Rats were provided by Vital River Experimental Technology Company Ltd. (Beijing, China). A total of 48 adult (age, 9–11 weeks) male Wistar rats, weighing 200–250 g, were housed 6 per cage at 20–25°C with a relative humidity of 50–60%, a 12 h light-dark cycle and free access to food and water. Rats were randomized to either receive vehicle (0.9% NaCl; 85 mg.kg-1.day-1) or D-galactose intraperitoneal injection (60 mg.kg-1.day-1) and AlCl₃ (25 mg.kg-1.day-1) for 90 days, to establish the aluminum neurotoxicity model. Following this, the learning and memory abilities of the rats were observed using a Morris water maze. The serum levels of LPS, TNF-α, and IL-1 were determined by the chromogenic Limulus amoebocyte lysate (CLAL) and radioimmunity assays. The level of DAO was measured using a Diamine oxidase colorimetric assay kit (Shanghai Bairui Biotech Co., Ltd., Shanghai, China) according to the manufacturer's instructions. The levels of Gln were measured using a Glutamine Colorimetric Assay kit (Biovision, Inc., Milpitas, CA, USA), according to the manufacturer's instructions, and glutaminase was measured using the Glutaminase ELISA assay kit (cat. no. 201233721; Cusabio Biotech Co., Ltd., Wuhan, Hubei, China) according to the manufacturer's instructions. All experimental animal procedures were approved by the Laboratory Animal Use and Care Committee of Shanxi Medical University (Taiyuan, China; permit no. SXMU-2012-18;) and the Ethics Committee of Animal Experiments of Shanxi Medical University (permit no. 20120623-1). All surgery was performed under sodium pentobarbital anesthesia and all possible efforts were made to minimize the suffering of the experimental rats. The rats were sacrificed via anesthetic overdose.

The Morris water maze was used to test the learning and memory of the rats for 5 days (n=10/group) and was performed 8 weeks following administration (10). For this experiment, a circular stainless steel tank (155 cm in diameter and 60 cm deep) containing water (24–26°C) to a depth of 40 cm was used; the extramaze cues, which consisted of four uniquely colored and textured inserts (coarse plastic mesh, metal screen, ridged plastic and sandpaper) placed on the arms of the plus-maze, were kept constant. The rats learned to escape from the water by searching for a submerged, and therefore hidden, Plexiglas platform (10 cm in diameter and 38.5 cm high), which was kept in a fixed position. To familiarize them with the experimental situation, prior to the experiment, the rats swam for 90 sec with the platform removed. During each trial, the animal was placed in the water facing the wall at one of five designated starting points. Animals that failed to find the platform within 60 sec were placed onto the platform and kept there for 30 sec, then the rats were removed from the maze until the next repeat of the experiment. The time taken to find the platform, termed the escape latency, was measured. This test was performed once for each of the 5 starting points per day for 5 days.

The CLAL test kit was obtained from Chromogenix (cat. no. 20120056; Molndal, Sweden). The analysis was conducted according to the manufacturer's protocol. The coefficient of variation was 8% (n=10/group).

