Dopamine D2 receptor and β-arrestin 2 mediate Amyloid-β elevation induced by anti-parkinson’s disease drugs, levodopa and piribedil, in neuronal cells

Although levodopa is the first-line medication for the treatment of Parkinson’s disease (PD) showing unsurpassable efficiency, its chronic use causes dyskinesia. Accordingly, dopamine agonists are increasingly employed as monotherapy or in combination with levodopa to reduce the risk of motor complications. It is well recognized that patients with PD often exhibit cognitive deficits. However, clinical and animal studies assessing the effects of dopaminergic medications on cognition are controversial. Amyloid-β (Aβ) is one of the major hallmarks of Alzheimer’s disease (AD), leading to progressive memory loss and cognitive deficit. Interestingly, the abnormal accumulation of Aβ is also detected in PD patients with cognitive deficits. Evidence indicated that levodopa induced a mild increase of Aβ plaque number and size in the brain of AD mouse. However, the underlying mechanism is unclear. Here we present that both levodopa and piribedil enhance the generation of Aβ and the activity of γ-secretase in human neuronal cells and primary neurons isolated from AD mouse. This effect was reduced by either the antagonism or the knockdown of dopamine D2 receptor (D2R). We further showed that in the cells expressing β-arrestin 2-biased D2R mutant, piribedil promoted cellular Aβ production to the extent comparable to the wild-type D2R whereas this activity was absent in those with G protein-biased D2R mutant. Moreover, the knockdown of β-arrestin 2 attenuated the increases of Aβ generation and γ-secretase activity mediated by levodopa or piribedil. Thus, our study suggests that targeting D2R-mediated β-arrestin function may have potential risk in the modulation of Aβ pathology.


ELISA for Aβ
HEK293/APPswe cells, SK-N-SH cells, and induced human neuronal cells were treated with chemicals at the indicated concentrations for 2 h or 24 h. The conditioned medium was then collected and subjected to a sandwich ELISA for the measurement of total Aβ level. The measurement was done according to the manufacturer's guidelines. ELISA kits for total human Aβ were obtained from ExCell Bio (Shanghai, China).

In vitro measurement of BACE1 and γ-secretase activities
Fluorogenic substrate assays were carried out as previously reported [15,16]. Briefly, total membrane fractions were extracted from SK-N-SH cells or induced human neuronal cells and re-suspended in reaction buffers (including 10 μM of specific fluorogenic substrate and vehicle or indicated chemicals). After incubation at 37˚C for 30 min (BACE1) or 120 min (γ-secretase), fluorescence of the cleaved substrates was measured by SpectraMax M5 spectrometer (Molecular Devices).

Cell viability measurement
Chemical-treated SK-N-SH or mouse primary cells were subjected to the CellTiter-Glo Luminescent Cell Viability Assay (Promega) following the manufacturer's instructions.

cAMP assay
The intracellular cAMP was measured using GloSensor™ cAMP assay following the manufacturer's instruction with minor modification (Promega). HEK293 cells were seeded in white 96-well plates (Costar) and co-transfected with pGloSensor™-22F cAMP plasmid and WT or mutant D 2 R using Effectene Transfection reagent. Before the cAMP assay, the medium was removed and replaced with the fresh medium containing 2% (v/v) of GloSensor™ cAMP reagent. After 90 min incubation at 37˚C, cells were equilibrated at room temperature (RT) for 15 min and treated with the ligands at indicated concentrations for another 15 min followed by the measurement of luciferase activity. The Nano-Glo substrate was added into each well following manufacturer's instruction (Promega). Donor emission (460 nm) and acceptor emission (610 nm) were measured as a basal signal by SpectraMax M5 spectrometer (Molecular Devices). The chemicals were then added as required and the plate was read.

Immunofluorescence microscopy
The induced human neuronal cells grown on cover-slip were fixed with 4% paraformaldehyde (PFA) in PBS for 20 min. Cells were permeabilized and blocked with PBS/0.2% Triton X-100/ 1% BSA for 30 min followed by the incubation with indicated primary antibodies for 2 h at RT. After washing with PBS/1% BSA for three times, cells were incubated with cy3-labeled donkey anti-goat IgG or Alexa Fluor 488-labeled donkey anti-mouse secondary antibodies in the dark for 1 h, washed with PBS/1% BSA, stained with DAPI (1 μg/ml, 10 min), and mounted on slides. Images were acquired using LAS SP8 confocal microscope (Leica, Germany) with a 63 ×/1.40 NA oil objective (Leica).

