Adiponectin controls the apoptosis and the expression of tight junction proteins in brain endothelial cells through AdipoR1 under beta amyloid toxicity

Alzheimer’s disease (AD) is the most common neurodegenerative disease, characterized by excessive beta amyloid (Aβ) deposition in brain, leading to blood–brain barrier (BBB) disruption. The mechanisms of BBB disruption in AD are still unclear, despite considerable research. The adipokine adiponectin is known to regulate various metabolic functions and reduce inflammation. Though adiponectin receptors have been reported in the brain, its role in the central nervous system has not been fully characterized. In the present study, we investigate whether adiponectin contributes to the tight junction integrity and cell death of brain endothelial cells under Aβ-induced toxicity conditions. We measured the expression of adiponectin receptors (AdipoR1 and AdipoR2) and the alteration of tight junction proteins in in vivo 5xFAD mouse brain. Moreover, we examined the production of reactive oxygen species (ROS) and the loss of tight junction proteins such as Claudin 5, ZO-1, and inflammatory signaling in in vitro brain endothelial cells (bEnd.3 cells) under Aβ toxicity. Our results showed that Acrp30 (a globular form of adiponectin) reduces the expression of proinflammatory cytokines and the expression of RAGE as Aβ transporters into brain. Moreover, we found that Acrp 30 attenuated the apoptosis and the tight junction disruption through AdipoR1-mediated NF-κB pathway in Aβ-exposed bEnd.3 cells. Thus, we suggest that adiponectin is an attractive therapeutic target for treating BBB breakdown in AD brain.

Alzheimer's disease (AD) is a neurodegenerative disease characterized by toxic plaques that consist of beta amyloid (Aβ) peptides generated from amyloid precursor protein (APP). [1][2][3] The excessive accumulation of these plaques in the brain impairs synaptic function and leads to neuronal death, ultimately manifesting in memory dysfunction. 1,4 Aβ has been known to trigger a synaptopathy of reactive oxygen species (ROS) production, neuronal cell death, glia activation, and tau hyperphosphorylation. 5 In addition to directly effecting neurons, Aβ is also known to be deposited on the walls of blood vessels and induce inflammation in endothelial cells. 6,7 A critical consequence of this is the disruption of the bloodbrain barrier (BBB) integrity 8,9 through ROS production and secretion of pro-inflammatory cytokines. 10 BBB comprises several cells including brain endothelial cells, interconnected by tight junctions consisting of the junctional adhesion molecule 1 (JAM-1), zona occludens 1 (ZO-1), occludin, and claudin. 11,12 In AD, BBB is damaged by Aβ accumulation; 13,14 its structure is changed by the disruption of tight junction proteins and the permeability of BBB is elevated during the progress of disease. 15,16 For these reasons, recent researchers have focused on understanding the BBB disruption-related mechanisms under Aβ accumulation in order to uncover effective solutions for alleviating AD pathology, [17][18][19] though a decisive target remains to be determined.
Adiponectin is a 244 amino acid polypeptide adipokine encoded by the ADIPOQ gene. 20 It binds to two receptors (AdipoR1 and AdipoR2), 21,22 which exist in the brain as well as other organs throughout the body. 23,24 Adiponectin is known to play key roles as an insulin sensitizer and an anti-inflammatory regulator, in addition to the regulation of glucose metabolism and fatty acid breakdown. 25,26 In the central nervous system, previous reports suggest that adiponectin modulates memory function and has a protective effect on neurons and neural stem cells against stress condition. 27,28 One study showed that serum adiponectin levels were lower in APP transgenic mice compared with control mice and outlined an association with inflammation and cognitive dysfunction in AD. 29 Moreover, adiponectin reduces the secretion of interleukin-6 (IL-6) from brain endothelial cells in response to oxidative stress, modulating BBB function. 30 Judging from previous evidences, adiponectin has the potential to play a cellular protective role in brain endothelial cells under Aβ-induced oxidative stress and attenuate the BBB disruption caused by Aβ accumulation in AD brain.
In the present study, we investigated whether adiponectin contributes to the apoptosis of brain endothelial cells and the loss of tight junction under Aβ toxicity condition. Our findings suggest that adiponectin may protect BBB disruption in the AD brain by alleviating the damage of brain endothelial cells caused by Aβ toxicity.

