Amyloid β Interaction with Receptor for Advanced Glycation End Products Up-Regulates Brain Endothelial CCR5 Expression and Promotes T Cells Crossing the Blood-Brain Barrier¹

Alzheimer’s disease (AD)⁴ is the most common form of age-related dementia. AD is characterized by the progressive deposition of amyloid β (Aβ) protein in the limbic and association cortices, where it accumulates to form highly aggregated and compacted (fibrillar) extracellular plaques (1, 2).

Extracellular deposits of highly aggregated Aβ fibrils trigger inflammatory responses that may play an important role in AD pathogenesis (3, 4). The local innate immunity of the CNS has been shown to be associated with the Aβ deposition and plaque formation (4, 5, 6). This innate immunity includes the activation of resident brain cells such as microglia and astrocytes, and involvement of a broad variety of inflammation-related proteins, such as complement factors, acute-phase proteins, proinflammatory cytokines, and chemokines. The local inflammatory response is thought to be responsible for the clearance of Aβ (4, 5). If microglial or astrocytic activation fails to clear the toxic forms of Aβ, the innate immune response will become chronic and neurotoxic (4, 5, 6).

Evidence demonstrated the ability of the immune system to generate Abs after immunization against Aβ that may promote removal of Aβ from the brain (7, 8, 9, 10, 11). Besides the Ab-mediated systemic adaptive response, increasing studies suggest that the T cell immune response may be involved in the inflammation process of AD (12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26). In vitro data showed a cross-talking between microglia and T cells; microglia can serve as APCs for Aβ-reactive T cells and, in turn, T cells themselves can influence microglial differentiation (12, 13, 14, 15). A number of reports suggest that peripheral T cells are activated in AD patients (16, 17, 18, 19, 20, 21, 22, 23). Besides periphery, activated T cells also exist as infiltrates in the brain of AD (19, 24, 25, 26). Particularly, increased occurrence of T cells was found in the brains of AD patients compared with subjects with non-AD degenerative dementias and aged-matched controls (26). However, how circulating T cells of AD patients penetrate the blood-brain barrier (BBB) that mainly consists of endothelial cells with tight junctions is not clear.

Recently, we found that peripheral T cells derived from AD patients overexpress CXCR2 and MIP-1α to enhance its transendothelial migration (TEM) (27, 28). Our previous study revealed that the MIP-1α interacts with CCR5 on brain endothelial cells to induce tight junction opening for T cell TEM, and that CCR5 expression in human brain microvascular endothelial cells (HBMECs) was up-regulated by Aβ in a time- and dose-dependent manner (28). The aim of this study is to explore the mechanism of CCR5 up-regulation by Aβ. Οur results show that Aβ up-regulates CCR5 expression and promotes T cell TEM via receptor for advanced glycation end products (RAGE) on brain endothelial cells.

Human brain microvascular endothelial cells were a gift from K. S. Kim (Johns Hopkins University, Baltimore, MD). They were cultured in RPMI 1640 medium, and supplemented with 10% FBS (HyClone), 10% Nu serum (BD Biosciences), 2 mM glutamine, 1 mM sodium pyruvate, 1× nonessential amino acid, and 1× MEM vitamin. Jurkat (ATCC TIB 152, human acute T cell leukemia) and 6T-CEM (ATCC CRL 8296, human acute lymphocytic leukemia) cell lines were maintained in RPMI 1640 medium, supplemented with 10% FBS. All cells were incubated at 37°C in a 5% CO₂/95% air humidified atmosphere.

Subjects with AD were recruited from the First and Second Affiliated Hospital, China Medical University, under an International Review Board-approved human studies protocol. The patients with AD have no immunological diseases and vascular risk factors such as hypertension, cardiac disease, and diabetes. The diagnosis of AD was based on National Institute of Neurological Disorders and Stroke-Alzheimer Disease and Related Disorders Association criteria (29) and included use of the Mini Mental State Exam (30). None of the subjects had experienced infection or taken immunoregulation drugs during the period of 6 mo before sample collection. Overall, 19 patients with AD (6 males and 13 females, aged 64–90 years, mean age 78.4 ± 8.3) were evaluated in this study. The peripheral T cells were separated with Fluorobeads Isolation Reagent (One Lambda). Above studies have been reviewed and approved by China Medical University Review Committee.

