Intranasal Carnosine Mitigates α-Synuclein Pathology and Motor Dysfunction in the Thy1-aSyn Mouse Model of Parkinson’s Disease

Abstract

Parkinson’s disease (PD) is a debilitating neurodegenerative disorder. Early symptoms include motor dysfunction and impaired olfaction. Toxic aggregation of α-synuclein (aSyn) in the olfactory bulb (OB) and substantia nigra pars compacta (SNpc) is a hallmark of PD neuropathology. Intranasal (IN) carnosine (2 mg/d for 8 weeks) was previously demonstrated to improve motor behavior and mitochondrial function in Thy1-aSyn mice, a model of PD. The present studies evaluated the efficacy of IN carnosine at a higher dose in slowing progression of motor deficits and aSyn accumulation in Thy1-aSyn mice. After baseline neurobehavioral assessments, IN carnosine was administered (0.0, 2.0, or 4.0 mg/day) to wild-type and Thy1-aSyn mice for 8 weeks. Olfactory and motor behavioral measurements were repeated prior to end point tissue collection. Brain sections were immunostained for aSyn and tyrosine hydroxylase (TH). Immunopositive cells were counted using design-based stereology in the SNpc and OB mitral cell layer (MCL). Behavioral assessments revealed a dose-dependent improvement in motor function with increasing carnosine dose. Thy1-aSyn mice treated with 2.0 or 4.0 mg/d IN carnosine exhibited fewer aSyn-positive (aSyn(+)) cell bodies in the SNpc compared to vehicle-treated mice. Moreover, the number of aSyn(+) cell bodies in carnosine-treated Thy1-aSyn mice was reduced to vehicle-treated wild-type levels in the SNpc. Carnosine treatment did not affect the number of aSyn(+) cell bodies in the OB-MCL or the number of TH(+) cells in the SNpc. In summary, intranasal carnosine treatment decreased aSyn accumulation in the SNpc, which may underlie its mitigation of motor deficits in the Thy1-aSyn mice.

Graphical Abstract

INTRODUCTION

Parkinson’s disease (PD) is the second-most common neurodegenerative disease^1,2 and is estimated to affect more than one million people in the United States.^1–3 About 90% of PD cases are sporadic and believed to be caused by gene-environment interactions.^1,4 The etiology of sporadic PD has yet to be elucidated,¹ but studies have suggested that PD manifestation is dependent on α-synuclein (aSyn) protein aggregation, mitochondrial dysfunction, and reactive oxygen species (ROS) production^5,6 that eventually lead to cell death in motor and nonmotor systems.^7,8

Preclinical prodromal stages of PD occur before classical parkinsonian motor dysfunction and before the population of dopaminergic neurons in the substantia nigra (SN) is significantly decreased.⁹ Early PD symptoms include hyposmia, decreased gastrointestinal motility, and mild motor dysfunction.^2,7,10 In sporadic PD, 90% of cases display clinically detectable olfactory impairment for years to decades before motor symptoms.¹¹ Degeneration of SN dopaminergic neurons begins at Braak stage 3⁷ and coincides with onset of hallmark PD motor symptoms.²

Aggregation of aSyn is a major component of PD neuropathology at all stages and is the primary component of Lewy body pathology, a key feature of PD.¹² Familial forms of PD in humans and in mouse models with overexpressed or mutated aSyn suggest abnormal aSyn accumulation plays a causal role in disease.¹³ The earliest stages of PD display aSyn pathology in the olfactory bulb (OB),^7,8 which precedes degeneration of dopaminergic neurons in the SN pars compacta (SNpc).⁷

Studies on individuals with PD^7,14,15 and in animal models of PD^16–19 have demonstrated that increased total aSyn protein expression is correlated with severity of symptoms and disease progression. In a dopaminergic cell model, aSyn overexpression decreased tyrosine hydroxylase (TH; the rate-limiting enzyme for catecholamine biosynthesis and the marker for dopamine in the SN) activity and dopamine production without changing TH protein content.²⁰ Therefore, increases in aSyn protein may be causative in dopaminergic malfunction observed in early stages of PD, potentially contributing to death of dopaminergic neurons.

The Thy1-aSyn transgenic mouse has been extensively characterized and displays phenotypic similarity to early PD. Prodromal symptoms of PD are present in the Thy1-aSyn transgenic mouse model as early as 6–8 weeks of age. These include hyposmia^18,21 and mild motor dysfunction.^17,21 This model displays region-specific aSyn aggregation in the midbrain and forebrain associated with both motor and nonmotor impairments observed clinically in early stage PD patients.^14,21,22 Moreover, Thy-aSyn mice display aSyn-aggregation pathology in the olfactory epithelium (OE),²¹ which resembles aSyn OE pathology of PD patients.²³ This model has previously been used to test potential PD interventions.^24,25

l-Carnosine (β-alanyl-L-histidine; carnosine) is a dipeptide and antioxidant that has been demonstrated to prevent protein aggregation²⁶ and exhibits low toxicity.^27,28 In a pilot study, PD patients treated with an oral supplement of 1.5 g/d carnosine for 30 days showed improved neurological symptoms and decreased oxidative byproducts in blood compared to the standard treatment alone (levodopa with dopamine receptor agonists).⁵ Previous work showed that Thy1-aSyn mice treated with 2.0 mg/d intranasal (IN) carnosine for two months demonstrated improvement in transcriptional dysregulation and mitochondrial function.²⁹ IN carnosine treatment also slowed the progression of motor deficits and reduced aSyn expression in the OE in the Thy1-aSyn mice.²¹ Moreover, a recent in vitro study reported that carnosine prevented 6-hydroxydopamine-induced cell death in GT1–7 cells by inhibiting the reactive oxygen species (ROS)-JNK signaling pathway.³⁰

In the present study, we hypothesized that IN carnosine in the Thy1-aSyn mice would reduce aSyn pathology and improve olfactory and motor function. The IN route is emerging as an effective route to deliver therapeutics to the brain in humans.^31,32 To test our hypothesis, we evaluated a higher dose of IN carnosine than was used in previous studies and demonstrated a greater beneficial effect on motor symptoms as well as reduced aSyn pathology in the brain. The results of this study provide preclinical data to justify future evaluation of a neuroprotective role for intranasal carnosine in clinical trials for PD patients.

