6.191, P<0.001), with the HD mice showing deficits by 10 weeks (P=0.012) and progressively worsening until 17 weeks (P<0.001). A significant main effect of treatment was also seen at the lower dose (F(1,39)=4.974, P=0.039).

NAC rescues gait abnormalities

The Digigait behavioral apparatus was used to assess if NAC rescues gait abnormalities present in R6/1 mice. Mice were assessed at 8, 10, 12 and 14 weeks of age. Later time points were not measured as HD mice were unable to perform the task due to the progressive degeneration. The session conducted at the 8-week time point was taken as a training session, so measurements were not included in the end analysis. At this time point all groups show movement patterns that are consistent with a motor learning phase (Supplementary Figures 1a and b). Overall, the duration of each phase of movement of the hind paws were found to be the more sensitive indicators of gait abnormalities in HD mice, therefore, the following data are specific to hind paws only.

The time spent in each phase of the stride can be broken down into the following: total stride, stance, swing, propulsion and brake time (as shown in Figure 1c). Significant interactions between age and genotype are present in stride, stance and swing duration (F(2,34)=7.841, P=0.002; F(2,34)=5.918, P=0.006; F(2,34)=4.941, P=0.010; Supplementary Figures 1d and f). Post hoc tests reveal that for each parameter, measurements for WT mice increased from 12 to 14 weeks (P=0.027; P=0.016; and P=0.027, respectively), while HD mice decreased in stride duration from 12 to 14 weeks (P=0.022). Within stance, no changes were found in the time spent propelling (data not shown), whereas a significant interaction of treatment and genotype was found in the time spent braking (F(1,35)=9.728, P=0.004; Supplementary Figure 1c). Post hoc comparisons revealed that saline-treated HD mice had lower braking time than controls (P=0.014), whereas NAC-treated HD mice showed no differences relative to controls. A significant interaction of age and genotype was also found (F(2,34)=4.399; P=0.010), with HD mice showing decreased brake duration at 14 weeks (P=0.012), but no difference at 10–12 weeks. A significant interaction of treatment and genotype was found in the propel/brake ratio (F(1,35)=9.841, P=0.003; Figure 1d). Post hoc tests revealed that saline-treated HD mice show an increased ratio compared with both NAC-treated HD mice (P=0.046) and controls (P=0.002), whereas NAC-treated HD mice showed no difference to controls. No changes were found in gait symmetry, paw overlap, paw positioning, stance width and the ataxia coefficient (data not shown).

NAC reduces immobility during forced swim

The time spent immobile during a period of forced swim was used as a measure of learned helplessness. No differences were found between genotypes. However, a significant effect of treatment was found (F(1,44)=5.739, P=0.021), with NAC-treated mice having lower immobility time than saline-treated mice (Figure 1e). This effect was replicated at a lower dose of 250 mg kg−1 (F(1,23)=5.992, P=0.022), with NAC-treated mice having lower immobility time than saline-treated mice (data not shown).

Change in brain weight, but not body weight, is rescued by NAC

A significant interaction of genotype and treatment was found in brain weight (F(1,43)=9.428; P=0.004). NAC-treated HD mice were found to have heavier brains than saline-treated HD mice (P<0.001), whereas both WT mice had heavier brains than both saline and NAC-treated HD mice (P<0.001 for each comparison; Figure 2a). A significant interaction of age and genotype was found (F(9,35)=16.763; P<0.001), with body weights of HD mice declining from 13 weeks (Figure 2b). However, no differences were noted due to treatment.

Figure 2
figure 2

NAC affects brain weight but not body weight of R6/1 mice. (a) A significant interaction between genotype and treatment was found in the brain weight of 17-week-old mice. Post hoc analyses revealed that saline-treated HD mice had significantly lower brain weights than WT mice, and NAC partially restored this. (b) A significant age × genotype interaction was found in body weight. Post hoc tests revealed that the genotype difference began at 12 weeks of age and progressively increased until 17 weeks. Error bars represent mean±s.e.m.; n=10–14; treatment effects: ***P<0.001. HD, Huntington’s disease; NAC, N-acetylcysteine; WT, wild type.

