Department of Neurology, Pittsburgh Institute for Neurodegenerative Diseases, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, USA
Department of Biomedical Sciences, Texas Tech University Health Sciences Center at El Paso, Paul L. Foster School of Medicine, El Paso, Texas, USA
Address correspondence and reprint requests to Ruth G. Perez, Graduate School of Biomedical Sciences, Center of Excellence in Neurosciences, Paul L. Foster School of Medicine, 5001 El Paso Drive, MSB 1, Suite 4002, El Paso, Texas 79905, USA. E-mail: email@example.com
Aging, the main risk factor for Parkinson's disease (PD), is associated with increased α–synuclein levels in substantia nigra pars compacta (SNc). Excess α-synuclein spurs Lewy-like pathology and dysregulates the activity of protein phosphatase 2A (PP2A). PP2A dephosphorylates many neuroproteins, including the catecholamine rate-limiting enzyme, tyrosine hydroxylase (TH). A loss of nigral dopaminergic neurons induces PD movement problems, but before those abnormalities occur, behaviors such as olfactory loss, anxiety, and constipation often manifest. Identifying mouse models with early PD behavioral changes could provide a model in which to test emerging therapeutic compounds. To this end, we evaluated mice expressing A53T mutant human (A53T) α–synuclein for behavior and α–synuclein pathology in olfactory bulb, adrenal gland, and gut. Aging A53T mice exhibited olfactory loss and anxiety that paralleled olfactory and adrenal α-synuclein aggregation. PP2A activity was also diminished in olfactory and adrenal tissues harboring insoluble α-synuclein. Low adrenal PP2A activity co-occurred with TH hyperactivity, making this the first study to link adrenal synucleinopathy to anxiety and catecholamine dysregulation. Aggregated A53T α–synuclein recombinant protein also had impaired stimulatory effects on soluble recombinant PP2A. Collectively, the data identify an excellent model in which to screen compounds for their ability to block the spread of α-synuclein pathology associated with pre-motor stages of PD.
Aging A53T α-synuclein mice develop behaviors common to premotor Parkinson's disease (PD) with synucleinopathy occuring in adrenal gland, olfactory bulb, and intestine. Insoluble aggregated A53T α-synuclein loses PP2A-stimulatory-effects, which normally down regulate dopamine synthesis. This is the first study to link adrenal synucleinopathy to catecholamine dysregulation and anxiety. The mice are an excellent model for testing emerging neuroprotective therapies for PD.
α-Synuclein (a-Syn) is a widely expressed pre-synaptic protein (Maroteaux et al. 1988). a-Syn gene mutations and multiplications cause early onset Parkinson's disease (PD) (Polymeropoulos et al. 1997; Kruger et al. 1998; Singleton et al. 2003; Chartier-Harlin et al. 2004; Zarranz et al. 2004), however, most PD is sporadic and linked to aging. Increases in a-Syn protein are noted in aging human Substantia nigra pars compacta (SNc) (Li et al. 2004; Chu and Kordower 2007), which can stimulate neuroinflammation and microglial activation (Croisier et al. 2005), leading to Lewy body formation. Families expressing A53T mutant a-Syn (A53T) develop early onset PD and Lewy bodies with highly phosphorylated a-Syn (Anderson et al. 2006). Cumulatively, these data suggest that having too much a-Syn protein in neurons is problematic.
PD motor symptoms emerge after extensive loss of SNc dopaminergic neurons (Bernheimer et al. 1973). Yet, non-motor symptoms precede motor onset by years, during a pre-motor phase of PD. Constipation is associated with low gut motility (Ashraf et al. 1997), dopaminergic defects (Singaram et al. 1995), and a-Syn accumulation in colonic neurons (Shannon et al. 2012). An impaired sense of smell affects many PD patients (Bohnen et al. 2007) and anosmia and hyposmia are common initial symptoms of pre-motor PD (Haehner et al. 2011). Olfactory impairment correlates with olfactory bulb (OB) Lewy body pathology (Beach et al. 2009), which occurs early in the course of PD (Braak et al. 2004). These findings suggest that gastrointestinal or OB assessment for biomarkers, coupled with behavioral tests could identify PD at a time when SNc remains intact (Doty et al. 1995; Savica et al. 2009).