RESULTS AND DISCUSSION

Motivation for This Work.

Carnosine [CASRN 305–84-0] is an extensively studied dipeptide and, as a result, is known to have beneficial effects such as improving mitochondrial function, serving as an intracellular metal chelator and preventing protein aggregation;^27,29 for a recent review, see ref 33. These features suggest that carnosine could have substantial beneficial effects in treatment of neurodegenerative diseases and specifically in reducing PD neuropathology and improving clinical outcomes in the Thy1-aSyn mouse model of PD. Further, the IN route of administration for carnosine, as used here, has been demonstrated to be superior to the oral route of exposure.²¹ In the present study, by showing abrogating effects on protein aggregation (specifically aSyn) in a key brain region (the SNpc) and by demonstrating a dose–response for improvement of motor function, we provide additional evidence for the beneficial effects of IN carnosine in a leading mouse model of PD that could be translatable to human medicine.

Study Design.

As shown in Figure 1, baseline behavioral assessments were performed when mice were approximately 8 weeks of age, followed by 8 weeks of daily doses of intranasal carnosine (2 or 4 mg/d). At the end of dosing, end point behavioral assessments were performed. No treatment-related clinical signs of toxicity were observed during intranasal carnosine treatment, and all mice gained weight at comparable rates throughout the study (see Supplemental File 1).

Figure 1.

Study timeline.

Neurobehavioral Assessments: Challenging Beam Traversal.

Challenging beam traversal was performed at baseline and after 8 weeks of daily IN carnosine treatment (end point) to measure nigrostriatal motor function (Figure 1). The challenging beam task is particularly robust for measuring motor dysfunction over time because habituation to the test conditions does not influence performance.³⁴ Latency to traverse the beam, number of steps, and number of errors were scored and averaged for each mouse at both time points (Figure 2A,B). The mean latency to traverse the beam was not different between genotypes at baseline (Figure 2C), but the latency was increased in the Thy1-aSyn mice at end point, (p = 0.010, two-way ANOVA, n = 35–39) (Figure 2D). Thus, the mean number of seconds to traverse the beam for wild-type and Thy1-aSyn mice was 9.9 ± 0.72 and 11.7 ± 0.82 at baseline and 10.4 ± 0.80 and 14.1 ± 0.95 at end point, respectively. There was no effect of carnosine treatment on latency to traverse the challenging beam at either dose level (p > 0.21).

Figure 2.

(A–I) Challenging beam traversal. Mice were scored for latency to traverse the beam, total right hindlimb steps, and total errors. (A) Mouse not making errors while traversing the beam. (B–H) Box and whisker plots indicate the 75th percentile (top of box), 25th percentile (bottom of box), and median (line inside of box), and the whiskers indicate the minimum and maximum for the data set. (B) Mouse making errors while traversing the beam. (C) There was no difference between the genotypes for latency to traverse the beam at baseline. (D) After 8 weeks of treatment (end point), Thy1-aSyn mice took more time to traverse the beam compared to wild-type animals. Carnosine treatment did not affect latency (p = 0.212). (E) At baseline, Thy1-aSyn mice made more total errors compared to wild-type mice. (F) At end point, there was a carnosine treatment by genotype interaction. Thy1-aSyn mice treated with 2.0 and 4.0 mg/d carnosine demonstrated a smaller increase in errors compared to vehicle-treated Thy1-aSyn mice. Vehicle-treated Thy1-aSyn mice exhibited a greater increase in errors end point compared to vehicle-treated wild-type and 4.0 mg/d carnosine-treated Thy1-aSyn mice. (G) Thy1-aSyn mice have a greater number of errors per step compared to wild-type mice at baseline. (H–I) The change in errors per step after 8 weeks of treatment demonstrated a treatment by genotype by time interaction. (H) Vehicle-treated Thy1-aSyn mice showed a greater increase in errors per step after 8 weeks of treatment compared to vehicle-treated wild-type mice and 4.0 mg/d carnosine-treated Thy1-aSyn mice. (I) Change in errors per step in Thy1-aSyn mice is dependent on carnosine dose level (Thy1-aSyn mean linear model R² = 0.9976). There is no dose–response in wild-type animals (R² = 0.0184). Solid line = mean dose values from data; dotted line = linear model of means; blue lines = wild-type; orange lines = Thy1-aSyn. ***p < 0.001, **p < 0.01, *p <0.05.

