Ascorbate deficiency decreases dopamine release in gulo^–/– and APP/PSEN1 mice

Abstract

Dopamine (DA) has important roles in learning, memory, and motivational processes and is highly susceptible to oxidation. In addition to dementia, Alzheimer’s disease (AD) patients frequently exhibit decreased motivation, anhedonia, and sleep disorders, suggesting deficits in dopaminergic neurotransmission. Vitamin C (ascorbate, ASC) is a critical antioxidant in the brain and is often depleted in AD patients due to disease-related oxidative stress and dietary deficiencies. To probe the effects of ASC deficiency and AD pathology on the DAergic system, gulo^–/– mice, which like humans depend on dietary ASC to maintain adequate tissue levels, were crossed with APP/PSEN1 mice and provided sufficient or depleted ASC supplementation from weaning until 12 months of age. Ex vivo fast-scan cyclic voltammetry, showed that chronic ASC depletion and APP/PSEN1 genotype both independently decreased dopamine release in the nucleus accumbens, a hub for motivational behavior and reward, while DA clearance was similar across all groups. In striatal tissue containing nucleus accumbens, low ASC treatment led to decreased levels of DA and its metabolites 3,4-dihydroxyohenyl-acetic acid (DOPAC), 3-methoxytyramine (3-MT), and homovanillic acid (HVA). Decreased enzyme activity observed through lower pTH/TH ratio was driven by a cumulative effect of ASC depletion and APP/PSEN1 genotype. Together the data show that deficits in dopaminergic neurotransmission due to age and disease status are magnified in conditions of low ASC which decrease DA availability during synaptic transmission. Such deficits may contribute to the non-cognitive behavioral changes observed in AD including decreased motivation, anhedonia, and sleep disorders.

1. Introduction

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder characterized by complex histopathology including amyloid-beta plaques, neurofibrillary tau tangles, and cerebral atrophy (Lane et al. 2018; Hyman et al. 2012; Jellinger and Bancher 1998). As the leading cause of dementia in older adults, the greatest risk factor for AD is age, suggesting that dysfunctions in neurotransmission are driven largely by age-related accumulation of oxidative stress damage and disease-specific neuropathology including β-amyloid aggregation and hyperphosphorylated tau neurofibrils (Grimm et al. 2016; Yao et al. 2009; Leuner et al. 2012). AD treatments have traditionally focused on rescuing cholinergic function (Wilcock et al. 1982; Muir 1997; H. Ferreira-Vieira et al. 2016), but current treatment options, such as acetylcholinesterase inhibitors, only boost diminishing signals and do not reverse or slow disease progression (Raina et al. 2008; Sun et al. 2008). Investigation of other neurotransmitter systems to identify novel preventative or treatment strategies is therefore warranted. More recently, GABAergic and glutamatergic systems have been explored as alternative mechanisms of cognitive decline and include targeted intervention strategies to stabilize neuronal hyperexcitability (Liedorp et al. 2010; Mi et al. 2018; Vossel et al. 2016). However, the role of dopamine (DA) and the dopaminergic (DAergic) system in AD remains understudied (D’Amelio and Nisticò 2018; Rosenberg et al. 2015; Martorana and Koch 2014) despite an emerging supportive role DA plays in learning and memory (Kulisevsky 2000; Puig et al. 2014).