Results
The expression of adiponectin receptors was reduced in 5xFAD mouse brain. To examine the expression of adiponectin receptors (AdipoR1 and AdipoR2) in 5xFAD mouse brain, we measured the expression of adiponectin receptors through western blotting (Figures 1a and b) and immunostaining (Figures 1f and g). In 5xFAD mouse brain, the protein level of AdipoR1 was significantly reduced compared with the control mouse brain (Con) (Figure 1a). The protein level of AdipoR2 also showed a slight, albeit non-significant decrease of protein level in 5xFAD mouse brain in comparison with the normal mouse brain (Figure 1b). The immunostaining images showed considerable reduction of AdipoR1 in 5xFAD mouse brain entorhinal cortex and striatum (Figure 1f). Figure 1g presents the reduction of AdipoR2 in 5xFAD mouse brain entorhinal cortex and striatum ( Figure 1g). These data suggest that levels of adiponectin receptors (AdipoR1 and AdipoR2) are altered in 5xFAD mouse brain (Figures 1a,b,f, and g). Figure 1h shows the PSD95 (postsynaptic protein95; considered as neuron) and AdipoR1, AdipoR2 colocalization in brain (Figure 1h). Based on our results of the colocalization of PSD95 and AdipoR1 or AdipoR2, we showed the expression of AdipoR1 and AdipoR2 in neuronal cells.
The increase of NF-κB phosphorylation and the loss of tight junction protein in 5xFAD mouse brain. To examine whether AD triggers the inflammatory signaling such as NF-κB, we assess the alterations of NF-κB phosphorylation Figure 1 The expression of adiponectin receptors and the activation of NF-κB and the decreased Claudin 5 protein in 5xFAD mouse brain. Western blotting showed the protein level of AdipoR1 (a) and AdipoR2 (b) in 5xFAD mouse brain. (a) The protein level of AdipoR1 was significantly reduced in 5xFAD mouse brain compared to the control mouse brain. (b) The protein level of AdipoR2 was slightly reduced in 5xFAD mouse brain compared to the control mouse brain. (c) Western blotting showed the protein level of p-NF-κB in mouse brain. The protein level of p-NF-κB was significantly increased in 5xFAD mouse brain compared to the control mouse brain. (d) Western blotting showed the protein level of Claudin 5 in mouse brain. The protein level of Claudin 5 was significantly reduced in 5xFAD mouse brain in comparison with the control mouse brain. (e) The protein level of CD31 was slightly reduced in 5xFAD mouse brain compared to the control mouse brain. Data are expressed as mean ± S.E.M., and each experiment conducted three repeats per conditions. β-actin was used as control. Differences were considered significant at *Po0.05, **Po0.01. Images showed the expression of AdipoR1 (Green) (f) and AdipoR2 (Green) (g). 5xFAD mouse showed the less expression of AdipoR1 (f) and AdipoR2 (g). (h) PSD95-positive cells matched with the AdipoR1 and AdipoR2 staining cell in brain. Con: control normal mouse, 5xFAD: 5xFAD mouse, Scale bar: 50 μm, 4',6-diamidino-2-phenylindole (DAPI): blue, AdipoR1: adiponectin receptor 1 (Green), AdipoR2: adiponectin receptor 2, p-NF-κB: phosphorylation of NF-kB, PSD95: postsynaptic density protein 95 ( Figure 1c). Also, to check the loss of tight junction proteins ( Figure 1d) that may occur as a result of AD pathology, we conducted western blotting analysis using Claudin 5 antibody by extracting the proteins in the 5xFAD mouse brain (Figures 1c and d). Figure 1e shows the protein level of CD31 as a vascular marker (Figure 1e). Western blotting data showed that the activation of NF-κB was markedly increased in 5xFAD mouse brain in comparison with the control mouse brain (Figure 1c), suggestive of the activation of NF-κB pathway in 5xFAD mouse brain ( Figure 1c). Also, western blotting data suggested that Claudin 5 junction protein was considerably reduced in 5xFAD mouse brain (Figure 1d). This result demonstrates a relative loss of tight junction protein Claudin 5 in 5xFAD mouse brain compared with the control mouse brain (Figure 1d).