At baseline, mean total errors were elevated in the Thy1-aSyn mice (8.2 ± 0.33 errors) compared to wild-type mice (1.8 ± 0.14 errors) (p < 0.001, Student’s t test, n = 35–39) (Figure 2E). After 8 weeks of treatment, vehicle-treated Thy1-aSyn mice increased mean total errors by 2.9 ± 0.67 from baseline, whereas the increase in total errors in Thy1-aSyn mice administered 2.0 mg/d carnosine or 4.0 mg/d carnosine was considerably less (0.8 ± 0.54 and 0.7 ± 0.70, respectively) (Figure 2F). At end point, wild-type mice treated with vehicle, 2.0 mg/d carnosine, or 4.0 mg/d carnosine had a change of total errors of −0.1 ± 0.21, 0.4 ± 0.23, and −0.1 ± 0.22, respectively. The total errors measurement was determined to have a carnosine-treatment by genotype by time interaction (p = 0.031, three-way RM mixed ANOVA, n = 6–16/treatment-group). Linear analysis of carnosine treatment found an overall significant reduction in the change in total errors from baseline to end point (change in total errors) at both doses in Thy1-aSyn mice (p < 0.001, general linear regression), whereas no effect of carnosine was observed in wild-type mice. Pairwise analysis of the change in total errors demonstrated that the vehicle-treated Thy1-aSyn mice had increased total errors compared to vehicle-treated wild-type mice (p < 0.001), but this increase was significantly reduced in the 4.0 mg/d carnosine-treated Thy1-aSyn mice (p = 0.024). Total errors did not increase over time in Thy1-aSyn mice treated with either dose of carnosine compared to their wild-type controls (p ≥ 0.819) (Figure 2F).

When total errors were normalized to number of steps (errors/step), the mean errors/step was found to be elevated in the Thy1-aSyn mice (0.47 ± 0.02 errors/step) compared to the wild-type mice (0.10 ± 0.01 errors/step) at baseline (p < 0.001, Student’s t test, n = 26–35) (Figure 2G). After 8 weeks of dosing, errors/step increased in Thy1-aSyn mice by 0.11 ± 0.03 in vehicle-treated mice and by 0.05 ± 0.03 in 2.0 mg/d carnosine-treated mice. The effect was even greater in Thy1-aSyn mice treated with 4.0 mg/d carnosine, with a change in errors/step of −0.01 ± 0.03. Wild-type mice treated with vehicle, 2.0 mg/d carnosine, or 4.0 mg/d carnosine had essentially no changes in errors/step after 8 weeks of treatment (0.00 ± 0.01, 0.02 ± 0.01, and −0.01 ± 0.01, respectively).

As observed with total errors, errors/step displayed a carnosine-treatment by genotype by time interaction (p = 0.009, three-way RM mixed ANOVA, n = 6–16/treatment-group) (Figure 2H,I). Carnosine-treatment had no effect on errors/step from baseline to end point in wild-type mice, whereas carnosine treatment had a dose-dependent effect on errors/step from baseline to end point in Thy1-aSyn mice. Thy1-aSyn mice treated with 4.0 mg/d carnosine had no increase in errors/step over time compared to vehicle-treated Thy1-aSyn mice (p = 0.004, Tukey HSD) and no difference over time compared to vehicle-treated wild-type mice (p = 0.999).

A key finding of this study was that the IN carnosine dose of 4.0 mg/d prevented an increase in errors per step after 8 weeks of treatment. Our analysis demonstrated a clear linear dose–response specific to the Thy1-aSyn mice, with the 4.0 mg/dose demonstrating no progressive motor dysfunction over the two months of treatment (Thy1-aSyn mean treatment linear model R² = 0.9976, p = 0.004, linear regression) (Figure 2I). There was no dose–response in wild-type animals (R² = 0.0184).

This study also reproduced the findings of Bermúdez et al.,²¹ that a 2 mg/d IN dose of carnosine has some benefit in preventing progression of motor deficits. aSyn pathology is an early component of PD progression,⁷ and the prevention of worsening motor dysfunction in an aSyn-overexpression model by IN carnosine may have important translational implications.

Neurobehavioral Assessments: Spontaneous Activity.

Spontaneous activity measurements included the number of forelimb steps, hindlimb steps, rears, and time spent grooming (Figure 3). The main effect of genotype was observed in all baseline measurements with Thy1-aSyn mice having fewer forelimb steps (Figure 3A), hindlimb steps (Figure 3D), rears (Figure 3F), and time spent grooming (Figure 3H) compared to wild-type mice (p < 0.001, Student’s t test, n = 37–54).

Figure 3.

(A–I) Spontaneous activity box and whisker plots indicate the 75th percentile (top of box), 25th percentile (bottom of box), and median (line inside of box), and the whiskers indicate the minimum and maximum for the data set. At (A) baseline and (B) end point, Thy1-aSyn mice displayed fewer forelimb steps compared to wild-type mice. (C) Carnosine treatment led to a dose-dependent increase in forelimb steps for both wild-type and Thy1-aSyn mice (mean treatment linear model R² = 0.9997, p < 0.001). Pairwise comparison demonstrated that mice treated with 4.0 mg/d IN carnosine exhibited increased forelimb steps compared to all vehicle-treated mice. At both (D) baseline and (E) end point, Thy1-aSyn mice displayed fewer hindlimb steps compared to wild-type mice, with no effect of treatment after 8 weeks of treatment (p = 0.216). (F) At baseline, Thy1-aSyn mice displayed deceased rearing compared with wild-type mice. (G) At end point, there was an interaction of carnosine treatment and genotype on cumulative rearing. Wild-type mice showed a dose-dependent decrease in rearing, whereas there was no effect of carnosine treatment on Thy1-aSyn mice. However, there was no significance when the wild-type rearing was assessed by linear regression (p = 0.059) and no pairwise effects. (H, I) Compared to wild-type mice, Thy1-aSyn mice spent less time grooming at baseline and at end point. No effect of carnosine treatment was observed at end point (p ≥ 0.658). Solid line = mean dose values from data; dotted line = linear model calculated from means; blue lines = wild-type; orange lines = Thy1-aSyn; gray lines = both genotypes combined. ***p < 0.001, *p < 0.05.

