2.1.

Subjects

Adult male Sprague–Dawley rats (180–220 g, obtained from Center of Laboratory Animal of Guangzhou University of Chinese Medicine, SCXK [Yue] 2008-0020, SYXK [Yue] 2008-0085) were pair-housed in plastic cages in a temperature-controlled (25°C) colony room on a 12/12-h light/dark cycle. Food and water were available ad libitum. All experiment protocols were approved by the Center of Laboratory Animal of Guangzhou University of Chinese Medicine. All efforts were made to minimize the number of animals used.

2.2.

Drugs and Treatment Procedures

The $\mathrm{A}{\beta }_{25–35}$ (A4559-1mg) was purchased from Sigma–Aldrich Inc. (St. Louis, Missouri, USA), dissolved in stroke-physiological saline solution to the concentration of $2\mathrm{g}/\mathrm{L}$ stored at $-20°\mathrm{C}$ and hatched at 37°C for 4 days before being used. All other reagents and chemicals used in this study were of analytical grade.

2.3.

Model Preparation and Treatment

The procedures were similar to those described previously.²⁷ Rats were intraperitoneally anesthetized with $30\mathrm{g}/\mathrm{L}$ pentobarbital sodium ($40\mathrm{mg}/\mathrm{kg}$, i.p.; Sigma–Aldrich) and placed in a stereotaxic frame (RWD Life Science Co., Ltd., Shenzhen, China). The hair was shaved, and the scalp was opened, and then holes were drilled with an electric dental drill (brushless motor, 30,000 rpm) according to a mouse brain atlas (AP-3.6 mm, $\mathrm{ML}\pm 2.5\mathrm{mm}$, DV3.0 mm). Then, 5 μL (10 μg) of $\mathrm{A}{\beta }_{25–35}$ (fibrillar state) was slowly injected into the CA1 region of the hippocampus over 5 min and the needle was kept in for 5 min. Then, the wound was sutured and penicillin ($30\mathrm{U}/\mathrm{kg}$) was injected intramuscularly to protect against infection. Then, the rats were isolated in a warm box until they recovered consciousness.

After 15 days, the rats were screened by water maze tests to select models that are up to the standard. Then, they were divided into seven groups randomly. The seven groups are as follows: treatments control group (no $\mathrm{A}{\beta }_{25–35}$ but distilled water was given as oral), sham-operation group (just do the surgery no $\mathrm{A}{\beta }_{25–35}$ was given, while distilled water was given as oral), model group ($\mathrm{A}{\beta }_{25–35}$ and distilled water were given as oral), positive group [donepezil HCl, $0.125\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$, as oral], low-dose group [BJ, $8\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$, as oral], medium-dose group [BJ, $24\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$, as oral], and high-dose group [BJ, $48\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$, as oral]. Every group had eight animals and the experiments lasted 25 days.

Rat spatial learning and memory abilities were tested in the Morris water maze (MWM, DMS-2, Chinese Academy of Medical Sciences Institute of Medicine) using procedures similar to those described previously.^28–30 The MWM consisted of a circular fiberglass pool (200-cm diameter) filled with water ($25\pm 1°\mathrm{C}$) that was made opaque by nontoxic black paint. The pool was surrounded by light blue curtains and three distal visual cues were fixed to the curtains. Four floor light sources of equal power provided a uniform illumination to the pool and testing room. A charge coupled device camera suspended above the center of the pool recorded animal swim paths, and video output was digitized by an EthoVision tracking system (Noldus, Leesburg, Virginia, USA). The water maze tests included three periods: initial spatial training, spatial reversal training, and the probe test.

2.4.

Preparation of Erythrocytes and Enzymatic Assays

After the rats have been treated for 15 days, fresh blood was obtained by arteriopuncture and then collected in a lithium-heparinized anticoagulant tube. The erythrocytes were obtained from the anticoagulant blood samples by centrifugation at 3000 rpm for 5 min at 4°C. The buffy coat was removed and then erythrocytes were washed three times with phosphate-buffered saline.