The expression of LYZ in the liver and Aβ1-40 in the brain (hippocampus) were detected via immunohistochemistry (11–13). Rabbit anti-rat/−rabbit LYZ and Aβ1-40 IgG antibodies (Abcam, Shanghai, China) were used. Following the completion of the Morris water maze experiments, 12 rats (6 rats/group) were sacrificed and the whole brain and liver were removed and dissected. Prior to sectioning, liver tissue and hippocampus samples were fixed in 10% formalin for 24 h at room temperature, then embedded in paraffin. The paraffin blocks were cooled to −10°C to ease sectioning. Using the Microm HM340 E microtome (Thermo Fisher Scientific Inc., Waltham, MA, USA), 4-µm sections were generated. The EnVision^™ staining kit was used according to the manufacturer's instructions, for immunohistochemical analyses [G|2 Doublestain System, Rabbit/Mouse (DAB+/Permanent RED); cat. no. K5361; Dako DK; Agilent Technologies, Inc., Santa Clara, CA, USA]. The optimal antigen retrieval method and concentration to use were determined per antibody. Following deparaffinization in dimethylbenzene twice at room temperature for 5 min each time, the slides were washed in PBS containing 0.05% Tween-20 prior to antigen retrieval in the microwave (for liver tissues; 2 min 780 W, 10 min 150 W and 10 min 80 W) or in 70% formic acid (for hippocampus samples). The slides were placed in cuvettes, as a part of the capillary gap staining method (Thermo Fisher Scientific, Inc.), which was performed as described previously (14). Endogenous peroxidase activity was blocked by the EnVision^™ dual endogenous enzyme block (Agilent Technologies, Inc.). Following washing with PBS containing 0.05% Tween-20, a second blocking step was applied using PBS, 0.1% bovine serum albumin (BSA; Sigma Aldrich; Merck KGaA, Darmstadt, Germany), 1% NHS and 0.2% Triton X-100. Then, the primary antibodies [rabbit anti-rat/−rabbit LYZ (cat. no. 20120789) and Aβ1-40 (cat. no. 20123349) IgG antibodies; Abcam, Shanghai, China; with a 1:3,000 dilution in 0.1% BSA/PBS] were added for 10 min at room temperature, followed by washing with PBS containing 0.05% Tween-20 and the addition of the EnVision^™ polymer/horseradish peroxidase (HRP) goat anti-rabbit secondary antibody (cat. no. 201213219; 1:500 dilution; Dako Cytomation; Agilent Technologies, Inc.) for 10 min at room temperature. Following washing with PBS, DAB+ (from EnVision^™ staining kit; Dako DK; Agilent Technologies, Inc.) was added to initiate the enzymatic reaction and allow for visualization of the elements. A total of 10 high magnification fields were randomly selected from each slice and the number of positive cells per high magnification field (upright light microscope; CX31-LV320; Olympus Corporation, Tokyo, Japan) were counted. Following this, the average number of positive cells per slice was calculated.

The S-100β level was measured using an S-100β ELISA kit (cat. no. 20120166; Neogen Europe, Ltd., Lansing, MI, USA) according to the manufacturer's protocol (n=10/group).

Protein was extracted from fresh brain tissue using a ProteoExtract Subcellular Proteome Extraction kit (Sigma Aldrich; Merck KGaA). Protein concentration in the lysates was evaluated using a bicinchoninic acid protein assay kit according to the manufacturer's instructions (Thermo Fisher Scientific, Inc.). Following 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (5 µg protein applied) and transfer onto a nitrocellulose membrane, the membranes were blocked with 5% BSA (Sigma Aldrich; Merck KGaA) and 0.05% Tween-20 in Tris-buffered saline (TBST) for 1 h at room temperature. This was followed by overnight incubation at room temperature with the following primary antibodies: Monoclonal rabbit anti-rat ZO-1 antibody (1:200; cat. no. 201212315; Abcam), monoclonal rabbit anti-rat APP antibody (1:500; cat. no. 20121232; Abcam), monoclonal rabbit anti-rat PS-1 antibody (1:1,000; cat. no. 20123343; Abcam), monoclonal rabbit anti-rat BACE1 antibody (1:500; cat. no. 201233561; Abcam) and monoclonal rabbit anti-rat β-actin antibody (1:1,000, cat. no. 201214231; Abcam). Following three washes with TBST, the membranes were incubated with a HRP-conjugated anti-rabbit antibody (1:5,000; Santa Cruz Biotechnology, Inc., Dallas, TX, USA) for 2 h at room temperature. The membranes were exposed to film (Kodak, Rochester, NY, USA) and developed. The films were scanned and analyzed using the Kodak IS440 Automated Digitizing System and image analytical software (ImageQuant TL v2005; GE Healthcare Life Sciences, Uppsala, Sweden) for densitometric analysis. Experiments were repeated 3 times.

Primers: APP (295 bp), forward 5′-GGATGCGGAGTTCGGACATG-3′ and reverse 5′-GTTCTGACTCTGCTCAAAG-3′; PS-1 (186 bp), forward 5′-AGATGCCTCCTCTGTCCTCA-3′ and reverse 5′-CCGTCTTTGGGCATACATCT-3′; BACE1 (309 bp), forward 5′-CACCGAGACCGACGAAGAGCC-3′ and reverse 3′-CTAATAGGCTATGGTCATGAGGGT-3′; GAPDH (307 bp), forward 5′-TTGCCAGTTGCTTTAGTGATA-3′ and reverse 5′-CTTTTTCCCCCATTTCATTTC-3′.