Preclinical studies have highlighted evidence for changes in DA release in the presence of β-amyloid accumulation. Cortical and subcortical levels of DA were decreased in the APP/PSEN1 mouse model of AD amyloidosis at 18 months of age (Von Linstow et al. 2017) and other reports found decreased striatal levels of DA and its metabolite DOPAC prior to 12 months of age (Perez et al. 2005; Kennard and Harrison 2014). Such decreases in tissue DA levels have largely been attributed to atrophy and neuronal loss in the ventral tegmental area (Liu et al. 2008; Nobili et al. 2017; Cordella et al. 2018), one of the primary DAergic hubs which projects to the hippocampus and nucleus accumbens (NAc) in the striatum. The NAc is a primary site of DA output for reward and motivation processing (Russo and Nestler 2013) and has received little attention in AD mouse models. Investigation of DA neurotransmission was performed in a single mutation AD mouse model, Tg2576, which showed decreased DA output in the NAc by amperometry at 6 months of age (Nobili et al. 2017; Cordella et al. 2018). 4–6 month old APP/PSEN1 mice showed long term DA replenishing deficits in caudate putamen by in vivo voltammetry (Kärkkäinen et al. 2015). DA is readily oxidized by reactive oxygen species (Zhang et al. 2019), and DAergic neurotransmission is, therefore, highly susceptible to the changes in oxidative stress associated with AD.

Vitamin C (ascorbate, ASC) is a critical antioxidant that supports neuronal health and may have a role in slowing the development of AD (Harrison 2012; Monacelli et al. 2017). ASC protects against oxidative stress accumulation in the brain by donation of electrons to neutralize reactive oxygen species (Dixit et al. 2017; Dixit et al. 2015). AD patient populations exhibit ASC deficiency despite adequate supplementation, suggesting that antioxidant reserves are consumed more rapidly under AD pathological conditions and subsequent oxidative challenge (Rivière et al. 1998; Schrag et al. 2013; de Wilde et al. 2017). ASC is also essential for DA synthesis and metabolism through its action as a cofactor for tyrosine hydroxylase, the enzyme which catalyzes the rate limiting step in DA synthesis (Fig. 1) (Opmeer et al. 2010; Daubner et al. 2011). ASC can also protect against oxidation of dopamine within the synapse (Rebec and Christopher Pierce 1994; Ballaz and Rebec 2019).

Figure 1 –

Role of ASC in DA synthesis and metabolism. Abbreviations: ASC – ascorbate (vitamin C), TH – tyrosine hydroxylase, p – phosphorylated, ERK1/2 – extracellular regulated kinases 1/2, AADC – aromatic amino acid decarboxylase, DOPAC – 3,4-dihydroxyohenyl-acetic acid, 3-MT – 3-methoxytyramine, HVA – homovanillic acid, MAO – monoamine oxidase, COMT – catechol-O-methyl-transferase, DBH – dopamine-β-hydroxylase. Enzymes shown as grey circles, kinases shown as blue circles.

The relationship between AD-related changes in dopamine release and ASC deficiency on DAergic signaling specifically in the NAc remains unknown. Here, we used ex vivo Fast Scan Cyclic Voltammetry (FSCV) to study the effects of brain ASC levels and APP/PSEN1 genotype on DA release in the NAc. We observed that both depleted ASC levels and APP/PSEN1 genotype interfered with proper DAergic synaptic signaling. Such deficits in DA synthesis and release could contribute significantly to AD related dysfunction and cognitive decline.

2. Methods

2.1. Subjects

Gulo^–/– mice were bred in-house from founders originally obtained from the Mutant Mouse Regional Resource Centers (mmrrc.org, RRID:MMRRC_000015-UCD) and were maintained on a C57BL/6J background (Jackson laboratories, RRID:IMSR_JAX:000664). Gulo^–/– mice, like humans, lack the gene encoding gulonolactone oxidase, the enzyme responsible for catalyzing the final step in ASC synthesis. They therefore depend on dietary ASC supplementation to prevent scurvy (Maeda et al. 2000; Harrison et al. 2008).

APP/PSEN1 mice (Borchelt et al. 1997; Savonenko et al. 2005) were bred in house from founders maintained on C57BL/6J background and obtained from Jackson laboratories (RRID:IMSR_JAX:005864). APP/PSEN1 mice begin to develop mild cognitive deficits and sparse β-amyloid aggregation from 5–6 months of age. A more severe neuropathological profile including β-amyloid plaque accumulation, neuroinflammatory response, and cholinergic dysfunction is observed by 12 months of age. (Lalonde et al. 2005; Machová et al. 2008; Reiserer et al. 2007; Zhu et al. 2017).

