Apolipoprotein A-II induces acute-phase response associated AA amyloidosis in mice through conformational changes of plasma lipoprotein structure

During acute-phase response (APR), there is a dramatic increase in serum amyloid A (SAA) in plasma high density lipoproteins (HDL). Elevated SAA leads to reactive AA amyloidosis in animals and humans. Herein, we employed apolipoprotein A-II (ApoA-II) deficient (Apoa2−/−) and transgenic (Apoa2Tg) mice to investigate the potential roles of ApoA-II in lipoprotein particle formation and progression of AA amyloidosis during APR. AA amyloid deposition was suppressed in Apoa2−/− mice compared with wild type (WT) mice. During APR, Apoa2−/− mice exhibited significant suppression of serum SAA levels and hepatic Saa1 and Saa2 mRNA levels. Pathological investigation showed Apoa2−/− mice had less tissue damage and less inflammatory cell infiltration during APR. Total lipoproteins were markedly decreased in Apoa2−/− mice, while the ratio of HDL to low density lipoprotein (LDL) was also decreased. Both WT and Apoa2−/− mice showed increases in LDL and very large HDL during APR. SAA was distributed more widely in lipoprotein particles ranging from chylomicrons to very small HDL in Apoa2−/− mice. Our observations uncovered the critical roles of ApoA-II in inflammation, serum lipoprotein stability and AA amyloidosis morbidity, and prompt consideration of therapies for AA and other amyloidoses, whose precursor proteins are associated with circulating HDL particles.

cardiovascular disease, and inflammatory processes remains controversial 14,15 . However, its adverse role has been established in the pathogenesis of AA amyloidosis. Sustained high levels of SAA result in tissue deposition of the N-terminal fragments of SAA as amyloid fibrils. In mice, AA amyloid deposition can be experimentally induced by multiple injections of silver nitrate (AgNO 3 ), casein or lipopolysaccharide (LPS), resulting in remarkable elevation and maintenance of high levels of plasma SAA 16 . Amyloid enhancing factor (AEF) and AA amyloid fibrils have been used to induce and/or accelerate AA amyloidosis in mice and other animals, such as hens and rabbits [17][18][19] .
In addition to the stimulation of reverse cholesterol transport from extra-hepatic tissue to the liver, HDL is known for its preventive roles in cardiovascular disease through anti-oxidant and anti-inflammatory effects 20,21 . Apolipoprotein A-II (ApoA-II) is the second most abundant protein component of HDL; however, its roles in HDL function and metabolism remain unclear 22 . ApoA-II is reported to be more hydrophobic than apolipoprotein A-I (ApoA-I), and is closely associated with modulation of HDL metabolism and alteration of HDL conformation by interacting with ApoA-I and other apolipoproteins [23][24][25] . In mice, ApoA-II is a serum precursor of amyloid fibrils (AApoAII) in age-associated systemic amyloidosis (AApoAII amyloidosis) 26,27 . Our previous study found that mouse SAA, ApoA-I and ApoA-II interact with each other during AA and AApoAII amyloidosis 28,29 . It has been reported that during APR, elevated SAA binds to HDL and decreases levels of ApoA-I and ApoA-II, leading to alteration of HDL particle size 11,30,31 .
To investigate the potential role of ApoA-II in lipoprotein particle distribution and progression of AA amyloidosis, we induced AA amyloidosis by co-injection of AA amyloid fibrils (AEF) and AgNO 3 (inflammation inducer) in wild type (WT), ApoA-II deficient (Apoa2 −/− ), ApoA-II overexpressing (Apoa2 c Tg) and ApoA-I deficient (Apoa1 −/− ) mice. We found that elevation in serum SAA and AA amyloid deposition was significantly suppressed in Apoa2 −/− mice. Moreover, ApoA-II deficiency resulted in dramatic alteration of lipoprotein particles and redistribution of apolipoprotein in lipoproteins. These results suggest an important role of ApoA-II in inflammation, lipoprotein metabolism and AA amyloidosis.