The supernatant fluid of fresh blood samples was stored at $-20°\mathrm{C}$ until they were used. The activities of malondialdehyde (MDA), superoxide dismutase (SOD), catalase (CAT), and glutathione reductase (GSH-Px) were measured with detection kits. Protein concentrations were determined using the Coomassie Brilliant Blue G250 assay. The kits were all purchased from Nanjing Jiancheng Bioengineering Institute (Nanjing, Jiangsu, PR China). The procedures were performed according to the manufacturer’s instructions. The levels were normalized to the protein concentration of each sample and expressed as a percentage of nontreated controls.

2.5.

Raman Spectra

The erythrocyte samples were placed on a silicon slice for Raman scanning. There were no extra Raman peaks in the region of $600–1700{\mathrm{cm}}^{-1}$ on the silicon slice. The Via Plus laser micro-Raman spectroscopy system was purchased from the Renishaw Company. The resolution of this instrument is $1{\mathrm{cm}}^{-1}$. All the Raman spectra were recorded for 10 s and 3 accumulations. A laser of 785 nm was chosen and 5% laser exposed which was much lower than the safe limit of exposure. All the data were collected under the same conditions and the instrument was calibrated by silicon at the $520{\mathrm{cm}}^{-1}$ band.

2.6.

Data Analysis

In order to compare the related spectrum changes, we chose the intensity of phenylalanine band ($1002{\mathrm{cm}}^{-1}$) to normalize the spectrum. At least five spectra were obtained from each sample. Then, the spectrum was baseline corrected by the software R2.8.1 (provided by Renishaw), smoothed, normalized, and averaged by Origin 8.0 (OriginLab Crop., Northampton, Massachusetts, USA). Statistical Package for the Social Science (SPSS 17.0, SPSS Inc., Chicago) was used for statistical analysis in this study.

The spectra recorded were statistically analyzed by principal component analysis (PCA). PCA is oriented toward modeling a variance–covariance structure of a data matrix from which the eigenvalues corresponding to principal components were extracted. Each principal component (PC) was a linear combination of the $n$ independent wavenumber variables ${\chi }_1,{\chi }_2,{\chi }_3,\dots ,{\chi }_n$. For example

The first PC (PC1) accounts for the greatest variance which corresponds to the largest eigenvalue. The second PC (PC2) is orthogonal to all those preceding and accounts for a decreasing proportion of the variance. In this paper, we chose the first three PCs which have accounted for $>75\%$ of the accumulative total contribution for analysis.

3.1.

Effect of Bajijiasu on Behavior

The initial spatial training results are shown in Fig. 1, which demonstrate that injecting $\mathrm{A}{\beta }_{25–35}$ into CA1 could cause memory loss and that the surgery did not cause any side effects. Compared with the model group, the incubation period for each BJ-treated group was significantly shorter. The variation in total swimming distance in each group was similar to the variation observed for the latency period. Compared to the model group, the differences of all BJ-treated groups were significant ($p<0.01$). However, the differences were not significant compared to the control group ($p>0.05$), and the average swimming speed was similar among all groups ($p>0.05$).

Fig. 1

Effect of bajijiasu on escape latency of $\mathrm{A}{\beta }_{25-35}$-treated rats in the MWM. The graph Control, control group; Sham, sham-operation group; Model, model group; Positive, positive group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\text{donepezil}\mathrm{HCl}$, $0.125\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]; BJ-8 mg, low-dose BJ group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\mathrm{BJ}$ $8\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]; BJ-24 mg, medium-dose BJ group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\mathrm{BJ}$ $24\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]; BJ-48 mg, high-dose BJ group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\mathrm{BJ}$ $48\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]. Values given are the $\mathrm{mean}\pm \mathrm{SD}$ ($n=8$).