Gulo^–/– mice were bred with APP/PSEN1^+/– mice to obtain gulo^–/– APP/PSEN1^+/– (G-AP) and gulo^–/– wild-type (G-WT) mice (Harrison et al. 2010). All mice used in this study were gulo^–/– (G) and were incapable of ASC synthesis. Group differences due to genotype hereafter refer to the effect of APP and PSEN1 mutations in APP/PSEN1 carriers (G-AP) compared to non-carriers (G-WT). Mice were group-housed by sex with two to five mice per cage, maintained in a room with a 12:12 hour light-dark cycle and provided food and ASC-supplemented water ad libitum. Genotypes (gulo^−/− and either wild-type or APP/PSEN1) were confirmed at weaning (21 days of age) and post-mortem by polymerase chain reaction. All mice were aged 12–15 months, and male and female mice were included in approximately equal numbers (7 males, 5 females for voltammetry; 14 males, 13 females for neurochemistry measures and western blots). All protocols were approved by the Vanderbilt University Institutional Animal Care and Use Committee under protocol number V1800098. All experiments were conducted in accordance with the NIH Guide for the Care and Use of Laboratory Animals.

2.2. ASC supplementation

Drinking water provided either 1.0 g/L (High, to maintain normal wild-type ASC levels) or 0.033 g/L (Low, to maintain consistently depleted ASC levels) ASC (Sigma, cat.no. A92902) stabilized with 0.01M EDTA. Supplemented water was replaced at least once per week. All breeding pairs received 1.0 g/L (High) ASC supplementation and pups were weaned to the High ASC level until 8 weeks of age. Cages were then randomly assigned to either maintain 1.0 g/L (High) or given 0.033 g/L (Low) ASC until used for experiments. Deficiency in synthesis due to gulo^–/– genotype does not result in a deficiency in tissue levels when sufficient supplementation is provided. The High supplementation of 1.0g/L is sufficient to fully saturate tissue ASC levels analogous to those of gulo^+/+ mice and excess ASC is excreted in urine. The Low supplementation level of 0.033 g/L is sufficient to maintain supra-scorbutic vitamin C levels, healthy weight gain with age, and normal cognition in G-WT mice, although it may modestly accelerate oxidative stress and amyloid accumulation in G-AP mice (Harrison et al. 2008; Harrison et al. 2010).

2.3. Fast-scan Cyclic Voltammetry

Mice were euthanized by decapitation, followed by removal of the whole brain into cold oxygenated artificial cerebral spinal fluid (aCSF) containing 126 mM NaCl, 2.5 mM KCl, 1.2 mM NaH₂PO₄, 2.4 mM CaCl₂, 1.2 mM MgCl₂, 25 mM NaHCO₃, and 11 mM glucose with pH adjusted to 7.4. Traditionally, ASC is added to aCSF to 0.4 mM, however, for the present studies, addition of ASC was intentionally omitted. The cerebellum and olfactory bulbs were removed, and remaining brain was mounted on a metal stage submerged in cold oxygenated aCSF for slicing. Multiple coronal slices 300 μm in thickness containing the nucleus accumbens (NAc) were prepared using a vibratome. Slices were equilibrated for 30 minutes in a holding chamber containing oxygenated aCSF at room temperature, then transferred to the testing chambers with bath aCSF flowing 1 ml/min and temperature maintained at 32 °C. A cylindrical carbon fiber microelectrode (100–150 μM length, 7 μM radius) and a bipolar stimulating electrode were placed in close proximity to each other into the core of the NAc, a region highly enriched in dopamine nerve terminals. Extracellular DA was measured by recording the current produced by DA oxidation when applying a triangular voltage waveform (−0.4 to +1.2 to −0.4 V vs Ag/AgCl, 400 V/s) every 100 ms at the carbon-fiber electrode. Currents measured at ~600 mV identified DA oxidation. Recording electrodes were calibrated prior to testing by recording the electrical current of a known concentration of dopamine (3 μM) using a flow injection system. Release of DA was evoked by a single, rectangular, electrical pulse (300 μA, 4 ms, monophasic) applied every 180 seconds until DA release reached three consistent (within 5%) baseline recordings, 5–10 total recordings per slice (Ferris et al. 2013; Calipari et al. 2012; Calipari et al. 2017).