Anxiety and depression (Dooneief et al. 1992) can also precede PD motor symptoms (Lauterbach and Duvoisin 1991; Shiba et al. 2000; Weisskopf et al. 2003), suggesting a neurochemical basis. Indeed, anxiety is associated with elevated adrenal catecholamines (Kvetnansky and Mikulaj 1970), as well as hyperactivity of tyrosine hydroxylase (TH; EC 188.8.131.52) in adrenal gland (Chobotska et al. 1998). Measuring behavior in combination with sensitive bioassays (Bidinosti et al. 2012) may help identify pre-motor PD cases.
Although a-Syn is implicated in PD neuropathology, it also contributes to normal physiology by interacting with key regulatory proteins in a chaperone-like manner (Perez and Hastings 2004; Sidhu et al. 2004; Geng et al. 2011). The catalytic subunit of protein phosphatase 2A (PP2A; EC 184.108.40.206) interacts with and is stimulated by soluble a-Syn in vitro and in vivo (Peng et al. 2005; Lou et al. 2010). Another enzyme that a-Syn modulates is TH which is inhibited (Perez et al. 2002; Peng et al. 2005; Lou et al. 2010). Too much or too little soluble a-Syn (Lou et al. 2010) or loss of soluble a-Syn by its aggregation, contributes to dysregulated TH and PP2A activity in brain (Alerte et al. 2008; Wu et al. 2012). However, whether a-Syn aggregation might affect TH or PP2A in the PNS is unknown.
In the current studies we assessed movement, olfaction, anxiety, gut pathology, and synucleinopathy in aging A53T homozygous mice and their non-transgenic (Non-Tg) littermates. We also measured phosphorylation of a-Syn serine 129 (PSer129), a modification abundant in Lewy bodies/Lewy neurites (Fujiwara et al. 2002); PP2A tyrosine 307 phosphorylation (PTyr307), a marker of low PP2A activity (Chen et al. 1992); and TH serine 40 (PSer40) phosphorylation, a marker of high TH activity by immunoblot and immunohistochemistry. In addition, we measured PP2A and TH activity in olfactory and adrenal homogenates to compare to behavioral data. Collectively, the results identify an excellent model for screening therapeutic compounds for efficacy for PD.
A53T a-Syn (B6; C3-Tg-Prnp/SNCA*A53T/83Vle/J) heterozygous breeders (Jackson Laboratories, Bar Harbor, ME, USA), produced our cohort of A53T homozygous (n =7) and Non-Tg littermates (n = 5). Genotyped mice (SeqWright-DNA-Technology, Houston, TX, USA) were housed in temperature and humidity-controlled rooms on 12-h light–dark cycles, with food and water ad libitum, except during olfactory testing (see below). Ethical treatment of animals followed AALAC, DLAR, ARRIVE, and NIH Animal Care Guidelines on IACUC approved protocols at the University of Pittsburgh.
All tests were performed in clean quiet test rooms at the University of Pittsburgh BST Rodent Behavioral Core.
Open field tests (12–14 month old mice) utilized TruScan™ activity monitors (Coulbourn Instruments, Whitehall, PA, USA). Mice acclimated 5–15 min in test rooms prior to placement in the center, then floor plane velocity (cm/s) and distance (cm) were monitored for 15 min using established methods (Roy et al. 2006; Hunt et al. 2013). Anxiety Test 1 - The open field arena is 26.67 cm2, with the margin being the area within 3.81 cm from the wall, and the center > 3.81 cm. Anxious mice have longer margin times (Crawley 1985), as demonstrated by margin/center time ratios (Pinna et al. 1997). Mice were tested in random order on three independent occasions.
