Age-dependent inverse correlations in CSF and plasma amyloid-β(1–42) concentrations prior to amyloid plaque deposition in the brain of 3xTg-AD mice

Amyloid-β (Aβ) plays a critical role as a biomarker in Alzheimer’s disease (AD) diagnosis. In addition to its diagnostic potential in the brain, recent studies have suggested that changes of Aβ level in the plasma can possibly indicate AD onset. In this study, we found that plasma Aβ(1–42) concentration increases with age, while the concentration of Aβ(1–42) in the cerebrospinal fluid (CSF) decreases in APPswe, PS1M146V and TauP301L transgenic (3xTg-AD) mice, if measurements were made before formation of ThS-positive plaques in the brain. Our data suggests that there is an inverse correlations between the plasma and CSF Aβ(1–42) levels until plaques form in transgenic mice’s brains and that the plasma Aβ concentration possesses the diagnostic potential as a biomarker for diagnosis of early AD stages.

Scientific RepoRts | 6:20185 | DOI: 10.1038/srep20185 treatment of patients in early AD stages would be possible, even before at risk individuals develop cognitive impairments. This possibility led us to search for plasma biomarkers with diagnostic potentials for early AD stage.
A significant correlation between Aβ concentrations in the CSF and plasma were identified when we measured the Aβ levels prior to the plaque formation in the brain of transgenic mice. However, such correlations gradually disappeared as our mice aged and developed Aβ plaques in their brains. We hypothesized that the concentration changes in plasma Aβ can be utilized as a reliable biomarker for early diagnosis of AD prior to the plaque formation. In order to evaluate our hypothesis, we selected APP swe , PS1 M146V and Tau P301L transgenic (3xTg-AD) mice, which imitate human AD pathophysiologies including age-dependent behavioral and cognitive alterations. The 3xTg-AD mice show cognitive deficits in the Morris water maze at 6 months of age and present visible Aβ plaque deposition in the brain at 6 months of age [17][18][19][20][21] . They are also known to develop Aβ plaques more slowly compared with other transgenic mice, making the 3xTg-AD model fit for studying the relationship between AD and the plasma and CSF Aβ level alterations before the plaque formation.
In this study, we first confirmed the absence of ThS positive amyloid plaques, dense-core plaques, in the brains of young 3xTG-AD mice with immunohistochemical staining. We then measured the changes in concentration of Aβ  and Aβ  in the CSF and plasma of 5-, 7-, 9-and 12-month-old female transgenic mice using sandwich-ELISA to define a relationship between the plasma Aβ and AD progression and, also, to confirm the diagnostic potential of the plasma Aβ prior to the plaque formation. We previously found that soluble monomeric Aβ  in the brain could pass the blood-brain barrier and can be found in the plasma 22 . To further confirm our hypothesis regarding the impact of plaque formation, we also assessed the permeability of the blood-brain barrier to insoluble fibrillary Aβ (1-42) in 7-week-old ICR mice and studied whether plasma Aβ (1-42) can reflect the condition of brain Aβ (1-42) even after the plaque deposition.

Results
Age-dependent accumulation of Aβ plaques in 3xTg-AD mouse brains. Here, female 3xTg-AD mice aged 5, 7, 9 and 12 months were selected to study changes in the plasma and CSF Aβ levels since ThSpositive Aβ plaques become detectable in the brain as early as 12 months of age (female; 5-month, n = 32; 7-, 9-and 12-month, n = 33) 20 . To confirm the absence of ThS-positive Aβ plaque in the brains of transgenic mice prior to Aβ measurements, brain samples from each age group were cyrosectioned and fixed with 4% paraformaldehyde for 72 hours, which is different from widely used protocols in other studies 21,23 . Then, they were stained with ThS for β -sheet-rich Aβ dense-core plaques, 6E10 antibody for both Aβ diffuse and dense-core plaques, DAPI for overall brain visualization and pS199 for hyperphosphorylated tau. We did not observe neither ThSnor 6E10-positive plaques in the brain of 5-month-old mice (Fig. 1a). 6E10 immunostaining showed diffuse plaques in the hippocampal regions of 7-, 9-and 12-month-old female 3xTg-AD mice (Fig. 1a,b). In addition to the extracellular 6E10-positive plaques, we observed 6E10-stained neural cells from 7-month-old 3xTg-AD mice (Fig. 1a). Of the ThS stained brain slices, only the 12 month-old 3xTg-AD mice showed ThS-positive, dense-core, Aβ plaque depositions in the brain (Fig. 1b). In addition, we observed accumulations of hyperphosphorylated tau tangles in the brain of 12-month-old 3xTg-AD mice.