All voltammetry data including Peak Current (nA), DA release per pulse ([DA]p, uM), Tau Decay Time (s), and Half-life (s) were collected and modeled using Demon Voltammetry and Analysis Software, RRID:SCR_014468 (Yorgason et al. 2011). Experimenter was blind to genotype and treatment group.

2.4. Western Blot Analysis

Mice were anesthetized with isoflurane (NDC, cat.no. 66794-017-25) to minimize suffering and sacrificed by decapitation. The brain was rapidly removed and sliced on ice using a 1 mm mouse brain matrix. Slices were flash frozen on dry ice and punches (1 mm diameter) of striatal tissue containing NAc were collected and stored at −80°C for further analysis. Tissue punches were homogenized on ice by adding 60 μl of homogenization buffer containing cOmplete protease inhibitor cocktail tablet (Roche, cat.no. 4693132001), 1:100 phosphatase inhibitor cocktail 3 (MilliporeSigma, cat.no. P0044), and 1 mM sodium orthovanadate per 10 ml of RIPA buffer (Thermo Fisher Scientific, cat.no. 89900). Protein extract was collected after incubating at 95 °C for 5 min and centrifugation at 15,000 rpm for 5 min. Samples were resolved using 4–12% NuPAGE Bis-Tris Plus gels (Thermo Fisher Scientific, cat.no. 23225) and transferred onto nitrocellulose membrane using the iBlot system (Thermo Fisher Scientific, cat.no. IB23001). For pERK1/2 and ERK1/2 (P-p44/42 MAPK and p44/42 MAPK, anti-rabbit, Cell Signaling Technologies, cat.no. 4370 and cat.no. 9102), 40 μg reactions were loaded. For pTH (Ser31) and TH (anti-rabbit, Cell Signaling Technologies, cat.no. 3370 and 2792) and DAT (anti-rat, Millipore Sigma, cat.no. MAB369), 20 μg reactions were loaded. Blots were blocked for 30 min with 5% milk then again for 30 min with 5% BSA in tris buffered saline plus 0.1% Tween (TBST). Primary antibody incubation was performed overnight at 4°C (1:1000 in 5% BSA in TBST) followed by secondary antibody incubation for 2 hours at room temperature (1:5000 in 5% BSA in TBST). Between sequential protein probes, blots were stripped with Restore stripping buffer (Thermo Fisher Scientific, cat.no. 21059) for 20 minutes followed by 3×10 min washes with TBST. Blots were visualized using Western Lighting Plus ECL (Perkin Elmer, cat.no. 103E001EA) followed by quantification using the NIH ImageJ gel analysis tool. All blots were run with two samples from each group, and all data was analyzed using unbiased methods. Each sample was normalized first to its own β-actin (anti-mouse, Millipore Sigma, cat.no. A2228) and then to the average High ASC G-WT control value for that blot (DiCarlo et al. 2019).

2.5. Neurochemistry measures

Tissue ASC concentrations were measured by ion-pair HPLC as described previously (Consoli et al. 2020; Harrison et al. 2008; May et al. 1998). Briefly, sample extracts were prepared by adding 10 μl extraction buffer (7:2 25% w/v metaphosphoric acid: 100 mM sodium phosphate, 0.05 mM EDTA pH 8.0) per mg of wet tissue, normalizing by weight. Following tissue homogenization and centrifugation, the supernatant was diluted and measured in triplicate with ion pair HPLC with electrochemical detection. For quantification of biogenic amines, striatal tissue containing NAc was collected from animals 12–15 months old as described for Western blots. Measurements of the concentration of biogenic amines were obtained using Mass Spectrometry by the Vanderbilt University Neurochemistry Core. For neurochemistry measures, the experimenter was blinded to genotype and treatment groups, and a separate individual performed data analysis.