Full size image

NAC does not affect glutamate transporters

Western blot analysis was carried out on striatal and cortical samples from 17-week-old mice. GLT-1 levels were found to be reduced in the striatum and cortex of HD mice (F(1,24)=6.032, P=0.029; and F(1,24)=15.03, P<0.001, respectively), however, this deficit was not rescued by NAC (Figures 3a and b). The light chain of System xc−, denoted xCT, confers substrate specificity,36 hence, protein levels of xCT were measured as an index of System xc− expression (Figures 3c and d). No changes in xCT were found across genotype or treatment groups in either the striatum or cortex.

Figure 3
figure 3

Western blot analysis of astrocytic glutamate transporter proteins in the striatum and cortex. (a–b) GLT-1 was significantly reduced in the striatum and cortex of HD mice, but was unaffected by NAC treatment. (c–d) Protein levels of xCT were unchanged across genotype and treatment. Error bars represent mean±s.e.m.; n=7–8; genotype effects: #P<0.05. HD, Huntington’s disease; NAC, N-acetylcysteine; WT, wild type.

Full size image

NAC has mitochondria-specific antioxidant effects

GSH is one of the primary endogenous antioxidants. When GSH is oxidized, it forms glutathione disulfide (GSSG). By looking at the ratio of GSH:GSSG, the antioxidant capacity of the cell can be inferred.9 To assess the antioxidant capacity of the striatum, GSH and GSSG were measured via HPLC. No changes in either GSH, GSSG or the GSH:GSSG ratio were found (Supplementary Figure 2).

As the effects of oxidative stress may differ across cellular compartments, the striatal tissue of 17-week-old mice was separated into mitochondrial and nuclear fractions. Protein carbonylation is a stable indicator of oxidative damage to proteins.37 Within the nuclear fraction of the striatum, no difference in protein carbonylation was found within either genotype or treatment groups (Figure 4b). In contrast, the analysis of protein carbonylation in striatal mitochondria revealed a significant interaction between genotype and treatment (F(1,26)=8.864, P=0.006; Figure 5a). Post hoc comparisons show that there is a two-fold increase in protein oxidation in HD mice (P<0.01), which NAC is able to completely rescue (P<0.05). No differences in protein carbonylation were found between the treatment or genotype groups, in either the nuclear or mitochondrial fraction of cortical tissue (Figures 4c and d).

Figure 4
figure 4

Oxidative stress as assayed by protein carbonyl content in fractions of striatal and cortical tissue. (a) There was a significant interaction between genotype and treatment in protein carbonyl content within the striatal mitochondria. Post hoc comparisons revealed that saline-HD mitochondria had higher protein carbonyl content than saline-WT and this was rescued in NAC-HD mice. (b–d) There were no differences in protein carbonyl content in the mitochondrial fraction of the cortex, or in the nuclear fractions from the striatum and cortex across either genotype or treatment groups. Error bars represent mean±s.e.m.; n=7–8; genotype and treatment effects: ##P<0.01 and *P<0.05, respectively. HD, Huntington’s disease; NAC, N-acetylcysteine; WT, wild type.

Full size image
Figure 5
figure 5

Functional analysis of mitochondria isolated from the striatum and cortex. (a) Overview of the results from the mitochondrial stress test performed on the striatal mitochondria, where it was found that: (b–d) State II, state III and state IV respiration, all displayed significant interactions. Post hoc tests all showed that saline-HD mitochondria had lower respiration than saline-WT. NAC-HD mitochondria did not differ from saline-WT mitochondria in any respiratory state. (e) A significant genotype effect was found in the state III:state II ratio, with HD mitochondria higher than WT. (f) No differences in RCR were found in the striatal mitochondria. (g) Overview of the results from the mitochondrial stress test performed on the cortical mitochondria, where it was found that: (h–l) state II, state III and state IV respiration, the state III:state II ratio and RCR, did not differ between groups. Error bars represent mean±s.e.m.; n=7–8; genotype effects: *P<0.05; ##P<0.01. HD, Huntington’s disease; NAC, N-acetylcysteine; RCR, respiratory control rate; WT, wild type.