Age-dependent Aβ(1-42) concentration changes in the CSF of 3xTg-AD mice. Decreased Aβ levels in CSF is a clinical indication of AD progression. To confirm the alterations of CSF Aβ levels in 3xTg-AD mice mimics those in AD patients, the levels of Aβ (1-42) in the CSF of 5-, 7-, 9-and 12-month-old 3xTg-AD mice were measured with sandwich-ELISA utilizing two anti-Aβ antibodies with different epitopes. The CSF was collected using laboratory-produced capillary tubes with tapered tips as previously described 24 and then we measured the levels of Aβ (1-42) and Aβ . The concentration of Aβ  in the CSF of 3xTg-AD mice showed an age-dependent decrease with a statistical significance before the age of 12 months (female; 5-month, n = 19; 7-month, n = 31; 9-month, n = 31; 12-month, n = 6, P < 0.0001 for 5-month-old vs. 7-month-old and 7-month-old vs. 9-month-old, Fig. 2a)(13 out of 32 brains from 5-month-old and 27 out of 33 brains from 12-month-old groups were excluded from the results due to saturated values during ELISA readings). After the formation of ThS-positive plaques, the CSF Aβ (1-42) level no longer showed a decrease in the trend. On the contrary, we did not observe distinguishable trend in the age-dependent alterations of CSF Aβ (1-40) levels (P = 0.0049 for 5-month-old vs. 9-month-old, P < 0.0001 for 7-month-old vs. 9-month-old, Fig. 2b). Our results agree with previous studies on the age-dependent decline of CSF Aβ (1-42) levels in both AD patients and animal models.
Age-dependent Aβ(1-42) concentration changes in the plasma of 3xTg-AD mice. In order to study if the plasma Aβ level changes in an age-dependent manner in the 3xTg-AD mice, we measured the concentration of plasma Aβ  in the aforementioned mice used in the CSF Aβ measurements. The blood was first transferred directly from the vena cava to EDTA-containing tubes and the plasma was isolated using centrifugation. Then we measured the levels of Aβ (1-42) and Aβ . In contrast to the CSF measurements, we found that plasma Aβ (1-42) levels increased from 5-to 9-month-old 3xTg-AD mice (female; 5-month, n = 32; 7-month, n = 32; 9-month, n = 32, Fig. 3a). Notably, the plasma Aβ (1-42) levels decrease in 12-month-old mice, of which brains only developed ThS-positive dense core plaques and hyperphosphoryated tau tangles (12-month, n = 31, Fig. 3a). All the results were statistically significant by one-way ANOVA followed by Bonferroni's post-hoc comparisons (P = 0.003 for 5-month-old vs. 7-month-old, P < 0.0001 for 7-month-old vs. 9-month-old, P < 0.0001 for 9-month-old vs. 12-month-old). We observed identical trends in the alterations of plasma Aβ (1-40) to those of plasma Aβ (1-42). Between 5-, 7-and 9-month-old groups, 3xTg-AD mice showed significantly increasing levels of Aβ  in the plasma, whereas the 12-month-old group showed substantial decline (P < 0.0001 for 5-month-old vs. 9-month-old, P < 0.0001 for 7-month-old vs. 9-month-old, Fig. 3b). To directly compare the age-dependent alterations in Aβ levels, the Aβ (1-42) and Aβ (1-40) concentrations in the plasma and CSF were plotted onto the same graph. The graph shows that the Aβ (1-42) concentrations in the plasma increases with age while the Aβ (1-42) concentrations in the CSF decreases, indicating inverse correlations between the Aβ (1-42) levels in the plasma and CSF (Fig. 4a). However, as there was no trend in CSF Aβ (1-40) levels, it was difficult to conclude any correlations between the Aβ (1-40) levels in the plasma and CSF (Fig. 4b).
Collectively, our study concludes that the plasma Aβ (1-42) increases in age-dependent manner in 3xTg-AD mice and has the diagnostic potential for detecting AD until ThS-positive Aβ plaques form in the brain.