Food pre-training for olfactory testing
Mice (12–14 month of age) were trained individually to eat the food pellets (~ 250 mg Cap'n Crunch Cereal; Quaker Oats Co., Chicago, IL, USA) placed on top of the bedding in the home cage. Training continued until all mice ate pellets in < 30 s. Anxiety Test 2 - Mice receiving familiar palatable food in a novel environment explore it extensively before eating, a behavior termed hyponeophagia (Merali et al. 2003). Mice, placed individually into clean cages for 5 min, had familiar palatable food placed on top of bedding while latency to approach, sniff, and begin eating the pellet was recorded in seconds.
Once mice ate food in a novel environment in < 30 s, buried pellet testing commenced (Nathan et al. 2004; Fleming et al. 2008; Lu et al. 2008). Testing began when mice were 13–15 months old (Session 1). Mice were retested at 14–16 months of age (Session 2), and again at 15–17 months (Session 3). To ensure motivation, food was removed 14–18 h prior to testing (Fleming et al. 2008), but weight loss never exceeded 10%. Room lighting was identical for each test, and mice were placed individually in clean holding cages for 5 min, then transferred individually to test cages for 2 min acclimation, then returned to the holding cage while a pellet was buried ~ 0.5 cm below the bedding in a random location to eliminate a learning component. After food placement, each mouse was placed in the center of the test cage and given 5 min to find the pellet while latency to sniff, dig up, and begin eating food was recorded using LimeLight™ (Coulbourn Instruments, Whitehall, PA, USA). Test cages were cleaned and fresh bedding applied before additional mice were tested.
Mice were killed by CO2 inhalation and decapitation. Brains - Brains with OBs were rapidly extracted and bisected longitudinally, freezing half for biochemistry and post-fixing half for immunohistochemistry. Adrenal - Adrenal glands were collected as previously described (Kolski-Andreaco et al. 2007), flash frozen, and stored at −80°C until analyzed. Tissues were homogenized in buffer that allowed measuring TH and PP2A activity from each sample. This buffer contained 52.5 mM HEPES, 1.05 mM AEBSF, 5.25 μg/mL aprotinin, 2.25 μg/mL leupeptin, 1.05 mM benzamidine, 10.5% glycerol. Adrenals were homogenized using Bullet Blender® (Next Advance Inc., Averill Park, NY) at speed 9, 4 min, 4°C. After homogenization, supernatants were split for TH and PP2A assays. Pellets were re-extracted in 50 mM Tris, 150 mM NaCl, 0.1% Triton-X 100, 0.2% sodium dodecyl sulfate for immunoblots. Gut - A dissecting microscope was used to prepare small intestine longitudinal muscle with myenteric plexus whole mounts as previously described (Drolet et al. 2009). Segments (5 cm) were isolated, bisected longitudinally and pinned flat in sylgard-coated dishes with mucosa facing down. Similar regions from each mouse were post-fixed overnight in 4% formaldehyde/sucrose.
Unperfused brains were sectioned sagittally (30 μm) and handled as previously described (Alerte et al. 2008; Lou et al. 2010).
TH (Aves Labs; MAB318, Millipore, Billerica, MA, USA), phospho-serine 19 on TH (AB5425, Chemicon), a-Syn (AB5334P, Chemicon; sc-7011-R, Santa Cruz, Santa Cruz, CA, USA), PP2A (sc-6110, Santa Cruz), PTyr307 PP2A (sc-12615, Santa Cruz). a-Syn aggregation was assessed with 1% Thioflavin-S (Thio-S, T1829, Sigma-Aldrich) on sections first immunolabeled for a-Syn and TH. As Thio-S signal is green, secondary antibodies for double labeling a-Syn and TH were Cy-5 and Cy-3, respectively. a-Syn Cy-5 signal was pseudocolored red on Thio-S stained sections to permit demonstrating a yellow co-localization signal.
Longitudinal muscle with myenteric plexus tissues were immunolabeled as previously described (Drolet et al. 2009). Quantification of Gut a-Syn - Optical densities of 10 fields/sample/mouse were generated using ImageQuant, (GE Healthcare, Waukesha, WI, USA). Quantitative analyses used baseline correction to normalize data by measuring the lightest-stained 15 nm2 region of each tissue, set as background. That value was subtracted from the signal for all other regions.
