LW-AFC, a new formula derived from Liuwei Dihuang decoction, ameliorates behavioral and pathological deterioration via modulating the neuroendocrine-immune system in PrP-hAβPPswe/PS1ΔE9 transgenic mice

Background Accumulating evidence implicates the neuroendocrine immunomodulation (NIM) network in the physiopathological mechanism of Alzheimer’s disease (AD). Notably, we previously revealed that the NIM network is dysregulated in the PrP-hAβPPswe/PS1ΔE9 (APP/PS1) transgenic mouse model of AD. Methods After treatment with a novel Liuwei Dihuang formula (LW-AFC), mice were cognitively evaluated in behavioral experiments. Neuron loss, amyloid-β (Aβ) deposition, and Aβ level were analyzed using Nissl staining, immunofluorescence, and an AlphaLISA assay, respectively. Multiplex bead analysis, a radioimmunoassay, immunochemiluminometry, and an enzyme-linked immunosorbent assay (ELISA) were used to measure cytokine and hormone levels. Lymphocyte subsets were detected using flow cytometry. Data between two groups were compared using a Student’s t test. Comparison of the data from multiple groups against one group was performed using a one-way analysis of variance (ANOVA) followed by a Dunnett’s post hoc test or a two-way repeated-measures analysis of variance with a Tukey multiple comparisons test. Results LW-AFC ameliorated the cognitive impairment observed in APP/PS1 mice, including the impairment of object recognition memory, spatial learning and memory, and active and passive avoidance. In addition, LW-AFC alleviated the neuron loss in the hippocampus, suppressed Aβ deposition in the brain, and reduced the concentration of Aβ1–42 in the hippocampus and plasma of APP/PS1 mice. LW-AFC treatment also significantly decreased the secretion of corticotropin-releasing hormone and gonadotropin-releasing hormone in the hypothalamus, and adrenocorticotropic hormone, luteinizing hormone, and follicle-stimulating hormone in the pituitary. Moreover, LW-AFC increased CD8+CD28+ T cells, and reduced CD4+CD25+Foxp3+ T cells in the spleen lymphocytes, downregulated interleukin (IL)-1β, IL-2, IL-6, IL-23, granulocyte-macrophage colony stimulating factor, and tumor necrosis factor-α and -β, and upregulated IL-4 and granulocyte colony stimulating factor in the plasma of APP/PS1 mice. Conclusions LW-AFC ameliorated the behavioral and pathological deterioration of APP/PS1 transgenic mice via the restoration of the NIM network to a greater extent than either memantine or donepezil, which supports the use of LW-AFC as a potential agent for AD therapy. Electronic supplementary material The online version of this article (doi:10.1186/s13195-016-0226-6) contains supplementary material, which is available to authorized users.


Preparation of LW-AFC and HPLC analysis
LW-AFC was prepared from an Liuwei Dihuang decoction (LW), a traditional Chinese medical prescription [31]. LW was prepared as previously described in Yang et al. [32], Zhang et al. [33,34], Cheng et al. [35], and Kusters et al. [36]. LW-AFC is composed of a polysaccharide fraction (LWB-B), a glycoside fraction (LWD-b), and an oligosaccharide fraction (CA-30). The LW was passed through a six-layer gauze filter, and the extracted solution was centrifuged. The supernatant was concentrated and then extracted in ethanol to produce LWD. The sediment was rinsed in deionized water and concentrated into a dried polysaccharide fraction (LWB-B). The LWD ethanol elution fraction was dissolved using macroporous adsorptive resins to obtain the glycosides component (LWD-b). The water elution fraction of the LWD was dissolved using an active carbon absorption column to obtain the oligosaccharide component . In this manner, the LW-AFC formula is composed of 20.3% of the polysaccharide component (LWB-B), 15.1% of the glycosides component (LWD-b), and 64.6% of the oligosaccharide component (CA-30) in a dry weight ratio. The LW-AFC components were analyzed using high-performance liquid chromatography (HPLC). Briefly, for the CA-30 and LWD-b mixture, chromatographic separation was obtained on a Diamond C18 column; there were 17 chromatogram peaks in the fingerprint of the CA-30 and LWD-b mixture. For LWB-B, the chromatographic separation was obtained on a NucleosilNH 2 100 Å column; five chromatogram peaks were observed, representing fructose, glucose, sucrose, mannotriose, and stachyose. The retention times of these peaks were 6.260 min, 6.829 min, 8.186 min, 18.305 min, and 21.506 min, respectively.

