APP/PS1 mice and wild-type littermates injected with CII in CFA revealed pathologic features of RA. Paw swelling and redness of the joint were manifest immediately after the second immunization, reached a peak 1 month later, and significantly persisted over the next 4 months when animals were sacrificed (Figure 1A, P < 0.05). In contrast to CIA induced paw swelling in all four paws in DBA/1 mice 23, paw swelling was confined to the injection site in wild-type and APP/PS1 mice. Plasma levels of the pro-inflammatory cytokines IL-1β and TNF-α were increased within 1 d after the second immunization and remained elevated over the next 4 weeks in wild-type and APP/PS1 mice (Figure 1B, P < 0.05). Paraffin-embedded joint sections stained with hematoxylin and eosin revealed severe bone erosion, cartilage/chondrocyte degradation, and proliferative synovitis in wild-type (Figure 1C) at 4 months after the second immunization. Such joint pathology was similar in APP/PS1 mice treated with CIA (data not shown).

We examined the effect of CIA on Aβ deposition in APP/PS1 mice. APP/PS1 mice received CII in CFA at 2 months of age before Aβ deposition was evolved. When Aβ deposition was analyzed 4 months after CIA, levels of Aβ deposition were moderate in APP/PS1 mice 24. Aβ plaques were widespread in the cerebral cortex and hippocampus of 6-month-old APP/PS1 mice (Figure 2A). Contrary to expectations, CIA substantially reduced Aβ plaque accumulation in APP/PS1 mice. Image analysis of Aβ deposition showed that Aβ plaques that were positive for 4G8 or thioflavine-S were significantly reduced in the cerebral cortex and hippocampus of 6-month-old APP/PS1 mice treated with CIA, compared with vehicle-treated APP/PS1 mice (Figure 2B). The effect of CIA on Aβ levels was analyzed in the opposite hemisphere using an ELISA that detects SDS-soluble and SDS-insoluble Aβ 40 and Aβ 42. Levels of soluble Aβ 40 and Aβ 42 appeared reduced by 48% (P = 0.056) and 20% (P = 0.026) in APP/PS1 mice 4 months after CIA, respectively, compared with vehicle-treated APP/PS1 mice. APP/PS1 mice treated with CIA revealed substantially significant reduction in levels of insoluble Aβ 40 and Aβ 42 (Figure 2C, P < 0.05).

CIA-induced systemic inflammation likely triggers brain inflammation that contributes to internalization and degradation of Aβ through activated microglia. Iba-1 or Mac-1-immunoreactive microglial cells were observed in the brains of wild-type mice and increased in the brains of APP/PS1 mice at 6 months of age (Figure 3A). Immunoreactive Mac-1 positive cells were transiently increased in the cortex and hippocampus of wild-type mice over the course of 1 d after the second immunization injection. However, by 2 weeks after CIA Mac-1 immunoreactivity had returned to levels that were comparable to the levels in vehicle-injected mice (Additional file 1). In APP/PS1 mice treated with CIA, the number of Iba-1 or 1-immunoreactive cells in the cerebral cortex and hippocampus increased 4 months later. CD45-positive activated microglia or macrophages were not observed in wild-type mice at 4 months after CIA (Figure 3A). Activated microglia or macrophages were observed in the cortex of 6-month-old APP/PS1 mice and increased in the cerebral cortex and hippocampus of APP/PS1 mice treated with CIA up to approximately 210% (the number of CD45-immunoreactivity) and 230% (the area of CD45-immunoreactivity) (Figure 3A, B, P < 0.05). Immunoreactivity to IL-1β and TNF-α was increased in microglia and astrocytes in the molecular layer of the dentate gyrus of APP/PS1 mice compared to wild-type as previously reported 252627. CIA further increased expression of IL-1β and TNF-α in APP/PS1 mice (Figure 3A). In CIA-treated APP/PS1 mice, Iba-1-positive or CD45-positive cells were observed adjacent to Aβ plaques (Figure 3C).

