The Effects of LW-AFC on the Hippocampal Transcriptome in Senescence-Accelerated Mouse Prone 8 Strain,a Mouse Model of Alzheimer’s Disease

Abstract

The senescence-accelerated mouse prone 8 (SAMP8) strain is considered a robust experimental model for developing preventative and therapeutic treatments for Alzheimer’s disease (AD), a neurodegenerative disease which cannot be effectively prevented, halted, or cured. Our previous studies showed that LW-AFC, a new formula derived from the classical traditional Chinese medicinal prescription Liuwei Dihuang decoction, ameliorates cognitive deterioration in PrP-hAβPPswe/PS1^ΔE9 transgenic mice and SAMP8 mice. This study aims to investigate the mechanism that mediates how LW-AFC improves cognitive deficit on the basis of the transcriptome. We conducted a genome-wide survey of gene expression in the hippocampus in mice from the senescence accelerated mouse resistant 1 (SAMR1) strain, from SAMP8 and from LW-AFC treated SAMP8. The results showed that LW-AFC reversed the transcriptome in the hippocampus of SAMP8 mice. The specific investigation of altered gene expression in subtypes defined by cognitive profiles indicated that the systemic lupus erythematosus pathway, spliceosomes, amyotrophic lateral sclerosis, and the insulin signaling were involved in the improvement of cognitive ability by LW-AFC. The expression of genes Enpp2, Etnk1, Epdr1, and Gm5900 in the hippocampus were correlated with that of LW-AFC’s ameliorating cognitive impairment in SAMP8 mice. Because LW-AFC is composed of polysaccharides, glycosides, and oligosaccharides, we infer that LW-AFC has direct or indirect effects on altering gene expressions and regulating pathways in the hippocampus of SAMP8 mice. These data are helpful for the enhanced identification of LW-AFC as new therapeutic modalities to AD.

Keywords

Alzheimer’s disease LW-AFC senescence-accelerated mouse prone 8 strain traditional Chinese medicine transcriptome

INTRODUCTION

Alzheimer’s disease (AD) is a progressive, neurodegenerative disorder, and an effective treatment for this disorder remains elusive. The spontaneous senescence-accelerated mouse prone 8 (SAMP8) strain harbors the behavioral and histopathological signatures of AD and is considered a robust experimental model for developing treatments for AD [1 –4]. LW-AFC is a new formula derived from the classical traditional Chinese medicinal prescription Liuwei Dihuang decoction and is composed of polysaccharides, glycosides, and oligosaccharides. Our previous studies showed that LW-AFC ameliorated cognitive deterioration, such as object recognition memory, spatial learning and memory, and active and passive avoidance, in PrP-hAβPPswe/PS1^ΔE9 transgenic mice [5] and SAMP8 mice [6 –8]. AD is a multifactorial, complex syndrome and thus, a multitarget-driven drug discovery may provide better therapeutic benefits than monotherapies for AD. There are many chemical entities (e.g., catalpol, pachymic acid, sweroside, benzoic acid, acteoside, loganin, morroniside, peoniflorin, 5-methoxyl-methyl-furaldehyde, ursolic acid, gallic acid, etc.) in LW-AFC. LW-AFC is one type of multitargeted drug and has beneficial effects in treating AD animal models due to its targeting several targets or pathways [5 –9]. Previous studies have indicated that LW-AFC may be a promising therapeutic medicine for AD.

Comprehensive understanding of SAMP8 mice pathogenesis and LW-AFC pharmacological mechanisms require a systemic view at the genomic level. In the current study, we primarily focus on the role of LW-AFC in the hippocampus transcriptome in SAMP8 mice. We conducted a genome-wide survey of gene expression in the hippocampus from the controls of SAMP8 mice, senescence accelerated mouse resistant 1 (SAMR1) strain mice, SAMP8 mice, and LW-AFC treated SAMP8 mice. To specifically investigate the altered gene expression in subtypes defined by cognitive profiles, we applied a weighted correlation network analysis (WGCNA) [10] to explore correlation patterns among genes in SAMP8 mice from a systemic biological perspective. We used this method to reveal coexpressed gene modules that are highly correlated with related cognitive behavior data. Finally, we identified the biological pathways and the genes associated with cognitive subtypes of SAMP8 mice and regulated by LW-AFC. This study provides evidence that LW-AFC ameliorates cognitive deterioration in SAMP8 mice due to the modulation of the brain transcriptome.