Results
AA amyloid deposition was suppressed in Apoa2 −/− mice. After co-injection of AgNO 3 and AA fibrils, tissue sections from various organs were stained with Congo red, and amyloid deposition (amyloid score: AS and amyloid index: AI) was subsequently determined by green birefringence under polarizing microscopy. The liver of WT and Apoa2 −/− mice and the spleen of WT mice showed AA amyloid deposition 12 h and 1 d after injection ( Fig. 1 and Supplementary Fig. 1). After 3 to 10 d of injection, AA amyloid deposition had expanded from the liver and spleen to the stomach, intestine, lung and kidney (Fig. 2). ASs in these organs were significantly less in Apoa2 −/− mice than in WT mice (Figs 1 and 2). The degree of amyloid deposition in the whole body (AI) was significantly reduced in Apoa2 −/− mice at 12 h, 1 d, 3 d and 10 d after treatment (Fig. 1).
Elevation of serum and hepatic SAA mRNA expression was suppressed in Apoa2 −/− mice. In WT and Apoa2 −/− mice, serum SAA levels were undetectable at 0 h, but increased dramatically and reached maximum levels 2 d after co-injection with AgNO 3 and AA fibrils, and then deceased rapidly until being undetectable at 10 d. In contrast, upregulation of serum SAA was significantly suppressed at 12 h and 1 d after inflammatory stimulus in Apoa2 −/− mice compared with WT mice (Fig. 3). On the other hand, Apoa1 −/− mice showed no difference compared with WT mice (Fig. 3).
To elucidate the mechanism leading to low serum SAA levels in Apoa2 −/− mice, real-time PCR was performed to assess hepatic Saa1/Saa2 mRNA in WT and Apoa2 −/− mice (Fig. 4). Prior to experimental manipulation,   . The serum concentrations of SAA were quantitated using a densitometric image analyzer with NIH Image, and serum SAA levels were represented as ratios to pooled standard serum isolated from C57BL/6 J mice 1 d after co-injection with AgNO 3 and AA fibrils (B). Each symbol represents mean ± SEM. **P < 0.01, ***P < 0.001, WT vs Apoa2 −/− , # P < 0.05, experienced less tissue damage and less inflammatory cell infiltrates compared to WT mice at 12 h during APR ( Fig. 5C and D). At 1 d during APR, Apoa2 −/− mice had less emphysematous changes than WT mice ( Fig. 5E and F). At 3 d during APR, the lungs of WT still exhibited infiltration of inflammatory cells, while Apoa2 −/− mice demonstrated recovery from damage ( Fig. 5G and H). At 10 d after APR, both WT and Apoa2 −/− mice showed recovery from inflammatory damage ( Fig. 5I and J). Significant differences in damage scores were noted between WT and Apoa2 −/− mice at 12 h to 3 d (Fig. 5K).  AA amyloid deposition and elevated serum SAA levels were accelerated in Apoa2Tg mice. We compared AA amyloid deposition and serum SAA levels in WT (R1.P1-Apoa2 c ) mice with those in Apoa2Tg mice. The serum concentration of ApoA-II in ApoA2Tg was 1.26 times that in WT mice 32 . In Apoa2Tg mice, AA immunohistochemistry (IHC) positive area was more abundant in the liver and spleen at 3 d and 10 d after co-injection compared with WT mice. Significant differences at 10 d were noted in the liver and spleen ( Fig. 6A and B). The amyloid deposition stained positively in IHC with anti-AA antiserum, but negatively with anti-AApoAII antiserum (data not shown). Apoa2Tg mice also showed significantly higher serum SAA levels at 12 h and 1 d after co-injection compared with WT mice ( Fig. 6C and D).
These results suggest that increased levels of serum ApoA-II may accelerate the APR associated elevation of serum SAA levels and accelerate AA deposition.