Total RNA from fresh brain tissue samples was extracted using the TRIzol® reagent (Takara Biotechnology Co., Ltd., Dalian, China) according to the manufacturer's instructions. RNA was reverse transcribed to cDNA using the Takara PrimeScript 1st Strand cDNA Synthesis kit, according to the manufacturer's instructions, and the obtained cDNA was used as a template to perform PCR amplification. Semi-quantitative RT-PCR was performed with a mixture of the cDNA product, primers, dNTP and Taq DNA polymerase (Invitrogen; Thermo Fisher Scientific, Inc.). The thermocycling conditions used were as follows: 95°C for 30 sec, 55°C for 30 sec and 72°C for 30 sec for 25 cycles, followed by a final extension at 72°C for 5 min. Results were normalized to the APP reference gene. Electrophoresis was performed on 1.5% agarose gel (10 mA; 100 V; 30 min). The images were scanned and then analyzed with Multi-Analyst software (Bio-Rad Laboratories, Inc., Hercules, CA, USA).

Data were analyzed using the SPSS software version 22.0 (IBM Corp., Armonk, NY, USA). All values are expressed as the mean ± standard deviation. Statistical analysis was performed using an unpaired Student's t-test, multiple-factor repetitive measurement and one-way repeated measures analysis of variance (for learning and memory abilities). P<0.05 was considered to indicate a statistically significant difference.

The rat escape latency in the Morris water maze from days 1–5 was 59±22, 52±18, 38±13, 30±11 and 25±9 sec, respectively, in the D-galactose + AlCl₃ group and 39±12, 30±10, 24±8, 15±7 and 12±5 sec, respectively, in the control group (Fig. 1A and B). These data are additionally presented in Fig. 1C. There were no differences in the two groups from days 1–3 (P>0.01). However, from days 4–5, the escape latency in the D-galactose + AlCl₃ group was increased compared with the control group (30±11 vs. 15±7 sec and 25±9 vs. 12±5 sec, respectively; P<0.01). From days 4–5, the swim path of the control group was altered from a random line to a straight line. However, the swim path of the test group remained random (Fig. 1B). The memory ability was decreased in the rats treated with D-galactose and AlCl₃.

The geometric means of the LPS, TNF-α and IL-1 levels in the normal rats were 0.76±0.08 EU/ml, 1.0±0.11 ng/ml and 23.96±3.39 pg/ml, respectively. However, those in the test group were 0.89±0.07 EU/ml, 1.22±0.16 ng/ml and 38.38±3.48 pg/ml, respectively, with a significant difference observed (P<0.01; Fig. 2).

There were numerous brown LYZ cells in the control rat livers (30.6±8.2; n=10), however fewer LYZ cells were present in the livers of the aluminum neurotoxicity model rats (18.1±5.1; n=10; P<0.01 vs. control group; Fig. 3). This observation indicated decreased Kupffer cell function.

The levels of DAO and Gln in the serum and intestinal mucosa of the model rats were significantly greater compared with control rats (P<0.01), whereas the level of glutaminase was decreased in the model rats compared with control rats (P<0.01; Table I).

The expression of S-100β in the serum (0.33±0.18) was increased in the model rats compared with control rats (0.11±0.03; n=10; P<0.01). The expression of ZO-1 in the brain was increased in the control rats (53.3±8.4; ratio of ZO-1/β-Actin) compared with the model rats (26.9±5.6; P<0.01; Fig. 4). These data demonstrated that the BBB function was decreased and permeability was increased in the model group.

As presented in Fig. 5, the addition of D-galactose and AlCl₃ resulted in a marked increase in the expression of APP, PS1 and BACE1 mRNA (72±12, 52±10 and 72±15%, respectively) that were ~two-fold increases compared with control rats (30±6, 32±7 and 36±9%, respectively). The protein expression levels of APP, PS1 and BACE1 in the model and control groups demonstrated a similar trend to the aforementioned mRNA results.

As presented in Fig. 6, the expression of Aβ1-40 in the hippocampal brain of the aluminum neurotoxicity model rats (24.1±5.1) was increased compared with control rats (8.6±2.2; n=10; P<0.05; Fig. 6C).