Experimental animals
The male APP/PS1 mice and wild-type (WT) mice were obtained from Beijing HFK Bioscience Co. Ltd., via the Jackson Laboratory (Bar Harbor, ME, USA). The mice were maintained at the Beijing Institute of Pharmacology and Toxicology under standard housing conditions, i.e., room temperature at 22 ± 1°C and humidity at 55 ± 5%, and a 12-h light/12-h dark cycle. The mice were separately given water and pellet food ad libitum (provided by the Animal Center of the Academy of Military Medical Sciences). All behavioral tests were performed between 19:00 and 6:00 (Beijing time).
Nine-month-old male APP/PS1 mice were randomly separated into four groups. Each group contained 10-11 mice. LW-AFC was dissolved in distilled water at 160 mg/mL, memantine (Ouhe Chemical Ltd., Beijing, China) at 1 mg/mL, and donepezil (Ouhe Chemical Ltd.) at 0.1 mg/mL. The drug-treated mice group was given an intragastric administration of memantine, donepezil, or LW-AFC (0.1 mL/10 g body weight) once a day for 150 days. APP/PS1 mice as a model group and agematched WT mice (15 males) as a control group were given an equal volume of deionized water. The mice were weighed every 3 days. Drug administration and behavioral tests were conducted according to the experimental timelines (Fig. 1). Following the behavioral experiments, the whole brain, hippocampus, cortex, hypothalamus, pituitary, spleens, and plasma of each mouse were collected for immunofluorescence, Nissl staining, soluble Aβ analysis, hormone determination, lymphocytes subsets analysis, and cytokine analysis.

Behavioral tests Locomotor activity test
The locomotor activity test was carried out according to Cheng et al. [37]. Motor tracking was performed using a video-based behavior monitoring system (Jiliang Software Technology Co. Ltd., Shanghai, China). Each mouse was placed in an aluminum-plastic panel locomotor activity chamber, and recorded for 20 min. The total distance traveled for each mouse was recorded to indicate its spontaneous motor activity.

Novel object recognition test
The object recognition test was performed as described by Chen et al. [38] and Bevins and Besheer [39]. The mice were familiarized with the testing environment, a black Plexiglas apparatus (30 × 30 × 30 cm), for 20 min per day for 2 days before the object recognition test. In the learning phase (day 3), the animal was placed in the apparatus with two similar black square (4 × 4 × 4 cm) Plexiglas objects, A and B, which were equidistant from the sides (5 cm) of the chamber. The animals were allowed to freely explore the chamber for 16 min. The exploration time refers to the duration that the animal spent exploring the object with their head orientated towards the object and their nose within 1 cm of the object. Then, the mice were returned to their home cage for 1 h, after which they entered a 4-min test phase, where a different object (C) replaced object A or B. The preferential index was calculated to assess the object recognition memory of mice.

Morris water maze test
The Morris water maze test was performed according to Vorhees and Williams [40]. During the trial, curtains with unique geometric figures were placed at all sides of the pool to avoid visual interference. The spatial learning phase consisted of four trials per day for 5 days, and one additional day (day 6) for a probe trial. In the spatial learning phase, each mouse was placed on the platform for 60 s before the first trial, and then released into the water to find the platform within 60 s. If the mouse found the platform within 60 s, it was allowed to stay there for 10 s. If not, the mouse was gently led to the platform and allowed to remain there for 10 s, the latency time being scored as 60 s. The latency time was recorded as a measure of spatial learning. For the spatial memory phase, the platform was removed, and the mouse was released into the water at a novel position and allowed to swim freely for 60 s. The dependent measure for the spatial memory was the time in the target quadrant and the number of platform crossings.
Step-down test The step-down test was carried out according to Fang et al. [41], Lou et al. [42], and Shi et al. [43]. On the first day, mice were allowed to acclimatize for 2 min without the platform. During the learning trial, if the mouse stepped down from the platform (error) with all four paws it received an aversive foot electric shock (36 V, AC), and the learning course was performed for 10 min. The number of errors and the number of times that mice did not step down were scored for a 3-min period. For the testing trials (days 2-7), the procedure was repeated at the same time, and the testing time was 3 min. The number of times that the mice did not step down was recorded for a 3-min period.