Proteins (10–20 μg) were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis, transferred to nitrocellulose, and incubated in primary antibody overnight, 4°C. a-Syn antibodies were sc-7011-R (Santa Cruz); Syn-1 (610787, BD Biosciences, San Jose, CA, USA), and a-Syn PSer129 (11A5, from John Anderson, Elan Pharmaceuticals). PP2Ac antibodies were 1D6 (05-421, Millipore/Upstate, EMD Millipore, Billerica, MA, USA) and sc-6110 (Santa Cruz), with β-actin (A5441, Sigma-Aldrich) used as a loading control. Blots were imaged using LiCor Odyssey (Lincoln, NE, USA).
Tissues were homogenized in imidazole buffer [20 mM imidazole-HCl, 2 mM EDTA, 2 mM EGTA plus protease inhibitors], 4°C, centrifuged to remove particulates followed by free phosphate removal with MicrospinTM-G-25 columns (GE Healthcare). Supernatants were incubated in 4-Nitrophenyl-phosphate (pNPP) buffer [50 mM Tris-HCl, pH 7.0, 0.1 mM CaCl2] with KRpTIRR substrate, 10 min, 30°C. Samples were assayed at 650 nm relative to fresh standards by malachite green assay (17-127; Millipore/Upstate). Inhibitors for specificity included: 50 nM protein-phosphatase-inhibitor 2 (PP1-specific), 3 nM fostriecin (PP2A/PP4-specific), 50 nM cantharidin (PP4-specific), and 10 nM okadaic acid as previously described (Peng et al. 2005; Lou et al. 2010; Wu et al. 2012).
Tissues were homogenized in simple buffer as described above with samples supplemented to [0.1 mM CaCl2].
a-Syn (gift of Dr. John Rosenberg, University of Pittsburgh) was incubated with PP2A (Cayman Chemicals, Ann Arbor, MI, USA; or Millipore), 4°C for 30 min in pNPP buffer then with substrate, 30°C, 10 min. Duplicate Samples were assayed in three independent experiments.
Tyrosine hydroxylase assay
TH activity utilized a colorimetric assay (Daubner and Fitzpatrick 1993). Adrenal tissue prepared in simple buffer was supplemented to [25 mM EDTA]. Samples were evaluated as previously described (Wang et al. 2009; Lou et al. 2010) spectrophotometrically at 490 nm (UV-1800, Shimadzu, Columbia, MD).
Recombinant A53T a-Syn aggregation
Protein was aggregated using established protocols (Giasson et al. 1999) which produces insoluble a-Syn pellets (P) and soluble a-Syn supernatants (S) exactly as previously described (Wu et al. 2012).
Data were transformed to log scale to induce normality (for distance, velocity, and latency to find food). Group effects (Non-Tg and A53T), time effects, and group x time interaction effects were evaluated by parametric repeated measures analysis of variance (RM-anova) with post hoc Bonferroni analysis. Data were also evaluated by non-parametric Friedman tests. Both RM-anova and Friedman tests utilized SPSS (v20, IBM, New York, NY), t-test and one-way anova utilized InStat (Graph Pad, San Diego, CA, USA). Data represent the mean ± SEM, with significance set to p < 0.05.
Aging A53T mice exhibit hyperactivity and anxiety
Homozygous A53T mice are a good model of synucleinopathy, but do not develop SNc-associated motor complications (Giasson et al. 2002; Lee et al. 2002; Tanji et al. 2010; Tsika et al. 2010). As movement is important for olfactory testing of mice, we confirmed their ability to move in an open field. The average distance moved in 15 min was 2323.9 ± 158.1 cm for A53T and 2016.5 ± 209.2 cm for wild type Non-Tg mice; demonstrating that A53T mice moved quite well. Data were log transformed to induce normality, adjusted for time × interaction effects and assessed by repeated measures analysis of variance (RM-anova). This revealed that A53T mice moved greater distances than Non-Tg littermates (A53T, 6.29 ± 1.53; Non-Tg, 5.74 ± 2.03; p = 0.009; RM-anova). The movement velocity of A53T mice was 7.6 ± 0.53 cm/s and for Non-Tg mice 6.6 ± 0.69 cm/s; when evaluated by RM-anova of log transformed data revealed significant differences (A53T, 0.55 ± 1.53; Non-Tg, 0.22 ± 1.79; p = 0.031). Thus, A53T mice were hyperactive. Hyperactivity has been reported for another strain of A53T a-Syn mice (Unger et al. 2006), raising the possibility that A53T mutant a-Syn expression may cause anxiety.