Limited blood-brain barrier transport of fibrillary Aβ. Because the blood-brain barrier is known to be permeable to soluble monomeric Aβ (1-42), plasma Aβ (1-42) is a viable candidate for early AD diagnosis. However, if the changes in plasma Aβ (1-42) levels can serve as a biomarker for AD only prior to the plaque formation in the brain, then the plasma should not be able to reflect the condition of brain Aβ (1-42), once the Aβ peptides become insoluble aggregates. Direct fibrillary Aβ injection to tissue enabled us to study the blood-brain barrier's permeability to insoluble Aβ by measuring the levels of fibrillary Aβ in the plasma. As we observed ThS-negative diffuse plaques did not have an effect on the plasma levels of Aβ in 3xTg-AD mice, it is important to inject ThS-postive Aβ aggregates. Thus, we performed ThS fluorescence assays to confirm formation of ThS-positive aggregates before the intracerebral (IC) injections (Fig. 5a). Then, we injected ThS-positive fibrillary Aβ  to the cortex of the brains of 7-week-old ICR mice (male, n = 5) by IC injection and assessed the permeability of the blood-brain barrier to fibrillary Aβ . We selected non-transgenic ICR mice for this study, since Aβ in the CSF of transgenic mice can potentially behave as a confounding factor. Additional 5 male ICR mice received IC injection with vehicle as controls. The blood was collected from the vena cava of the mice 30 minutes after the IC injection to allow sufficient time for fibrillary Aβ to pass the blood-brain barrier. After analyzing the plasma with sandwich-ELISA, no statistically significant differences in the plasma soluble Aβ concentrations of fibrillary Aβ -injected mice and vehicle-injected mice were found (Unpaired student's t-test, P = 0.629, Fig. 5b). Therefore, the brain-blood barrier was not freely permeable to fibrillary Aβ (1-42), and the plasma Aβ (1-42) could not reflect the condition of brain Aβ (1-42) once they become insoluble plaques.

Discussion
In this study, we found that the plasma Aβ (1-42) concentration increases in an age-dependent manner, while the level of CSF Aβ (1-42) decreases, indicating an inverse correlation between the plasma and CSF Aβ (1-42) levels in 3xTg-AD mice before dense-core Aβ plaque depositions appear in their brains. On the contrary, we did not observe such correlation in the alterations of Aβ (1-40) concentration between CSF and plasma. These results suggest that measuring the plasma Aβ (1-42) levels can function as an early diagnostic marker of AD. One previous comparison study involving the APP23 transgenic mice and the TG2576 mice also reported increasing plasma Aβ concentrations in age-dependent manner 25 , and another comparison study reported that in the amyloid precursor protein transgenic mouse models, Aβ concentrations in CSF decreases when Aβ deposition starts to appear 15 , further supporting our inverse correlation of Aβ levels. Although other additional studies have investigated Aβ levels in the CNS, they did not directly compare the Aβ levels in the CSF and plasma before and after plaque formation. Our study has further determined the aforementioned issue.
Our results indicate that the plasma Aβ (1-42) concentration possesses a diagnostic potential as a biomarker for early diagnosis of AD when there is no ThS positive Aβ plaque depositions in the brain. This finding is important because it could provide a possible explanation for controversial results from previous studies on the Aβ measurements in blood. Although the plasma Aβ was thought to be related to AD progression and was suggested as a potential target for AD diagnosis, previous controversial results made it difficult to draw a meaningful conclusion. Our study shows that the plasma Aβ (1-42) levels increases with age and, therefore, it can be used as an early marker for AD progression. However, the plasma Aβ may not be an ideal biomarker for AD diagnosis in later developing stages of AD, since amyloid plaque formation makes the plasma Aβ measurements unreliable. The plasma Aβ (1-42) levels may become unable to reflect the AD prognosis after the plaque deposition because insoluble Aβ (1-42) cannot pass the blood-brain barrier as seen in this study. However, one study reported possible endothelial damages to the blood-brain barrier by Aβ (1-42) peptides, which will impair the barrier's function and increase its permeability 26 . Therefore, further study is recommended to warrant the diagnostic potential of plasma Aβ (1-42) levels after the plaque deposition in the brain. It is not clear yet how the formation of amyloid plaques in the brain affects the age-dependent correlations between the plasma and CSF Aβ levels that we observed. Moreover, we only observed such correlations in female transgenic mice, which are known to develop amyloid plaques earlier than males 27,28 . Thus, the underlying mechanism causing the inverse correlation between the plasma and CSF Aβ (1-42) levels needs to be further studied, and investigations with the Dominantly Inherited Alzheimer Network are desired for more clinical data 29 . Nonetheless, this study opens up more therapeutic strategies for AD patients and could potentially lead to development of more convenient and economic diagnosis of AD in its early stages. Animals. B6;129-Psen1tm1Mpm Tg (APP swe , Tau P301L ) 1Lfa/Mmjax mice (3xTg-AD) were obtained from Jackson Laboratory (USA) and then bred in a laboratory animal breeding room at the Korea Institute of Science and Technology. The mice were housed in groups of four per cage and maintained at constant temperature with an alternating 12-hour light-dark cycle. Food and water were available ad libitum. 129 mice were assessed in this study; 5-month-old 3xTg-AD (n = 32), 7-month-old 3xTg-AD (n = 33), 9-month-old 3xTg-AD (n = 33) and 12-month-old 3xTg-AD (n = 33). All animal experiments were performed in accordance with the National Institutes of Health guide for the care and use of laboratory animals (NIH Publications No. 8023, revised 1978). The animal studies were approved by the Institutional Animal Care and Use Committee of Korea Institute of Science and Technology (AP-2011L1015).