METHODS

Preparation of LW-AFC, animals and drug administration, and behavioral tests

LW-AFC was prepared from a Liuwei Dihuang decoction which included a polysaccharide fraction, a glycoside fraction, and an oligosaccharide fraction. LW-AFC was prepared as previously described in Wang’s studies [5 –9]. Original SAMR1 and SAMP8 mice were kindly provided by Dr. T. Takeda at Kyoto University, Japan. The mice were maintained in the Beijing Institute of Pharmacology and Toxicology under standard housing conditions (SPF) with a 12-h light/12-h dark cycle and were allowed free access to water and food. The 6-month-old male mice were randomly separated into 3 groups with 5 mice per group. Each mouse in the drug-treated groups was given an intragastric administration of LW-AFC (dose was 1.6 mg/kg body weight) once a day during the test period of 5.5 months. SAMP8 mice served as negative controls and SAMR1 mice served as positive controls; both groups were given with equal volume of distilled water, respectively. After administrating LW-AFC for 90 consecutive days, animals were, one by one, assessed with a Morris-water maze test, a novel object recognition test and a shuttle-box test which were the same as those in Wang’s study [7]. The animal treatment, husbandry, and experimental protocols in this study were approved by the Institute of Animal Care and Use Committee (IACUC) of the National Beijing Center for Drug Safety Evaluation and Research (NBCDSER).

Total RNA preparation

Total RNA was isolated from the hippocampus using TRIzol reagent (Cat. No. R1020012, Exprogen) according to the manufacturer’s instructions. Quantification and quality assessment of total RNA isolated from the hippocampus were performed using a Qubit^® RNA BR assay kit and Agilent RNA 6000 Nano kit, respectively. Some of the total RNA was used for RNA-sequencing, and the rest of the total RNA was reverse-transcribed into first-strand cDNA using MMLV reverse transcriptase (Cat No. M1705, Promega), oligo (dT15) primers (Cat.# C1101, Promega) and a dNTP mixture (Cat.# U1515, Promega). The cDNA was stored at –20°C for real-time fluorescence quantitative PCR analysis.

RNA-seq and data analysis

The TruSeq RNA Sample Prep kit v2 (Illumina) was used to prepare sequencing libraries following the manufacturer’s protocol. All samples were then sequenced (paired ends, 100 bp) on the Illumina HiSeq 2500 sequencer (TruSeq^® Rapid PE Cluster Kit, TruSeq^® Rapid SBS Kit). In the current study, we reported RNA-seq analysis for the SAMR1, SAMP8 and LW-AFC treated SAMP8 mouse groups with each group containing five samples.

Tophat [11] was used to align transcript sequences obtained from RNA sequencing to the UCSC reference genome hg19. Cufflinks was then used to estimate the transcript levels (FPKM) of 38923 Refseq genes. Differentially expressed genes were identified using Cuffdiff [11] with default parameters at FDR ≤0.05 for the three groups.

Gene set enrichment analysis

Gene set enrichment analysis (GSEA) was performed with the built-in GseaPreranked tool in GSEA v2.2.2 obtained from Broad Institute. Gene sets were defined as the up- and downregulated genes between the LW-AFC treated SAMP8 and the SAMP8 groups, and between the SAMP8 and SAMR1 groups. The gene list was ranked based on FDR value calculated by Cufflinks. The number of permutations was set as 1000, and the classic method (no weight) was used to calculate enrichment score.