Probe test results showed that there was no significant difference ($p>0.05$) among the groups with regard to total swimming distance or speed. Compared to the control group, the differences of all BJ-treated groups were significant ($p<0.01$), as shown in Figs. 2, and 3, the swimming trajectory of the BJ-high dose group was obviously denser in the NW quadrant than the others. All the results suggest that bajijiasu could ameliorate $\mathrm{A}{\beta }_{25–35}$-induced learning and memory dysfunction in the rat model.

Fig. 2

The swimming time in the platform quadrant during the spatial probe test. The graph Control, control group; Sham, sham-operation group; Model, model group; Positive, positive group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\text{donepezil}\mathrm{HCl}$, $0.125\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]; BJ-8 mg, low-dose BJ group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\mathrm{BJ}$ $8\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]; BJ-24 mg, medium-dose BJ group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\mathrm{BJ}$ $24\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]; BJ-48 mg, high-dose BJ group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\mathrm{BJ}$ $48\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]. Values given are the $\text{mean}\pm \mathrm{SD}$ ($n=8$), #$p<0.01$ versus control group, *$p<0.05$ versus model group, **$p<0.01$ versus model group.

Fig. 3

The swimming trajectory of each group during the probe test in $\mathrm{A}{\beta }_{25–35}$-induced model rats. CP, control group; SP, sham-operation group; MP, model group; PP, positive group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\text{donepezil}\mathrm{HCl}$, $0.125\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]; BLP, low-dose BJ group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\mathrm{BJ}$ $8\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]; BMP, medium-dose BJ group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\mathrm{BJ}$ $24\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]; BHP, high-dose BJ group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\mathrm{BJ}$ $48\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$].

3.2.

Raman Spectra of Normal Erythrocyte

A typical Raman spectrum of normal erythrocyte is shown in Fig. 4. We had known that the Raman peaks of erythrocyte mostly come from the intracellular hemoglobin and erythrocyte membrane. The spin state marker region of hemoglobin appeared at 1547, 1566, 1585, and $1605{\mathrm{cm}}^{-1}$. Various pyrrole ring vibration modes of hemoglobin were observed at 896, 1398, 1374, and $1343{\mathrm{cm}}^{-1}$. The vinyl in-plane C–H bending mode appeared at $1305{\mathrm{cm}}^{-1}$. The ${\mathrm{CH}}_2/{\mathrm{CH}}_3$ deformation modes, primarily from amino acid side chains of membrane protein, appeared at $1451{\mathrm{cm}}^{-1}$. The peak at $1002{\mathrm{cm}}^{-1}$ is assigned to C–C twist and symmetric C–C stretch of phenylalanine ring. The complete Raman frequency assignment of spectra recorded is presented in Table 1.^31–39

Fig. 4

Raman spectrum of the normal control erythrocytes which were smoothed, normalized, and averaged from eight samples and five spectra of each sample.

Table 1

Raman frequencies and assignments of the erythrocyte of normal control group rats.

3.3.

Raman Spectra of Model Erythrocyte in Rats Induced by $A{\beta }_{25–35}$

Oxidative stress is defined as a disturbance of the balance between the production of reactive oxygen species (ROS) and antioxidant defense systems. Excessive ROS production is known to cause oxidative damage to major macromolecules in cells, including DNA, lipids, and proteins, which disrupts cellular functions and integrity^40,41 and can result in cell death and tissue damage. ROS are implicated in several diseases, including cancer, diabetes, cardiovascular diseases, and aging. Previous studies indicated that oxidative damage is an early factor in ADs.^42–44 Therefore, it may be possible to stop AD progression with antioxidants and amyloid treatment causing a significant increase in ROS.^45–46 In this study, $\mathrm{A}{\beta }_{25–35}$ caused a marked accumulation of the lipid peroxidation product MDA, and reduced GSH-Px, CAT, and SOD expression ($p<0.01$, compared with and expressed as percentage of normal control controls), as shown in Fig. 5.