To assess this we used established methods in which the amount of time mice spent in the margin versus center of an open field during 15 min trials was recorded. A53T a-Syn mice spent more time in the margin than in the center (Fig. 1a). Anxiety was also assessed by comparing the latency to approach and begin eating a familiar palatable food, placed on top of the bedding in a novel environment, a measure of anxiety called hyponeophagia. Mice quickly ate the familiar palatable in their home cages, but A53T mice were significantly slower to approach and taste the food in a new cage, further demonstrating anxiety (Fig. 1b, A53T, 137.1 ± 23.4 s; Non-Tg, 55.2 ± 14.8 sec; p = 0.03, t-test).
Anxiety is associated with catecholamine synthesis (Leduc 1961; Kvetnansky and Mikulaj 1970; Anisman and Zacharko 1986), and stress increases adrenal catecholamine levels (LeBlanc and Ducharme 2007). Furthermore, Lewy bodies occur in PD adrenal medulla (Fumimura et al. 2007; Wakabayashi et al. 2010). Therefore, we next assessed adrenal a–Syn solubility, and TH and PP2A activity as described below.
Aging A53T adrenal glands have synucleinopathy and dysregulated enzymatic activity
Soluble a-Syn normally regulates TH and PP2A activity in opposite directions (Perez et al. 2002; Peng et al. 2005), while loss of soluble a-Syn by its aggregation, impairs PP2A activity in human brain (Wu et al. 2012). However, no one has evaluated the impact of a-Syn aggregation in adrenal gland in any model. We measured a-Syn levels in soluble adrenal extracts and found low amounts of monomeric a-Syn in Non-Tg mice, and as expected, higher levels of a-Syn monomers in A53T adrenal tissue (data not shown). When we re-extracted pellets, we noted high and low molecular weight insoluble a-Syn species in A53T adrenal but little signal in Non-Tg adrenal homogenates (Fig. 2a). a-Syn oligomers quantified by densitometry, were significantly elevated in A53T mice compared to Non-Tg littermates, demonstrating synucleinopathy in aging A53T mouse adrenal gland (Fig. 2b; p < 0.001, t-test).
Phosphorylated TH is the active form of the protein (Kumer and Vrana 1996) and adrenal catecholamine levels are significantly elevated in anxious animals (Kvetnansky and Mikulaj 1970). We therefore measured adrenal TH activity first indirectly by quantifying serine 40 phosphorylation (PSer40) levels by immunoblot (Bobrovskaya et al. 2007), and then directly using established methods (Lou et al. 2010). Aging A53T adrenal gland had 27% higher TH PSer40 levels compared to Non-Tg adrenal (Fig. 2c, p < 0.001, t-test), a finding that paralleled significantly higher TH activity in A53T adrenal homogenates (Fig. 2d; p < 0.01, t-test).
Dephosphorylation of TH PSer40 is mediated by PP2A (Haavik et al. 1989; Kumer and Vrana 1996), therefore, we also measured PP2A activity, first indirectly by quantifying PP2A PTyr307, a marker of less active PP2A by immunoblot; and then directly by PP2A assay (Peng et al. 2005; Lou et al. 2010). A53T adrenal PP2A had 24% more PTyr307 than Non-Tg adrenal (Fig. 2e, p < 0.001, t-test). This paralleled significantly less PP2A activity in A53T compared to Non-Tg adrenal (Fig. 2f; p < 0.01, t-test), further suggesting that a-Syn-mediated-stimulation of PP2A activity becomes impaired when a-Syn aggregates. This is the first demonstration that adrenal a-Syn pathology dysregulates TH and PP2A in a manner to induce anxiety in aging A53T mice.
