Arrestins contribute to amyloid beta-induced cell death via modulation of autophagy and the α7nAch receptor in SH-SY5Y cells

Amyloid β-protein (Aβ) is believed to contribute to the development of Alzheimer’s disease (AD). Here we showed that Aβ25-35 rapidly caused activation of autophagy, subsequently leading to reduction of autophagy associated with cellular apoptosis. Further investigation revealed that the accumulation of β-arrestin 1 (ARRB1) caused by Aβ25-35 contributed to the induction of autophagic flux. The depletion of ARRB1 led to decreases in the expression of LC3B, Atg7, and Beclin-1, which are essential for the initiation of autophagy. ARRB1 depletion also reduced downstream ERK activity and promoted Aβ25-35-induced cell death. As with ARRB1, transient upregulation of ARRB2 by Aβ25-35 was observed after short treatment durations, whereas genetic reduction of ARRB2 caused a marked increase in the expression of the α7nAch receptor at the cell surface, which resulted in partial reversal of Aβ25-35-induced cell death. Although expression of both ARRB1 and ARRB2 was reduced in serum from patients with AD, the levels of ARRB1 were much lower than those of ARRB2 in AD. Thus, our findings indicate that ARRB1/2 play different roles in Aβ25-35 cytotoxicity, which may provide additional support for exploring the underlying molecular mechanism of AD.

Given the results indicating that apoptosis was initiated after prolonged exposure to Aβ [25][26][27][28][29][30][31][32][33][34][35] in SH-SY5Y cells, we reasoned that a transient protective mechanism might occur before apoptosis in response to Aβ 25-35induced stress. Since autophagy exerts a dual role in controlling cell survival and death, we examined level of the microtubule-associated protein 1 light chain 3B (LC3B-I) and the conversion of LC3B-I to lipidated LC3B-II (membrane-bound form), which is an autophagy marker. Figure 1D shows that levels of LC3B-I and LC3B-II were significantly increased in SH-SY5Y cells after 30 min treatment, and sharply dropped down at 1 h thereafter. There was a detectable level of basal LC3B-II in PC12 cells, which remained almost unchanged during treatment (Fig. 1E). Immunofluorescence staining further supported the observations that LC3B markedly increased after Aβ 25-35 stimulation for 1 h in SH-SY5Y cells and was eliminated after 8 h treatment (Fig. 1F).
Depletion of arrb1 enhances cytotoxicity of Aβ 25-35 in SH-SY5Y cells. Since ARRB1 functions as a regulator of autophagic capacity in response to Aβ [25][26][27][28][29][30][31][32][33][34][35] , the role of β-arrestins in Aβ 25-35 -mediated cell death was investigated. The mRNA levels of both arrb1 and arrb2 were significantly reduced in human blood samples obtained from AD patients (N = 27) when compared to healthy age-matched controls (N = 27) (Fig. 4A). It appeared that the decreases in arrb1 were more evident in AD patients compared to arrb2 (Fig. 4A). Most AD patients develop Aβ accumulation and deposition, prompting us to test the possibility that ARBB1 and ARRB2 might also be downregulated in HEK293-APPwt cells that constitutively overexpress APP and produce Aβ. As shown in Fig. 4B, ARBB1 expression was slightly reduced in cells that overexpressed Aβ, whereas the expression of ARRB2 was lower than that of ARRB1 in APPwt cells, and was associated with a decreased conversion of LC3B-II/LC3B-I (Fig. 4B). These results indicated that persistent overexpression of Aβ led to a reduction in β-arrestin expression, which subsequently led to the inactivation of autophagy. As a downstream target of β-arrestins, phosphor-ERK was downregulated, while Akt was predominantly activated in APPwt cells, suggesting that Akt signaling may be more critical in APPwt cell survival and proliferation (Fig. 4B).