Shuttle box test
The shuttle-box test was performed according to Cheng et al. [44]. Working memory was evaluated using a shuttle Fig. 1 Schematic diagram of the experimental procedure box apparatus (VT 05448, Med Associates Inc., East Fairfield, VT, USA). The training session began with 2 min of acclimatization to the chambers followed by 30 trials, and the inter-trial interval was 30 s. A tone (60 dB) and light (8 W) were presented as the conditioned stimulus, for 10 s, and followed by the unconditioned stimulus, an electrical foot shock (0.2 mA), for 5 s. If the mouse moved to the opposite chamber during the presentation of the conditioned stimulus, no electrical foot shock was presented and an active avoidance response was recorded. The shuttle-box procedure was performed for 5 consecutive days. On day 6, all of the mice were submitted to another session (no shock) to evaluate learning and memory, and the number of active avoidances was recorded.

Biochemical and histochemical analyses Immunofluorescence
Mouse brains were removed and one hemisphere was fixed via immersion in 4% paraformaldehyde in phosphatebuffered saline (PBS) (pH 7.4) at 4°C overnight, and then fixed in 10% buffered formalin, and paraffin embedded. Serial 5-μm thick sections were prepared, deparaffinized, hydrated, rinsed with PBS, and pretreated with 0.01 M citric acid for 15 min for antigen retrieval, and then with the blocking solution (2% fetal bovine serum in PBS) for 30 min. Subsequently, sections were incubated with mouse anti-β-amyloid (clone: 6E10, 1:100; Biolegend, San Diego, CA, USA) overnight at 4°C. After rinsing, the sections were incubated with goat anti-mouse IgG HRP (1:1000; ZSGB-Bio, Beijing, China) for 2 h at room temperature, then incubated with Opal520 working solution (PerkinElmer, Waltham, MA, USA) for 10 min, and mounted with DAPI-containing medium. The tissue sections were photographed, and the images were digitized with a fluorescence lifetime imaging microscope (Vectra 2, PerkinElmer-Caliper LS, Waltham, MA, USA). The area of Aβ deposits in each slice section was quantified using Image Pro Plus 6.0 software.

Nissl staining
The sections were stained using 0.5% cresyl violet acetate (Beyotime, Beijing China). Stained sections were scanned using a transmission electron microscope (H-7650, Hitachi, Tokyo, Japan). The integrated optical density (IOD) of Nissl bodies in the CA1 and CA3 regions was quantified using the Image Pro Plus 6.0 software.

Soluble Aβ analysis
The Aβ AlphaLISA assay was carried out according to Cheng et al. [45] and Tesseur et al. [46]. The hippocampus and cortex from one brain hemisphere of each mouse was sequentially extracted. The extracted tissue was separately homogenized using an ultrasonic disintegrator in 50 mmol/ L Tris-HCl, pH 8.0, and 5 mol/L guanidine hydrochloride.
After 3 h at room temperature, the suspension was diluted in Dulbecco's phosphate-buffered saline (0.03% Tween-20, 5% fetal bovine serum, PBS, pH 7.4) and complete proteinase inhibitor cocktail (Roche, Indianapolis, IN, USA), centrifuged at 16,000 g for 20 min at 4°C, and the supernatants containing the soluble fraction were used for measuring soluble Aβ. Plasma was prepared from the obtained supernatant by centrifuging the blood plus 4% EDTA-Na 2 at 3000 g for 15 min at 4°C. The Aβ 1-40 and Aβ 1-42 content in the hippocampus, cortex and plasma were determined using the AlphaLISA technique and the AlphaLISA® human amyloid beta 1-40 (high specificity) (AL275C, PerkinElmer, Waltham, MA, USA) and Alpha-LISA® human amyloid beta 1-42 (high specificity) (AL276C, PerkinElmer) Kits according to the manufacturers' instructions.