Lipoprotein particles exhibited smaller sizes in Apoa2 −/− mice. To elucidate the effect of ApoA-II deficiency on lipoprotein particles and the distribution of major apolipoproteins containing SAA, we analyzed HDL particle size by non-denaturing gradient PAGE with serum pre-stained by Sudan Black B (Fig. 7A). Further, we performed Western blot analysis of serum amyloidogenic apolipoproteins (SAA, ApoA-I, ApoA-II and ApoE) ( Fig. 7B-E). We marked the size classes of HDL 1 , HDL 2 , and HDL 3 based on the distributions of ApoA-I in WT mice according to our previously reported results 29,32 . The predominant form of HDL in WT mice was HDL 3 . The HDL particle size increased within 24 h of APR, before returning to the pre-inflammatory size by 72 h. In pre-inflammatory Apoa2 −/− mice, the HDL band was weak, broader and smaller corresponding to HDL 3 (Fig. 7A), while during APR, the density, size and distribution of HDL particles increased. Without inflammatory stimulation, SAA was not detected by Western blot in WT and Apoa2 −/− mice (Fig. 7B). During APR, SAA increased dramatically and was mainly present within HDL 3 , HDL 2 and HDL 1 , with the HDLs smaller than HDL 3 in WT mice. However, SAA was found in HDL 2 and HDL 1 in Apoa2 −/− mice, as well as in HDLs smaller than HDL 3 and larger lipoproteins. ApoA-I was mainly found within HDL 3 and HDL 2 for WT mice, and partially in HDL 1 , as determined by PAGE. During APR, the level of ApoA-I decreased gradually and was largely Figure 6. AA amyloid deposition and serum SAA levels were accelerated in Apoa2Tg mice. The positive areas of amyloid deposits in the liver and spleen were determined as IHC positive area with antiserum against AA at 3 and 10 d after co-injection with AgNO 3 and AA fibrils (A and B). Three areas in each of the liver and spleen sections were randomly captured under x200 magnification, and the positive areas (ratio to whole liver and spleen) were calculated with the ImageJ software. Each column represents the mean at 12 h, 1 d, 3 d, and each column and bar represents the mean ± SEM. The tables below the figures represent the number of mice used. *P < 0.05, WT vs. Apoa2Tg (unpaired Student's t-test). WT and Apoa2Tg mice were co-injected with AgNO 3 and AA fibrils. Serum was obtained at 0 h, 12 h, 1 d, 2 d, 3 d, 5 d, 7 d and 10 d after injection, and serum SAA levels were determined by Western blot analysis (C and D). Serum SAA levels are represented as a ratio to pooled standard serum isolated from C57BL/6 J mice. Each symbol represents mean ± SEM. *P < 0.05, **P < 0.01, WT vs Apoa2Tg (unpaired Student's t test). The table below the figure represents the number of mice used.
SCIeNtIfIC REPORTS | (2018) 8:5620 | DOI:10.1038/s41598-018-23755-y found in HDL 3 and HDL 2 , and the particle size only increased slightly. In Apoa2 −/− mice, the amount of ApoA-I was dramatically decreased, and was mainly found in HDL particles smaller than HDL 3 in normal and APR states (Fig. 7C). ApoA-II was distributed mainly in smaller HDL 3 compared with ApoA-I containing HDL 3 in WT mice. The amount of ApoA-II decreased and the size of HDL containing ApoA-II increased slightly during APR (Fig. 7D). There was no ApoA-II in Apoa2 −/− mice, as expected. ApoE was distributed widely, and found in lipoproteins ranging from HDL 1 to larger lipoprotein particles in both WT mice and Apoa2 −/− mice. These observations suggest that ApoA-II deficiency resulted in the disruption of HDL structure and redistribution of elevated SAA and ApoA-I during APR. There was no obvious influence on the distribution of ApoE in this study.