Fig. 5

Effect of SOD, MDA, CAT, and GSH-Px levels on the serum of $\mathrm{A}{\beta }_{25–35}$-induced model rats. The graph Normal, normal control group; Sham, sham-operation group (animals underwent surgery but did not receive $\mathrm{A}{\beta }_{25–35}$); Model, model group (10 μg $\mathrm{A}{\beta }_{25–35}$ was given), Positive, positive group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\text{donepezil}\mathrm{HCl}$, $0.125\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]; BJ-8 mg, low-dose BJ group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\mathrm{BJ}$ $8\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]; BJ-24 mg, medium-dose BJ group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\mathrm{BJ}$ $24\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]; BJ-48 mg, high-dose BJ group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\mathrm{BJ}$ $48\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]. Values given are the $\text{mean}\pm \mathrm{SD}$ ($n=8$) and expressed as percentage of normal control controls, #$p<0.01$ versus control group, *$p<0.05$ versus model group, **$p<0.01$ versus model group.

Free radicals are the product of metabolism in cells. Under normal circumstances, the free radicals can be used by the immune system to kill the pathogen. However, in some conditions (drug decoys, ultraviolet radiation, high temperature exposure, and so on), the content of free radicals will increase sharply. This will induce oxidation pressure in cells and then damage cells’ structure. The Raman spectrum of normal erythrocyte and the model erythrocytes of the rats induced by $\mathrm{A}{\beta }_{25–35}$ were shown in Fig. 3, which showed that under oxidative pressure, the s–s makers of the model erythrocyte around the region $650{\mathrm{cm}}^{-1}$ were increased obviously compared with the normal. This result is consistent with the fact that the oxidative stress leads to the increase of concentration of disulfide bridges within the erythrocytes. Meanwhile, the formation of protein–ss–glutathione mixed disulfide is expected to be formed inside the cells as a response to oxidative stress.⁴⁷ On the other hand, there is a decrease of the intensity in the high spin (deoxygenated-Hb) marker band at $1605{\mathrm{cm}}^{-1}$, and an increased intensity at the peak $1374{\mathrm{cm}}^{-1}$ which arises from $\nu 4$ as a sign of concentration of ${\mathrm{O}}_2$. The results suggested that the oxidative damage of the blood from the model rats induced by $\mathrm{A}{\beta }_{25–35}$ can be detected by the Raman spectra, and showed the same results as in Fig. 6.

Fig. 6

The Raman spectra of the erythrocyte on the $\mathrm{A}{\beta }_{25–35}$-induced model rats. The graph Normal, normal control group; Sham, sham-operation group (animals underwent surgery but did not receive $\mathrm{A}{\beta }_{25–35}$); Model, model group (10 μg $\mathrm{A}{\beta }_{25–35}$ was given), while (a) is the normal control compared with the model and (b) is the sham-operation compared with the normal control.

Erythrocyte membrane lipids are constituted of phospholipid bilayer and together with the unsaturated fatty acid, ensures the liquidity of membrane. The liquidity of membrane relates to the deformability and ionic permselectivity of the erythrocytes. This function can ensure the normal work of erythrocyte even if in blood capillary. Therefore, erythrocytes can carry oxygen to every tissue of body and then provide energy. Some of the Raman spectra of erythrocytes come from the membrane lipids and membrane protein, as the Raman peaks at 1123 and $1033{\mathrm{cm}}^{-1}$ belong to the trans stretching vibrations of CAC skeleton and the peak at $1076{\mathrm{cm}}^{-1}$ is assigned to the gauche vibrations of CAC skeleton^48,49 which are very sensitive to the change of the conformation of membrane lipids. The more content the of the trans conformation the higher the order of vertical chain. And more content of the trans conformation also results in less liquidity of the erythrocyte membrane. As shown in Fig. 6(a), the Raman peaks that belong to the membrane lipids or membrane protein of the model group were different from that of the normal control, which suggested that the structure and performance of the erythrocyte membrane had changed. All the data showed the same results as in Fig. 5, the lipid peroxidative products of MDA, GSH-Px, CAT, and SOD expression were changed.