Radioimmunoassay of hypothalamic and hypophyseal hormones
The hypothalamuses and pituitaries were weighed and boiled in 1 mL of saline for 5 min. Peptides were extracted by homogenizing the hypothalamuses and pituitaries in 0.5 mL of 1 mol/L glacial acetic acid followed by centrifuging the mixture at 3000 rpm for 30 min. Supernatants were stored at -20°C. Concentrations of adrenocorticotropic hormone (ACTH), luteinizing hormone (LH), follicle-stimulating hormone (FSH), corticotropin releasing hormone (CRH), and gonadotropin-releasing hormone (GnRH) in the supernatants were determined with a 125 I-ACTH RIA kit (North Institute of

Enzyme-linked immunosorbent assay
The corticosterone (CORT) level in the plasma of the mice was measured using a precoated corticosterone enzymelinked immunosorbent assay (ELISA) kit (EC3001-1, ASSAYPRO, Charles, MO, USA) according to the manufacturer's instruction. The absorbance was measured at 450 nm with a reference wavelength of 570 nm using Enspire™ multilabel reader 2300 (Perkin Elmer, Turku, Finland).

Immunochemiluminescence assay
The level of testosterone (T) in the plasma of the mice was measured using an Access Immunoassay System (Beckman Coulter, Brea, CA, USA), access testosterone (33560, Beckman Coulter), and access testosterone calibrators (33565, Beckman Coulter). The entire measurement was automatically processed according to the scheduled program.

Statistical analysis
All data are expressed as the mean ± SEM. GraphPad Prism 6.0 (GraphPad Software, Inc., La Jolla, CA, USA) was used to plot and analyze the data. Data between two groups were compared using a Student's t test. Comparison of the data from multiple groups against one group was performed using a one-way analysis of variance (ANOVA) followed by a Dunnett's post hoc test or a two-way repeated-measures analysis of variance with a Tukey multiple comparisons test. P < 0.05 was considered statistically significant.

LW-AFC improves the cognitive impairment of APP/PS1 mice
The locomotor activity test evaluated the spontaneous motor activity of APP/PS1 mice; no significant difference was observed between the treatment groups (Fig. 2a). This result indicates that the spontaneous locomotor activity of mice did not influence the results of the other behavioral experiments.
The novel object recognition test evaluated the object recognition memory of mice. LW-AFC treatment in APP/PS1 mice significantly decreased the preferential index (Fig. 2b), indicating that the object recognition memory deficit of APP/PS1 mice was ameliorated after LW-AFC treatment; this effect was superior to that of memantine or donepezil.
The Morris water maze test evaluated the spatial learning and memory of APP/PS1 mice. For the learning task, APP/PS1 mice had longer escape latencies than WT mice on the final test day, and the latencies of the LW-AFC-or memantine-treated APP/PS1 mice were significantly longer than those of the mice in the nontreated group (Fig. 2c1). These results indicate that both LW-AFC and memantine ameliorated the spatial learning impairment of APP/PS1 mice. For the probe trial, the escape latency was longer (Fig. 2c2), the number of plate crossings decreased (Fig. 2c3), and the time in the target quadrant was shorter (Fig. 2c4), but swimming speed was not significantly different (data not shown) in APP/PS1 mice compared with WT mice. The escape latencies decreased and the number of plate crossings increased in LW-AFC-or memantine-treated mice, while the time in the target quadrant was elevated only in mice treated with LW-AFC. These results indicate that LW-AFC and memantine administration significantly improved the spatial learning and memory deficits of APP/PS1 mice.
The step-down test evaluated passive avoidance in APP/PS1 mice. Compared with the WT mice, the number of errors and training time in APP/PS1 mice significantly increased ( Fig. 2d1 and d2), and the memory retention of APP/PS1 mice tended to decrease (Fig. 2d3). LW-AFC administration significantly decreased the number of errors and training time. These results indicate that LW-AFC improved the passive avoidance impairment of APP/PS1 mice; this effect was greater than that observed for either memantine or donepezil.
The shuttle-box test evaluated active avoidance in APP/PS1 mice. Significantly fewer successful avoidance times were observed for APP/PS1 mice than for WT mice after the third day of the training phase, but significantly increased after treatment with either LW-AFC, memantine, or donepezil after the third day (Fig. 2e1). Significantly shorter successful avoidance times were observed for APP/PS1 mice than for WT mice in the testing phase, but increased after LW-AFC, memantine, or donepezil administration (Fig. 2e2). These results indicate that the deteriorated active avoidance response of APP/PS1 mice was ameliorated after LW-AFC, memantine, or donepezil administration.
LW-AFC treatment decreases neuronal loss in the hippocampus of APP/PS1 mice Nissl staining revealed typical neuropathological changes in the CA1 and CA3 region of the hippocampus in APP/ PS1 mice compared to WT mice, including neuron loss and nucleus shrinkage or disappearance (Fig. 3a). Furthermore, significantly lower Nissl body numbers were observed in the whole brain, hippocampus, and CA1 and CA3 regions of APP/PS1 mice than in those regions in WT mice (Fig. 3b-e). LW-AFC and memantine treatment significantly decreased these neuropathological changes and increased the density of healthy neurons in the hippocampus and CA3 region of APP/PS1 mice ( Fig. 3c and e). These findings indicate that LW-AFC and memantine protected against neuronal loss in the hippocampus of APP/PS1 mice.