The total RNA was treated with RNase-free DNase I (Takara Bio) and reverse transcribed with avian myeloblastosis virus reverse transcriptase (Promega). Real-time PCR was performed on ABI 7500 real-time PCR system (Applied BioSystems) with an Ex Taq R-PCR (Takara Bio), according to the manufacturer’s protocol. The sequences of probe and primers for human CCR5 are CTGAACTTCTCCCCGACAAAGGCA (probe), GACGCACTGCTGCATCAA (forward), and CATTTGCAGAAGCGTTTGG (reverse). The probe and primers for human GAPDH are described previously (27). The amplification conditions were as follows: 95°C for 10 s, and 40 cycles of 95°C for 5 s and 64°C for 34 s. Recombinant pGEM T-vectors (Promega) containing CCR5 or GAPDH cDNA were constructed and used to establish the standard curves. The amount of CCR5 transcripts of individual samples was normalized to GAPDH.

Cells were washed twice with ice-cold PBS and prepared with RIPA buffer (50 mM Tris-HCl, 150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholate, and 0.1% SDS) containing protease inhibitor mixture (Roche). The samples were separated by SDS-PAGE, and then transferred to the polyvinylidene difluoride membrane (Millipore) using SemiDry Transfer Cell (Bio-Rad). The polyvinylidene difluoride membrane was blocked with 5% nonfat milk and incubated with the first Ab at 4°C overnight. The blots were incubated with a HRP-conjugated secondary Ab (Santa Cruz Biotechnology) for 1 h at room temperature. Immunoreactive bands were visualized by the SuperSignal West Pico chemiluminescent substrate (Pierce) using LAS3000mini (Fuji Film). The Abs against AKT/p-AKT, ERK/p-ERK, JNK/p-JNK, and P38/p-P38 were from Cell Signaling Technology. CCR5, RAGE, Egr-1, and GAPDH Abs were obtained from Santa Cruz Biotechnology.

The CCR5Δ32 associated with a loss of function is a 32-nt deletion in the open reading frame (31). The CCR5Δ32 was generated by ligation of two amplified fragments at PstI sites. The primer sequences of the first fragment (1–553) were AAGCTTTCATGGATTATCAAGTGTCAAGTC (forward) and CTGCAGGTGTAATGAAGACCT (reverse). Primers for the second fragment (531–1056) were CTGCAGCTCTCATTTTCCATACATTAAAGATAGTCATCTTGGGGCT (forward) and CTGCAGCTCGAGCAAGCCCACAGATATT (reverse). CCR5Δ32 was subcloned into pcDNA3.1/myc-his A (Invitrogen) vector using HindIII and XhoI. The full-length RAGE cDNA was obtained by RT-PCR from HBMECs, and was cloned into pcDNA3.1/myc-his B (Invitrogen) using HindIII and XbaI. N-terminal truncated RAGE (351–1212) cDNA and C-terminal truncated RAGE cDNA (1–1080) were obtained by PCR from full-length RAGE. Primer sequences for N-terminal truncated RAGE were CCGGAATTCCACCATGGTCTACCAGATTCCTGGGAA (forward) and CCGACTCGAGCGAGGCCCTCCAGTACTCT (reverse); primer sequences for C-terminal truncated RAGE were CCGGAATTCCACCATGGCAGCCGGAACAGCAGT (forward) and CCGACTCGAGCGCCACAAGATGACCCCATGA (reverse). Both kinds of cDNA fragments were cloned into pcDNA3.1/myc-his B using EcoRI and XhoI. The full-length cDNA of PI3Kγ was a gift from K. Wu (University of Texas Medical School, Houston, TX). The dominant-negative PI3Kγ (p110γ Δ948–981) (32) was generated by PCR overlap extension method, as described (33), and subcloned into BamHI/XhoI sites of pcDNA3.1/Myc-his A vector. The primers of the first fragment (1–2964) are GCGGATCCACATGGAGCTGGAGAACTATAAAC (forward) and GGTTAGCACAAATGGCACTCTTCTGTCGCCTATTCCAAGAAC (reverse). The primers of the second fragment (2821–3309) are GTTCTTGGAATAGGCGACAGAAGAGTGCCATTTGTGCTAACC (forward) and CTCGAGGGCTGAATGTTTCTCTCCTT (reverse). All the cDNAs were analyzed by DNA sequencing to ensure there were no other mutations.