Given the role of ARRB1 in the activation of autophagy, which was decreased in AD patients, we next sought to determine whether the downregulation of ARRB1 would affect cell viability in the presence of Aβ. Transient transfection of siRNA specifically targeting arrb1 led to a reduced amount of ARRB1 protein in cells at48 h after transfection compared to control cells transfected with nontargeting scramble siRNA, as measured by western blotting (Fig. 4C). Cell viability was reduced after silencing the expression of ARRB1, and depletion of ARRB1 facilitated Aβ-induced cell death (Fig. 4C). These results are consistent with the observation that interfering with ARRB1-activated autophagy attenuated its cytoprotective effect in response to Aβ. Parallel experiments examining changes in cell morphology again showed an increase in apoptosis of cells treated with Aβ 25-35 after depletion of ARRB1 (Fig. 4D). Moreover, activation of cleaved PARP was observed in cells with reduced ARRB1 in response to Aβ 25-35 compared to the group treated with Aβ 25-35 alone (Fig. 4E). We also transiently overexpressed ARRB1 in cells to validate the protective role of ARRB1 in Aβ 25-35 -mediated cytotoxicity. Ectopic expression of ARRB1 alone had no significant effect on cell proliferation ( Fig. 4F and G). However, Aβ 25-35 induced cell death, and this cytotoxic effect was partially reversed by overexpression of ARRB1 ( Fig. 4F and G). Together, these data demonstrate that interfering with the ARRB1-induced autophagic response could exacerbate Aβ 25-35 -induced cell death.
Recognizing the role of β-arrestins in the internalization of G protein-coupled receptors (GPCRs), we decided to investigate the expression of the α7 nicotinic acetylcholine receptor (α7nAChR), a subtype of nicotinic acetylcholine receptors (nAChRs) that are widely distributed in the postsynaptic membrane of the neuron in the brain and offer protection against Aβ-induced toxicity 35,36 . The mRNA of α7nAChR was markedly decreased in the peripheral blood of AD patients compared to healthy controls (Fig. 5D). This finding is consistent with previous reports indicating that the expression of α7nAChR decreases early in AD and correlates well with cognitive dysfunctions [37][38][39] .
Since nicotine is a α7nAChR agonist that can activate α7nAChR and offer protective effects from Aβ toxicity 36 , we first validated the effects of α7nAChR activity in the presence of Aβ [25][26][27][28][29][30][31][32][33][34][35] . As expected, compared to the control cells, knockdown of α7nAChR accelerated cell death, and cell viability as a result of nicotine treatment did not show significant recovery due to the loss of the α7nAChR (Fig. 5E). Moreover, Aβ 25-35 induced cell death, and this cytotoxic effect was exacerbated by the depletion of α7nAChR (Fig. 5E), consistent with previous reports 36 .
Flow cytometry analysis revealed that Aβ treatment caused an increase in the expression ofα7nAChR at the cell membrane (Fig. 5F). We next asked whether the ARRB1/2-mediated effect was dependent on α7nAChR expression on the cell membrane. There was a small increase in the expression of α7nAChR in the cell membrane lacking ARRB1, as detected by flow cytometry (Fig. 5G). The cell viability assays revealed that silencing of arrb1 exacerbated Aβ-induced toxicity, and nicotine pretreatment also did not improve the survival of cells exposed to Aβ when ARRB1 expression  was low in cells (Fig. 5H). In contrast, down-regulation of ARRB2 significantly enhanced the abundance of α7nAChR (Fig. 5I), which in turn leading to improved survival, and enhancement of nicotine-mediated protection against Aβ (Fig. 5J). These findings suggest that ARRB2, but not ARRB1, played a critical role in the acceleration of Aβ toxicity, at least in part via the ability of ARRB2 to modulate α7nAChR expression on the cell membrane.