Full size image

NAC rescues functional deficits in HD mitochondria

Mitochondria were isolated from the striatum and cortex and a mitochondrial stress test was performed to assess their function. Mitochondria isolated from the striatum show differences in state II, state III and state IV respiration, and the state III:state II ratio (Figure 5a). Specifically, a significant interaction was found in state II respiration (F(1,18)=8.078, P=0.011), with post hoc tests revealing that saline-treated HD mitochondria had lower state II respiration than saline-treated WT mitochondria (P<0.05). A significant interaction was also found in state III and state IV respiration (F(1,18)=9.409, P=0.007; and F=6.738, P=0.018, respectively), with post hoc tests revealing that saline-treated HD mitochondria had lower state III and state IV respiration than saline-treated WT mitochondria (P<0.05). Furthermore, a genotype effect was found on the state III:state II ratio (F(1,18)=10.85, P=0.004), with HD mitochondria showing a higher ratio than WT. No changes were found in the respiratory control rate. Mitochondria isolated from the cortex show no differences across treatment or genotype groups on any measures of respiration (Figures 2g and l).

Discussion

Motor dysfunction is a primary clinical symptom of HD. In this study, we show that NAC is able to delay the onset, and slow the progression, of motor symptoms in a transgenic mouse model of HD. Depressive symptoms are also common in HD38 and we show that NAC displays antidepressant-like effects in both HD and WT mice. NAC was found to rescue oxidative stress within striatal mitochondria of HD mice. These changes were associated with rescued functional deficits in the mitochondria. This supports the hypothesis that NAC rescues motor dysfunction by ameliorating mitochondrial dysfunction and inhibiting oxidative stress.

We have employed a comprehensive analysis of R6/1 motor deficits to assess the domains that are rescued by NAC. Uncontrolled and reflexive paw-clasping is a primary motor endophenotype shown by the R6 transgenic mice.28 NAC was able to substantially delay the onset of paw-clasping in HD mice, and significantly reduce the severity of this abnormal phenotype. Similar effects were also seen in rotarod performance.

The rotarod is the most widely used tool to test motor performance in HD mice. The rotarod primarily tests locomotor coordination, with R6/1 mice displaying a progressive deficit in their performance.39 R6/1 mice were treated with NAC after the onset of this deficit; however, the drug was still able to significantly inhibit the progressive loss of locomotor coordination. It is striking that NAC was able to take effect even in animals that had already undergone some loss of striatal function. In contrast, NAC has been shown previously to have detrimental effects in fly and zebrafish models of HD40 when administered before the development of HD phenotypes. This was shown to be due to an inhibition of autophagy clearance of mutant HTT aggregates. It may be that the therapeutic benefit of NAC outweighs possible detriments after the pathological hallmarks have developed,41 although it is difficult to directly compare dosing regimes across such divergent model species. Similar dichotomous effects have also been seen in HD models using other paradigms such as stress, where corticosterone treatment was found to hasten the onset of cognitive symptoms in presymptomatic R6/1 mice42 whereas dexamethasone (a synthetic glucocorticoid) administration was able to increase heat-shock response, reduce mutant HTT aggregate load and slow the progression of motor dysfunction in the more severe R6/2 model of HD.43 Furthermore, evidence for the therapeutic effects of NAC in HD mice expressing a human transgene is likely to be more clinically relevant than the data from transgenic fish and flies. Thus, NAC has advantages for potential clinical application of this therapy, where diagnosis and treatment are typically commenced after the development of analogous motor symptoms.