Weighted correlation network analysis

WGCNA was performed on all expressed genes for all groups as previously described [10]. We first checked data for excessive missing values and outlier microarray samples. Genes with no expression were removed first, and 21997 genes were left after preprocessing (Supplementary Table 1). To construct a weighted gene network, the adjacency matrix was calculated by raising the absolute values of the correlation matrix to a soft-thresholding power β (β= 4) based on the criterion of approximation scale-free topology. The topological overlap (TO), which measured node similarity (how close the neighbors of gene 1 were to the neighbors of gene 2), was then calculated. Next, the genes were hierarchically clustered using 1-TO as the distance measure and modules were determined by using a dynamic tree-cutting algorithm. All analyses were performed with the WGCNA package in R version 3.1.1 (https://labs.genetics.ucla.edu/horvath/htdocs/CoexpressionNetwork/). Once the modules were identified, the module eigengene (ME, i.e., first principal component of the expression values across subjects) was calculated using all genes in each module. The module eigengenes were then correlated with related traits (cognitive behavior data in the current study) using Pearson correlation to identify modules that were significantly associated with the trait. We also quantified associations of individual genes with traits (cognitive behavior data) by calculating gene significance (GS) as (the absolute value of) the correlation between the gene and the trait.

Real-time fluorescence quantitative PCR

The cDNA was used as template for PCR in the real-time qPCR Master Mix (R1060011, Exprogen). The reaction was performed at 94.0°C for 5 min, 35 cycles (94.0°C for 30 s, 60°C for 30 s, 72°C for 30 s), 72°C for 7 min, with a thermal cycler (9600, Applied Biosystems) and primers for genes (Table 1). The melting curve analysis verified the presence of primer dimers and specificity in the PCR product. The assay software was the ddCt (relative quantitation) plate and ddCt (relative quantitation) of the SDS v2.2 software package. The endogenous control was actin. The calculations were performed based on the Ct method. The level of target gene expression was calculated using the formula 2^–ΔCT.

Statistical analysis

All data were expressed as the mean±S.D. GraphPad Prism 6.0 (GraphPad Software, Inc., La Jolla, CA, USA) was used to plot and analyze data. Data between two groups were compared by Student’s t-test. p < 0.05 was interpreted as statistically significant. Correlation was performed by one-tailed Pearson analysis with 95% confidence intervals.

Ethics

The animal treatment, husbandry and experimental protocols in this study were approved by the Institute of Animal Care and Use Committee (IACUC) of the National Beijing Center for Drug Safety Evaluation and Research (NBCDSER).

RESULTS

LW-AFC reversed the transcriptome in the hippocampus of SAMP8 mice

Based on the RNA-seq analysis result (Cuffdiff with FDR ≤0.05), 286 genes are differentially expressed between the SAMP8 and SAMR1 groups (we represent this gene set as P8–R1), of which 108 genes are overexpressed and 178 genes are downexpressed when compared with those in the SAMR1 group (Supplementary Table 2). 82 genes are differentially expressed between the LW-AFC treated SAMP8 and SAMP8 groups (we represent this gene set as LW-P8), of which 49 genes are upregulated and 33 are downregulated when compared with the SAMP8 group (Supplementary Table 2). Simple overlap analysis revealed that LW-AFC treatment could reverse the transcriptome in SAMP8 mice (Supplementary Table 2). The actual overlap number of overexpressed genes in the SAMP8 group and downregulated genes in the LW-AFC treated SAMP8 group was 18, and the overlap number of downexpressed genes in the SAMP8 group and upregulated genes in the LW-AFC treated SAMP8 group was 44 (Supplementary Table 2). This directly indicated LW-AFC treatment could modulate the abnormally expressed genes in SAMP8 mice to a normal direction. This was further validated by the GSEA [12] result (Fig. 1). The up- and downregulated genes between the LW-AFC treated SAMP8 group and SAMP8 group were highly negatively and positively correlated with over- and downexpressed genes between the SAMP8 group and SAMR1 group, respectively (Fig. 1A and Fig. 1B).

WGCNA identified modules of co-expressed genes in all transcriptomes

We performed WGCNA on the 15 transcriptomes from the SAMR1, SAMP8, and LW-AFC treated SAMP8 groups to complement the former differential expression analysis and to find co-expressed genes that were biologically related. Using this method, we can further find genes which are highly correlated with cognitive behavior data, and thus, may shed light on the underlying genes that regulate cognitive abilities. WGCNA first calculates an adjacency matrix by raising the absolute values of the correlation matrix to a soft-thresholding power β. The adjacency matrix is then used to calculate topological overlap (TO), a measure which takes into account the shared neighbors of each gene pair in the network. Genes are hierarchically clustered using 1-TO as the distance measure and modules are then defined.