HBMECs were transfected with the constructed plasmids using Lipofectamine 2000 (Invitrogen) and selected in the presence of G418 (300 μg/ml; Invitrogen). Single clone cells were selected and confirmed by Western blot analysis.

The small interfering RNA (siRNA) sequence targeting human Egr-1 corresponded to the coding regions, TCAGTGGCCTAGTGAGCATGA (743–763), was built by GenScript’s siRNA design center (http://www.genscript.com/rnai.html). A nonsilencing siRNA sequence (TTCTCCGAACGTGTCACGT) (27) was used as control. The siRNA sequences were inserted into pRNA-U6.1/Neo vector (GenScript). Recombinant siRNA plasmids were transfected into HBMECs, and single clones were selected by G418.

For ERK and JNK, RNAi experiments were performed according to the recent publications (34, 35).

ChIP-IT chromatin immunoprecipitation kit was from Active Motif. Aβ_1–42 and reverse peptide Aβ_42–1 were purchased from AnaSpec. The Aβ peptides were prepared, as described (36), by dissolving the peptide first in 35% acetonitrile/0.1% trifluoroacetic acid, diluted to 2 mM with sterile water, and then to 1 mM through the dropwise addition of PBS. The Aβ solution was treated at 37°C for 24 h to prepare fibrillar aggregates. HBMECs (4.5 × 10⁷) were treated by cell signaling inhibitors, including LY294002, PD98059, or SP600125 (Calbiochem) for 30 min, and then were treated by Aβ_1–42 for 30 min. The cells were chemically cross-linked thereafter with 1% formaldehyde. ChIP analysis was performed, according to the manufacturer’s protocols, using Egr-1 Ab. PCR was done to amplify CCR5 promoter region (from −777 to −564) using primers (CTCGATGATTCGCTTGTCCTT, forward; GATATGACAAGACCGAAAAGGGTTC, reverse). The PCR products were analyzed by PAG electrophoresis.

HBMECs (2 × 10⁵) were seeded on the upper chamber of Transwell insert with 5 μm pore size (Corning-Costar) in 24-well plates. The integrity of the HBMEC monolayer was monitored by daily transendothelial electrical resistance (TEER) measurements, using a Millicell-ERS endothelia volt-ohmmeter (World Precision Instruments). Experiments were conducted 4 days after plating when TEER was >200 Ω × cm². T cells (5 × 10⁵) from AD patients or 6T-CEM cells were loaded into the upper chamber of the Transwell insert; Aβ was added to the lower chamber. For blocking experiments, the T cells were pretreated with neutralization Ab against MIP-1α (R&D Systems), CCL5 (R&D Systems), and CCR5 (BD Pharmingen) before their application to the upper chamber. In some experiments, anti-RAGE neutralization Ab (R&D Systems) was added to the lower chamber. After incubation for 20 h, the cells that had transmigrated into the lower chamber were harvested and counted in a hemocytometer.

The cDNA of rat sRAGE was obtained by chemical synthesis in Shanghai RealGene Bio-tech, and was cloned into pFastBac1 with BamHI and XhoI. The construct was transposed into Escherichia coli DH10 Bac. Recombinant Bacmid was transfected into Sf9 cells using Cellfectin Reagent (Invitrogen). The conditioned medium was dialyzed against 50 mM Tris-HCl, 300 mM NaCl, and 5% glycerol (pH 6.5) at 4°C overnight and loaded on the nickel column, and the protein was eluted by 0–250 mM imidazole (50 mM Tris-HCl, 5% glycerol (pH 6.5)). The eluted protein was dialyzed in 50 mM Tris-HCl, 5% glycerol (pH 7.5), and then loaded on a SP Sepharose fast-flow column (Pharmacia) and eluted with a salt gradient (0.1–0.5 M NaCl in 50 mM Tris-HCl and 5% glycerol (pH 6.5)). The fractions from SP Sepharose fast-flow column were combined and dialyzed in 50 mM Tris-HCl and 5% glycerol (pH 7.5), and loaded on heparin column. HPLC analysis was used to detect the purity of the protein.