LW-AFC treatment alleviates Aβ deposition in the brain of APP/PS1 mice
Aβ deposition in the brain is a typical pathological sign of AD in patients and animal models. Our results show that APP/PS1 mice developed a significant number of Aβ plaques in the brain at 14 months, while Aβ plaques were not observed in the WT mice (Fig. 4a). LW-AFCor memantine-treated mice had a significantly smaller area of Aβ deposits in the whole brain and hippocampus, while donepezil had a less prominent influence on Aβ plaque formation in APP/PS1 mice (Fig. 4b and c). These results indicate that the Aβ deposition in the brain of APP/PS1 mice was alleviated after LW-AFC or memantine administration. Fig. 2 The effect of LW-AFC on the cognitive deterioration of APP/PS1 mice. Shown are: a the total distance in a spontaneous locomotor activity test; b the preferential index (time on novel object C/( time on novel object C + time on sample object A) × 100% ) in a novel object recognition test; the escape latency of c1 learning task and c2 probe trial, c3 number of times that the mice crossed the removed platform, and c4 swimming time in the target quadrant in the Morris water maze test; The d1 number of errors, d2 training time, and d3 memory retention in the step-down test; and the successful avoidance times during e1 training and e2 testing phases in the shuttle-box test. The values are mean ± SEM; n = 11-15. *P < 0.05, ***P < 0.001, versus the wild-type (WT) mouse group by unpaired Student's t test; # P < 0.05, ## P < 0.01, ### P < 0.001, versus the APP/PS1 (Tgs) mouse group by one-way ANOVA analysis followed by Dunnett's post hoc test and two-way repeated-measures analysis of variance with Tukey multiple comparisons test

LW-AFC treatment decreases Aβ 1-40 and Aβ 1-42 levels in the brain and blood of APP/PS1 mice
The results of the AlphaLISA assay showed that the concentrations of Aβ 1-42 and Aβ 1-40 in the hippocampus and plasma of APP/PS1 mice were significantly higher than that of WT mice (Fig. 5). LW-AFC or memantine treatment led to significantly lower Aβ 1-42 levels in the hippocampus of APP/PS1 mice than in those of WT mice, and Aβ 1-42 levels in the plasma were decreased only after LW-AFC administration. These findings indicate that LW-AFC downregulates Aβ 1-42 levels in the brain and blood of APP/PS1 mice, while similar effects for memantine were observed only in the brain.