CIA has been shown to increase BBB permeability in DBA/1 mice 28. CIA likely causes the infiltration of monocytes and activation of microglia/macrophages that contribute to clearance of Aβ peptide in APP/PS1 mice 29. Immunoreactivity to a macrophage mannose receptor (CD206) antibody was observed sparsely in the perivascular and leptomeningeal regions of wild-type and APP/PS1 mice. CD206-positive macrophages were increased in the vicinity of vessels in wild-type and APP/PS1 mice subjected to CIA (Figure 4A), suggesting the possibility that CIA induces the migration of macrophages into the brain parenchyma through the disrupted BBB. In support of this hypothesis, levels of immunoglobulin G (IgG) were significantly increased in the brains of wild-type and APP/PS1 mice 1.5 and 4 months after CIA compared with relevant vehicle (Figure 4B and 4C).

Finally, we examined whether CIA would cause cerebrovascular pathology. Cerebral microvessels were immunolabeled with an anti-collagen type IV antibody, a marker of the endothelial basement membrane. CIA significantly reduced vascular density and length in the cerebral cortex and hippocampus of APP/PS1 mice 4 months later (Figure 5, A and 5B). CIA also induced cerebral microvascular pathology, such as vascular segments and atrophic string vessels (Figure 5C). At 4 months after the injection procedure, survival of vehicle-treated APP/PS1 mice was approximately 93.3%, while the survival of CIA-treated mice was reduced to 76.9% (Additional file 2).

All experiments were performed in accordance with the Guideline for Animal Experiments of Ajou University and GNT Pharma. APP/PS1 mice (APPswe/PS1dE9) were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). These mice were generated by co-injection of chimeric APPswe (K595N, M596L) and an exon-9-deleted PS1 variant (dE9) vector controlled by a mouse prion promoter 55. Male APP/PS1 transgenic mice were cross-bred with B6C3F1/J background females. APP/PS1 mice (Tg+) and wild-type littermates (Tg-) were determined by PCR analysis of tail DNA.

Bovine CII (Chondrex, Redmond, WA, USA) was dissolved in 0.1 M acetic acid at a 4 mg/ml by stirring overnight at 4°C, added to an equal volume of CFA (Chondrex), and homogenized as described previously 23. To induce CIA, 2-month-old female Tg+ and Tg-mice were injected intradermally at the tail base with 0.1 ml CII in CFA or 0.1 ml mineral oil as a vehicle. The animals received another injection of CII in CFA or mineral oil at the right hind paw 14 d after the primary immunization. Paw edema was measured using a plethysmometer (Ugo Basile, Comerio, Italy) monthly after the second immunization. After 4 months, 6-month-old mice were sacrificed and examined for AD pathogenesis on the basis of amyloid plaques and cerebrovascular pathology.

Plasma cytokine levels were analyzed at various time points after the secondary injections of CII in CFA. Blood was collected via cardiac puncture and then transferred into a heparin-coated tube (Becton Dickinson, USA). The tube was placed on ice and centrifuged at 5000 rpm (1500 g) for 3 min. The supernatant was collected and stored at -80°C. Plasma levels of IL-1β and TNF-α were measured with a commercial ELISA kit (Biosource, Camarillo, CA, USA) according to the manufacturer's instructions.

Animals were anesthetized with an intraperitoneal injection of chloral hydrate (400 mg/kg) and transcardially perfused with 0.9% NaCl. The right hind paw injected with CII in CFA was removed, fixed in 10% formalin solution at 4°C for 24 h, and decalcified in 10% EDTA solution at 4°C for 3 weeks. The tissues were then embedded in paraffin and cut into 8-μm-thick sections. The right brain hemisphere was immersion-fixed with 4% paraformaldehyde for 24 h, cryoprotected in 30% sucrose for 48 h, and cut into 18-μm-thick coronal cryosections. The hippocampal formation and cortex were dissected out from the left hemisphere and frozen at -80°C for protein assay.