We first removed genes that were not expressed in all three groups (FPKM = 0), and then used WGCNA to find co-expressed modules. 25 modules of genes with high TO were identified (Fig. 2A, a full list of genes for each module is in Supplementary Table 3). Genes in each module have more similar expression patterns to each other when compared to genes outside of the module, and thus may share the same biological process. We can characterize these modules with existing functional analyses such as GO [13] and KEGG [14], etc.

To relate each module with cognitive behavior data, we obtained the module eigengene (ME) using principal component analysis (PCA). In our study, the ME represents the gene expression of a module. Furthermore, modules with similar gene expressions will have similar MEs. To quantify co-expression similarity of the 25 modules, we further clustered the eigengenes based on Pearson correlation (Fig. 2B). The correlation of any two eigengenes was less than 0.8, so we did not merge the modules.

Based on the WGCNA identified modules of genes with similar gene expression patterns, we first summarized the differentially expressed genes of P8–R1 and LW-P8 in each module to characterize the modules (Fig. 3). In these two groups, most modules contain differentially expressed genes. Module 1 and module 18 own the most differentially expressed genes, which indicated the important roles of these two modules.

We visualized these two modules using Cytoscape software. The edge was weighted by topological overlap, and thus, different weight thresholds would generate different graphs. For module 1, we kept edges with topological overlaps of no less than 0.4 (Fig. 4A). For module 18, we kept edges with topological overlaps of no less than 0.1 and nodes with degrees of no less than 10 (Fig. 4B). For module 1, most of the differentially expressed genes in P8–R1 and LW-P8 group were retained, and owned high intramodular connectivity (large degree). These genes were hub genes in this module. For module 18, many brown nodes (differentially expressed genes in both groups) were central in the graph, which indicated the important role of these genes.

The top 10 genes with highest degrees in module 1 are Gm5514, Gm15920, Gm5900, Etnk1, 6030458C11Rik, Epdr1, D430019H16Rik, Caly, Gm2564, and 9630033F20Rik. 9 out of the 10 genes are differentially expressed (except gene Gm15920). Note that 5 genes are differentially expressed in both groups. The top 10 genes in module 18 are F5, Tmprss11a, mt-Nd4l, Enpp2, 4933421A08Rik, Slc16a8, Sh3rf2, Col8a1, Gm16066, and Aqp1. Enpp2 is differentially expressed (Supplementary Table 4).

Modules correlated with cognitive ability

The cognitive abilities of SAMR1, SAMP8, and LW-AFC treated SAMP8 group were observed and assessed in Wang’s study by cognitive behavior tests (i.e., novel object recognition test, Morris water maze test, shuttle-box test) [9]. These data can be used as indications of recognition abilities in mice. We focused on three behavior indicators related to learning and memory abilities (i.e., the number of platform crossings in the probe trial of the Morris water maze test indicated spatial learning and memory abilities, the successful avoidance times in the testing phase of the shuttle-box test indicated the active avoidance response capability, the discrimination index in the object recognition memory test indicated object recognition memory) (Supplementary Table 5).

To find the core transcriptome modules correlated with these cognitive abilities, we analyzed the relationship between the 25 modules and these cognitive behavior indicators using WGCNA. Here we use the module instead of the individual gene since cognition is a complicated process which can hardly be regulated by a single gene. We first quantified the module-cognitive behavior associations by correlating eigengenes with all cognitive behavior data using Pearson correlation. A student asymptotic p-value was also calculated. The results (Fig. 5) showed that modules were related with three cognitive behavior indicators. The ability of spatial learning and memory is positively related with module MEred (correlation –0.64, p-value 0.01). The active avoidance response capability is negatively related with module MEgreenyellow (correlation –0.56, p-value 0.03). The object recognition memory is negatively related with modules MEdarkred (correlation –0.59, p-value 0.02) and MElightgreen (correlation –0.59, p-value 0.02).