At end point, the effect of genotype was detected in forelimb steps (Figure 3B), hindlimb steps (Figure 3E), and grooming seconds (Figure 3I) (p < 0.001, mixed ANOVA). After 8 weeks of carnosine treatment, the number of forelimb steps was increased in wild-type and Thy1-aSyn mice in a dose-dependent manner (p = 0.014, 2-way mixed ANOVA, n = 6–26) (Figure 3C). All mice showed a dose–response when modeled with linear regression (mean treatment linear model R² = 0.9997, p < 0.001, linear regression, n = 14–39/treatment). Pairwise comparisons showed mice receiving 4.0 mg/d carnosine took more forelimb steps compared to vehicle-treated mice (p = 0.042, Tukey HSD). There was no effect of carnosine treatment on the number of hindlimb steps (Figure 3E) or time spent grooming (Figure 3I). There was a carnosine treatment by genotype interaction for rearing after 8 weeks of treatment. Wild-type mice exhibited a dose-dependent decrease in rearing, whereas there was no effect of carnosine treatment on Thy1-aSyn mice (p = 0.037, two-way mixed ANOVA, n = 6–25) (Figure 3F). However, there was no significance when the wild-type rearing was modeled by linear regression (p = 0.059) and no pairwise effects were detected. Although the Thy1-aSyn mice displayed fewer rears compared to wild-type at baseline, this difference was not present at end point.

Consistent with Bermúdez et al.,²¹ we observed no effect of 2.0 mg/d IN carnosine on grooming behavior and additionally no effect of 4.0 mg/d IN carnosine on this end point measurement (Figure 3I). Rearing, grooming, forelimbs steps, and hindlimb steps do not display accelerated functional decline in Thy1-aSyn mice at two and four months of age.³⁴ Therefore, we conclude that these repeated measurements of sensorimotor behavior do not detect progressive decline in motor function in this mouse model.

Neurobehavioral Assessments: Buried Food Pellet Test.

Comparable to the results of Bermúdez et al.²¹ with Thy1-aSyn mice treated with 2.0 mg/d carnosine, we found no effect of 4.0 mg/d carnosine in latency to retrieve the buried food pellet. However, we modified the method of Bermúdez et al.²¹ and recorded the latency to first dig over the buried food pellet location; this measurement was more sensitive to detect olfactory dysfunction in the Thy1-aSyn mice at 8 weeks of age. There were no differences in the latency to find the surface pellet. Therefore, the results of the experiment were not confounded by motor effects.

At baseline, the latency to dig at the buried pellet was significantly greater for Thy1-aSyn mice (Figure 4A). After 8 weeks of treatment, there was a genotype by 4.0 mg/d carnosine treatment interaction (p = 0.033, 2-way ANOVA, n = 13–23) (Figure 4B). End point latency to dig at the buried food pellet after vehicle and 4.0 mg/d carnosine treatment showed no significant differences by pairwise analysis. There were no significant differences in the latency to find the surface pellet at both baseline (p = 0.267, Student’s t test, n = 31–36) (Figure 4C) and end point (p ≥ 0.064, two-way ANOVA, n = 13–23) (Figure 4D).

Figure 4.

(A–D) Buried food pellet test. Box and whisker plots indicate the 75th percentile (top of box), 25th percentile (bottom of box), and median (line inside of box), and the whiskers indicate the minimum and maximum for the data set. (A) Thy1-aSyn mice demonstrate an increased latency to dig at the buried food pellet location compared to wild-type mice. (B) At end point, there was a carnosine treatment by genotype interaction (p = 0.033). There were no significant differences detected by pairwise analysis. (C, D) There was no difference in latency to locate the surface food pellet at both (C) baseline and (D) end point. **p = 0.002.

α-Synuclein and Tyrosine Hydroxylase Immunohistochemistry and Stereology: Substantia Nigra Pars Compacta. Details of immunostaining features of aSyn-positive and TH-positive neurons can be found in Supporting Information section S2.

A key feature of human PD pathology is the Lewy body, a cytosolic protein aggregate composed mainly of aSyn and is found in the nerve cell body. In post-mortem PD patient brains, aSyn is also seen as diffuse aggregates in the cytosol.^7,15

In mouse brain, Taguchi et al.³⁵ reported that aSyn was expressed ubiquitously in the synaptic terminals and processes, whereas aSyn expression in the neuronal cell bodies was observed in specific brain regions including the OB and the SNpc of healthy mice. Furthermore, Taguchi et al.³⁵ suggested somatic localization of aSyn in the healthy brain may be a risk factor for aSyn pathology in PD.

Here, we report the first stereological quantification of total numbers of aSyn(+) cell bodies in the SNpc of Thy1-aSyn mice. Other studies have used stereology to quantify total numbers of aSyn cell bodies in the SNpc of rats³⁶ and mice,³⁷ in the SN of monkeys,³⁸ and in the dorsal motor nucleus of the vagus nerve.³⁹ As determined using design-based stereology, there was a main effect of genotype and carnosine treatment for the number of aSyn(+) cell bodies in the SNpc (p ≤ 0.001, two-way mixed ANOVA, n = 24–25/genotype, n = 13–20/treatment) (Figure 5). There was a main effect of genotype and carnosine treatment for the number of aSyn(+) cell bodies in the SNpc (p < 0.001, two-way mixed ANOVA, n = 24–25/genotype, n = 13–20/treatment) (Figure 5G). Pairwise comparisons demonstrated both 2.0 mg/d carnosine and 4.0 mg/d carnosine decreased the number of aSyn(+) neurons in the SNpc of the Thy1-aSyn mice (p < 0.001, Tukey HSD).