Discussion
This study presents a novel role of ARRB1 and ARRB2 in Aβ 25-35 -induced neuronal cell death. Our results reveal that upregulation of ARRB1 and ARRB2 was an early event after Aβ 25-35 exposure, associated with an induction of autophagy. Downregulation of arrb1 led to the inactivation of autophagic flux and exacerbation of Aβ 25-35 -mediated cell death, whereas depletion of arrb2, to some extent, reversed the cytotoxicity of Aβ [25][26][27][28][29][30][31][32][33][34][35] . Our study further showed that, unlike ARRB1, which was critical for activation of autophagy, ARRB2 preferentially regulated α7nAch receptor expression on the membrane, which mediates the neuroprotective effect of nicotine. Knockdown of arrb2 enhanced the expression of the α7nAch receptor at the plasma membrane, which in turn attenuated Aβ 25-35 toxicity. β-arrestins have initially identified as desensitizers of canonical GPCR signaling, the β-arrestins are now recognized as regulators of G protein-independent signaling 40 41 . We also demonstrated that ARRB2 was able to regulate α7nAChR, because downregulation of ARRB2 facilitated the expression of α7nAChR on cell membrane, leading to the enhancement of nicotine on cell proliferation. Further investigation is required for elucidating the mechanism by which ARRB2 desensitizes α7nAChR activity.Aβ can suppress neuronal autophagy by impairing AMPK 23,42 , leading to the reduced autophagic clearance of some aggregation-prone proteins and enhancing neuron cell death 24 . Basal levels of autophagy are essential for the removal and repurposing of damaged cytoplasmic contents and aggregated proteins, which is critical for cellular homeostasis. Therefore, many efforts have been made to ascertain whether the induction of autophagy would be beneficial for stimulating the clearance of Aβ, reducing the toxicity of Aβ in AD. For example, Justicidin A, an arylnaphthalide lignin, reduces Aβ 25-35 -mediated neuronal cell death by inhibiting hyperphosphorylation of tau and inducing autophagy in SH-SY5Y cells 43 . Cilostazol stimulates CK2/SIRT1 activation, resulting in upregulation of autophagy and a decrease in Aβ expression in neurons 44 . In contrast, Hung et al. demonstrated that extracellular Aβ induces a strong autophagic response in both SH-SY5Y cells and mice that overexpress LC3 45,46 . These findings also indicate that α7nAChR binds with extracellular Aβ and the complex internalizes into the cytoplasm, subsequently inhibiting Aβ-induced neurotoxicity via autophagic degradation. We also found that Aβ 25-35 treatment transiently activated autophagic flux in SH-SY5Y cells and eliminated autophagy after 4 h. Disruption of the autophagic process led to an increase in Aβ 25-35 -induced cell death, indicating that clearance of Aβ by the autophagosome provides neuroprotection. In addition, we demonstrated that incubation of cells with cytotoxic concentrations of soluble Aβ 25-35 resulted in a rapid increase in the number of lysosomes and caused lysosomal damage after 4 h of incubation, consistent with observations previously reported 35 . We conclude that suppression of the autophagic process by longer treatments with Aβ may be a consequence of interrupted autophagosome maturation due to Aβ-induced damage of the lysosomal membrane.