Reverse transcription and quantitative real-time PCR
Total RNA was extracted with TRI Reagent (T9424; Sigma) according to the manufacturer's instructions. Random hexamer primer and MMLV Reverse Transcriptase (M5301; Promega) were used for reverse transcription. All gene transcripts were quantified by quantitative realtime PCR performed with 2 × HotStart SYBR Green qPCR Master Mix (ExCell Bio, Shanghai, China) on a Stratagene Mx3000P (Agilent Technologies). The primers used for the detection of mRNA levels of human DRD1-5 and mouse Drd1-5 were listed as below: DRD1 sense:

Primary neuronal cell culture
The preparation of mouse primary neuronal cells was performed according to the standard protocols [19,20] with minor modification. Briefly, after dissection of the cortices and hippocampi from APP/PS1 P0 pups, cells were trypsinized, dissociated, and then seeded in 96-wellplates. The neuronal cells were maintained in 1 Ã B27 and 1 Ã Glutamax (Gibco, 35050)-containing Neurobasal medium. The medium was half refreshed every 4 days. And the compound treatment was performed on DIV8.

Statistical analysis
All data were analyzed by Prism 6.0 (GraphPad Software Inc., San Diego, CA). Concentration-response curves were analysed using a three parameter non-linear regression analysis. Unpaired Student's t-test (two-tailed) was applied for the comparisons of two datasets. Oneway or Two-way analysis of variance (ANOVA) with Bonferroni's post-test was used where more than two datasets or groups were compared. Statistical significance was accepted at p < 0.05.

Results
Levodopa and piribedil were identified to enhance endogenous Aβ generation and γ-secretase activity in SK-N-SH cells In human neuroblastoma cell line SK-N-SH cells, both levodopa and piribedil significantly promoted endogenous Aβ generation in a dosage-dependent manner whereas carbidopa had little effect ( Fig 1A). Bromocriptine, another dopaminergic PD drug, showed a mild tendency to promote Aβ generation with low efficacy. Notably, the expression of APP or the level of soluble APPα (sAPPα) in the culture medium was not affected by either levodopa or piribedil ( Fig 1B). Levodopa at micromolar concentrations has been reported to show cytotoxicity in neuronal cells [21,22]. Here, we showed that levodopa increased Aβ generation with an EC max at 100 nM which has little effect on cell viability ( Fig 1C). The effect of levodopa on Aβ generation was effectively prevented by the addition of L685,458, a highly selective inhibitor of γsecretase ( Fig 1D). The increased Aβ production could have resulted from the change of secretase activity. To test this, SK-N-SH cells were challenged with the chemicals followed by membrane preparation and measurement of γ-secretase or BACE1 activity. L685,458 and BSI IV were used as positive controls and significantly inhibited the γ-secretase and BACE1 activity respectively (Fig 1E and 1F). Data showed that stimulation with levodopa or piribedil significantly promoted γ-secretase activity whereas bromocriptine had little effect which is consistent with the results of Aβ generation (Fig 1E). On the other hand, levodopa or piribedil did not change the activity of BACE1 (Fig 1F). Notably, the expressions of γ-secretase components and BACE1 were not altered by the treatments (Fig 1G), suggesting the increased γ-secretase activity is independent of its expression.
Levodopa and piribedil stimulate Aβ generation and γ-secretase activity through D 2 R Levodopa is the precursor of dopamine while piribedil is an agonist of D 2 R and dopamine D 3 receptor (D 3 R). We examined the mRNA level of dopamine receptors in SK-N-SH cells and found that D 2 R was the most abundant one among all the subtypes (Fig 2A). Therefore, we hypothesized that the effect of levodopa or piribedil on the promotion of Aβ production could be mediated by D 2 R in neuronal cells. To examine this, a selective D 2 R antagonist L741,626 was used to treat the cells before levodopa or piribedil stimulation. Pre-treatment with L741,626 (1 μM) significantly reduced the potency of piribedil-mediated cAMP response indicating the antagonist was effective under this condition (Fig 2B). We further showed that L741,626 (1 μM) numerically but not significantly reduced endogenous Aβ production in SK-N-SH cells while it prevented the elevation of Aβ level induced by levodopa or piribedil ( Fig 2C). The endogenous D 2 R expression in SK-N-SH cells was detected at above 50 KD and the infection of gene specific shRNA resulted into an obvious reduction of D 2 R level ( Fig 2D). Knockdown of D 2 R did not obviously affect the basal Aβ level but abolished levodopa or piribedil-increased Aβ production (Fig 2E). It did not impair forskolin (FSK)-mediated Aβ response indicating the specificity (Fig 2F). Consistently, levodopa or piribedil-stimulated γsecretase activity was reduced in the cells with D 2 R knockdown ( Fig 2G). All these suggest that levodopa and piribedil increase Aβ generation and γ-secretase activity through D 2 R.  3a), which showed biased signal transduction property for G protein and β-arrestin 2 respectively [4]. Neve's lab showed that the alanine substitution of residues 213-215 ( [3A] D 2 R) in the 3 rd intracellular loop (IC3, Fig 3a) did not affect G protein activation but impaired β-arrestin 2 recruitment [23]. All these mutants showed little change on cellular expression or distribution compared with the wild-type (WT) receptor ( [WT] D 2 R) [4,23].
To distinguish whether piribedil stimulates Aβ production through D 2 R-mediated G protein or β-arrestin signal pathway, we systematically assessed the function of these D 2 R mutations compared with the WT receptor. HEK293 cells have little endogenous expression of dopamine receptors and are easy to be transfected with plasmids. Gs protein-mediated cAMP response and β-arrestin 2 recruitment in response to the treatment were measured using Glo-Sensor™ cAMP and NanoBRET assay respectively to confirm the functions of these D 2 Rs. The cells were transfected with relatively low amount of C-terminal HaloTag protein-tagged [ D 2 R (Fig 3E). Compared to the D 2 R, the expression of D 1 R in HEK293/APPswe cells did not influence the Aβ generation after treatment with piribedil (S1B Fig). The evoked cAMP accumulation in response to dopamine indicates the successful expression of D 1 R (S1A Fig) whereas piribedil failed to mediate cAMP generation suggesting its poor affinity for D 1 R.