Wistar rats (male, 250–300 g) were obtained from the Lab Animal Center of China Medical University. The rats were anesthetized with chloral hydrate (30 mg/kg; Sigma-Aldrich) and mounted in a small-animal stereotaxic instrument. Aβ_1–42 or reverse peptide Aβ_42–1 was prepared, as described above. A total of 2.5 μl of Aβ_1–42 solution (0.5 mM in PBS) was injected stereotaxically into CA1 region of the right lateral hippocampus (posterior from bregma, 3.0 mm; lateral, 2.2 mm; and depth, 2.8 mm). To block RAGE in rat brain endothelial cells, sRAGE intracerebroventricular injection (i.c.v.) was performed, as described (27, 37). Rats were anesthetized and stereotaxically implanted with a cannula aimed at the left lateral ventricle (posterior from bregma, 1.0 mm; lateral, 1.2 mm; and depth, 3.5 mm). The cannula was fixed to the skull with dental cement. At the end of the surgical procedure, a dummy cannula was inserted into the cannula to prevent blockage. The rats were then allowed to recover from surgery for 5 days, and injected Aβ stereotaxically into CA1 region of right hippocampus, as above. The experimental groups were as follows: group controls, with 2.5 μl of PBS or Aβ_42–1 injected into the hippocampus; group Aβ, with 2.5 μl of Aβ_1–42; group Aβ plus sRAGE, with 5 μl of sRAGE injected i.c.v., using a microsyringe per day after Aβ_1–42 injection; and group Aβ plus rat serum albumin (RSA), with 5 μl of RSA injected i.c.v. per day after Aβ_1–42 injection. For the sake of comparison, the rats of group controls and group Aβ were injected i.c.v. with 5 μl of PBS per day similarly. The rats were maintained for 10 days, and administered as above. Cardiac perfusion was performed with 4% paraformaldehyde, and then the brains were postfixed in 4% paraformaldehyde and cryosections were prepared at 8 μm. All the animal studies have been reviewed and approved by China Medical University Review Committee.

T cells in brain sections were identified using anti-CD3 mAb (BD Bioscience Pharminogen), followed by incubation with rhodamine-conjugated secondary Ab. Counterstaining was performed with 4′,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich). Double staining of the brain cryosections was conducted using anti-rat CCR5 Ab (Santa Cruz Biotechnology) and anti-rat RECA-1 Ab (AbD Serotec), and then incubated with FITC-conjugated and rhodamine-conjugated secondary Ab (Santa Cruz Biotechnology). Counterstaining was performed with DAPI. Slides were visualized with an immunofluorescence microscopy (Olympus BX51).

Statistical significance between groups was analyzed by Student’s t test and ANOVA using SPSS software. Two-way ANOVA was used to compare multiple groups. Pairwise comparisons were performed using the Tukey method. Differences were considered statistically significant at <0.05.

We previously found that CCR5 expression in HBMECs was up-regulated by Aβ in a time- and dose-dependent manner (28), but the mechanism is not known. To determine the signal pathway of Aβ-dependent CCR5 expression, we first examined the effect of pharmacological inhibitors of MAPK and PI3K on CCR5 expression. As shown in Fig. 1,A, pretreatment of HBMECs with SP600125, an inhibitor of JNK, or with U0126, a specific inhibitor of MEK (MAPK/ERK kinase), significantly reduced the CCR5 expression in HBMECs. However, CCR5 expression was not changed when HBMECs were pretreated with SB203580, an inhibitor of P38 MAPK. In addition, pretreatment of HBMECs with LY294002, an inhibitor of PI3K, significantly reduced the CCR5 expression. Subsequently, we examined whether Aβ treatment could activate ERK, JNK, and PI3K in HBMECs. The phosphorylation state of AKT was used to check the activation of PI3K. Our results showed that Aβ could activate ERK, JNK, and PI3K, but it cannot activate P38 (Fig. 1, B–E). To further confirm the role of ERK and JNK in Aβ-dependent CCR5 expression, ERK and JNK siRNA duplexes were transfected into HBMECs, and the results showed that Aβ failed to up-regulate CCR5 expression when ERK and JNK were knocked down (Fig. 1, F–I). We also constructed the HBMEC cell line overexpressing dominant-negative PI3K, named as ΔP110 (Fig. 1,J). As shown in Fig. 1 K, the Aβ-induced CCR5 expression was abrogated in ΔP110 cell line. These results suggested that ERK, JNK, and PI3K activation were involved in Aβ-induced CCR5 expression in HBMECs.