Amide III is a sensitive mode in the conformation of the main chain in the membrane protein. Some of its peaks had change after induced by $\mathrm{A}{\beta }_{25–35}$, as shown in Fig. 6; the Raman peaks around $1321{\mathrm{cm}}^{-1}$ which assigned to amide III was discussed⁴⁹ and the quantity decreased, the peak at $1659{\mathrm{cm}}^{-1}$ that belongs to protein amide I showed the same reaction. At the same time, the Raman intensity of peak at $1254{\mathrm{cm}}^{-1}$ assigned to random coil increased, which indicated that the secondary structure of the main chain in membrane protein has been destroyed.

Not only the main chain, but also the side chain of the erythrocyte membrane protein had been damaged by $\mathrm{A}{\beta }_{25–35}$. In the normal control erythrocyte membrane, there was one obvious Raman peak at $1451{\mathrm{cm}}^{-1}$, which belongs to ${\mathrm{CH}}_2/{\mathrm{CH}}_3$ stretch in the side chain. The structure and the content of chemical groups in the side chain of the membrane will affect the tertiary structure of protein. The decreased content of ${\mathrm{CH}}_2/{\mathrm{CH}}_3$ indicated that the CAH bond has been damaged, which was induced by $\mathrm{A}{\beta }_{25–35}$. And the previous studies showed that erythrocyte anion exchange of citrate is abnormal in AD, and the erythrocyte sedimentation rate is significantly associated with the clinical traits of AD,^15–16 above all, the side chain of membrane protein of the rats induced by $\mathrm{A}{\beta }_{25–35}$ had been changed.

3.4.

Principal Component Analysis

As shown in Fig. 7, we can see that the Raman spectra fingerprints of the erythrocyte of model group were obviously different from those of the normal control. In order to distinguish the Raman spectra of erythrocyte variants of normal control group and model group, PCA was employed. The results were shown in Fig. 7, which depicts the plot between scores of the PC1 and the PC2 26 Raman spectra of control group and 27 of the model group that were used for PCA. The plots showed that almost all the Raman spectra of the control group were located at the bottom left, while those of model group were at the opposite, and the analysis results showed a good classification compared with the other method.

Fig. 7

Scores plots of primary component (PC), PC1 versus PC2.

3.5.

Protective of Bajijiasu on the Erythrocyte

Based on PCA of the full Raman range, samples with varying percentages of erythrocyte with oxidative damage could be separated and identified. After being treated with bajijiasu, the Raman spectra fingerprints of the erythrocyte were close to those of the normal control, and the BJ-48 mg groups’ were almost the same as those of the normal control, as shown in Figs. 8 and 5. The results of this study suggested that the method to detect the variants of the erythrocyte in a rat model of $\mathrm{A}{\beta }_{25–35}$-induced neurotoxicity based on micro-Raman spectroscopy was useful, and bajijiasu could be used as a potential drug treatment for AD.

Fig. 8

The Raman spectra of the erythrocyte on the $\mathrm{A}{\beta }_{25–35}$-induced model rats. The graph Normal, normal control group; BJ-8 mg, low-dose BJ group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\mathrm{BJ}$ $8\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]; BJ-24 mg, medium-dose BJ group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\mathrm{BJ}$ $24\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]; BJ-48 mg, high-dose BJ group [$\mathrm{A}{\beta }_{25–35}$ $20\mu \text{g}+\mathrm{BJ}$ $48\mathrm{mg}/(\mathrm{kg}·\mathrm{d})$]. Values given are the $\text{mean}\pm \mathrm{SD}$ ($n=8$).