The therapeutic effects of NAC were also mirrored in the gait analysis of R6/1 mice. The Digigait analysis has only been published once previously in a HD mouse model, in which Pallier et al.44 conducted a cross-sectional analysis of gait in late-stage R6/2 mice (17 weeks). We are the first to report a progressive analysis of gait in a HD animal model, beginning at the time of motor symptom onset. Overall, R6/1 mice show an emerging deficit in two components of stance: time spent braking and the brake-to-propel ratio. Specifically, HD mice display a decrease in the time spent braking, with a subsequent increase in the propel-to-brake ratio. These findings are consistent with Pallier et al., who found deficits within ‘stance’ components only. Importantly, we found that NAC was able to rescue these deficits. Thus, with the treatment of NAC, we were able to rescue deficits in R6/1 mice from three separate domains of motor function: reflexive paw-clasping, locomotor coordination and gait. NAC also had antidepressant-like effects.

Depression is the highest occurring psychiatric comorbidity in HD,38 with recent studies suggesting that it is an inherent symptom of the disorder.45 Immobility time during the forced swim test is the gold standard for screening the efficacy of antidepressants in rodents.46 We found that NAC was able to reduce the time spent immobile in both R6/1 mice and their wild-type littermates. This has implications for both HD and wider research on major depression, suggesting that NAC may be an effective antidepressant. This supports data from previous clinical studies on bipolar patients, which have shown NAC to have positive effects on the depressive symptoms of the disorder.47 Taken together, NAC was able to display antidepressant-like effects, while also rescuing an array of motor deficits in the R6/1 mice. Our findings are echoed by a recent publication displaying the ability of cysteine supplementation to reverse clasping and Rotarod deficits within the R6/2 model of HD—although gait and affective-like behaviors were not assessed.14 In this study, cysteine supplementation was given using a cysteine-rich diet in combination with NAC. Our study is the first to display the therapeutic effect of NAC alone in a transgenic animal model.

NAC may be acting in the brain through modulation of glutamatergic excitotoxicity or levels of oxidative stress. Glutamate-dependent toxicity is an established mechanism of cellular dysfunction within both HD patients and animal models.7, 48, 49, 50 Glutamate can be transported out of the synapse by a variety of excitatory amino-acid transporters, including the astrocytic transporter, GLT-1.24 GLT-1-dependent glutamate uptake is impeded by the presence of a high extracellular glutamate-to-cysteine ratio.24, 25 Therefore, the increase in synaptic glutamate may explain why GLT-1 is decreased in HD.19, 20, 51 Our findings replicate previous studies19, 52 showing a deficit in striatal and cortical GLT-1, however, NAC did not rescue this change.

NAC is also capable of acting on a second astrocytic glutamate transporter, System xc−.23 System xc− is responsible for the transport of glutamate from astrocytes in a 1:1 exchange for extracellular cysteine.22, 53 Cystine is produced by creating a disulfide bond through the oxidation of two cysteine groups, and is thus dependent on extracellular levels of cysteine.24 Previous studies have shown that NAC is able to upregulate glutamate–cysteine exchange, and this has a therapeutic effect in treating drug withdrawal in animals and humans.26, 54 The light chain of System xc−, xCT, confers substrate specificity, thus protein levels of xCT were analyzed to evaluate functional changes in System xc−. No changes were found in levels of xCT. Considering GLT-1 and System xc− are the only glutamate transporters dependent on the presence of extracellular cysteine, we can conclude that the mechanism of action of NAC in R6/1 mice is not directly related to the reduction of glutamate-dependent toxicity. We, therefore, hypothesized that NAC is functioning by reducing oxidative stress.

GSH is the primary endogenous antioxidant.55 When GSH is oxidized, it forms GSSG.55 By looking at the ratio of GSH:GSSG, the antioxidant capacity of the cell can be inferred.9 Previous studies in HD have shown that GSH is reduced in the serum of HD patients.10 However, other studies have shown that GSH and GSSG are unchanged in R6/1 mouse striatum at the onset of motor symptoms.33 NAC could potentially stabilize the GSH:GSSG balance by donating cysteine for the synthesis of new GSH.56 We observed no changes in either GSH, GSSG or the GSH:GSSG ratio in unfractionated tissue homogenates. However, the levels of these key elements of the antioxidant system are differentially regulated across cellular organelles.