To further confirm these results, pooled sera from WT and Apoa2 −/− mice isolated at 0 h, 12 h, and 1 d after co-injection were analyzed with a HPLC gel filtration system (Fig. 8A). Under normal conditions, HDL cholesterol levels were markedly decreased in Apoa2 −/− mice, while low density lipoprotein (LDL) cholesterol levels were increased compared with WT mice. During APR at 1 d, WT and Apoa2 −/− mice showed an increase in lipoprotein cholesterol levels in LDL and very large HDL. Western blot analysis of isolated HPLC fractions showed that SAA protein was clearly associated with very small to very large HDL in WT mice, as well as very small LDL. However, in Apoa2 −/− mice, SAA distributed more widely, and was found in very small to very large HDL, very small LDL, very low density lipoprotein (VLDL) and chylomicron (CM), compared with WT mice (Fig. 8B).

Discussion
In this study, we investigated whether apoA-II could influence the metabolism of SAA and HDL and formation of AA amyloidosis in mice. Our previous study showed that serum ApoA-II and SAA interact with AApoAII and AA amyloid fibrils, and facilitate amyloid formation in R1.P1-Apoa2 c mice 28 . SAA protein cross-reacts with pre-existing AApoAII amyloid fibrils and complements the seeding effect of AA fibrils to SAA, increasing AA amyloidosis 28 . On the other hand, increased SAA expression during APR reduces the serum concentration of ApoA-II and suppresses AApoAII amyloidosis. However, there are no studies demonstrating the effects of deficiency and overexpression of ApoA-II on SAA metabolism and AA amyloidosis.
We employed Apoa2 −/− and Apoa2Tg mice to explore the mechanism by which ApoA-II influences SAA metabolism and AA amyloidosis. After co-injection of AgNO 3 (inflammation inducer) and AA fibrils (seed), we showed a significant effect of ApoA-II on elevation of serum SAA levels and AA amyloidosis (Figs 1, 2, 3 and 5). To further elucidate the mechanism of SAA suppression in Apoa2 −/− mice, we analyzed hepatic SAA mRNA expression by real time PCR. Expression of hepatic SAA mRNA in Apoa2 −/− mice was significantly lower than in WT mice in both pre-inflammatory and inflammatory states (Fig. 4). During inflammation, hepatic synthesis of SAA is induced by the macrophage-secreted cytokines IL-6, IL-1, and TNFα via an Nf-κB-dependent pathway. Pathological investigation showed that Apoa2 −/− mice experienced less lung tissue damage and inflammatory cell infiltration during APR. Moreover, the lungs of Apoa2 −/− mice showed quicker recovery from inflammatory damage than WT mice (Fig. 5A to K). However, the impact of elevated SAA on inflammation and tissue damage remains controversial 8 . It has been shown that ApoA-II plays multiple and controversial roles in influencing inflammation 23,34 . Some studies have shown that overexpression of ApoA-II could confer pro-inflammatory properties to HDL by reducing protection against LDL oxidation and stimulating lipid hyperoxidation and monocyte transmigration in mice 34 and that ApoA-II may maintain host responses to lipopolysaccharide (LPS) by suppressing inhibitory activity of LPS binding protein 35 . In contrast, other studies have suggested an anti-inflammatory effect of ApoA-II in humans 36 . Macrophages and monocytes activated by inflammation may cleave the C-terminal part of SAA and induce AA amyloid fibril formation 37 . Macrophages may also resorb amyloid deposits [37][38][39] . Thus, inflammation may change the microenvironment of tissues in which amyloid deposits form, and may modulate Figure 7. ApoA-II deficiency resulted in the disruption of HDL structure and redistribution of SAA during APR. HDL particle size and distributions of major apolipoproteins among HDL species were analyzed by nondenaturing 5-15% PAGE. To detect HDL particle size, 5 μl of pooled serum obtained from WT or Apoa2 −/− mice during APR at 0 h, 12 h, 1 d and 3 d after co-injection with AgNO 3 and AA fibrils were pre-stained with Sudan Black B and applied to non-denaturing 5-15% PAGE (A). SAA (B), ApoA-I (C), ApoA-II (D), and ApoE (E) protein distributions among lipoprotein species were detected by Western blot analysis with specific antisera following non-denaturing 5-15% PAGE of 1 μl pooled serum of mice. All blots were obtained under the same experimental conditions, and cropped images of the blots are shown.