The reduction of cell viability and the increase of nitrite oxide production in Aβ-treated brain endothelial cells. To investigate the toxicity of Aβ in the brain endothelial cells, we measured cell viability in bEnd.3 cells by MTT assay (Figure 2a). The cell viability of brain endothelial cells was approximately 70% in 10 μM Aβ-treated group and 60% in 20 μM Aβ-treated group (Figure 2a). We treated 20 μM Aβ for 24 h in bEnd.3 cells to study the effect of adiponectin in brain endothelial cells against Aβ-induced stress condition. When we treated 20 μM Aβ for 24 h in bEnd.3 cells, we observed a marked increase of NO production in bEnd.3 cells. Pretreatment of Acrp 30 (as an adiponectin globular form) 31 10 μg/ml leads to the considerable decrease of NO production in bEnd.3 cells, despite Aβ-induced stress condition ( Figure 2b). In addition, we checked the mRNA level of inducible nitric oxide synthases (iNOS) (as an inducer of NO production) 32  Acrp30 rescues cell death and inhibits the production of ROS in bEnd.3 cells under Aβ-induced oxidative stress condition. To examine whether adiponectin contributes to the cell death and the production of ROS in Aβ-exposed brain endothelial cells, we checked the mRNA levels of Bax   (Figures 5c and d). Pre-treatment of Acrp 30 reversed Aβinduced decrease of AdipoR1, but not of AdipoR2 in bEnd.3 cells (Figures 5a-d).
Acrp30 changes the expression of RAGE and LRP-1 in bEnd.3 cells under Aβ toxicity. To check the expression of receptor for advanced-glycation end products (RAGE) and low-density lipoprotein receptor-related protein 1 (LRP-1) in brain endothelial cells under Aβ toxicity, we measured the mRNA levels of RAGE and LRP-1 in Aβ-treated bEnd.3 cells by quantitative real-time PCR (Figure 6). Aβ triggered the Acrp 30 protects the tight junction integrity and attenuates the inflammatory responses through AdipoR1. First, we checked the protein level of Claudin5 in AdipoR1 and AdipoR2 knockout condition (Figures 7a and b). We used appropriate non-silencing control siRNA and confirmed the efficiency of the siRNA knockdown in this study. We confirmed almost the same protein level of claudin5 in AdipoR1 and AdipoR2 knockout condition (Figure 7a). Also, we measured the protein level of Claudin 5 in AdipoR1 and AdipoR2 knockout cells under Aβ treatment (Figure 7b). We found that the protein level of Claudin 5 was not largely changed in AdipoR1 and AdipoR2 knockout condition under Aβ treatment condition (Figure 7b). To assess whether adiponectin boosts NF-κB signaling (known as the inflammatory pathway) in Aβ-exposed brain endothelial cells through AdipoR1, we used siRNA AdipoR1 for knock-down of AdipoR1 in brain endothelial cells (Figure 7c). We observed the activation of NF-κB in AdipoR1 knockdown group despite pre-treatment of Acrp 30 in Aβ-treated bEnd.3 cells (Figure 7c). In addition, to investigate whether adiponectin influences tight junction integrity through AdipoR1, we used siRNA AdipoR1 for knockdown of AdipoR1 in brain endothelial cells (Figure 7d). We observed loss of Claudin 5 by Aβ toxicity in AdipoR1 knockdown group despite Acrp 30 pretreatment (Figure 7d). When we suppress the expression of AdipoR2 using siRNA AdipoR2, we found a little bit change of Claudin 5 protein level in brain endothelial cells under Aβ with Acrp30 treatment condition (Figure 7e). Moreover, we measured the production of NO by Griess reagent assay (Figure 7f) and the production of TNF-α by ELISA assay (Figure 7g

Discussion
AD, a neurodegenerative disorder, is characterized by abnormal accumulation of Aβ and neurovascular dysfunction. 33 The excessive Aβ deposition in brain endothelial cells aggravates the increase of BBB permeability by impairing BBB transport systems in the AD brain. 14,34 Adiponectin acts by binding with specific receptor AdipoR1 and AdipoR2, 21 which exist in various organs including brain. 23,24 Considering that our results showed the AdipoR1 and AdipoR2 detected cells were neuronal cells in brain, we assume that the adiponectin receptors could be involved in the excitability of neurons 35 and the suppression of neuronal damage against oxidative stress. 36 Also, adiponectin has been reported that it is present in the cerebrospinal fluid (CSF) of rodents 37,38 and humans, 39,40 although the concentration of it in CSF is less than the concentration of it in plasma and is controversial whether or not it could cross the BBB. 30 Recent studies suggested the protective potential of adiponectin on BBB breakdown in AD, 29,30 but the specific mechanisms remained to be fully characterized. In the present study, we found a reduction of AdipoR1 and AdipoR2 expression and the loss of tight junction protein Claudin 5 in models of AD pathology. These effects were concomitant with the activation of NF-κB. Taken together, these data suggest that decreased expression of adiponectin receptors in the AD brain may be associated with BBB disruption and activation of the inflammatory response.