2.6. Statistics

This study was not pre-registered. Normal distribution of data was confirmed using the Anderson-Darling test, and data were subsequently analyzed using parametric statistics. A multivariate ANOVA analysis, which included sex as an additional variable, showed no main effects of sex on any of the key outcomes (peak current, clearance, neurochemistry measures), so male and female groups were combined for all subsequent analysis. Baseline voltammetry and neurochemistry data were compared using two-way analysis of variance (ANOVA) using Graphpad Prism 8 statistical software (RRID:SCR_002798) with genotype (gulo^–/– APP/PSEN1^–/–, “G-WT” or gulo^–/– APP/PSEN1^+/–, “G-AP”) and ASC treatment (High 1.0g/L or Low 0.033g/L) as the independent variables. Significant omnibus ANOVA (P < 0.05) were followed by Sidak’s multiple comparisons post hoc analyses to test differences between groups. Predicted sample sizes were estimated based on previous experiments with similar variability between groups. All available subjects that survived to 12 months of age were included and distributed to treatment groups equally to balance for genotype and sex. Prior to the 12-month timepoint, one G-AP mouse was sacrificed due to prolapse, one G-WT for bullying, and one G-AP mouse was found dead in the cage at 8 months of age. Numerical outliers that likely reflected experimental error were identified and removed using ROUT (Q = 5%). This applies to 3 values from High ASC G-AP and 3 values from Low ASC G-AP in peak current and DA per pulse, as well as 1 value from High ASC G-WT and 1 value from Low ASC G-WT in tau and half-life values. Error bars are shown as standard error of the mean (SEM).

3. Results

3.1. Decreased DA output in NAc in gulo^–/– and APP/PSEN1 mice

Fast Scan Cyclic Voltammetry (FSCV) was used to measure changes in DA neurotransmission ex vivo between each of the four groups: G-WT High ASC, G-AP High ASC, G-WT Low ASC, and G-AP Low ASC (Fig. 2A). Brain ASC was depleted by approximately 60% in Low ASC treated mice (F_(18,27) = 399.89, P < 0.001; Fig. 2B) with no difference according to APP/PSEN1 genotype (F_(1,27) =0.424 P = 0.520; Interaction F_(1,27) =0.599 P = 0.446). Both APP/PSEN1 genotype and Low ASC treatment significantly decreased dopamine release determined by peak current compared to G-WT High ASC control mice (Fig. 2C–F; Genotype: F_{(1, 60)} = 39.10, P < 0.001; Treatment: F_{(1, 60)} = 32.69, P < 0.001; Interaction: F_{(1, 60)} = 7.079, P = 0.010). Low ASC decreased DA release in both G-WT and G-AP mice (Ps < 0.012) and APP/PSEN1 genotype decreased DA release regardless of ASC level (Ps < 0.04). The greatest decrease in DA release compared to controls was observed in low ASC G-AP mice, however, this trend toward a cumulative effect of ASC and genotype was not statistically significant (P = 0.057).

Figure 2 –

Experimental design and voltammetric measurement of DA release. (A) Treatment and genotype groups. (B) Cortical ASC levels. n = 31 mice. ^###P < 0.001 main effect of ASC treatment by two-way ANOVA. Dotted line represents mean age-matched cortical ASC level in wild-type gulo^+/+ mice. (C) Slicing protocol and voltammetry setup. (D) Peak amplitude reported as current in nanoamps from each slice. (E) Representative color plots for each group. (F) Representative current versus time traces for each group. n = 64 slices from 12 mice. ***P < 0.001 different from G-WT High ASC by Sidak’s multiple comparisons following significant interaction in the omnibus ANOVA.