Combined cortical and hippocampal tissues were homogenized in 0.5 ml Tris-buffered saline (TBS) containing protease inhibitor cocktail (Calbiochem, San Diego, CA, USA) and centrifuged at 14,000 rpm for 30 min at 4°C. The supernatant was collected and stored at -80°C for Western blot analysis of IgG. The pellet was sonicated in 0.5 ml of 2% SDS in TBS and centrifuged at 14,000 rpm for 40 min at 4°C. The supernatant was used for analysis of soluble Aβ. For analysis of insoluble Aβ, the pellet was sonicated with 0.5 ml 70% formic acid in water and then neutralized by 1:40 dilution into 1 M Tris-phosphate buffer (pH 11.0). Levels of Aβ 40 and Aβ 42 were analyzed using an ELISA kit (Biosource, Camarillo, CA, USA) according to the manufacturer's instructions.

Brain sections were washed in phosphate-buffered saline, incubated in washed in 0.3% H₂O₂ and 0.25% Triton X-100 for 10 min at room temperature, and reacted with 10% horse serum for 1 h. Sections were then reacted overnight at 4°C with one of the following primary antibodies: mouse monoclonal anti-4G8 (1:1000, Signet), rat monoclonal anti-CD45 (1:500, Serotec), rabbit polyclonal anti-Iba-1 (1:1000, WAKO), rat monoclonal anti-Mac-1 (1:500, Serotec), goat polyclonal anti-IL-1β (1:200, R&D), goat anti-polyclonal anti-TNF-α (1:200, R&D), rat monoclonal anti-CD206 (1:500, Serotec), and rabbit polyclonal anti-collagen type IV (1:500, Millipore). For 4G8 immunostaining, sections were pretreated with 70% formic acid for 10 min to expose the Aβ epitope. The sections were reacted with anti-mouse, anti-rat, or anti-rabbit IgG fluorescent or biotin-conjugated (Vector, Burlingame, CA, USA) secondary antibody for 2 h. The biotin-labeled sections were incubated with avidin-biotin-peroxidase complex (ABC Elite kit, Vector) for 1 h and then visualized using 3,3'-diaminobenzidine tetrahydrochloride dehydrate (DAB, Vector).

Brain sections containing the hippocampal formation and cortex were immunolabeled with an antibody to 4G8 (amyloid plaque burden), CD45 (activated microglia/macrophage), or collagen type IV (vascular pathology). To identify amyloid plaques using thioflavine-S staining, the brain sections were incubated in 1% aqueous thioflavine-S (Sigma) for 10 min, followed by dehydration with 70% ethyl alcohol, and coverslipped with mounting medium. Three sections on the anatomically comparable plane for each animal were analyzed using MetaMorph 6.1 software (Universal Imaging Corp, Downingtown, PA, USA). The immunostained sections were captured using an Olympus BX 51 microscope (Olympus, Japan) at a magnification of 200×. The areas positive for 4G8, thioflavine-S, or CD45 in the hippocampal formation and cortex were subtracted from nonspecific signal from the same section and scaled to the vehicle group (= 100). The vascular pathology was analyzed in the area including the CA1 pyramidal layer and cortex by measuring vessel density and length immunolabeled with collagen type IV.

BBB permeability was investigated by monitoring levels of IgG. Combined cortical and hippocampal samples (see above) were electrophoresed on an 8-10% SDS polyacrylamide gel and transferred to a nitrocellulose membrane. The membrane was pre-incubated with 5% nonfat dry milk and reacted with a biotinylated anti-mouse lgG (1:500, Vector) overnight at 4°C. The membrane was then reacted with avidin-biotin-peroxidase complex (ABC Elite kit, Vector) for 2 h and then with enhanced chemiluminescence reagents (Amersham, Buckinghamshire, UK) on X-ray film. Band intensity was quantified using Image Gauge 3.12 (Fuji PhotoFilm Co). The blot was stripped and probed with rabbit polyclonal anti-actin (1:1000, Sigma) to normalize the protein load among samples.

All values are expressed as mean ± S.E.M. An independent-samples t-test was used to compare two samples. Analysis of variance (ANOVA) and the Student-Newman-Keuls test were used for multiple comparisons. All analyses were calculated using SPSS version 12.0 for Windows. Statistical significance was assumed at P < 0.05.