Excessive production of ROS can cause severe cell damage 41 and can also increase vascular endothelial permeability and leukocyte adhesions. 16 In addition, a large amount of ROS triggers loss of endothelial cell interactions 42 and affects BBB integrity by disturbing tight junctions. 43,44 One study showed that adiponectin prevents the inflammation of vascular endothelial cells by reducing the secretion of ROS. 45 Considering previous results and our in vitro results, we suggest that adiponectin may suppress Aβ toxicity-induced Previous study demonstrated that Aβ increases the expression of iNOS gene. 46 iNOS produces NO and triggers the inflammatory response. 47 Excessive production of NO in endothelial cells promotes pro-inflammatory signaling and the process of APP. 48,49 Adiponectin has been known to control the production of cytokines by regulating the expression of eNOS. 50 In addition, adiponectin inhibits fibroblast migration through AdipoR1-AMPK-iNOS pathway in inflammatory condition. 51 Based on our results, we infer that adiponectin contributes to the expression of cytokines and inflammation signaling by blocking the production of NO against Aβ-induced toxicity. Several studies demonstrated that adiponectin reduces the production of IL-6, 30 IL-8, vascular endothelial growth factor and matrix metalloproteinases (MMPs) in endothelial cells. 52 In addition, the overexpression of adiponectin receptors promotes the anti-inflammatory response in vascular endothelial cells. 53 Also, adiponectin inhibits vascular endothelial hyperpermeability through cAMP/PKA signaling. 54 Regarding our results, we suggest that adiponectin reduces the expression of pro-inflammatory cytokines including IL-6, TNF-α and MCP-1 in brain endothelial cells under Aβ-induced oxidative stress conditions.
In AD, the accumulation of Aβ in endothelial vessel walls leads to endothelial dysfunction 55 and impaired BBB integrity. 56 The loss of tight junction proteins such as occludin, ZO-1, and Claudin 5 by Aβ aggravates the increase of barrier permeability and apoptosis of vascular endothelial cells. 57,58 The alteration of tight junction and ROS production causes BBB leakage in AD brains. 59 Recent study reported that the activation of NF-κB by Aβ accumulation disrupts the integrity of BBB by decreasing Claudin 5 and increasing receptor for RAGE. 60 In the present study, we found that adiponectin suppresses the loss of tight junction proteins and the increase of RAGE expression in brain endothelial cells under Aβinduced toxicity.
In the AD brain, Aβ could be transported into the brain across the BBB and it is regulated by the BBB receptors and transporters. 61,62 Aβ accumulation in AD brain results from decreased clearance from BBB and increase of uptake from the circulatory system. 61,62 RAGE transports Aβ from the circulation into the brain, 63,64 whereas LRP-1 is related with rapid clearance of Aβ from the brain to blood. [65][66][67][68] The dysfunction of LRP-1 weakens the ability of BBB to clearing Aβ. 67 Several clinical studies have demonstrated that the onset and procession of AD is associated with low LRP-1 levels and high RAGE levels, leading to the accumulation of Aβ peptides in the brain parenchyma. 69,70 One clinical study has observed that the expression of LRP-1 is reduced and the expression of RAGE is increased in AD patients. 71 According to the current study, 5xFAD mice have decreased levels of LRP-1 receptor and lower levels of Aβ in plasma, with an increase of Aβ in the brain. 72 Aβ oligomers are known to cause upregulation of RAGE expression in endothelial cells. 73 This may have important consequences given that RAGE promotes the expression of MMP-2 related with permeability. 74,75 One study has reported that Aβ triggers the increase of permeability and the disruption of tight junction protein ZO-1 and the secretion of MMP in brain endothelial cells by interacting with RAGE. 15,34 In recent years, an inverse relationship was found to exist between the expression of adiponectin and RAGE. 76,77 Given that our results showed low mRNA level of RAGE and high mRNA level of LRP-1 in Aβexposed bEnd.3 cells by Acrp 30 treatment, we assume that adiponectin contributes to the transportation of Aβ into brain and the clearance of Aβ by enhancing the expression of LRP-1 and suppressing the level of RAGE in brain endothelial cells.