Based on our cell model, using SH-SY5Y cells treated with Aβ 25-35 , we demonstrated that Aβ 25-35 rapidly increased ARRB1 expression after short-term treatment, which contributed to the activation of autophagy, and the impairment of arrb1 exacerbated Aβ-mediated cell death. In line with the reports that ARRB1 interacts with Beclin 1 and promotes activation of autophagy, deletion of arrb1, but not arrb2, aggravates neuronal injury in cerebral ischemia 30 . The findings from our study and others support a neuroprotective role of ARRB1 against neuronal injury in the regulation of autophagosome formation. We noted that mRNA levels of both arrb1 and arrb2, particularly arrb1, were markedly reduced in blood samples of AD patients (Fig. 4A). However, β-arrestin levels have been correlated with Aβ toxicity in brains of AD patients and animal models 31,32,47 . These conflicting findings concerning the expression of β-arrestins in AD indicate that β-arrestin expression varied in tissue and blood samples. Further investigation is required to validate levels of the two proteins in a large sample, and to examine the correlation of β-arrestins with AD progression. AMPK is also activated by the exposure of cells to Aβ. Thornton et al. provided evidence that Aβ 1-42 activates AMPK via the N-methyl-D-aspartate (NMDA) receptor, which in turn leads to the hyperphosphorylation of tau, a hallmark of AD 48 , although the authors did not show changes in autophagy upon the activation of AMPK in response to Aβ. It seems that the effect of Aβ on AMPK is context-dependent; under certain conditions, AMPK can be activated by Aβ, but in other contexts, AMPK is inhibited 49 . Research is still needed to clarify the role of AMPK in the autophagic process in AD. In addition to a regulatory effect of Aβ on ARRB1-induced autophagy, the expression of ARRB2 was enhanced after short treatment with Aβ. However, unlike ARRB1, genetic silencing of arrb2 reduced the toxicity of Aβ in cell culture. In further experiments, knockdown of arrb2 significantly increased the expression of the α7nAChR at the cell membrane, partially increased nicotine-mediated cytoprotective effects and rescued cells from Aβ 25-35 -induced cell death. Silencing of arrb1 slightly facilitated α7nAChR expression, and the effect of arrb2 on α7nAChR expression was more predominant in cells. To our knowledge, this is the first report demonstrating the role of the ARRB2/α7nAChR complex in Aβ 25-35 -mediated cytotoxicity. Huang et al. provided evidence that overexpression of LC3 in SH-SY5Y cells was associated with higher α7nAChR expression, which facilitated cell survival via internalization and autophagic degradation of extracellular 45,46 . Given the role of β-arrestins in the recruitment of membrane receptors into intracellular compartments, it is reasonable to assume that the membrane α7nAChR could be recycled with the help of ARRB2, but further investigation is required to define this mechanism of action. Moreover, α7nAChR protein is reduced in the cortex and hippocampus in patients with AD 50 . Increased α7nAChR expression at the cell membrane could be beneficial for AD treatment 51 . Therefore, genetically or pharmacologically targeting ARRB2 with or without the use of nicotine or nicotinic ligands may have therapeutic potential in AD treatment 36 .
In this report, we observed the changes in ARRB1 and ARRB2 expression in response to Aβ [25][26][27][28][29][30][31][32][33][34][35] and demonstrated the roles of ARRB1 and ARRB2 in Aβ 25-35 -mediated toxicity. We showed a neuroprotective role of ARRB1 in activating autophagy, which further confirms the importance of autophagy in neuroprotection. Moreover, our results indicate that therapies aimed at reducing ARRB2 may offer a promising approach for AD treatment because downregulation of ARRB2 enhances therapeutic effects mediated by α7nAChR.
Cell viability assay. Cells were seeded in 96-well plates and were treated with Aβ 25-35 for 16 h. Each test dose was performed in triplicate on each plate. The treated cells were then incubated with 10 μl MTT for 4 h at 37 °C, and the cell growth response to Aβ was detected by measuring the absorbance at 570 nm on a plate reader (Bio-Rad, USA). Three replicates were performed for each measurement.
Transient transfection of plasmids and siRNAs. Cells were transfected with an ARRB1 expression plasmid using Lipofectamine 3000 (Invitrogen Life Technologies). After 24 h of transfection, cells were exposed to Aβ for an additional 16 h and subjected to further analysis. Control cells were transfected with the empty vector pcDNA3.1 under the same conditions. For the siRNA assay, ARRB1, ARRB2, LC3B, Beclin-1 or CHRNA7 siRNAs (Invitrogen Life Technologies) were transfected into cells for 48 h. Scrambled siRNA served as a control. After transfection, cells were exposed to chemicals as indicated and subjected to the cell viability assay or lysed for the western blot assay. At least three independent experiments were performed. The siRNA sequences were as follows: sense: 5′-GAGACGCCAGUAGAUACCAAUCUCA, anti-sense: 5′-UGAGAUUGGUAUCUACUGGCGUCUC for ARRB1; sense: 5′-GACCGACUGCUGAAGAAGUTT, anti-sense: 5′-ACUUCUUCAGCAGUCGGUCTT for ARRB2; sense: 5′-GCACCUUCGAACAAAGAGUTT, anti-sense: 5′-ACUCUUUGUUDGAAGGUGCTT for LC3B;sense: 5′-UGAAAUUUCAGACCCAUCUUAUUGG, antisense: 5′-CCAAUAAGAUGGGUCU GAAAUUUCA for Beclin-1;sense: 5′-GCUGGUCAAGAACUACAAUTT, anti-sense: 5′-AUUGUAGUUCUU GACCAGCTT for CHRNA7, the negative-controlsilencing RNA was used as a control. All of the experiments were performed intriplicate wells and repeated at least three times.