Figure 5.

(A–G) α-Synuclein immunohistochemistry and stereology in the substantia nigra pars compacta and representative images of aSyn(+) neurons in the SNpc (100× magnification). (A–C) aSyn(+) neurons in the SNpc of wild-type mice treated with vehicle, 2.0 mg/d carnosine, or 4.0 mg/d carnosine. (D–F) aSyn(+) neurons in the SNpc of Thy1-aSyn mice treated with vehicle, 2.0 mg/d carnosine, or 4.0 mg/d carnosine. (G) Box and whisker plot indicates the 75th percentile (top of box), 25th percentile (bottom of box), median (line inside of box), and the whiskers indicate the minimum and maximum for the data set. Quantification of aSyn(+) cell bodies in the SNpc using design-based stereology. There was a main effect of genotype and treatment. Thy1-aSyn mice contain more aSyn(+) cell bodies compared to wild-type mice. Mice treated with 2.0 and 4.0 mg/d IN carnosine had fewer aSyn(+) cell bodies compared with vehicle-treated mice. Pairwise comparisons of genotype-treatment groups demonstrated that Thy1-aSyn mice treated with 2.0 and 4.0 mg/d carnosine have fewer aSyn(+) cell bodies compared to vehicle-treated transgenic mice. Thy1-aSyn mice treated with 2.0 and 4.0 mg/d carnosine have equivalent numbers of aSyn(+) cell bodies compared to vehicle-treated wild-type (p ≥ 0.257). Vehicle-treated Thy1-aSyn mice treated exhibit more aSyn(+) cell bodies compared to vehicle-treated wild-type mice. Scale bar in (A) is 50 μm for (A)–(F). V = aSyn(+) cell body. ***p ≤ 0.001.

Moreover, the number of nigral aSyn(+) cells in the Thy1-aSyn mice treated with 2.0 mg/d or 4.0 mg/d IN carnosine was not statistically different from that of the vehicle-treated wild-type mice (p ≥ 0.257, Tukey HSD) (Figure 5G). Stated in another way, both IN doses of carnosine had the impact of reducing the number of aSyn positive neurons in the SNpc of Thy1-aSyn mice to the levels noted in wild-type mice. However, due to the fact that our data show marked motor function improvements in the Thy1-aSyn at 4 mg/d, compared to 2 mg/d, while aSyn(+) neurons were equivalent in number at both doses, other mechanisms should be investigated in the future to understand the greater benefit of the 4 mg/d dose on motor function.

Our results demonstrated no difference in the numbers of nigral TH(+) neurons between the vehicle-treated Thy1-aSyn mice and wild-type mice (Figure 6), consistent with previous findings.⁴⁰ Moreover, there was no effect of carnosine treatment in either genotype, indicating that IN treatment with carnosine does not perturb normal TH expression in the SNpc. Thus, in this study, carnosine reduced aSyn accumulation in nigral neurons without adversely affecting TH expression.

Figure 6.

(A–G) Tyrosine hydroxylase immunohistochemistry and stereology in the substantia nigra pars compacta and representative images of aSyn(+) neurons in the SNpc (100× magnification). (A–C) TH(+) neurons in the SNpc of wild-type mice treated with vehicle, 2.0 mg/d carnosine, or 4.0 mg/d carnosine. (D–F) TH(+) neurons in the SNpc of Thy1-aSyn mice treated with vehicle, 2.0 mg/d carnosine, or 4.0 mg/d carnosine. (G) Box and whisker plot indicates the 75th percentile (top of box), 25th percentile (bottom of box), and median (line inside of box), and the whiskers indicate the minimum and maximum for the data set. Quantification of TH(+) neurons in the SNpc with design-based stereology demonstrated no effect of genotype or carnosine treatment (p ≥ 0.119). Scale bar in (A) is 50 μm for (A)–(F). V = TH(+) cell body.

α-Synuclein Immunohistochemistry and Stereology: Olfactory Bulb.

As in the SNpc, this is the first reported stereological quantification of numbers of aSyn(+) cell bodies in the OB mitral layer of Thy1-aSyn mice. Mitral cells were evaluated because the total number of mitral cell and tufted cells, the output neurons of the OB, is reportedly decreased in the OBs of PD patients.⁸ Tufted cells were not evaluated because there were not enough aSyn(+) cells in wild-type animals to generate stereological estimates.

aSyn was immunostained diffusely throughout the OB and required a hematoxylin counterstain to clearly identify the Syn(+) mitral cells in the OB-MCL (see Supporting Information section S2). aSyn(+) cell bodies in the entire OB-MCL of one randomly selected hemibrain per animal were counted. Representative staining is shown in Figure 7A–D. After 8 weeks of treatment with vehicle or 2.0 mg/d carnosine, the estimated number of aSyn(+) cell bodies in the OB-MCL was approximately 2-fold higher in Thy1-aSyn mice versus wild-type mice, respectively (Figure 7E). The number of aSyn(+) mitral cell bodies was not affected by IN carnosine treatment (Figure 7E). These results suggest downregulation of aSyn by IN carnosine may be brain region dependent. Stereology was not performed on the animals with the 4.0 mg/d dose because there was no 4.0 mg/d carnosine-dependent improvement detected in olfactory behavior.