To explore the mechanism involved in CCR5 protein up-regulation in HBMECs, real-time RT-PCR was done to analyze the level of CCR5 mRNA expression in Aβ-treated HBMECs. As shown in Fig. 2, A and B, the level of CCR5 mRNA in HBMECs was increased by Aβ in a time- and dose-dependent manner. Giri et al. (38) had reported the involvement of ERK, JNK, and Egr-1 in Aβ-induced CCR5 up-regulation in THP-1 cell line (monocytic cells). Therefore, we performed ChIP assay to examine whether Egr-1 is associated with Aβ-induced CCR5 expression in HBMECs. Our results showed that Aβ could enhance the binding of Egr-1 to CCR5 promoter in HBMECs (Fig. 2,C). This increased binding of Egr-1 to CCR5 promoter was abolished by ERK inhibitor (PD98059), JNK inhibitor (SP600125), or PI3K inhibitor (LY294002). To further confirm the role of ERK, JNK, PI3K, and Egr-1 in Aβ-induced CCR5 expression, we constructed the Egr-1 knockdown HBMEC cell line, and found that Aβ-induced CCR5 up-regulation was abolished (Fig. 3, A and B), but the Aβ-induced ERK, JNK, and PI3K activation were not affected (Fig. 3, C–E). These results indicated that Egr-1 is a downstream effector of ERK and JNK, and PI3K contributed to Aβ-induced CCR5 up-regulation in HBMECs.

RAGE, a member of the Ig superfamily, is a cell surface receptor expressed by many cell types, including cerebral endothelial cells (39, 40, 41). Evidence showed that Aβ could interact with RAGE (39, 40, 41). Additional studies suggested that ERK and JNK are downstream signal molecules coupled to RAGE (39). There are also reports about involvement of PI3K in intracellular RAGE signaling (42, 43). Therefore, we tried to explore whether RAGE is involved in Aβ-dependent CCR5 expression in HBMECs. We found that RAGE expression in HBMECs was increased by Aβ treatment in a time- and dose-dependent manner (Fig. 4, A and B). Neutralizing experiment showed that RAGE Ab could abrogate Aβ-induced CCR5 expression (Fig. 4^,C). It is clear that RAGE has an extracellular region containing one V-type domain and two C-type domains (42, 44, 45, 46). N-terminal V-type domain is essential for ligand binding, and N-terminal truncated RAGE overexpression in ECV304 cells (a HUVEC line) leads to attenuated function of RAGE (47). The C-terminal cytoplasmic domain of RAGE is essential for intracellular signaling (42, 44, 45, 46). Low-level expression of RAGE variant proteins, RAGE, N-terminal truncated RAGE, and C-terminal truncated RAGE was detected in HBMECs (supplemental Fig. S1).⁵ To further confirm the role of RAGE in Aβ-up-regulated CCR5 expression, we established three stable HBMEC cell lines expressing full-length RAGE, N-terminal truncated RAGE, and C-terminal truncated RAGE, respectively. The stable cell lines were identified by Western blot (Fig. 4,D). The subsequent results showed that overexpression of full-length RAGE in HBMECs strengthened Aβ-induced CCR5 expression (Fig. 4,E), whereas overexpression of N-terminal or C-terminal truncated RAGE in HBMECs abrogated Aβ-induced CCR5 up-regulation (Fig. 4, F and G). These data indicated that RAGE is required for Aβ-induced CCR5 expression in HBMECs.

Our results demonstrated that RAGE is involved in Aβ-induced CCR5 expression in HBMECs. To identify whether RAGE is associated with the activation of ERK, JNK, and PI3K contributing to CCR5 expression in HBMECs, we administered RAGE neutralization Ab to HBMECs, and found that RAGE Ab significantly inhibited the activation of ERK, JNK, and PI3K induced by Aβ (Fig. 5, A–C). In addition, Aβ failed to activate ERK, JNK, and PI3K in stable HBMEC cell line expressing N-terminal truncated RAGE or C-terminal truncated RAGE (Fig. 5, D–I). All of these data indicated that the RAGE of brain endothelial cells is a receptor for Aβ to induce CCR5 expression in HBMECs.