3.2. DA clearance is unchanged in gulo^–/– and APP/PSEN1 mice

The DA decay time, Tau, was similar among groups and a trend toward higher values in the Low ASC animals was not significant suggesting that while there were release differences, there was no effect on DA clearance or uptake mechanisms (Fig. 3A–B; Genotype: F_{(1, 64)} = 0.275, P = 0.602; Treatment: F_{(1, 64)} = 3.87, P = 0.053; Interaction: F_{(3, 64)} = 0.182, P = 0.671). The recorded half-life of DA was also not significantly different between groups (Fig. 3C; Genotype: F_{(1, 64)} = 0.202, P = 0.655; Treatment: F_{(1, 64)} = 3.512, P = 0.066; Interaction: F_{(1, 64)} = 0.252, P = 0.618). In striatal tissue punches containing NAc (Fig. 3D), DAT expression was unchanged across groups when evaluated by Western Blot (Fig. 3E; Genotype: F_{(1, 19)} = 0.011, P = 0.918; Treatment: F_{(1, 19)} = 0.055, P = 0.817; Interaction: F_{(3, 19)} = 0.400, P = 0.535).

Figure 3 –

No change in DA clearance. (A) Representative normalized clearance traces. (B) DA decay time, Tau. (C) DA half-life. (D) Slice and punch tissue sampling. (E) DAT expression by Western Blot. Representative blot is presented in the same order shown in the chart, G-WT High ASC, G-AP High ASC, G-WT Low ASC, G-AP Low ASC. n = 23 mice. Error bars shown as SEM.

3.3. DA Metabolism and Synthesis

Tissue concentrations of DA and its three metabolites (Fig. 1) 3,4-dihydroxyphenylacetic acid (DOPAC), 3-methoxytyramine (3-MT), and homovanillic acid (HVA) were significantly decreased in Low ASC treated mice (Fig. 4A–B; Main effect of ASC Treatment: DA F_{(1, 28)} = 13.97, P < 0.001; DOPAC F_{(1, 28)} = 12.61, P = 0.0014; 3-MT F_{(1, 28)} = 12.54, P = 0.0015; HVA F_{(1, 27)} = 19.70, P < 0.001) but were not otherwise affected by APP/PSEN1 genotype (Fs < 1.854, Ps > 0.184). Ratios of DA metabolites to DA (DOPAC/DA, 3-MT/DA, HVA/DA) in paired samples were not significantly different due to treatment or genotype (Fs < 2.687, Ps > 0.1124; data not shown).

Figure 4 –

Changes in DA metabolism. Concentrations of (A) DA and (B) DA metabolites DOPAC (left), 3-MT (middle), and HVA (right) in striatal tissue punches containing NAc. n = 32 mice. (C) Expression of pTH and TH and pTH/TH ratio. (D) Expression of pERK1/2 and ERK1/2 and pERK/ERK ratio. n = 23 mice. Error bars shown as SEM. ^###P < 0.001 ^##P < 0.01 main effect of ASC treatment by two-way ANOVA; *P < 0.05 significant difference by Sidak’s multiple comparisons following significant interaction in the omnibus ANOVA.