Figure 7.

(A–E) α-Synuclein immunohistochemistry and stereology in the olfactory bulb mitral cell layer and representative images of aSyn(+) neurons in the OB mitral cell layer (OB-MCL) (100× magnification). (A, B) Wild-type mice treated with vehicle or 2.0 mg/d carnosine. (C, D) Thy1-aSyn mice treated with vehicle or 2.0 mg/d carnosine. (E) Box and whisker plot indicates the 75th percentile (top of box), 25th percentile (bottom of box), and median (line inside of box), and the whiskers indicate the minimum and maximum for the data set. Quantification of aSyn(+) cell bodies in the OB-MCL in each treatment group. Thy1-aSyn mice exhibited more aSyn(+) cell bodies compared to wild-type mice, with no effect of carnosine treatment. Scale bar is 50 μm for (A)–(D). V = aSyn(+) cell body. ***p < 0.001.

Presently, the specific cellular mechanism(s) underlying IN carnosine’s prevention of progressive nigrostriatal motor dysfunction is unclear, but previous studies from this lab suggest that improved mitochondrial function may underlie carnosine’s beneficial effect.²⁹ It is also conceivable that carnosine may elicit enhanced release of nigrostriatal dopamine which could contribute to the improvement in motor behavior, a possibility that warrants further investigation. Future studies may evaluate S129 phospho-aSyn, which is found elevated in the Thy1-aSyn mice and in humans, as well as PP2A, a protein phosphatase that is activated by aSyn and inhibits TH.^41,42 Future studies should also evaluate intranasal carnosine at higher doses and/or with multiple daily doses to test improvement of nonmotor end points, such as olfactory function. Additionally, intranasal carnosine should be evaluated in other models of PD such as the mutant human aSyn A53T transgenic rodent models, which recapitulate a human genetic basis for PD. Assessing carnosine’s neuroprotective potential in models of idiopathic PD is also warranted.

In summary, intranasal carnosine treatment resulted in dose-dependent improvements in motor behavior in mice overexpressing human aSyn using the challenging beam traversal and spontaneous activity behavioral assessments. These results suggest that intranasal carnosine may improve motor symptoms that are commonly observed in PD.² Importantly, we found that intranasal carnosine decreased the number of aSyn(+) cell bodies in the SNpc, a key brain region associated with motor dysfunction in PD patients, without perturbing TH expression. Furthermore, 4.0 mg/d carnosine provided greater improvement of motor function compared to 2.0 mg/d. Whereas these studies did not address mice at a more advanced stage of disease, increasing the daily dosing to 2–3 times per day might be a feasible way to increase the dose of carnosine (administering larger volumes of carnosine solution in a single treatment session would raise animal health/welfare concerns). Due to carnosine’s very low toxicity, its previous oral use in a PD patient clinical trial, and the urgent demand for a treatment that prevents PD progression, intranasal carnosine should be considered for evaluation by clinical trial to evaluate safety and efficacy in individuals with PD.

METHODS

Animals.

Mice were bred in-house under barrier conditions. Line 61 Thy1-aSyn mice were obtained from Dr. Eliezer Masliah (University of California, San Diego). Overexpression of human aSyn in these mice is driven by the mouse Thy1 promoter, with the transgene inserted on the X-chromosome.¹⁹ Thy1-aSyn female mice on a BDF-1 hybrid background were bred with wild-type BDF-1 male mice (Charles River strain 099). Males were used exclusively in these studies because females display a less robust phenotype.¹⁷

At 4 weeks of age, male mice were genotyped (see Supporting Information section S3), weaned, and randomly assigned to treatment groups. Mice were housed 2–4 per cage in lixit polysulfone shoebox cages with corncob bedding, with nestlets and shredded paper as enrichment. Mice were housed under an inverted 12 h light–dark cycle (lights on at 21:00 h and off at 09:00 h) with a temperature range of 22–24 °C. Drinking water and rodent chow [Teklab diet 7922 (NIH-07)] were provided ad libitum except during food restriction. Baseline behavior testing was conducted at 6–8 weeks of age and repeated after 8 weeks of IN carnosine treatment. Treatment began within two weeks of baseline behavior and continued for two months (Figure 1). Animals were monitored for body weight, mortality, and clinical symptoms throughout the study. Experiments were conducted under protocols approved by the University of Cincinnati Institutional Animal Care and Use Committee and in full accordance with the 8th Guide for the Care and Use of Laboratory Animals and consistent with ARRIVE guidelines.⁴³

Treatment.

Carnosine (CASRN 305–84-0, 98% pure) was supplied by Acros Organics (New Jersey, USA). Carnosine was stored at 4 °C, and dosing solutions were prepared daily by dissolving in sterile ultrapure water at concentrations of 0.20 and 0.27 mg/μL for the 2.0 mg/d and 4.0 mg/d groups, respectively. After baseline behavioral assessment, unanesthetized mice were treated daily via intranasal administration of carnosine for 8 weeks. Wild-type and Thy1-aSyn mice from the same litter were included in each treatment when possible. Intranasal dosing consisted of 10 μL (2.0 mg/d dose) or 15 μL (4.0 mg/d dose) of treatment solution per day, administered into both nostrils. Control mice received equivalent volumes of ultrapure sterile water (vehicle).

Neurobehavioral Assessments.