The HBMEC monolayer cultured on Transwell insert has been broadly used as an in vitro BBB model (28, 48, 49). Based on the BBB model, we previously showed that Aβ-induced CCR5 expression was related to T cells crossing the HBMEC monolayer (28). In this study, to mimic AD brain’s microenvironment, the Aβ-containing in vitro BBB model was constructed. The results showed that the migration rate of AD patients’ T cells had an obvious elevation in BBB model when Aβ was added to the lower chamber (Fig. 6,A), whereas the overexpression of dominant-negative CCR5 (CCR5Δ32) in HBMECs significantly inhibited T cell TEM (Fig. 6, B and C). Our previous studies showed that MIP-1α, a CCR5 ligand, is overexpressed in AD patients’ T cells contributed to T cell TEM (28). It is known that CCR5 and CCL5, another CCR5 ligand, are expressed on T cells (50). To demonstrate the role of MIP-1α, CCL5, and CCR5 in T cell migration through the Aβ-containing in vitro BBB model, T cells were pretreated with neutralizing Abs against MIP-1α, CCL5, and CCR5, respectively, before their application to the upper chamber. The results showed that T cell TEM could be blocked by the Ab against MIP-1α, but not CCL5 or CCR5 Ab (Fig. 6 D). These results further indicated that the MIP-1α in T cells and the CCR5 on HBMECs are essential for AD patients’ T cell TEM. Subsequently, we examined whether the signaling pathway of Aβ-induced brain endothelial CCR5 expression is associated with T cell migration through HBMEC monolayer.

MIP-1α is expressed in 6T-CEM cells similar to AD patients’ T cells (28); thus, the 6T-CEM cells were used in TEM assay in parallel with AD patients’ T cells. As shown in Fig. 7, A and B, RAGE neutralization Ab effectively blocked both T cell migration through the in vitro BBB model. Furthermore, three HBMEC cell lines, overexpressing full-length RAGE, N-terminal truncated RAGE, and C-terminal truncated RAGE, were, respectively, seeded in Transwell inserts for TEM assay. The results showed that full-length RAGE overexpression in HBMECs increased AD patients’ T cell migration through the HBMEC monolayer (Fig. 7,C). Unexpectedly, the TEM of 6T-CEM cells through RAGE-overexpressed HBMECs was not coincident with that of AD patients’ T cells (Fig. 7,D), which may be due to the difference between the two cell types. However, Aβ failed to promote T cell TEM when N-terminal truncated RAGE or C-terminal truncated RAGE was overexpressed in HBMECs (Fig. 7, E–H). These data indicated that RAGE receptor on HBMECs is associated with T cell TEM in AD.

Next, we examined the effect of ERK, JNK, and PI3K on T cell TEM. HBMECs cultured on Transwell insert were transfected with ERK or JNK siRNA duplexes, and the experiments of T cell TEM were conducted when TEER was >200 Ω × cm². As shown in Fig. 8, A–D, ERK and JNK RNAi abolished T cell TEM induced by Aβ. Subsequently, we seeded HBMECs overexpressing dominant-negative PI3K in Transwell insert, and found that Aβ-promoted T cell TEM was blocked (Fig. 8, E and F). To further examine the effect of Egr-1 on T cell TEM, Egr-1 knockdown HBMECs were seeded in Transwell insert, and the results showed that Aβ-induced TEM of T cells was abrogated (Fig. 8, G and H). These results suggested that the ERK/JNK/PI3K-Egr-1 signaling pathway is required for AD patients’ T cell TEM.

Aβ injection in rat hippocampus has been used to study the pathogenesis of AD (51). Our previous study has shown that the Aβ injection in rat hippocampus could induce higher expression of CCR5 and T cell TEM (28). To further confirm the role of RAGE on brain endothelial cells in Aβ-induced CCR5 expression and T cell TEM, we tried to study the effects of blockage of intracerebral RAGE on T cell infiltration in the brain primed by Aβ. It is well documented that sRAGE is cleaved from cell surface RAGE by matrix metalloproteinase, which is the extracellular ligand-binding domain of RAGE. sRAGE binds with ligands of RAGE to block their interaction with and activation of RAGE (39, 42). To corroborate the role of RAGE in Aβ-primed T cell TEM in AD, rat sRAGE was expressed and purified (Fig. 9,A), and the sRAGE was injected into the ventricle of rats with Aβ deposition in hippocampus. Immunofluorescence showed that CCR5 was seldom expressed in brain endothelial cells of the rats injected with sRAGE compared with controls (Fig. 9,B). More importantly, the number of T cells in the brain parenchyma of sRAGE-injected rats was significantly less than that of controls (Fig. 9, C and D). These results indicated that RAGE may be required for T cell infiltration in the brain in AD.