Additionally, we quantified associations of individual genes in each module with a cognitive behavior indicator by defining gene significance (GS) as (the absolute value of) the correlation between the gene and the cognitive behavior data. For each module, a quantitative measure of module membership (MM) was also defined as the correlation of the module eigengene and the gene. With GS and MM, we found the most influential genes in each module that was related with the cognitive behavior data.

Finally, for the three cognitive behavior indicators, we found the most related modules based on correlation and p-value. In each module, GS was calculated for each gene. Genes with p-values no more than 0.01 were grouped in the final candidate list. This list represented the most important genes that might affect specific cognitive ability. We also annotated the gene list with differentially expressed genes in the P8–R1 and LW-P8 groups (Supplementary Table 6).

We found 200 candidate genes in module MEred (module 6 in Fig. 3) which were highly related with the abilities of spatial learning and memory, 2 of which were differentially expressed in the P8–R1 group (3110047P20Rik, Ift80). 40 genes in module MEgreenyellow (module 11 in Fig. 3) were highly related with active avoidance response capability. 24 genes in module MEdarkred (module 21 in Fig. 3) and 59 genes in module MElightgreen (module 18 in Fig. 3) were highly related with object recognition memory, 5 of which were differentially expressed in the P8–R1 group (Clic6, Pla2g5, Myo7a, Rnf11, Lpl) and 2 of which were differentially expressed in the LW-P8 group (Myo7a, Rnf11).

To further find the specific pathways associated with the candidate genes in each module, we performed function analysis using DAVID [15]. Mus musculus genes in each module were first converted to nomenclature symbols with Mouse Genome Informatics [16] (Supplementary Table 6). We used Mus musculus as a background species in the DAVID analysis. The results indicated that the module of spatial learning and memory (MEred) was enriched in the KEGG pathway of systemic lupus erythematosus (mmu05322) and in the spliceosome (mmu03040). The module of active avoidance response (MEgreenyellow) was enriched in amyotrophic lateral sclerosis (mmu05014). The module of object recognition memory (MElightgreen) was enriched in the insulin signaling pathway (mmu04910) (Supplementary Table 7).

Validation of differentially expressed genes in the transcriptomes by the real-time fluorescence quantitative PCR

Eight genes, including Enpp2, Lpl, Rnf11, Gm5514, Etnk1, Epdr1, Gm5900, and Myo7a, were chosen to be studied for their expressions in the hippocampus. The results showed that genes Lpl, Rnf11, Gm5514, and Myo7a could not be detected, whereas genes Enpp2, Etnk1, Epdr1, and Gm5900 were successfully detected (Supplementary Table 5). The mRNA expression of Enpp2 in the hippocampus of SAMP8 mice was higher when compared with that in SAMR1 mice, and the treatment of LW-AFC could downregulate Enpp2 mRNA expression in SAMP8 mice (Fig. 6). The mRNA expressions of genes Etnk1, Epdr1, and Gm5900 in the hippocampus of SAMP8 mice were low, while the administration of LW-AFC could upregulate mRNA expression of Etnk1 in SAMP8 mice (Fig. 6).

In addition, the Pearson analysis indicated that mRNA expressions of Etnk1, Epdr1, and Gm5900 in the hippocampus of SAM mice were correlated with the ability of spatial learning and memory (Fig. 7). The mRNA expressions of Enpp2, Etnk1, Epdr1, and Gm5900 in the hippocampus were correlated with both active avoidance response capability (Fig. 8) and object recognition memory (Fig. 9).

These results were consistent with the result of sequencing differentially expressed genes. This proved the credibility of sequencing results and indicated the validity that the transcriptome in SAMP8 mice was regulated by LW-AFC.

DISCUSSION

Array technology was used to analyze the transcriptional patterns in the hippocampus of SAMP8 mice as an animal model of AD [17 –19]. These studies identified the aberrant gene expression in the hippocampus and led to an improved understanding of the mechanisms underlying AD-like progression in SAMP8 mice [17 –20]. The results of these studies indicated that the expression of genes associated with the stress response, xenobiotic metabolism [18], extracellular matrix maintenance [19], and the regulation of synaptic transmission and apoptosis [17, 21] played potential roles in the cognitive degeneration of SAMP8 mice. Sequencing technology was used in the current study to identify gene expression differences between SAMP8 mice and SAMP8 mice treated with LW-AFC, and revealed potential mechanisms that may mediate how LW-AFC improves learning and memory abilities in the AD mouse model. Our results showed that LW-AFC reversed the transcriptome in the hippocampus of SAMP8 mice. It also identified that the genes Enpp2, Etnk1, Epdr1, and Gm5900 in the hippocampus were correlated with how LW-AFC administration ameliorated the cognitive impairment in SAMP8 mice. We found that genes regulated by LW-AFC were not consistent with the findings from array technology in SAMP8 mice.