Immunofluorescence. Cells were seeded on coverslips in 24-well plates. After transfection and/or treatment, cells were fixed with a mixture of methanol and acetone (1:1) and permeabilized in phosphate-buffered saline (PBS) containing 3% BSA and 0.1% Triton X-100 for 20 min. After washing with PBS, cells were probed with primary antibodies overnight. The cells were incubated with peroxidase-conjugated secondary antibodies, and images were acquired using an LSM-700 confocal fluorescence microscope (Carl Zeiss, Germany).
Flow cytometry. After transfection of ARRB1 or ARRB2 siRNAs for 48 h, cells were washed and re-suspended in PBS containing 5 mM EDTA and 0.2% BSA for 15 min. A monoclonal antibody against the nicotinic AChRα7 subunits (Abcam, UK) was added to the buffer, and the cells were incubated for 30 min on ice. The cells were then probed with FITC-conjugated secondary antibody for 30 min on ice in the dark, and analyzed on a FACScan flow cytometer (Becton Dickinson, USA).

Real-time quantitative PCR.
To evaluate changes in the mRNA levels of arrb1 and arrb2 in patients with AD, a total of 54 blood samples were collected from Chinese subjects (27 with AD and 27 healthy controls). The subjects were recruited from the Second Affiliated Hospital of Shandong University, Jinan. AD was clinically diagnosed according to the diagnosis guidelines spearheaded by the Alzheimer's disease and the National Institute on Aging (NIA) of the National Institutes of Health (NIH). Informed consent was obtained from all participants, and all the experimental protocols involving human participants were approved by the medical ethics committee of Shandong University School of Medicine and Second hospital of Shandong University. And all the experimental protocols conducted in accordance with the ethica guidelines of the Declaration of Helsinki of the World Medical Association. Total RNA was isolated from venous blood mixed with ethylenediamine Scientific RepoRts | 7: 3446 | DOI:10.1038/s41598-017-01798-x tetra-acetic acid (EDTA) utilizing the RiboPure ™ -Blood Kit (ThermoFisher) in accordance with the manufacturer's instructions. Complementary DNA was synthesized by reverse transcription using ReverTra Ace qPCR RT Kit (Toyobo, Japan). Quantitative PCR analysis of cDNA was performed using SYBR Green (Toyobo) on a real-time PCR system (Eppendorf International, Germany). The mRNA levels of the desired genes were normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH). The following primer pairs were used: arrb1 primers: 5′-AAAGGGACCCGAGTGTTCAAG-3′, 5′-CGTCACATAGACTCTCCGCT-3′; arrb2 primers: 5′-TCCATGCTCCGTCACACTG-3′, 5′-ACAGAAGGCTCGAATCTCAAAG-3′; gapdh primers: 5′-GGAGCGAGATCCCTCCAAAAT-3′, 5′-GGCTGTTGTCATACTTCTCATGG-3′.
Statistical analysis. All experiments were performed at least three independent times in triplicate. Results are expressed as mean ± standard deviation (SD). The two-tailed Student's t-test was performed to assess differences between the experimental and control groups. ARRB1, ARRB2, and cell viability under conditions of depletion and treatment were measured and presented as mean ± standard error of the mean (SEM), and values were compared by two-way analysis of variance (ANOVA). P < 0.05 was considered statistically significant; P ≤ 0.001 was considered highly significant. The fluorescence intensity of images were analyzed by Image J software.