LW-AFC restores the NIM network in APP/PS1 mice
To further investigate whether LW-AFC affected the hypothalamic-pituitary-adrenal (HPA) axis and the hypothalamic-pituitary-gonadal (HPG) axis in APP/PS1 mice, the concentration of GnRH and CRH in the hypothalamus, and ACTH, FSH, and LH in the pituitary, were measured using a radioimmunoassay. The concentration of T and CORT in the plasma was measured using a chemiluminescence assay and ELISA, respectively. The results show that, within the HPA axis, the concentration of CRH, ACTH, and CORT were significantly higher in APP/PS1 mice than in WT mice (Fig. 6a-c). LW-AFC significantly decreased the CRH level (Fig. 6a), while both LW-AFC and memantine reduced ACTH (Fig. 6b). Within the HPG axis, the concentration of GnRH, FSH, and LH increased in APP/PS1 mice (Fig. 6d-f), but T (data not shown) was not significantly different between the APP/PS1 and WT mice. LW-AFC significantly decreased the concentration of GnRH (Fig. 6d), while both LW-AFC and memantine reduced FSH and LH levels ( Fig. 6e and f) in APP/PS1 mice. These data indicate that both LW-AFC and memantine had an ameliorative effect on the endocrine system in APP/PS1 mice, especially the HPA and HPG axes.
A Pearson correlation analysis was performed to identify altered endocrine hormone expression after LW-AFC treatment that was associated with cognitive impairment, neuron loss, and Aβ deposition in APP/PS1 mice. The results show that among the endocrine hormones with altered expression after LW-AFC treatment, CRH levels were correlated with neuron loss and Aβ deposits in APP/ PS1 mice. ACTH levels were correlated with object recognition memory, spatial memory, and an active avoidance response. GnRH levels were correlated with object recognition memory, an active avoidance response, and neuron loss. LH levels were correlated with an active avoidance response and neuron loss. FSH levels were correlated with object recognition memory, a passive avoidance response, an active avoidance response, and neuron loss in APP/PS1 mice (see Additional file 1: Table S1).

LW-AFC modulates the impairment of lymphocyte subsets in APP/PS1 mice
To investigate the effect of LW-AFC on the expression of lymphocyte subsets in APP/PS1 mice, the expression of CD3 + CD4 + T cells, CD3 + CD8 + T cells, CD8 + CD28 + T cells, CD3 + CD25 + Foxp3 + T cells, CD19 + B cells, and CD19 + CD80 + B cells were detected using flow cytometry. Significantly fewer CD8 + CD28 + T cells (Fig. 7a) and significantly more CD3 + CD25 + Foxp3 + T cells (Fig. 7b) were observed in APP/PS1 mice than in WT mice. Expression of the other lymphocyte subsets in APP/PS1 mice (data not shown) was not significantly different. In APP/PS1 mice, the expression of CD8 + CD28 + T cells increased after memantine, donepezil, or LW-AFC treatment (Fig. 7a), while the expression of CD4 + CD25 + Foxp3 + T cells decreased after memantine or LW-AFC treatment (Fig. 7b). These results indicate that memantine or LW-AFC treatment partially restored normal lymphocyte expression in APP/PS1 mice.
These results indicate that cytokine secretion in APP/PS1 mice was abnormal and that administration of LW-AFC and memantine restored this aberrant immune function in APP/PS1 mice.
The results of a Pearson correlation analysis showed that among the cytokines with altered expression after Fig. 4 Suppressive effects of LW-AFC on Aβ deposits in the hippocampus of APP/PS1 mice. a Representative immunofluorescence staining images showing amyloid-β (Aβ) deposits (green and indicated by white arrows) in the hippocampus and brain of wild-type (WT) and APP/PS1 (Tgs) mice. Quantification of Aβ deposits in the brain (b) and hippocampus (c) of WT and Tgs mice by Image Pro Plus 6.0 software. Red arrows indicate false positive result of Aβ deposits. The values are mean ± SEM; n = 11-15. **P < 0.01, ***P < 0.001, versus the WT mice by unpaired Student's t test; # P < 0.05, ## P < 0.01, versus the Tgs mice by one-way ANOVA analysis followed by Dunnett's post hoc test LW-AFC administration, the levels of IL-1β, GM-CSF, TNF-α, TNF-β, and eotaxin were correlated with object recognition memory, spatial learning and memory, passive avoidance response, active avoidance response, neuron loss, and Aβ deposits in APP/PS1 mice. IL-2 levels were correlated with object recognition memory, spatial learning and memory, an active avoidance response, and neuron loss. IL-6 and G-CSF levels were correlated with object recognition memory, spatial memory, a passive avoidance response, an active avoidance response, neuron loss, and Aβ deposits. IL-23 levels were correlated with object recognition memory, spatial memory, a passive avoidance response, and an active avoidance response. IL-4 levels were correlated with object recognition memory, spatial memory, a passive avoidance response, an active avoidance response, and Aβ deposits (see Additional file 1: Table S1 and Additional file 2: Figure S1).