The protein encoded by the gene ectonucleotide pyrophosphatase/phosphodiesterase 2 (Enpp2) undergoes proteolytic processing to generate a mature protein with lysophospholipase D activity, which then catalyzes the cleavage of the choline group from lysophosphatidylcholine to produce lysophosphatidic acid (LPA). LPA is a pleiotropic growth factor-like lysophospholipid, acts as a phospholipid mediator [9, 10], and activates at least six G-protein coupled receptors [11–13 , 22–24]. The main LPA-induced signaling pathways and their downstream effects include cytoskeletal remodeling and cell migration, calcium mobilization and PKC activation, cell proliferation, cell survival, neural development, changes in inducing cAMP and the production of growth factors and cytokines [24 –27]. Enpp2 also can act on the sphingosylphosphorylcholine producing sphingosine-1-phosphate (S1P), a modulator of cell motility which is also involved in several motility-related processes such as angiogenesis and neurite outgrowth. Gene ethanolamine kinase 1 (Etnk1) encodes an ethanolamine-specific kinase, which functions in the first committed step of the phosphatidylethanolamine (PE) biosynthesis pathway [28 –30]. The cytosolic enzyme Etnk1 is highly specific for ethanolamine phosphorylation and exhibits negligible kinase activity on choline. PE has diverse roles including the definition of the membrane architecture and participation in the respiratory complexes in the inner membrane of mitochondria [31]. The protein encoded by the gene ependymin-related protein 1 (Epdr1) is a glycosylated type II transmembrane protein that plays a role in calcium-dependent cell adhesion because it is similar to protocadherins and ependymins [32, 33]. The function of the gene Gm5900 is currently not known. In the current study, we showed that ENPP2, Etnk1, Epdr1, and Gm5900 in the hippocampus contributed to cognitive deterioration and were regulated by LW-AFC in the AD mouse model.

Additionally, we discovered that a systemic lupus erythematosus pathway, spliceosomes, amyotrophic lateral sclerosis, and the insulin signaling pathway might be involved in improvement of cognitive ability by LW-AFC. From previous studies, we know that patients with juvenile systemic lupus erythematosus (JSLE) show severe cognitive disorders (e.g., impairment of attention, concentration, learning, memory, information processing, and executive functions, etc.) and poor academic outcomes and that cognitive dysfunction is one of the most common manifestations of JSLE [34 –39]. The spliceosome, an RNA-based enzyme, removes intronic sequences from primary transcripts to generate a functional messenger and long noncoding RNAs [40, 41]. The spliceosome, consisting of five snRNAs and more than 200 proteins, contributes to the regulation of alternative splicing, a prevalent process that contributes to cell differentiation, homeostasis, and disease [41 –44]. The core spliceosome is a target and an effector of the cellular response to transcription-blocking DNA damage and non-canonical ataxia telangiectasia mutated signaling [45]. Amyotrophic lateral sclerosis is a multiple-system neurodegenerative disease. Cognitive impairment, including impairment of mental flexibility, verbal and nonverbal fluency, abstract reasoning, memory for both verbal and visual materials, social cognition, and episodic memory, may be evident in up to 60% of all patients [46 –49]. Insulin signaling is required for neuronal growth and survival and is necessary for maintaining synaptic plasticity and sustaining critical brain functions such as learning and memory. When defective, it contributes to general neurodegenerative processes and aging [50 –57].