ASC is an essential co-factor for tyrosine hydroxylase (TH) the enzyme catalyzing the rate limiting step in DA synthesis. We therefore hypothesized that decreased tissue concentrations of DA could be due to DA synthesis dysfunction as a result of low ASC availability. TH is activated when phosphorylated by the MAP kinase, ERK1/2. To probe whether ASC depletion resulted in changes in the DA synthesis pathway, we examined expression of TH, ERK1/2, and their phosphorylated forms pTH and pERK1/2 by immunoblot. A main effect of Low ASC on phosphorylation of TH and pTH/TH ratio (Fig. 4C; pTH: F_{(1, 19)} = 8.013, P = 0.011; pTH/TH: F_{(1, 19)} = 17.56, P < 0.001) was driven by a cumulative effect of ASC depletion and APP/PSEN1 genotype on enzyme activity (Fig. 4C; APP/PSEN1 pTH F(_1,19) = 4.74, P = 0.042; Interaction pTH: F_{(1, 19)} = 11.49, P = 0.003; pTH/TH: F_{(1, 19)} = 10.75, P = 0.0039). While no changes in overall levels of pERK1/2 or ERK1/2 were observed (Fs < 2.770, Ps > 0.1144), some variability among groups was observed in pERK/ERK ratio, (Fig. 4D; Interaction: F_{(1, 18)} = 4.61, P = 0.046). Nevertheless, there was apparent unequal variance in the data, particularly comparing G-AP high and low ASC supplemented groups. Thus, although these results suggest that differences in phosphorylation of tyrosine hydroxylase by phosphorylated ERK1/2 may partially contribute to the observed differences in DA and metabolites in G-AP mice according to ASC status, this hypothesis will need to be investigated further in future studies.

4. Discussion

We investigated the role of ASC deficiency in DAergic neurotransmission in the context of AD-related amyloid pathology. Cortical ASC levels were similar among G-WT and G-AP mice and were directly determined by dietary ASC treatment. Previous work from our lab showed that chronic low dietary ASC in G-AP mice accelerated β-amyloid deposition at this age, but that genotype did not further impact global ASC levels in cortex (Dixit et al. 2015; Harrison et al. 2009). In the present study, we have shown that ASC depletion and APP/PSEN1 genotype both independently decrease dopaminergic release in the NAc. There is a specific role for ASC in DAergic function through its action as a co-factor for the TH enzyme. While DA clearance was similar across all groups, DA release was adversely impacted by both APP/PSEN1 mutations and Low ASC. This likely impacts DAergic function by limiting the amount of DA available in the synapse during synaptic function. Such a difference is particularly detrimental in aged mice in which release is already much diminished compared to levels reported in younger animals, roughly 25–45% of recorded values in 2 month old mice (Calipari et al. 2012; Calipari et al. 2017; Ferris et al. 2013). In line with lower global DA levels, DA metabolites DOPAC, HVA and 3-MT were also decreased in Low ASC mice. Since the ratios of DA metabolites to global DA levels were unchanged, decreased DA levels are likely dependent on synthesis deficits, such as decreased enzymatic activity or limited ASC availability, rather than increased breakdown into DA metabolites. pTH was lowest in Low ASC APP/PSEN1 mice suggesting an interaction with disease status, however. The strong effects of chronically depleted ASC levels may also have masked a significant interaction between low brain ASC and APP/PSEN1 genotype on DAergic function, which might have been more evident at less dramatic depletion levels, or following a shorter period of depletion. While a strong cumulative effect was not observed in this study, it is relevant that depletion of ASC led to the same functional deficits as APP/PSEN1 mutations. The same cumulative effect was not clearly visible for pERK. While there is no specific action for ASC on the ERK enzyme as there is for the TH enzyme, no clear changes in ERK expression further support the hypothesis that ASC depletion is decreasing DA synthesis specifically through decreased TH activity. Together the data highlight that DA transmission is likely affected by age and disease status and this may be magnified in conditions of low ASC.