Levodopa and piribedil enhance Aβ generation and γ-secretase activity in induced human neuronal cells
To confirm the effects of levodopa and piribedil in a more relevant system, human neuronal cells were differentiated from iPSC-derived NSCs. The induced neuronal cells were stained level of Aβ produced by SK-N-SH cells in response to vehicle (control) or levodopa at 30 nM either with or without L685,458 pre-treatment. Data are mean + s.e.m., n = 3-4. #p < 0.05 versus the control within the group; ***p < 0.001 versus the corresponding treatment without L685,458. (E and F) The measurements of γ-secretase (E) and BACE1 (F) activities after the treatment with vehicle (control) or the indicated chemicals. Data are mean + s.e.m., n = 3-5. ***p < 0.001 versus control. (G) Representative image of Western-blot showing the expressions of BACE1 and γ-secretase components (NCT, PS1-NTF, APH1aL, and Pen2) after the treatment with vehicle (control) or indicated chemicals. Actin was used as loading control.
doi:10.1371/journal.pone.0173240.g001 positive for DCX, a neuronal marker, while the expressions of nestin and Sox2, two NSC markers, were markedly reduced (Fig 5A). Moreover, the differentiated cells showed low ability in proliferation as indicated by reduced expression of ki67, a proliferation marker. Further, the transcription level of nestin and Sox2 were down-regulated in the induced neuronal cells (Fig 5B) and meanwhile microtubule-associated protein 2 (MAP2), synapsin 1 (SYN1), and microtubule-associated protein Tau (MAPT) were all up-regulated, indicating the formation of neurons in the cell population ( Fig 5C). Furthermore, the genes encoding D 2 R was also upregulated while differentiation (Fig 5D). In these cells, levodopa and piribedil consistently enhanced Aβ generation ( Fig 5E) and γ-secretase activity (Fig 5F) while had no effect on BACE1 activity ( Fig 5G). As controls, L685,458 reduced both the cellular generation of Aβ ( Fig  5E) and γ-secretase activity (Fig 5F) while BSI IV inhibited BACE1 activity (Fig 5G). The expressions of sAPPα, APP, and secretases were not changed upon the treatments (Fig 5H). Additionally, both levodopa and piribedil stimulated Aβ generation in NSC which expresses other D 2 -like family receptors with a relatively high expression of D 4 R and a week expression of D 3 R (S2A Fig). The knockdown of D 2 R completely prevented the increase of Aβ after the treatment with levodopa or piribedil indicating that D 4 R or D 3 R may not be responsible for this effect (S2B Fig). We further examined the effect of levodopa and piribedil in mouse primary cultures derived from APP/PS1 mouse cortice and hippocampi. Compared to the WT mouse, the expression of D 2 R was higher in APP/PS1 mouse brain on postnatal day 0 (data not shown) and in hippocampus at the age of 2.5 month (Fig 6A). The primary hippocampal and cortical neuronal cells were isolated from postnatal day 0 APP/PS1 mouse cells and cultured for 7 days in neuronal basal medium. Levodopa and piribedil consistently stimulated Aβ generation ( Fig  6B) while cell viability was unchanged indicating consistent cell number (Fig 6C).