All behavior experiments were performed during the dark-cycle, under low-light conditions, between 15:00 and 21:00 h. Animals were acclimated for at least 2 days to the inverted light-cycle. On the first day of behavioral assessment, challenging beam traversal was performed, immediately followed by spontaneous activity. Following completion of the spontaneous activity test, food restriction was initiated for the buried food pellet test, with the goal of a 10% body weight reduction to motivate the mice. After 4 days of food restriction, mice underwent the buried food pellet evaluation. Daily intranasal treatments started after the last baseline behavioral measurement. After 8 weeks of treatment, the behavioral procedures were repeated. The challenging beam traversal and spontaneous activity experiments were video recorded and were scored by an investigator blinded to genotype and treatment.

Challenging Beam Traversal.

Experiments were performed as previously described.^21,22,34 Prior to baseline behavioral testing, mice were trained for 2 days to traverse a beam that gradually narrowed in width across four sections of the beam (25 cm in length each). The mice begin on the widest section of the beam (3.5 cm), with the end of the narrowest section of the beam (1.0 cm) resting in the home cage. On test day, the beam was fitted with a wire-grid mesh that rested 1 cm above the beam. Latency was recorded as the number of seconds required for the mouse to reach the end of the beam. We used the placement of the paw on the last square mesh as the latency stop point. Errors were defined as the paw falling at least halfway between the wire grid and the beam while the mouse was moving forward. The number of footsteps taken per beam section was counted for the right hindlimb. Latency, total errors, and errors per step were averaged for five trials.

Spontaneous Activity.

Following the challenging beam traversal test, mice were assessed for spontaneous activity in a cylinder (14 cm diameter, 12 cm height), as previously described.^21,22,34,44 Mice were placed in a clean, transparent plexiglass cylinder positioned on top of a glass sheet for 3 min and recorded by a camera from below. Mice were scored live for cumulative rears and time spent grooming. Mice were scored for forelimb and hindlimb steps using video recordings.

Olfactory Function.

The buried food pellet test was used to assess olfactory function. Following the spontaneous activity task, mice were weighed, and food was restricted (2 g of food/day/mouse) for 4 days such that the mice maintained 90% of their initial body weight. On the test day, mice were individually placed in a cage containing clean corncob bedding (∼5 cm deep). Mice were acclimated for 45 min, and then the mice were briefly removed for placement of a 1.5 g food pellet. The pellet was buried just under the surface of the bedding or placed on the surface of the bedding. Mice were allowed a maximum time of 300 s to retrieve the pellet. The time to first dig at the buried food pellet location and the time to find the food pellet were recorded. The location of the pellet and mouse was changed each time the test was performed (modified from ref 18).

Tissue Collection.

Necropsies were performed between 14:00 and 16:00. Animals were euthanized using pentobarbital overdose and transcardially perfused with 10 mL of sterile saline containing 10 U/mL heparin followed by 20–30 mL of 4% paraformaldehyde (PFA) in sterile saline. Brains were removed from the skull and postfixed in 4% PFA for 1 day, then equilibrated in sterile 30% sucrose in saline and stored at 4 °C until sectioning.

Immunohistochemistry.

Preserved brains were blocked at approximately Bregma levels 1.0 mm and −2.0 mm, frozen in Tissue-Tek optimal cutting temperature (OCT) compound on dry ice, and stored at −80 °C. The SN and OB were sectioned coronally at 50 μm thickness using a Leica freezing microtome (stage temperature −30 °C). The SN was cut in 1/6 series (6–7 sections/series),⁴⁵ and the OB was cut in 1/8 series. The OB was sectioned from the rostral start of the olfactory limb of the lateral ventricle⁴⁶ until the end of the OB, yielding 9–10 hemibrain sections per series. Sections were placed in cryoprotectant solution (300 mL of ethylene glycol, 300 g of sucrose, 500 mL of phosphate buffer [PB; 0.1 M, pH = 7.4], ultrapure water to 1 L) and stored at −20 °C until processed for immunohistochemistry.

Anti-aSyn antibody (1:4000 dilution, mouse anti-aSyn clone 42, BD Biosciences, catalog numbers 610786 and 610787) was selected because of previously determined high specificity and selectivity for human and mouse aSyn reactivity. This antibody was found to lack immunoreactivity in aSyn knockout mice⁴⁷ and detected post-translationally modified human and mouse aSyn.^20,48 The anti-TH antibody (1:8000 dilution, Millipore, mouse anti-TH monoclonal antibody, catalog number MAB318) was chosen because it is highly specific for TH in the rodent SNpc⁴² and was previously used to detect TH in the Thy1-aSyn mice.⁴⁹

Mouse on mouse (MOM) Elite Peroxidase kit (Vector, PK-2200, containing blocking reagents and secondary antibody) reagents were diluted according to the manufacturer’s specifications using PB, except for the MOM diluent, which was made with 150 μL per 2.5 mL of PB or PB with 0.2% Triton X-100 for the primary antibody incubations. Sections were developed in 3,3′-diaminobenzidine (DAB) (Vector, SK-4100) to chromatically visualize the immunoreactivity. Sterile PB was used to rinse between incubation steps before incubation with avidin/biotin (ABC; Vector, PK-2200) reagent, and sterile Tris buffer (0.1 M, pH = 7.4) was used after incubation with the ABC reagent. Between incubations, sections were washed three times, 5 min each in either PB or Tris.

