Role of microglial amylin receptors in mediating beta amyloid (Aβ)-induced inflammation

Two different types of microglial, primary cultures of human fetal microglia and mouse microglial BV2 cell line, were used for in vitro testing of amylin receptor function. This approach provided confirmation of the presence of microglial amylin receptors across species and cross-validated our observations on the function of amylin receptors in both primary cell cultures and a frequently used cell line. Primary mixed glial cultures were prepared from 12- to 15-gestational-week fetuses as previously reported [21, 22]. Briefly, the meninges and blood vessels were removed and the brain tissue was washed in minimum essential medium and mechanically dissociated by repeated trituration through a 20-gauge needle. Cells were centrifuged at 1500g for 10 min and re-suspended in minimum essential medium with 10% fetal bovine serum. The cultures were grown in a 5% CO₂ humidified incubator at 37 °C. After mixed glial cultures were completely confluent, human fetal microglia (HFM) were isolated by shaking flasks at 100 rpm (IKA KS-260, IKA Works, Wilmington, NC) for 1 h at 37 °C. The media was then collected and centrifuged at 2500 RCF for 5 min at 4 °C. Cell pellets were re-suspended in microglia plating media (MEM + 10% FBS + 1% penicillin/streptomycin). HFM were plated at a 2 × 10⁵ cells per milliliter density, and microglia were allowed to attach overnight before further use for experiments. BV2 cells (kindly provided by Dr. T. Trang, University of Calgary) were cultured in DMEM/F12 media with 10% FBS. For all in vitro experiments, primary human fetal microglial cultures and BV2 cell cultures, each experiment was performed on a fresh batch of cell cultures and repeated a minimum of four times. Statistical analysis was performed on mean data.

In order to characterize the antagonist activity of cAC253 at subtypes of amylin receptors, we compared hAmylin-generated cAMP responses in stably expressed AMY1, AMY2, and AMY3 and CTR receptors in HEK293 cells and wild-type HEK (HEK-WT) cells (Additional file 1).

Frozen brain tissues or cultured cells were homogenized in cold RIPA buffer with protease inhibitors and proteins were quantified with BCA assay (BioRad, Mississauga, ON, Canada). Proteins were loaded at 50 μg per lane on a 12% polyacrylamide gel. Proteins were transferred to nitrocellulose membrane and then blocked with LiCor blocking buffer. Blots were further incubated with primary antibodies overnight at 4 °C on a shaker. Primary antibody used for Iba1 (1:500, rabbit, Wako), CD68 (1:500, mouse monoclonal, Dako), caspase-1 (1:1000, rabbit, Abcam), NLRP3 (1:1000, rabbit, Millipore), and β-actin (1:10,000 mouse, Sigma-Aldrich). IRDye 800CW goat anti-rabbit and IRDye 680CW goat anti-mouse (LiCor, 1:10,000) were used as secondary antibodies. Blots were imaged using LiCor Odyssey image system.

The PCR array for TNFα, IL1β, and IL23 gene expression used the RT² Profiler PCR Array system (96-well assay plate, SABiosciences) following product instruction. Briefly, HFM were plated in a six-well plate for 24 h, pre-treated with/without AC253 10 μM for 24 h, and then exposed to either hAmylin (1 μM) or Aβ_1–42 (1 μM) for 6 h. RNA was isolated with Rneasy Mini kit (Qiagen) as per the product instructions. For each PCR Array, 4 μg of total RNA were used to prepare cDNA with the appropriate first strand kit from SABiosciences. The cDNA was characterized on the iCycler® iQ Real-Time PCR System (Bio-Rad Laboratories) using the RT2 Profiler PCR Arrays. The resulting raw data were then analyzed using the PCR Array Data Analysis Template and expressed as gene expression fold-change after treatment compared to the control untreated HFM cultures.

For cultured cell immunofluorescent staining, cells were plated on Lab-TEK chamber slide (ThermoFisher) and fixed with 4% paraformaldehyde (PAF) in PBS, then permeabilized with 0.3% Triton X-100 in PBS and blocked with 2% BSA + 10% goat serum for 1 h. Primary antibodies used Iba-1 (1:100, Dako), CTR (1:50, ThermoFisher, Catalog # PA1-84457, Lot # QJ2098825 and RI2275863), RAMP3 (1:100, Abcam), Aβ (6E10, 1:300, Covance), and NLRP3 (1:300, Millipore) followed fluorescent secondary antibody (goat anti-mouse Alexa Fluor-546 and goat anti-rabbit Alexa Fluor-488) and were nuclear-stained with DAPI (0.1 μg/ml, ThermoFisher). Fluorescent dyes, DyLight-594 Lectin (Vector), HiLyte 555-labled Aβ_1–42 (AnaSpec), and caspase-1 (FAM-FLICA caspase 1 kit, ImmunoChemistry Technologies) were used following product instruction. Fluorescent microscopy images were acquired with an Axioplan-2 fluorescence microscope with AxioVision software (Carl Zeiss Ltd., Toronto, ON, Canada).