Discussion
Dopaminergic medications such as levodopa and dopamine receptor agonists are prescribed to improve motor deficits in PD whereas their effects on cognition are complex [7][8][9][10]. Dementias is frequent in PD with evidences indicating abnormal Aβ plaques accumulation in the brain [24][25][26]. Although levodopa showed protective effect in learning and memory deficits in an AD mouse model, both the Aβ plaque number and size was numerically increased [11]. Here we consistently observed increased Aβ generation after drug treatment in primary AD mouse neuron. We suspect in addition to Aβ generation, dopaminergic medications may play other functions to coordinately modulate cognition. The present study presents that levodopa and dopamine receptor agonist piribedil promote Aβ generation possibly via enhancing γ-secretase activity. Molecular mechanism study revealed the involvement of D 2 R and βarrestin 2 implying that targeting D 2 R-mediated β-arrestin 2 pathways may have impact on Aβ pathology. We recently reported that the other class of anti-PD drugs targeting to adenosine A 2A receptor could also promote the generation of Aβ both in neuronal cells and in the brain of AD mouse [27]. Together with the present study, these indicate that the neurological medications targeting to certain GPCRs could have multiple effects in the CNS. the group. (D) The protein level of D 2 R in SK-N-SH cells with the infection of scrambled or D 2 R gene specific shRNA. The overexpression of human D 2 R in HEK293 cells indicates that the band at over 50 KD is D 2 R. endo, the endogenous D 2 R in SK-N-SH cells. Actin was used as a loading control. (E and F) Measurement of Aβ level after the stimulation with vehicle (control), levodopa or piribedil at 30 nM (E) or FSK at 1 μM (F) in the cells infected as described in (D). Data are mean + s.e.m., n = 5-6. **p < 0.01; ***p < 0.001 versus the control within the group. FSK, forskolin. (G) Measurement of γ-secretase activity after the stimulation with vehicle (control), levodopa or piribedil at 30 nM in the cells infected as described in (D). Data are mean + s.e.m., n = 3. ***p < 0.001 versus the control within the group.
doi:10.1371/journal.pone.0173240.g002   Although age-dependent reduction of D 2 R has been observed in the brain of healthy individuals [5,6], the change of D 2 R expression with AD is debatable. An early study showed that D 2 R was reduced in the striatum of AD patients even in the absence of parkinsonian symptomatology [28] while a following evidence presented the loss of striatal D 2 R in AD with Parkinsonism but not PD or AD patients [29]. A latter microarray correlation analysis of human brain samples revealed an up-regulation of D 2 R gene expression in the patients with AD [30] and an increase of D 2 R mRNA level in the hippocampus of APP/PS1 transgenic mouse is observed in the current study. It is likely that dopamine receptors in different brain regions response differently to the disease development. Aβ is toxic to neurotransmitter pathways and can cause the impairment of dopaminergic system [31]. To compensate the reduction of dopamine release, D 2 R may up-regulated in the early stage of AD to improve dopamine function. However as the disease advances, a progressive loss of dopamine neuron results into the impairment of compensatory mechanism leading to the reduction of D 2 R and the development of disease complications such as PD. Thus, D 2 R is correlated with both PD and AD pathogenesis. Mounting evidences have shown that GPCRs modulate APP processing via multiple mechanisms. Especially for β 2 -adrenergic receptor and G protein-coupled receptor 3, β-arrestins are required. As reported, β-arrestin 2 interacts with γ-secretase complex and modulates the complex re-distribution toward detergent-resistant membranes leading to increased secretase catalytic activity [14]. In another study, β-arrestin 1 regulates the maturation of γ-secretase complex by the interaction with γ-secretase component [13]. Accordingly, it is likely that piribedil stimulates the activation of D 2 R leading to the recruitment of β-arrestin 2 to the receptor, which may further allow for the re-distribution of γ-secretase. However, β-arrestin 2 recruitment may not necessarily lead to the increased secretase activity and Aβ generation. The final outcome may also depend on receptor signaling and/or the induced subcellular distribution of β-arrestin 2/secretase complex which could be receptor or ligand specific.
Dopaminergic medications generally have multiple targets. Apart from D 2 R, levodopa could target other dopamine receptor subtypes such as D 1 R and D 3 R in vivo [32][33][34]. Piribedil could also activate D 3 R. Therefore, their effects on Aβ generation could depend on the expression pattern of these dopamine receptor subtypes in a specific cell type or brain region. Particularly, receptor expression could be altered by disease progression or drug treatment [10]. Accordingly, the cognitive effects of levodopa are influenced by the exposition duration to levodopa [35], disease stage, motor fluctuations [36], and genetic polymorphisms in levodopa metabolic pathway [37]. Another dopaminergic agent apomorphine targets both dopamine receptors and serotonin receptors with high affinity. The latter has been shown to regulate Aβ deposition [38]. We consistently found that apomorphine reduced cellular Aβ generation (data not shown). All these imply the complicated effects of dopaminergic medications on cognition in vivo.