Discussion
In AD patients, cognitive impairments and psychological symptoms are associated with an early dysfunction of the HPA [47] and HPG axes [48][49][50][51]. However, few studies have investigated neuroendocrine function in APP/PS1 mice or analogous models, although one study reported that plasma CORT levels were increased in 8month-old male APP/PS1 mice [52]. In addition, the neuroendocrine system is disturbed in Aβ knockin-or injection-induced mouse models [23,[53][54][55][56]. Several lines of evidence also indicate that hormones in the neuroendocrine system regulate pathogenic Aβ accumulation in AD animal models and patients [20,22,[57][58][59][60]. In the present study, the HPA and HPG axes were significantly disturbed in APP/PS1 mice as compared with those of age-matched WT mice. These disturbances might contribute to the acceleration of Aβ pathogenesis, neuron loss, and cognitive impairment. Our results show that long-term treatment with LW-AFC restored the neuroendocrine system of APP/PS1 mice, while memantine had little impact on the function of the HPA and HPG axes.
The immune system has a fundamental role in the development and progression of AD. AD patients show altered lymphocyte expression that includes CD4 + and CD8 + T cells [61][62][63][64][65][66], CD8 + CD28 + T cells [65], CD4 + CD25 + Foxp3 + T cells [67], and B cells [66,68,69]. AD patients have increased levels of CD8 + CD28 + T cells and reduced levels of cytotoxic CD28cells, which leads to T helper cell unresponsiveness [65]. Nevertheless, we found lower levels of CD8 + CD28 + T cells in APP/PS1 mice than in WT mice. This discrepancy might be due to the overexpression of a Swedish variant of the gene encoding hAPP along with the overexpression of a mutant PS1 gene in APP/PS1 mice [70], but not a multifactorial inducer in AD patients. The values are mean ± SEM; n = 11-15. **P < 0.01, ***P < 0.001, versus the WT mice by unpaired Student's t test; # P < 0.05, ## P < 0.01, versus the Tgs mice by one-way ANOVA analysis followed by Dunnett's post hoc test CD4 + CD25 + Foxp3 + T cells were also elevated in AD patients [64,65,71], which may be related to AD progression, as patients with a mild cognitive impairment (MCI) show lower CD4 + CD25 + Foxp3 + T-cell activity than AD patients [67]. Our results show that LW-AFC or memantine administration ameliorated this change in CD4 + CD25 + Foxp3 + T cells in APP/PS1 mice.
Cytokines play a critical role in brain inflammation in AD, and are involved in complex cognitive processes [72]. An aberrant cytokine concentration is found in AD Fig. 6 The effect of LW-AFC on the hormone secretion of the HPA and HPG axes in APP/PS1 mice. The hormones of the HPA axis including a corticotropin releasing hormone (CRH) in the hypothalamus, b adrenocorticotropic hormone (ACTH) in the pituitary, and c corticosterone (CORT) in the plasma. The HPG axis including d gonadotropin-releasing hormone (GnRH) in the hypothalamus, e luteinizing hormone (LH), and f follicle-stimulating hormone (FSH) in the pituitary. The values are mean ± SEM; n = 9-11. *P < 0.05, **P < 0.01, ***P < 0.001, versus the wildtype (WT) mouse group by unpaired Student's t test; # P < 0.05, ## P < 0.01, versus the APP/PS1 (Tgs) mouse group by one-way ANOVA analysis followed by Dunnett's post hoc test Fig. 7 The effect of LW-AFC on the subsets of spleen lymphocytes in APP/PS1 mice. Flow cytometric analysis of CD8 + CD28 + T cells (a) and CD4 + CD25 + Foxp3 + T cells (b) in the spleen supernatant of mice. The values are mean ± SEM; n = 3. **P < 0.01, ***P < 0.001, versus the wild-type (WT) mouse group by unpaired Student's t test; ## P < 0.01, ### P < 0.001, versus the APP/PS1 (Tgs) mouse group by one-way ANOVA analysis followed by Dunnett's post hoc test patients [73], and contributes to the impairment of learning and memory [72]. IL-1β, IL-2, and IL-6 modulate synaptic transmission and plasticity in the hippocampus [72], inhibit long-term potentiation [74,75], affect