Constipation is a common complaint occurring long before motor onset in PD. Animals with gastrointestinal slowing consistent with constipation, develop a-Syn aggregates in enteric neurons that innervate the gut (Drolet et al. 2009). Although gut motility was not evaluated in our mice, we used established methods to assess a-Syn aggregation in their small intestines. Proteinase K (PK) treatment removes soluble, while leaving insoluble proteins. When PK treated gut tissues were evaluated immunohistochemically, only low levels of PK resistant a-Syn were present in Non-Tg myenteric neurons (Fig. 3a, at arrows), in stark contrast to high level PK resistant a-Syn in A53T myenteric neurons (Fig. 3b). Quantitative analyses, with baseline correction, revealed significantly more PK resistant a-Syn in aging A53T mouse myenteric neurons compared to Non-Tg mice (Fig. 3c; A53T, 2.15 ± 0.15 M; Non-Tg, 0.49 ± 0.14 M; p < 0.001, t-test), nearly identical to data from rotenone treated rats with low gut motility (Drolet et al. 2009).
Aging A53T mice develop progressive olfactory loss
Olfaction was assessed using a buried pellet test (Fleming et al. 2008), performed during three sessions over 8 weeks. Results, evaluated by repeated measures analyses on log transformed data (Fig. 4), confirmed interaction effects between groups and sessions. The group effect was examined at each session using unpaired t-tests as well as by Wilcoxon rank sum tests after Bonferroni correction. In Session 1, when mice were 13–15 months old, both groups found the buried pellet with similar latencies (A53T, 4.45 ± 0.65; Non-Tg, 4.01 ± 0.78; p = 0.412, Wilcoxon rank sum test). By Session 2, when mice were 14–16 months old, A53T a-Syn mice were significantly slower at finding food (A53T, 5.10 ± 0.44; Non-Tg, 3.74 ± 0.74; p = 0.024, Wilcoxon rank sum test). In Session 3 when mice were 15–17 months old, olfaction remained significantly impaired in A53T mice (A53T, 5.17 ± 0.42, Non-Tg, 4.08 ± 0.63; p = 0.012, Wilcoxon rank sum test). Prior studies, using mice that express wild-type human a-Syn, showed OB a-Syn aggregates with a parallel loss of olfaction (Fleming et al. 2008), but the impact of synucleinopathy on OB PP2A has never been assessed.
Aging A53T olfactory bulb with a-Syn aggregation has impaired PP2A activity
Highly phosphorylated a-Syn Ser129 is present in Lewy bodies (Saito et al. 2003; Anderson et al. 2006), and Lewy bodies are common in OB of individuals with PD (Braak et al. 2004). To determine if our aging A53T mice that developed progressive olfactory loss, had Lewy-like pathology in OB, we used Thio-S staining and a-Syn immunohistochemistry. Non-specific Thio-S signal was noted in blood vessels of our unperfused mice, similar to what was reported by others using unperfused animals (Sun et al. 2002). The non-specific labeling appeared as linear groupings of less intense Thio-S signal in both A53T and Non-Tg OB and likely represents staining of red blood cells. In contrast, large irregularly shaped Thio-S stained aggregates were found in A53T OB (Fig. 5a, bottom left panel), while Non-Tg Thio-S signal appeared at background levels (Fig. 5a, top left panel). In A53T OB, Thio-S and a-Syn signals overlapped (Fig. 5a, bottom middle panel), as can be appreciated in the yellow merged image (Fig. 5a, bottom right) confirming Lewy-like a-Syn aggregation in aging A53T OB. Non-Tg OB had uniform a-Syn signal (Fig. 5a, top middle panel) that did not overlap Thio-S (Fig. 5a, top right panel). Lewy-like phosphorylated Ser129 (PSer129) a-Syn immunoreactivity was also strong in a-Syn stained profiles in A53T OB (Fig. 5b, bottom panels). When quantified in FluoView (Olympus America Inc., Center Valley, PA, USA), PSer129 signal intensity was significantly greater in A53T compared to Non-Tg littermate OB (p < 0.0001, t-test). Similar total PP2A immunoreactivity was apparent in A53T a-Syn and Non-Tg OB (Fig. 5c, left panels), yet PP2A phosphorylated on Tyr307 which labels less active PP2A, was significantly greater in A53T OB (Fig. 5c, middle panels), as confirmed by quantitation (Fig. 5c right panels; p < 0.001; t-test). Despite equal amounts of PP2A catalytic subunit on immunoblots (Fig. 5d), PP2A activity was significantly lower in A53T OB (Fig. 5e), further demonstrating that aggregated A53T a-Syn is a poor stimulator of PP2A activity. To directly assess this, we next compared PP2A activation in response to soluble and insoluble recombinant A53T a-Syn.