In the present study decreased DA was observed in striatal tissue containing NAc in Low ASC mice but did not differ according to APP/PSEN1 genotype. We have previously reported that severe acute ASC depletion decreased several neurotransmitter levels including DA in cortex, but not in striatum in young (less than 6 months old) gulo^–/– mice (Ward et al. 2013), suggesting chronic ASC depletion may be needed to trigger decreases in DA in striatum. It is possible that the prolonged ASC depletion (10–12 months) utilized here drives homeostatic changes in the DA synthesis pathway, notably downregulation of pTH/TH, leading to decreased DA availability. Such homeostatic mechanisms may be more notable in the gulo^–/– background that lack the ability to synthesize ASC to ensure a dynamic response to antioxidant challenge which is otherwise observed in gulo^+/+ mice. We have also shown decreased monoamine levels in the NAc overall with age and not determined by APP/PSEN1 genotype that were not rescued by acute intravenous ASC treatment (Kennard and Harrison 2014). We confirmed that DA levels in striatal tissue were unaffected by APP/PSEN1 genotype when maintained on High ASC supplementation, despite clear deficits in evoked DA release. The deficits in DA release measured by FSCV in G-AP mice may be attributable to neuronal loss in the VTA, as has been reported in a similar mouse model of amyloidosis, Tg2756 (Nobili et al. 2017; Cordella et al. 2018), although this was not directly tested here. It is therefore possible that reduced evoked DA release in High ASC G-AP mice despite normal total DA tissue levels is instead due to a release mechanism impairment. Chronic restoration of High ASC levels in ASC deficient mice may ultimately rescue deficits in DA release observed in G-WT mice. However, it is unlikely that similar deficits in DA release could be rescued by ASC treatment in the G-AP mice since those deficits appear to be driven by different homeostatic changes relating to genotype and related neuropathology. Since we observed no changes in DA clearance across groups, it is likely that these deficits in DA neurotransmission are not driven by changes in DA clearance by DAT and are instead driven by DA availability through synthesis dysregulation in ASC deficient mice.

Preclinical studies in AD mouse models show beneficial cognitive effects when targeting the DAergic system. Optogenetic stimulation of DAergic fibers boosted hippocampal functions such as synaptic plasticity and memory (McNamara et al. 2014), and a study targeting DAergic systems in AD mouse models showed decreased memory deficits with age (Ambrée et al. 2009). Further, our previous studies in APP/PSEN1 indicate higher brain ASC levels amidst β-amyloid pathology are associated with better performance on cognitive behavioral tasks (Dixit et al. 2015; Harrison et al. 2010). The data presented in this study suggest greater alterations in DAergic neurotransmission driven by the combination of amyloid accumulation and related neuropathology with ASC depletion. As such, DA targeted therapies could be beneficial in rescuing cognitive impairments present in these mice at this age.

The NAc is rich in DA nerve terminals and heavily implicated in goal-directed behavior, reinforcement, reward, and drug addiction (Ikemoto and Panksepp 1999; Saddoris et al. 2015; Di Chiara 2002). Decreased motivation and reward have been extensively documented in AD patients (Russo and Nestler 2013; Perry and Kramer 2015; Mitchell et al. 2011; Forstmeier and Maercker 2015). The NAc has recently been shown to regulate sleep through the integration of motivational stimuli (Valencia Garcia and Fort 2018; Oishi et al. 2017), and sleep disorders are common in AD patients (Sterniczuk et al. 2013; Hennawy et al. 2019). The DAergic system in the brains of AD patients shows many of the AD neuropathological hallmarks including neurofibrillary tangles, amyloid plaques, and neuron loss (Burns et al. 2005; Bozzali et al. 2019) and imaging studies have shown decreased DA receptor availability in the hippocampus correlate with worsened memory performance in AD patients (Kemppainen et al. 2003). Further, the administration of DA agonists to AD patients has been shown to have positive effects on cortical neurotransmission and synaptic plasticity mechanisms (Martorana and Koch 2014; Koch et al. 2014). Our findings support the notion that decreased DA release specifically in the NAc may be at least partially responsible for these symptoms, and that ASC depletion in AD patient populations may contribute to the apathy, anhedonia, and sleep disorders observed in AD.

Abbreviations:

AD

Alzheimer’s disease

ASC

ascorbate (vitamin C)

DA

dopamine

DAergic

dopaminergic

FSCV

fast scan cyclic voltammetry

NAc

nucleus accumbens

TH

tyrosine hydroxylase

RRID

Research Resource Identifier