SCIeNtIfIC REPORTS | (2018) 8:5620 | DOI:10.1038/s41598-018-23755-y the progression of AA amyloidosis. Our study demonstrates that ApoA-II provokes inflammation and increases cytokines and macrophages. Further research is required to validate these results.
In mouse AA amyloidosis, amyloid deposition occurs independent of inflammation, while the time and degree of amyloid deposition is determined by plasma SAA concentration in transgenic mice 40,41 . In human AA amyloidosis, plasma SAA concentration is a major factor determining amyloid deposition 37 . In other systemic amyloidosis, circulating concentrations of precursor protein determine amyloid deposition. In Apoa2Tg mice, an increase in serum ApoA-II levels of 1.26× was shown to lead to accelerated AApoAII amyloid deposition 32 , and reduced ApoA-II concentration was associated with decelerated amyloid deposition by treatment with calorie-restriction 42 . Thus, we believe that a lower SAA concentration suppresses AA amyloidosis in Apoa2 −/− mice, and increased SAA accelerates amyloid deposition in Apoa2Tg mice. The genetic background of Apoa2Tg mice is the R1.P1-Apoa2 c strain, and elevation of SAA concentration during APR in this strain was less obvious compared with C57BL/6 J mice that have a genetic background of Apoa2 −/− (Figs 3 and 6). Due to the different genetic background and the relatively smaller number of Apoa2Tg mice examined compared with Apoa2 −/− mice, we plan to examine transgenic ApoA-II mice more intensively in future studies to provide additional support for our findings. We believe that a decrease in serum SAA levels in Apoa2 −/− mice was associated with suppression of amyloid deposition and tissue damage. However, there is a possibility that structural changes in AA amyloid fibrils of Apoa2 −/− mice could affect amyloid deposition, and will be the topic of future amyloidosis research.
During APR, plasma proteins increase by at least 25% and the major acute-phase proteins include C-reactive protein (CRP), SAA and fibrinogen 43 . Increased SAA circulates with HDL and results in redistribution of HDL and its apolipoproteins. HDL contains several apolipoproteins, such as SAA, ApoA-I, ApoA-II, ApoA-IV, C-II, C-III and ApoE, which are currently considered to be amyloidogenic or amyloid-associated proteins 3,44 . The most abundant HDL apolipoprotein that is impacted during APR is ApoA-I [45][46][47] . A number of reports have found marked decreases in serum ApoA-I during APR [48][49][50][51] , and acute phase HDL is larger in size than normal HDL 3 , with the radius extending into the HDL 2 range 52 . In contrast, other studies reported no changes in HDL cholesterol or ApoA-I levels when SAA was increased to levels comparable to those during infection or inflammation 46,53 . In our study, HDL particle size was increased and ApoA-I and ApoA-II decreased during APR (Fig. 7). Some reports have revealed decreased hepatic expression and serum concentrations of ApoA-II during the acute phase 28,43 . Interestingly, our observation showed that the increased SAA was widely distributed during APR, from LDL, HDL 1 , HDL 3 to very small HDL, while ApoA-I and ApoA-II were not found in LDL and very small HDL (Figs 7 and 8).