various forms of hippocampal-dependent memory [76], and impair planning [77]. The upregulation of IL-23 contributes to age-associated brain dysfunction, the modulation of Aβ, and neuronal loss [24,78]. High levels of GM-CSF in the plasma of AD transgenic mice correlate with the expansion of regulatory T cells that suppress the effector T-cell response to Aβ 1-42 [79]. Moreover, increasing the peripheral eotaxin concentration in young mice inhibited adult neurogenesis and learning and memory [80]. AD model mice treated with G-CSF show a significant reversal of cognitive deficits, and decreased Aβ deposition and soluble Aβ levels in the periphery and hippocampus [21]. The contribution of IL-23 to AD pathogenesis is not fully understood [81]. Previous studies demonstrated that the cytokine network in APP/ PS1 mice is indiscriminate as compared to that in WT mice [82][83][84][85][86]. In the present study, we measured 17 cytokines in the plasma of APP/PS1 mice and found that longterm administration of LW-AFC ameliorated the secretion of 10 of these cytokines. The association of many of these Fig. 8 The effect of LW-AFC on the cytokines in the plasma of APP/PS1 mice. Concentrations (pg/mL) of a interleukin-1β (IL-1β), b interleukin-2 (IL-2), c interleukin-6 (IL-6), d interleukin-23 (IL-23), e granulocyte-macrophage colony stimulating factor (GM-CSF), f tumor necrosis factor α (TNF-α), g tumor necrosis factor β (TNF-β), h eotaxin, i interleukin-4 (IL-4), and j granulocyte colony stimulating factor (G-CSF) in the blood plasma of wild-type (WT) and APP/PS1 (Tgs) mice were detected using Luminex® X-MAP® technology. The values are mean ± SEM; n = 9-15. **P < 0.01, ***P < 0.001, versus WT mouse group by unpaired Student's t test; # P < 0.05, ## P < 0.01, ### P < 0.001, versus Tgs mouse group by one-way ANOVA analysis followed by Dunnett's post hoc test cytokines (including eotaxin, G-CSF, IL-2, IL-23, and TNF-β) with AD-like learning and memory deficits in APP/PS1 mice is a novel observation.
Our study shows that, in APP/PS1 mice, the concentrations of TNF-β and IL-1β in the blood are correlated with object recognition memory, spatial learning and memory, a passive or active avoidance response, neuron loss, and Aβ deposits, and could be regulated by LW-AFC treatment. Under normal immunological conditions, TNF is essential for learning and memory [87]. The cognitive dysfunction induced in mice after overexpression of TNF might be due to decreased nerve growth factor (NGF) levels [88] and/or the modulation of synaptic plasticity [89][90][91], impaired long-term potentiation (LTP), and neurodegeneration [92]. The TNF receptor is crucial for Aβ-mediated cell death [93,94]. Elevated cerebrospinal fluid (CSF) levels of soluble TNF receptors is observed in patients with MCI and in AD patients [95], and chronic administration of thalidomide, a well-known TNF-α inhibitor, reduced Aβ load, plaque formation, and BACE1 levels and activity in APP23 transgenic mice [96]. Our results show that chronic administration of LW-AFC decreased the elevated levels of TNF-α and TNF-β in the blood of APP/PS1 mice. Some studies report that overexpressing IL-1β leads to an increase in the microglia and astrocytes surrounding Aβ plaques in AD patients and animal models [97,98]. Moreover, patients with MCI and AD exhibited a significant increase in peripheral IL-1β levels compared to controls [99]. Sustained hippocampal IL-1β overexpression exacerbated tau phosphorylation and tangle formation via aberrant activation of p38 mitogen-activated protein kinase and glycogen synthase kinase 3 [100,101], and impaired long-term contextual and spatial memory [102]. Some agents can alleviate AD symptoms via inhibiting IL-1β release, such as nimodipine [103], Gossypium herbaceam L. extracts [104], linalool [105], melatonin [106], and anti-IL-1R blocking antibody [100]. The present study shows that LW-AFC treatment reverses the decrease in IL-1β levels in the blood of APP/ PS1 mice.