Free-floating brain sections were removed from the cryoprotectant solution and thoroughly rinsed with PB. Endogenous peroxidase was blocked with 3% hydrogen peroxide in PB for 60 min before and after primary antibody incubation. Sections were blocked with the MOM mouse anti-immunoglobulin blocking reagent for 3.5 h. Sections were incubated with anti-aSyn antibody at room temperature or anti-TH antibody at 4 °C for 55–65 h. Sections were incubated with MOM biotinylated anti-mouse IgG secondary antibody for 40 min, incubated in ABC reagent for 30 min, and then developed with DAB. After thorough rinsing, sections were mounted on slides from Tris buffer and dried for 4–6 h. Then sections were dehydrated with 70% and 95% ethanol, cleared with xylenes, and coverslipped with Permount mounting medium (Fisher Chemical). Sections immunostained for aSyn were additionally counterstained with hematoxylin.

Stereology.

The total number of aSyn(+) and TH(+) cells was quantified in the SNpc. Mitral cells in the OB mitral cell layer (OB-MCL) were only counted for aSyn immunoreactivity because OB-MCL does not express TH. Wild-type and Thy1-aSyn mice were included for all assessments, and the investigator was blinded to treatment and genotype. Design-based stereological procedures were adapted using the optical dissection method,⁵⁰ and counting was performed with MBF Biosciences StereoInvestigator software version 10⁴⁴ and version 2019. Both hemispheres of the SN and one hemisphere of the OB were used for stereology.

Contours of the SNpc and OB were traced at 2× and 4× magnification, respectively, and were counted at 60× magnification. Analysis was performed in 1/6 series for the SNpc (6–7 sections/animal) and 1/8 series for the OB-MCL (9–10 hemibrain sections/animal). The coefficient of error (CE) was ≤0.10 for all stereological estimations.

α-Synuclein Cell Counts.

Six to ten or five to six animals per genotype-treatment group were included for the SNpc or OB-MCL, respectively. Sections were counted with a final thickness of 20 μm after processing and mounting. The probe height was 16 μm with 2 μm guard zones on the top and bottom of each section. The SNpc was counted using a 135 μm × 135 μm grid and 60 μm × 60 μm optical dissection probe. Neurons were counted as aSyn(+) only if a hematoxylin(+) nucleus and axon hillock were present, and immunoreactivity was present throughout the entire cell body. The axon hillock was used as a cytoarchitectural feature specific to neurons (Supporting Information section S2). Each section was counted in its entirety for mitral cells in the OB-MCL.

Tyrosine Hydroxylase Cell Counts.

Six to nine animals per genotype-treatment group were included for the SNpc. The SNpc sections were counted with a final thickness of 27 μm after processing and mounting. The probe height was 23 μm with 2 μm guard zones on the top and bottom of each section. The SNpc was counted using a 135 μm × 135 μm grid and 60 μm × 60 μm optical dissection probe.⁵¹ Cells were counted as TH(+) only if a nucleus was visible and immunoreactivity was observed throughout the entire cell body with an immunonegative nucleus.

Statistical Analysis.

Absolute body weight and body weight gain were assessed using a two-way analysis of variance (two-way ANOVA), and mortality was assessed with the Z-score test for two population proportions. Statistical significance was set with α = 0.05. Outliers were identified and removed from analysis if values were above or below the mean ± (2× standard deviation). Results are presented as the mean ± standard error of the mean (SEM).

Multiple comparison procedures were run for pairwise comparisons using the Holm–Šídák and Tukey HSD test when the ANOVA omnibus p was ≤0.05. Statistical tests were performed using R statistical analysis software (x64 version 3.3.1 or version 3.6.0). For challenging beam traversal, behavioral effects of treatment, genotype, and time point (baseline or end point), as well as their interactions, were assessed using a three-way repeated measures mixed-model ANOVA (three-way RM mixed ANOVA) for the omnibus F-statistic. Additionally, the dose–response for challenging beam traversal was assessed for each genotype by fitting the mean values of each carnosine treatment level to a linear model and comparing the fit using R².

Power analysis was conducted for the challenging beam traversal experiment using SigmaPlot 13.0. For challenging beam traversal, n = 15 per genotype treatment group is sufficient to achieve a power = 0.8. Three-way repeated measures ANOVA (three-way RM ANOVA) was used for spontaneous activity time spent grooming. Two-way mixed-model ANOVA was used for end point spontaneous activity forelimb steps. For all other behavioral tests, Student’s t test or two-way ANOVA was used. Stereology was analyzed using two-way mixed ANOVA for effects of treatment and genotype and their interactions. For aSyn(+) neurons in the SNpc, n = 6–7 is sufficient for power = 0.9.

ABBREVIATIONS

ANOVA

analysis of variance

aSyn

α-synuclein protein

aSyn(+)

α-synuclein-positive

A53T

α-synuclein protein containing the mutation A53T

CE

coefficient of error

DAB

3,3′-diaminobenzidine

IN

intranasal

MOM

mouse-on-mouse

n

sample size

OB

olfactory bulb

OB-MCL

olfactory bulb mitral cell layer

OE

olfactory epithelium

OCT

optimal cutting temperature

PB

phosphate buffer

PD

Parkinson’s disease

PFA

paraformaldehyde

PP2A

protein phosphatase 2A

RM

repeated measures

ROS

reactive oxygen species

SN

substantia nigra

SNpc

substantia nigra pars compacta

SNpr

substantia nigra pars reticulata

S129

α-synuclein phosphorylated at residue S129

TH

tyrosine hydroxylase

TH(+)

tyrosine hydroxylase-positive

Thy1

thymocyte antigen 1

Thy1-aSyn

Line 61 Thy1-aSyn transgenic mouse model

Tris

Tris-HCl