Aggregated A53T a-Syn has less PP2A stimulating capacity
Soluble a-Syn interacts with PP2A catalytic subunit and stimulates its activity in vitro and in vivo (Peng et al. 2005; Lou et al. 2010). In contrast, aggregated a-Syn is less able to stimulate PP2A activity in human brain and in vitro (Wu et al. 2012). Here, we compared soluble and aggregated mutant human A53T recombinant a-Syn for its ability to stimulate the activity of soluble recombinant PP2A. A 19 kDa a-Syn band on Coomassie stained gels, was insoluble when present in the pellet (Fig. 6a), a fraction that contained 68% ± 3% of total A53T a-Syn protein after the aggregation step. When we assayed PP2A activity, recombinant PP2A efficiently cleaved PO4 from the substrate at baseline, and soluble A53T a-Syn strongly stimulated PP2A. In contrast, aggregated A53T a-Syn was much less able to stimulate PP2A activity (Fig. 6b), further demonstrating that a-Syn aggregation causes a loss of function of A53T a-Syn for PP2A stimulation, as also noted for wild type a-Syn (Wu et al. 2012).
Aging is the major risk factor for PD (Collier et al. 2011) and is highly correlated with increasing a-Syn levels in human brain (Li et al. 2004; Chu and Kordower 2007). Such age-dependent increases in a-Syn likely contribute to Lewy body formation and eventually to PD motor symptoms. Before motor onset, however, constipation, olfactory impairment, and anxiety are behaviors common to the pre-motor, prodromal phase of PD (Shiba et al. 2000; Weisskopf et al. 2003; Pellicano et al. 2007). Anxiety also manifests after PD onset, however, that may occur primarily in response to fluctuations associated with levodopa therapy (Maricle et al. 1995). In this study, our data reveal a strong association between accumulation of insoluble a-Syn and pre-motor behavioral markers of PD.
Regarding mice, the original A53T M83 mice developed a motor phenotype by ~ 8 month and had no a-Syn pathology in OB (Giasson et al. 2002). Their rapid demise may have prevented OB a–Syn accumulation. A53T M83 mice that are now commercially available remain healthy as they age (Oaks et al. 2013). Using this same model, we evaluated homozygous A53T a-Syn mice for synucleinopathy and pre-motor PD behaviors with aging. Like most a-Syn transgenic mice, these mice do not model nigrostriatal damage (Blesa et al. 2012) and as expected, their movement was also unimpaired (Fig. 1). However, they developed age onset abnormalities in enzymatic activity that could underlie early behavioral and physiological changes of PD. These mice could thus be useful for pre-clinical evaluation of novel therapeutics for PD.
This is important because levodopa, the current major therapy for PD, does nothing to slow progression and may even contribute to pathogenesis (Simuni and Stern 1999). With society aging worldwide, there is an urgent need to find therapies that will halt PD progression. One strategy for accomplishing this is to elucidate the normal function(s) of proteins implicated in PD, especially those that affect dopaminergic/catecholaminergic physiology. For example, a-Syn contributes to dopamine handling at many levels. We discovered that soluble a-Syn controls dopamine synthesis by binding to and regulating dopamine regulatory proteins including the rate limiting enzyme, TH, the next biosynthetic enzyme, aromatic amino acid decarboxylase, and the catalytic subunit of PP2A. PP2A is the main phosphatase that dephosphorylates and inhibits TH and aromatic amino acid decarboxylase (Perez et al. 2002; Peng et al. 2005; Tehranian et al. 2006; Lou et al. 2010). Soluble a-Syn differentially regulates PP2A (activates) and TH (inhibits) under normal conditions. Furthermore, soluble a-Syn also controls dopamine release and reuptake (Sidhu et al. 2004; Sulzer 2010). This implies that therapies that sustain soluble a-Syn at low levels will protect against dopamine mishandling and its toxic consequences (Caudle et al. 2007).