Circulating immune cells, including T cells, enter the CNS as part of normal immune surveillance (52, 53). In AD patients, activated T cells were found in the brain parenchyma (19, 24, 25, 26); however, little was known about the mechanism. In this study, we showed that intracerebral Aβ interaction with RAGE on brain endothelial cells up-regulated CCR5 expression and triggered T cell infiltration in the brain. Our results suggest for the first time that the intracerebral Aβ deposition signal can be transmitted to systemic immune T cells by RAGE at BBB.

CCR5 is a seven-transmembrane G protein-coupled receptor, which is selectively expressed on different leukocytes (54). In CNS, CCR5 expression has been identified in microglia and brain endothelial cells (28, 55, 56, 57, 58, 59). We and others had shown that CCR5 was normally expressed on brain endothelial cells at lower level (28, 57). Moreover, we had found that CCR5 expression in HBMECs was up-regulated by Aβ in a time- and dose-dependent manner (28). In this study, we explored the mechanism and significance of Aβ-induced CCR5 up-regulation in brain endothelial cells.

In the peripheral blood monocytes and human monocytic THP-1 cell line, it has been reported that activation of ERK and JNK is involved in Aβ-induced CCR5 expression and that Aβ can increase DNA-binding activity of Egr-1 to CCR5 promoter (38). However, the mechanism of Aβ-induced CCR5 expression in brain endothelial cells remains unknown. In the present study, we demonstrated that Aβ-induced ERK and JNK activation leads to increased Egr-1 binding with CCR5 promoter and up-regulated CCR5 expression in HBMECs. Meanwhile, our results also revealed the involvement of PI3K in Egr-1 activation, causing CCR5 up-regulation in brain endothelial cells.

Aβ-triggered PI3K activation has been implicated in the expression of chemokines in microglia (60, 61). However, there is no report about the PI3K signaling pathway coupled to chemokine receptor CCR5 expression. In the present study, we showed that PI3K inhibitor and overexpression of dominant-negative PI3K in HBMECs abolished Aβ-induced brain endothelial CCR5 expression. This suggests that Aβ-induced CCR5 expression in the brain endothelial cells requires intracellular PI3K activation. It is known that the nuclear Egr-1 protein is a downstream signaling molecule of PI3K; however, PI3K activation showed opposing effects on Egr-1 expression in different situations (62, 63, 64, 65). In this study, we showed for the first time that Aβ-triggered endothelial PI3K activation promoted Egr-1 binding with CCR5 promoter to induce CCR5 up-regulation, and we also found that PI3K activation up-regulated the expression of Egr-1 in HBMECs (data not shown). Interestingly, Aβ treatment could activate ERK, JNK, and PI3K in HBMECs, but the peak times of ERK, JNK, and PI3K were 15, 5, and 15 min, respectively (Fig. 1, B–D). The control experiments (vector) in Fig. 5, D–I, also showed the similar results. These results indicated that the activation of JNK in HBMECs is earlier than ERK and PI3K after Aβ stimulation. Future studies would be necessary to explore the signaling cascade of ERK, JNK, and PI3K activation events in Aβ-induced brain endothelial CCR5 expression.

Why does Aβ trigger endothelial ERK, JNK, and PI3K activation, and is there any receptor on brain endothelial cells contributing to Aβ-induced CCR5 expression? In this study, we identified the role of RAGE in Aβ-induced brain endothelial CCR5 expression. RAGE is a member of the Ig superfamily of cell surface molecules with a diverse repertoire of ligands, including advanced glycation end products, S100 proteins, amphoterin, Aβ peptide, β-sheet fibrils, and the leukocyte β₂-integrin MAC-1 (39, 42). In the brain, RAGE is normally expressed by endothelial cells, microglia, and neurons at low levels. However, in AD and AD models, RAGE expression increases by several-fold in affected cerebral vessels and in microglia and neurons (66). We showed in this study that RAGE expression on HBMECs was up-regulated by Aβ peptide in a time- and dose-dependent manner. Compared with our previous findings (28), we found that the Aβ-induced RAGE expression preceded CCR5 up-regulation in HBMECs. Overexpression of dominant-negative form of RAGE in HBMECs abolished CCR5 expression induced by Aβ, whereas CCR5Δ32 overexpression in HBMECs had no effects on the Aβ-induced RAGE expression (data not shown). Also, our results showed that both anti-RAGE Ab and overexpression of dominant-negative RAGE in HBMECs significantly decreased Aβ-triggered ERK, JNK, and PI3K activation, which is required for Aβ-induced CCR5 expression in brain endothelial cells. Our in vivo data showed that blockage of intracerebral RAGE abolished the brain endothelial CCR5 up-regulation induced by Aβ injection in rat hippocampus. These suggest that RAGE on brain endothelial cells is a receptor that contributes to Aβ-triggered CCR5 expression. Therefore, we hypothesized that Aβ-RAGE interaction on brain endothelial cells would thus lead to a positive feedback activation, which further increases RAGE expression, and then up-regulate endothelial CCR5 expression by ERK/JNK/PI3K-Egr-1 pathway.