For human brain immunohistochemical staining, 20-μm-thick brain sections were cut from 1-cm-thick coronal block of the human inferior temporal gyrus (IFG) of an AD brain (Braak Stage 6) obtained from the Netherlands Brain Bank. Prior to immunohistochemistry, the sections were pretreated with absolute methanol and 3% hydrogen peroxide for 10 min to avoid visualizing endogeneous peroxidase activity, followed by three 10 min (3 × 10 min) washes in TBS (Tris-buffered saline). Then, they were incubated overnight at 4 °C with the rabbit polyclonal anti-CTR antibody (ThermoFisher Scientific) at a working dilution 1:20 in Supermix (TBS containing 0.3% Triton X-100 and 2.5 mg/ml gelatin). The next day, the sections were washed in TBS (3 × 10 min) and then incubated in biotinylated goat anti-rabbit IgG, diluted 1:400 in Supermix for 2 h at room temperature. Following incubation, the sections were washed in TBS (3 × 10 min) and processed by the avidin-biotin complex (ABC) (Vector Laboratories, Burlingame, CA) for 2 h at room temperature. After washing in TBS (3 × 10 min), reaction products of dark blue color were visualized with 3,3′-diaminobenzidine-4HCl (20 mg/100 ml, Sigma) in TBS containing 0.2% ammonium nickel sulphate and 3% H₂O₂. Several sections were immunostained for CTR as mentioned above and after final washing in TBS were incubated overnight at 4 °C with either Iba-1 antibody (1:100, Dako) or RAMP3 rabbit polyclonal antibody (Abcam, Catalog # ab78017, Lot # 816327; Catalog # ab197372, Lot # GR2042843) at working dilution of 1:500 in Supermix. The next day, the sections were washed in TBS (3 × 10 min) and then incubated in biotinylated goat anti-rabbit IgG diluted 1:400 in Supermix for 2 h at room temperature. Further, the sections were washed 3 × 10 min in TBS and processed with ABC for 2 h at room temperature. After washing 3 × 10 min in TBS, the sections were treated with 3,3′-diaminobenzidine-4HCl (20 mg/100 ml, Sigma) in TBS containing 3% H₂O₂, resulting in the reaction product of yellow color. Thus, after CTR + Iba-1 dual immunostaining of the IFG sections, the neuronal final reaction product of dark blue color was associated with the presence of the CTR whereas deposit of yellow color was coupled with Iba-1 or RAMP3. Brightfield images were acquired on Axioplan 2 microscope (Zeiss) with AxioVision software (ver 4.8.2, Zeiss). The ability of the CTR and RAMP3 antibodies to detect these proteins was verified using AMY3-HEK293 transfected cells and wild-type HEK293 cells (Additional file 1 and Additional file 2: Figure S1A). We also performed siRNA RAMP3 transfections of BV2 microglial cells and compared detection of the RAMP3 protein in these cells compared to non-transfected cells (Additional file 1 and Additional file 2: Figure S1B).

For in-cell Western blot cyclic adenosine monophosphate (cAMP) quantification, mouse monoclonal anti-cAMP (R&D Systems) was used as a primary antibody and IRDye 700 goat anti mouse antibody (LI-COR) was used as a secondary antibody. Plates were imaged using an Odyssey Infrared Imaging System (LI-COR), and the integrated intensity was normalized to the total cell number on the same well as previously described [23].