In humans, the a-Syn A53T mutation induces Lewy body/Lewy neurite pathology and early onset PD (Polymeropoulos et al. 1997). In our current study, we assessed the impact of the A53T a-Syn mutation on pre-motor PD-like symptoms in mice. Evaluating A53T a-Syn mice over time allowed us to correlate progressive synucleinopathy with behavioral markers of early PD. The A53T mice had good mobility but displayed anxiety (Fig. 1). In exploring the source of anxiety we found that a-Syn aggregation contributes to abnormal TH and PP2A activity in the adrenal gland (Fig. 2), which may provide novel biomarkers for early PD detection. We further established that aging A53T mice develop gut pathology associated with a-Syn over-expression (Fig. 3), as well as olfactory loss (Fig. 4) and Lewy-body-like pathology in OB (Fig. 5a, b). OB synucleinopathy was paralleled by impaired PP2A activity (Fig. 5e). These data suggest that with aggregation, insoluble a–Syn loses normal functionality toward both PP2A and TH, producing effects exactly opposite to the effects of soluble a–Syn. This was confirmed in vitro using recombinant A53T a-Syn and PP2A (Fig. 6).
The etiology of anxiety in PD is complex and data from mouse models vary, which may be because of differences in the ages of the mice evaluated and the levels of a-Syn expression. For instance, animals expressing only mouse a-Syn or entirely lacking a-Syn have no anxiety (Peña-Oliver et al. 2010). Homozygous A53T a-Syn mice at 2 month of age are less anxious than controls (George et al. 2008), while A53T heterozygous mice are less anxious than controls at 12 month (Graham and Sidhu 2010). We, however, evaluating older homozygous A53T a-Syn mice. Some have noted that a loss of olfaction can produce anxiety in mice (Glinka et al. 2012), though our mice had normal olfaction (Fig. 4) when anxiety was first observed (Fig. 1), implying an a-Syn effect. We noted a–Syn aggregation comparable to PD (Fumimura et al. 2007) in A53T adrenal glands with altered enzymatic activity (Fig. 2), making ours the first report to link adrenal synucleinopathy with catecholamine dysregulation and anxiety.
In conclusion, PP2A appears to be a rational therapeutic target for PD. While there is no evidence directly linking PP2A to PD, a large PD-linkage-region on human chromosome 5 encompasses genes for both the PP2A catalytic subunit (5q31.1) and a PP2A B subunit (5q32) (Scott et al. 2001; Lill et al. 2012), suggesting that further evaluation is warranted. Furthermore, a-Syn mice (Fleming et al. 2008) respond well to treatments aimed at stimulating PP2A activity (Lee et al. 2011). Because aging is a major contributing factor to PD, our aging A53T model should be particularly useful for testing medications aimed at slowing synucleinopathy. The model will also be appropriate for evaluating a-Syn imaging agents that are currently in development. As the diagnosis of pre-motor PD improves and novel PD therapies are developed, “cures” for this devastating neurodegenerative disorder may be on the near horizon.
This work is dedicated to M. J. Fox, R. Byer, J. Cordy, and in memory of L. “Rusty” Lanelli. We thank A. Dwivedi and I. Mallawaarachchi for statistical assistance, G. Vidal and J. Vargas for editorial advice, and L. Bermudez for technical assistance. We also thank Dr. T. Iwatsubo for generously sharing his a-Syn PSer129 antibodies. Funding was provided by NINDS [NS42094, RGP] and the Ministry of Education of China [BS2010YY036, LH]. The agencies played no role in study design, data collection, analysis, interpretation, writing, or choosing where to publish. The authors declare no conflict of interest regarding financial, personal or other relationships that could have biased the research.