Despite these results, the role of ApoA-II in HDL metabolism remains unclear 22 . We sought to elucidate the effect of ApoA-II deficiency on the distribution of lipoprotein particles and other apolipoproteins during APR. In Apoa2 −/− mice, lipoproteins were dramatically decreased and widely distributed, from LDL to very small HDL (Figs 7 and 8). ApoA-I was also dramatically deceased, with its distribution ranging from HDL2 to very small HDL. HDL cholesterol in Apoa2 −/− mice exhibited a smaller size compared with that of WT mice (Fig. 8A). ApoA-II-deficient mice have been reported to have dramatically decreased and smaller HDL 54 , consistent with our observations. In contrast, Apoa1 −/− mice showed slight decreases in ApoA-II levels and a larger HDL size in our previous study 29 . During APR, Apoa2 −/− mice showed an increase in lipoprotein levels, mainly from LDL and very large HDL at 24 h (Fig. 8A). Surprisingly, we observed that SAA in Apoa2 −/− mice was more predominantly located in CM, LDL, HDL and very small HDL, as assessed by PAGE and HPLC gel filtration analysis (Fig. 8B). However, ApoA-I was found mainly in very small HDL during APR in Apoa2 −/− mice. Our previous study showed that ApoA-I plays a key role in maintaining ApoA-II distribution and HDL particle size 29 . On the other hand, it has been reported that ApoA-II is an important factor regulating HDL size and the ratio of ApoA-I to ApoA-II in plasma 55 . ApoA-II is more hydrophobic and modulates more strongly the conformation of HDL than ApoA-I 56-58 . Our results revealed that, compared with ApoA-I, ApoA-II is as a stronger regulator of lipoprotein metabolism and apolipoprotein stabilization. Conformational changes of lipoprotein particles may alter the binding strength of SAA to lipoprotein. The extracellular matrix bound by SAA includes heparan sulfate proteoglycan and anti-inflammatory effects of lipoproteins. It has been suggested that conformational changes of lipoproteins may increase the susceptibility to AA amyloidosis 59 .
Our study showed that A-II plays a critical role in the pathogenesis of AA amyloidosis in mice. Important factors associated with the pathogenesis of AA amyloidosis include: (1) serum concentration of SAA, (2) SAA associated circulating lipoprotein structure, and (3) the microenvironment in which SAA forms fibrils and deposits. Recent studies showed that ApoA-II plays multiple roles in regulating the metabolism of plasma HDL and its apolipoproteins, and in prevention or acceleration of cardiovascular disease 23,54,60 . This study sheds light on the relationship between ApoA-II and SAA, and provides new information regarding the pathogenesis of amyloidosis associated with HDL-related proteins. ApoA-II may serve as a therapeutic target for AA. Additional studies are warranted to further explore this important issue.

Materials and Methods
Mice and induction of AA amyloidosis. C57BL/6 JJmsSlc mice were obtained from Japan SLC Inc.
Isolation of amyloid fibrils. AA fibrils were isolated from the pooled livers and spleens of C57BL/6 J mice with severe AA amyloidosis. Amyloid fibril fractions were isolated by Pras' method with some modification 32 . Isolated amyloid fibrils were suspended in deionized/distilled water (DW) at a concentration of 1.0 mg/ml and were stored at −70 °C until use. We placed 1 ml of this mixture into a 1.5 ml Eppendorf tube and sonicated on ice for 30 seconds (s) with an ultrasonic homogenizer VP-5S (Tietech Co., Ltd., Tokyo, Japan) at maximum power. This procedure was repeated 5 times at 30 s intervals. Sonicated AA fibril samples were used immediately.

Detection of amyloid deposition and inflammation in mice.
Mouse organs embedded in paraffin were sliced into 4 μm sections and stained with Mayer's hematoxylin/eosin (HE) or Congo red. Amyloid deposition was identified under polarizing microscopy by green birefringence in tissue sections stained by Congo red 61 . AA fibrils were also identified by IHC analysis using the avidin-biotin horseradish peroxidase complex method with specific rabbit antiserum against mouse AA, which was produced against guanidine hydrochloride-denatured AA fibrils in our laboratory 62 . The antiserum reacts specifically with AA amyloid deposits in IHC analysis and serum SAA protein in Western blot analysis 32,63,64 . For quantitative IHC analysis, 3 areas in each liver and spleen section were randomly captured and the ratios of positively stained areas with anti-AA antiserum to the whole section areas were calculated using an image processing program (NIH Image J software, version 1.61). The intensity of the AA deposit was also determined semi-quantitatively using the amyloid score (AS) and amyloid index (AI). As described previously, the AI is the average of the AS graded from 0 to 4 in the 7 organs examined (liver, spleen, skin, heart, stomach, small intestine and tongue) in sections stained with Congo red and observed under polarizing microscopy 63 . Two blinded observers, who had no information regarding the tissues, graded the AS.