Ca²⁺ concentration were monitored with confocal microscopy as described previously [23, 24]. Briefly, HFMs were plated on glass coverslips pre-coated with poly-l-ornithine and incubated at 37 °C for 12–36 h. Then, incubated with 5 μM of membrane-permeant fluorescent Ca²⁺-sensitive dye Fluo-8L-AM (AAT Bioquest, Inc., Sunnyvale, CA) for 40 min at room temperature (20–23 °C) within < 2 h before imaging. For Ca²⁺ imaging, the superfusate of ion content similar to extracellular brain fluid thus contain the following: 130 mM NaCl, 4 mM KCl, 1 mM MgCl₂, 2 mM CaCl₂, 10 mM HEPES, and 10 mM glucose (pH 7.35) was applied at a rate of 3 ml/min using a roller pump (Watson-Marlow Alitea, Sin-Can, Calgary, AB, Canada). Aβ_1–42 and hAmylin were dissolved in sterile bidistilled water at 1 mM stock solution and incubated at room temperature for 10 min before dilution with superfusate for the final used concentrations. Fluorescence intensity was monitored with a FV-300 laser-scanning confocal microscope (Olympus FV300, Markham, Ontario, Canada) equipped with an argon laser (488 nm) and excitation/emission filters (FF02-520/28-25; Semrock, Inc.) for an emission wavelength at 514 nm, measured with a numerical aperture of 0.95 20× XLUMPlanF1 objective (Olympus). Images were acquired at scan rates of 1.25–1.43 per second using a 2–3× digital zoom at full-frame (512 × 512 pixel) resolution. Regions of interest were drawn around distinct cell bodies, and analysis of time courses of change in fluorescence intensity were generated with FluoView software (version 4.3; Olympus).

AC253 peptide was cyclized to improve its stability and brain penetration when administered systemically [15]. To cyclize the AC253 peptide (cAC253) via the flanking d-cysteines, the crude peptide (61.2 mg) was dissolved in 0.1 mM Tris buffer pH 8.3 having 20% of DMSO to accelerate disulfide bond formation, and the mixture was stirred at room temperature in an open flask for 48 h and further purified on RP-HPLC (reversed phase high-performance liquid chromatography).

Soluble oligomeric Aβ_1–42 or the inverse sequence peptide Aβ_42–1, hAmylin, and AC253 were prepared according to published protocols [14, 25]. Human amylin and AC253 were purchased from American Peptide (Sunnyvale, CA) and Aβ peptides from rPeptide (Bogart, GA). The human amylin was prepared in 1 mM stock solution in water and further diluted to the final application concentration with cell culture media as previously described [14].

For in vivo experiments, intraperitoneal injection (ip) administration of cAC253 was carried out in transgenic 5XFAD mice [15], and wild-type littermate control mice (both male and female) were obtained from Dr. David Westaway (University of Alberta). These mice were equally and randomly distributed into four groups, Tg-NS (n = 10), Tg-cAC253 (n = 10), Wt-NS (n = 10), and Wt-cAC253 (n = 10). Mice received ip either normal saline (NS) or cAC253 (200 μg/kg) three times a week starting at 6.5 months of age for 5 weeks. Mice were housed under standard laboratory conditions (12/12-h light/dark cycle, lights on at 0600 h) with a room temperature of 21 °C. Water and food were available ad libitum. After completion of treatments, all mice were sacrificed with an overdose of isoflourane anesthetic and perfused transcardially with saline, and the brains were harvested. The right hemisphere was frozen for biochemical analysis (Western blot, ELISA), and the left hemisphere was fixed with PAF for 4 h at 4 °C. These brain tissues were further processed with modified CLARITY protocol (http://​www.​chunglabresource​s.​com/​cl1#cl-protocol). Briefly, the fixed brain tissue was transferred to hydrogel monomer solution (4% acrylamide, 4% PFA, and 1xPBS (phosphate-buffered saline)) at 4 °C for 24 h and, subsequently, to a 24-well plate, merged in fresh hydrogel solution, and the tissue was brought to 37 °C till formation of the gel. Thick sagittal slices (400 μm) were cut on an HR2 Slicer (Sigman Electronic, Germany). The thicker brain sections were cleared with 8% SDS in PBS for 24 h, followed by 0.3% Triton X-100 in PBS for 24 h, blocking the section in 2%BSA-10% goat serum for 4 h. A modified thioflavin S staining was used for detecting Aβ plaques. Briefly, the brain sections were rinsed with distilled water, dropped with thioflavin S (0.0125% in 50% ethanol) for 5 min, further washing with 50% ethanol and water. Then, these sections were incubated with CD68 antibody (1:50, DAKO) at 4 °C for overnight. After washing with PBS, these sections were incubated with goat-anti mouse Alexa-fluor546 (1:200, Invitrogen) at RT for 4 h. Sections were cleared with 8% SDS for 24 h followed by a 24-h wash with PBS-Triton-X100 solution. Images were visualized using fluorescence microscopy (Axioplan-2, Carl Zeiss Ltd). Amyloid plaque size and area were analyzed with Image J software.