GSH is produced in the cytosol and transferred to other intracellular components, such as nuclei and mitochondria.57, 58 Thus, although the total GSH/GSSG pools may be unchanged at the level of the whole tissue, it may be that oxidative stress is occurring in specific organelles of the cell. To investigate the possibility of compartmentalized oxidative stress, the striatal and cortical tissue of 17-week-old mice were fractionated into mitochondrial and nuclear components. Given the volatile nature of the GSH-GSSG redox cycle, as well as the reactive nature of GSH to oxidation, it may be that subtle differences in GSH:GSSG are quickly lost in the postmortem brain.59 Therefore, protein carbonylation was assessed as a more stable indicator of oxidative damage to proteins.37 It was found that this key marker of damage resulting from oxidative stress was increased in mitochondrial, but not nuclear, fractions of striatal tissue only. Importantly, NAC was able to rescue this region- and organelle-specific oxidative stress.

It has been shown previously that mitochondria are more prone to oxidative damage than nuclei within immortalized striatal cells from HdhQ111/Q111 mice.60 This is to be expected given that mitochondria are the primary source of reactive oxygen species within the cell.61 Mitochondria are more likely to leak reactive oxygen species when they are not functioning efficiently.61 Multiple studies have shown mitochondrial processes to be dysfunctional in HD, including reductions in the activity of electron transport chain complexes II, III and IV,6, 48, 62, 63 as well as reduced activity of aconitase,50 pyruvate dehydrogenase64 and creatine kinase.65 Each of these studies are beneficial in showing reductions in activity of specific enzymes, however, the electron transport chain contains dynamic protein complexes, and isolation of specific enzymes cannot elucidate the overall mitochondrial function. Thus, we sought to look at the respiratory function of intact mitochondria isolated from the striatum of R6/1 mice. We hypothesized that the oxidative stress specific to the R6/1 mitochondria would be associated with an overall reduction in mitochondrial oxidative metabolism, which would be rescued by NAC treatment.

To assess mitochondrial function, oxygen consumption rates were assessed after the administration of various mitochondrial substrates and inhibitors. This analysis showed that NAC enhanced baseline respiration (state II), maximal ADP-stimulated respiration (state III) and respiration after consumption of ADP (state IV), in HD mice, suggesting that NAC could enhance mitochondrial capacity in this model. This provides functional evidence to support the finding that ATP pools are depleted in HD models.66 This also has potential implications for the generation of mitochondrial reactive oxygen species, a putative pathogenic factor in HD. Lower capacity for ADP-stimulated respiration, in the face of constant energy demand and substrate supply, increases the mitochondrial membrane potential. Electron leak from the respiratory chain in the form of O2− is a compensatory response to limit further increases in membrane potential, resulting in an oxidative stress response.67 Indeed, the enhanced respiratory capacity of HD mice treated with NAC was associated with lower protein carbonylation in mitochondrial fractions from HD mice and could have protected from further mitochondrial oxidative damage and dysfunction.

Taken together, NAC was found to delay the onset and progression of motor dysfunction in the R6/1 transgenic mouse model of HD, while displaying antidepressant-like effects on both R6/1 and WT mice. NAC was also shown to rescue mitochondrial dysfunction and reduce oxidative stress within the R6/1 striatum, identifying a likely mechanism of action. Considering that NAC is established as safe for clinical use,68 our data suggest that it should be considered as a candidate for trial as either a primary or adjunctive treatment for symptomatic HD patients. The question of whether NAC would be effective in presymptomatic stages of the disease is also a high priority for investigation, and it is suggested that trialing NAC on a slower progressing model, such as knock-in or BACHD mice, would provide more insight, as there are no disease-modifying therapies currently available for this devastating disorder.