Saa1 and Saa2 mRNA expression in the liver. Total RNA was extracted from the livers of C57BL/6 J and Apoa2 −/− mice at 0 h, 12 h and 1 d after co-injection of AgNO 3 and AA amyloid fibrils, using TRIZOL Reagent (Invitrogen, Carlsbad CA), followed by treatment with DNA-Free reagent (Ambion, Austin TX) to remove contaminating DNA. Total RNA was then subjected to reverse transcription using an Omniscript RT kit with random primers (Applied Biosystems, Foster CA). Quantitative real-time PCR analysis was carried out using an ABI PRISM 7500 Sequence Detection System (Applied Biosystems Foster CA) with SYBR Green (Takara Bio, Tokyo, Japan), and values were normalized with respect to β-actin. We designed primers which detect both Saa1 and Saa2 mRNA. The following primers were used: Saa1/2-F: 5′-AGTGGCAAAGACCCCAATTA-3′, Saa1/2-R: 5′-GGCAGTCCAGGAGGTCTGTA-3′; β-actin-F: 5′-GACAGGATGCAGAAGGAGATTACT-3′ and β-actin-R: 5′-TGATCCACATCTGCTGGAAGGT-3′. Serum lipoprotein quantity, particle size, and distribution of apolipoproteins. To determine HDL particle size, serum (5 μl for C57BL/6 J mice and Apoa2 −/− mice) samples were pre-stained for lipids with Sudan Black B and electrophoresed on a non-denaturing PAGE gel with a 5 to 15% linear polyacrylamide gradient. Electrophoresis was carried out at 25 mA for 2 h 32,33 . The distribution of ApoA-I, ApoA-II, ApoE, and SAA protein among the various lipoprotein particles was determined by Western blot analysis of 1 μl serum separated by non-denaturing PAGE. Antibodies used included: anti-mouse AA (1:3000) 62 , rabbit anti-mouse ApoA-I (1:3000) 65,66 , rabbit anti-mouse ApoA-II (1:200) (sc-366255, Santa Cruz Biotechnology, Dallas, TX) and goat anti-apoE antiserum (1:200) (sc-6384, Santa Cruz Biotechnology). Subsequently, membranes were incubated with anti-rabbit IgG solution (1:3000) or donkey anti-goat IgG-peroxidase conjugated solution (1:3000) (sc-2020, Santa Cruz Biotechnology). To further determine the cholesterol profiles in serum lipoproteins, pooled sera from five C57BL/6 J mice and six Apoa2 −/− mice treated with AgNO 3 and AA fibrils were analyzed by dual-detection high-performance liquid chromatography (HPLC) (Liposearch System, Skylight Biotech, Inc., Akita, Japan) 67 . In addition, isolated HPLC fractions with different particle diameters (ranging from 7.6 nm to >80 nm) from 30 µl pooled serum were concentrated. SAA protein in each fraction was detected by Western blot analysis after separation by 16.5% SDS-PAGE. Statistical analysis. The SPSS 16.0 software package (Abacus Concepts, Berkley, CA) was used to analyze the data. All data are presented as mean ± SEM. Significant differences in real time PCR and AA IHC positive areas were examined by unpaired Student's t-test or the Tukey-Kramer method for multiple testing. Because the AI and damage score are nonlinear indexes, significant differences were examined using the nonparametric Mann-Whitney U-test. A two-tailed P value of <0.05 was considered to be statistically significant.