In primary HFM cultures, the microglia population comprises more than 90% of cells as judged by staining with microglial markers, Iba-1, and DyLight 594 lectin (Fig. 1a). To establish AMY expression in these microglial cells, we first immunohistochemically identified the presence of CTR and RAMP3 (Fig. 1b). We next assessed functionality of AMYs using live-cell calcium imaging (with Fluo-8) and confocal microscopy of identified microglia (Fig. 1c). Bath applications of either monomeric human amylin (hAmylin) or soluble oligomeric Aβ_1–42 (1 μM) to HFM for 30 s produced a robust Ca²⁺ increase within 1 min after entry of the peptide within the imaging chamber. These Ca²⁺ increases displayed a sharp peak and returned to base line within 2 min (Fig. 1e, f), and they are similar to responses evoked by these peptides in HEK293 cells expressing the AMY3 receptor subtype [23]. The transient elevations in Ca²⁺ due to hAmylin or Aβ_1–42 applications were abolished by prior application of the amylin receptor antagonist, AC253 (Fig. 1h, i), suggesting that these responses are AMY receptor-mediated. AC253 applied alone did not evoke any alterations in intracellular Ca²⁺ levels (Fig. 1g). Pooled data shown in Fig. 1d are from 123 HFM in 12 culture wells from four different batches of HFM cultures. Finally, we also confirmed, using inferior temporal gyrus sections from autopsied AD patient brain, that adult human microglia were immunopositive for CTR and RAMP3 (Fig. 2a–e). In addition, BV2 cells, a mouse microglia cell line, also expressed amylin receptors (Fig. 2f, g). In these cells, exposure to Aβ_1–42, (1 μM) resulted in an uptake of the peptide into the microglia and formation of plaque-like structures on the cell surface as determined by thioflavin S staining (Fig. 2g).

We next sought to examine the potential role of microglial AMY receptors in Aβ-induced activation of the intracellular inflammasome cascade. We observed levels of NLRP3, the best studied of inflammasomes in the context of AD, to be low under basal conditions but increased significantly after hAmylin or Aβ stimulation (Fig. 3a). This stimulated increase in the NLRP3 expression was attenuated in the presence of the AMY antagonist, AC253, thus identifying a role for AMY receptors in activation of the inflammasome cascade (Fig. 3b). Western blot analysis further confirmed AMY receptor-mediated changes in NLRP3 expression (Fig. 3c, d).

Since inflammasomes function as intracellular sensors for danger signals and in response trigger activation of caspase-1 and subsequent cleavage and release of cytokines, we also examined the role of AMY receptors in these downstream events. Both Aβ and hAmylin increased caspase-1 expression in HFM and BV2 cell cultures, a response that was blunted by AC253 (Fig. 4a, b). Additionally, using RT-PCR assay, we also identified Aβ and hAmylin-mediated increases in IL1β, and TNFα, but not IL23, that were significantly attenuated by AC253 (Fig. 4c).

We have recently synthesized a cyclized form of AC253 (cAC253), which retains the neuroprotective properties of the linear form of the peptide but is proteolytically stable and highly brain-penetrant when administered intraperitoneally [15]. The pKa = 6.942 for the cAC253 peptide was calculated from Molecular Operating Environment (MOE) from Chemical Computing Group (Montreal, QC, Canada). The cAC253 also demonstrated a higher level of antagonist activity at AMY receptors with IC50, 0.3 μM, versus AC253, 0.85 μM, using in-cell western blot detection [15]. In HEK293 transfected cells, cAC253 demonstrated similar activity as AC253 and blocked hAmylin activation of AMY3 and AMY1 receptors (Additional file 3: Figure S2). We also tested the ability of cAC253 (at different concentrations) to competitively antagonize human amylin-generated cAMP responses across full range of concentrations. We examined the effects of chronic systemic administration of cAC253 on changes in amyloid pathology and inflammatory markers in a transgenic mouse model of AD, 5xFAD. Six and a half-month-old 5xFAD mice that received 5 weeks of cAC253 treatment (ip, three times a week) demonstrated significant reduced amyloid plaque formation, and we observed less number of activated microglial cells compared to 5xFAD mice treated with normal saline (NS) injections (Fig. 5a, b). Aβ deposition was significantly reduced as measured by either the number of Aβ plaques, or total area of Aβ-positive profiles in the cAC253-treated 5xFAD animals compared to those receiving NS (Fig. 5b). Wt mice showed no Aβ deposits. Protein expression levels of microglial markers Iba-1 and CD68 (activated microglia), inflammasome NLRP3, and caspase-1 were significantly reduced (approximately 30%) in cAC253-treated 5xFAD group compared to NS treatment (Fig. 5c, d). Additionally, levels of cytokines IL-1β and TNFα measured by ELISA were also significantly reduced after ip cAC253 treatment (Fig. 5e).