Even though we did not find a direct mechanistic connection between adiponectin and Aβ-RAGE or Aβ-LRP-1 interaction in this study, we highlight the necessity of further study concerning the action of adiponectin related with Aβ transporters.
Moreover, previous studies demonstrated that adiponectin negatively regulates the production of pro-inflammatory cytokines including MCP-1, IL-6 in endothelial cells by modulating the activation of NF-κB. 78,79 Based on our results, we assume that adiponectin may regulate the expression of pro-inflammatory cytokines and inflammatory responses by inhibiting the activation of NF-κB in brain endothelial cells under Aβ toxicity. In particular, we found that AdipoR1 mediates the effect of adiponectin such as the suppression of NF-κB phosphorylation and the protection of tight junction protein loss.
Taken together, we assume that adiponectin may alleviate the BBB disruption in AD (1) by inhibiting apoptosis of brain endothelial cells, (2) by protecting tight junction integrity, and (3) by mediating the balance of Aβ transporters in brain endothelial cells. Furthermore, AdiopoR1 may be crucial in the inflammatory response through NF-κB and the loss of tight junction protein in brain endothelial cells. Hence, we suggest that the action of adiponectin through AdipoR1 may alleviate AD pathogenesis by protecting BBB disruption.
Aβ oligomer preparation. Oligomeric Aβ were prepared following previous report. 81 Synthetic Aβ 42 peptide (American Peptide, Sunnyvale, CA, USA) was dissolved to 1 mM hexafluoroisopropanol (Sigma-Aldrich). The solution was evaporated for 2 h and subsequently peptide film was resuspended in dimethyl sulfoxide (DMSO, Sigma-Aldrich) to produce a 1 mM solution. Later, to prepare oligomeric Aβ 42, this solution was diluted to 100 μM in DMSO medium (Gibco) and incubated for 12 h at 4°C.
Reverse transcription PCR. RNA in bEnd.3 cells was isolated using Trizol Reagent (Gibco) following the manufacturer's instructions. RT-PCR reaction was performed by using Invitrogen One step III Reverse Transcription PCR kit (Invitrogen). cDNA synthesis from mRNA and sample normalization were performed. PCR was performed using the following thermal cycling conditions: 95°C for 10 min; 40 cycles of denaturing at 95°C for 15 s, annealing at 58°C for 30 s, elongation at 72°C for 30 s; final extension at 72°C for 5 min; and holding at 4°C. PCR was performed using the following primers (5′ to 3′); Bcl2 (F): TACCGTCGTGACTTCGCAGAG, (R): GGCAGGCTGAGCAGGGTCTT, GAPDH (F): GAC AAG CTT CCC GTT CTC AG, (R): GAG TCA ACG GAT TTG GTC GT. PCR products were electrophoresed in 1% agarose gels and stained with cyber green. Each sample was normalized with GAPDH.
Immunohistochemistry. Brain sections were cut (20 μm) onto coated glass slides (Thermo Scientific, Waltham, MA, USA), and fixed in acetone for 20 min at − 20°C. The slides were first washed in TBS and then incubated with methanol. To block nonspecific labeling, sections were incubated in 5% bovine serum albumin (Sigma-Aldrich) diluted in PBS for 1 h before incubation with primary and secondary antibodies. Primary antibodies for AdipoR1 (1:500, Abcam), AdipoR2 (1:500, Abcam), postsynaptic density protein 95 (PSD95) (1:500, Cell Signaling) were applied to the samples for 16 h at 4°C, followed by 1 h incubation with appropriate florescence secondary antibody (1:500, Invitrogen) and three times washes in PBS for 5 min each. After three washes in 0.1 % PBS with Tween-20, the sections were incubated with each secondary antibody for 1 h in the dark at room temperature. Later, all sections were incubated with 1 μg/ml DAPI (Sigma-Aldrich) for a counter staining. Brain tissues were then visualized under a confocal microscope (Carl Zeiss, Oberkochen, Germany).