Many studies suggest that RAGE may play an important role in the regulation of inflammation (39, 42, 43, 66). Through RAGE, some ligands can bind to mononuclear phagocytes and modify their functions, including expression of key inflammatory mediators (67, 68, 69, 70). Inhibition of RAGE in T cells has been found to markedly decrease its infiltration in the brain, leading to suppression of experimental autoimmune encephalomyelitis (71). RAGE also functions as an endothelial adhesion receptor promoting leukocyte recruitment (72). One report showed that the Aβ-mediated migration of monocytic cells (THP-1 and HL-60) and peripheral blood monocytes through HBMEC monolayer was inhibited by RAGE Ab (73), but the mechanism remains unknown. We previously found that peripheral T cells derived from AD patients overexpress MIP-1α, which binds to CCR5 on brain endothelial cells, and the MIP-1α-CCR5 interaction triggers the endothelial tight junction opening for T cell TEM (28). In the present study, we show that, similar to the results of CCR5 deletion mutation in HBMECs, the MIP-1α-expressing T cell migration through in vitro BBB model was effectively blocked by anti-RAGE Ab or overexpression of dominant-negative RAGE in HBMECs. In particular, blockage of intracerebral RAGE abolished the brain endothelial CCR5 up-regulation and the increased T cell infiltration in the brain induced by Aβ injection in rat hippocampus. These data suggest that the RAGE-mediated brain endothelial CCR5 up-regulation caused by Aβ is associated with T cell entry into the brain, which may be one reason that increased occurrence of T cells was found in the brains of AD patients compared with age-matched controls. We conclude that RAGE on brain endothelial cells may function as a sensor for the intracerebral Aβ deposition, and hence, transmit the signal of Aβ deposition to systemic immune system, including circulating T cells by up-regulating endothelial CCR5 expression. To date, the role of T cells in AD pathogenesis remains unknown. This study gives a new target for blocking T cell entry into the brain, which is helpful to answer the question of whether T cells play an important role in the AD pathological process.

In AD, the mechanisms underlying neural-immune interaction are still not known. CCR5, particularly RAGE, has been implicated in the pathogenesis of AD. Evidence showed that the chemokine receptor CCR5 is up-regulated in AD brain and plays a role in the recruitment and accumulation of microglia in senile plaques (55, 74). The major features of RAGE activation in contributing to AD result from its interaction with Aβ. RAGE interaction with Aβ results in perturbation of neuronal function, amplification of glial inflammatory response, elevation of oxidative stress and amyloidosis, and increased Aβ influx at the BBB (75, 76, 77, 78). Our results suggest that Aβ-RAGE interaction at the BBB links the intracerebral Aβ deposition to systemic immune T cells by up-regulating CCR5 on brain endothelial cells. This study may provide a new insight into the understanding of inflammation in the progress of AD.

In summary, this study provides clear evidence that intracerebral Aβ interacting with RAGE on brain endothelial cells triggers intracellular ERK, JNK, and PI3K activation, and thus leads to nuclear protein Egr-1 binding with CCR5 promoter and promotes endothelial CCR5 expression for the circulating MIP-1α-expressing T cells crossing the BBB. Our results suggest for the first time that the intracerebral Aβ deposition signal can be transmitted to systemic immune T cells by RAGE at BBB. The determinants RAGE and CCR5 expressed by brain endothelial cells could be potential research and therapeutic targets for AD.

We are grateful to Drs. Monique Stins and Kwang Sik Kim (Department of Pediatrics, Johns Hopkins University School of Medicine) for providing HBMECs.

The authors have no financial conflict of interest.