Materials
CSF and plasma collection. CSF collection was performed according to the method described previously 24 . Mice were anesthetized with a blend of tiletamine.HCl, zolazepam.HCl (80 mg/kg, IP, Zoletil 50 ® , Virbac, France) and xylazine (20 mg/kg, IP, Rompun ® , Bayer Pharma, Germany). The anesthetized mouse   concentrations between CSF and plasma of 5-, 7-and 9-month-old female 3xTG AD mice (CSF: r = 0.7030, P < 0.005/Plasma: r = 0.9791, P < 0.0001). The error bars represent the SEMs. Oneway ANOVA followed by Bonferroni's post-hoc comparisons tests were performed in all statistical analyses (*P < 0.05, **P < 0.01, ***P < 0.001; other comparisons were not significant). was placed horizontally, and its cisterna magna was surgically open. The opened meninges were stabbed with laboratory-produced capillary tube that had a tapered tip and achieved CSF. After CSF collection, the mouse was placed supinely, and its abdominal cavity was opened. Blood sampling from the vena cava was injected to EDTA tube with protease inhibitor cocktail (Roche Diagnostics, Switzerland, cat# 11836170001) and shaken gently. Plasma was separated from the blood. After centrifugation (3000 rpm, 15 minutes, 4 °C), CSF and plasma samples were stored at − 80 °C freezer until use.
ThS staining and immunohistochemistry. Animal were perfused with 0.9% NaCl, and the brains were fixed with 4% paraformaldehyde for 3 days and then were immersed in 30% sucrose solution for 3 days 23,30 . The brain samples were sliced at 35 μ m using Cryostat (Microm HM 525, Thermo Scientific, USA) and mounted onto glass slides. Aβ plaques in cryo-sectioned brains were visualized by ThS staining. ThS was suspended in 50% of ethanol at 500 μ M and brain sections were stained for 7 minutes. Then, to remove non-specific binding of ThS dye, the sections were rinsed through 100, 95 and 90% of ethanol for 10 seconds each and moved into PBS in succession.
To detect diffused plaque, brain slices were incubated in 0.3% PBS-T for permeabilization. After washing with PBS, brain slices underwent blocking step by 5% BSA for 1 hour and then incubated with primary antibody against Aβ (1:200, 6E10 clone, Covance, USA) overnight at 4 °C. Next day, brain slices were stained with Cy3 conjugated secondary antibody (1:400, Jackson ImmunoResearch, USA). To detect endogenous tau phosphorylation, brain slices were stained with primary antibody against pSer199 (1:200, Abcam, UK). Alexa fluor 633 congugated secondary antibody (1:500, Abcam, UK) was used for fluorocente detection. All slices were stained with Hoechst for counter staining. The images were taken with Leica DM2500 fluorescence microscope and the Cari Zeiss LSM700 confocal microscope.  For all sample dilutions during the analysis, Standard Diluent Buffer was used. Diluted CSF and plasma samples were pipetted into each kit that NH 2 -terminus of human Aβ specific monoclonal antibody coated wells, and co-incubated with each Aβ specific monoclonal antibody. The intensity of this color was directly proportional to the concentration of human Aβ present in the CSF and plasma. All procedures were performed according to the manufacturer's instructions same as both ELISA kits. (1-42). ICR mice were obtained from ORIENT BIO Incorporated (Korea). Fibrillary Aβ (1-42) were prepared from the laboratory-produced synthetic Aβ (1-42) with 10% dimethyl sulfoxide (DMSO) that allow aggregation for 2 weeks at 37 °C and then centrifuged at 15,000 rpm for 5 minutes. After centrifugation, we discarded the supernatant and dissociated the pellet in 10% DMSO as the initial volume. Five 7-week-old ICR mice were IC injected with 5.85 μ g fibrillary Aβ (1-42) in 1.3 μ L according to previously published procedures 31,32 . 30 minutes after the IC injection, blood samples from the vena cava were collected and the plasma was separated from the blood. Soluble Aβ (1-42) levels in the plasma were quantified by Aβ 42-ultra-sandwich-ELISA kit (Invitrogen, cat# KHB3544). The procedure was performed according to the manufacturer's instructions as described previously. Statistical analysis. Graphs were obtained with GraphPad Prism 5 and statistical analyses were performed with one-way ANOVA followed by Bonferroni's post-hoc comparisons and Student's t-test (*P < 0.05, **P < 0.01, ***P < 0.001; other comparisons were not significant). The error bars represent the SEMs. Results from statistical analyses are included in the supporting information (Table S1, S2).