LW-AFC is composed of a polysaccharide fraction (LWB-B), a glycosides fraction (LWD-b), and an oligosaccharide fraction (CA-30) and is derived from the Liuwei Dihuang decoction. Our phytochemical study showed that the glycosides fraction contained more than 30 compounds which could be separated into five categories: iridoid glycosides, peoniflorin, phenylpropionic acid and phenethanol-glycosides, 5-hydroxymethyl-furaldehyde and derivant, and others. There were levidulinose, TMAN, and stachyose in the oligosaccharide fraction. The polysaccharide fraction contained more than 16 compounds which could be divided into four categories: polygalacturonic acid, rhamno-galacturonic acid polysaccharide, arabinogalactan, and dextran. In the glycosides fraction, catalpol [58, 59], pachymic acid [60], sweroside [61, 62], benzoic acid [63], and acteoside [64] can pass through the blood-brain barrier. Therefore, we inferred that LW-AFC had direct or indirect methods of altering the gene expression of Enpp2, Etnk1, Epdr1, and Gm5900 in the hippocampus and regulating the systemic lupus erythematosus pathway, spliceosomes, amyotrophic lateral sclerosis, and insulin signaling pathway. Future studies are warranted to clarify the relationships between the activities of the chemical components in LW-AFC and brain cognitive function in SAMP8 mice. Furthermore, future studies will be needed to address the precise mechanisms by which the genes Enpp2, Etnk1, Epdr1, and Gm5900 in the hippocampus contribute to how LW-AFC ameliorates cognitive impairment in SAMP8 mice.

Footnotes

ACKNOWLEDGMENTS

This work was supported by the National Science and Technology Major Project (No. 2013ZX09508104) and the National Natural Science foundation of China (No. 81473191).

Authors’ disclosures available online ().

Appendix

References

Pallás

(2012) Senescence-accelerated mice P8: A tool to study brain aging and Alzheimer’s disease in a mouse model. ISRN Cell Biol 2012, 1–12.

Morley

, Armbrecht

, Farr

, Kumar

(2012) The senescence accelerated mouse (SAMP8) as a model for oxidative stress and Alzheimer’s disease. Biochim Biophys Acta 1822, 650–656.

Pallas

, Camins

, Smith

, Perry

, Lee

, Casadesus

(2008) From aging to Alzheimer’s disease: Unveiling “the switch” with the senescence-accelerated mouse model (SAMP8). J Alzheimers Dis 15, 615–624.

Cheng

, Zhou

, Zhang

(2014) The behavioral, pathological and therapeutic features of the senescence-accelerated mouse prone 8 strain as an Alzheimer’s disease animal model. Ageing Res Rev 13, 13–37.

Wang

, Lei

, Cheng

, Zhang

, Liu

, Cheng

, Xu

, Zeng

, Zhou

, Zhang

(2017) 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. Alzheimers Res Ther 8, 57.

Wang

, Zhang

, Cheng

, Liu

, Xu

, Zeng

, Qiao

, Zhou

, Zhang

(2017) LW-AFC, a herbal medicine for Alzheimer’s disease, ameliorates cognitive deterioration and modulates neuroendocrine-immune system in senescence-accelerated mouse prone 8 strain. Curr Alzheimer Res 14, 18.

Wang

, Ye

, Cheng

, Zhang

, Liu

, Ni

, Qiao

, Zhou

, Zhang

(2016) The effects of LW-AFC on intestinal microbiome in senescence-accelerated mouse prone 8 strain, a mouse model of Alzheimer’s disease. J Alzheimers Dis 53, 907–919.

Wang

, Cheng

, Zeng

, Yuan

, Wang

, Zhou

, Zhang

(2017) LW-AFC effects on N-glycan profile in senescence-accelerated mouse prone 8 strain, a mouse model of Alzheimer’s disease. Aging Dis 8, 14.

Wang

, Cheng

, Zhang

, Cheng

, Xu

, Zeng

, Zhou

, Zhang

(2016) The anti-aging effects of LW-AFC via correcting immune dysfunctions in senescence accelerated mouse resistant 1 (SAMR1) strain. Oncotarget 7, 26949–26965.

10.

Zhang

, Horvath

(2005) A general framework for weighted gene co-expression network analysis. Stat Appl Genet Mol Biol 4, Article 17.

11.

Trapnell

, Roberts

, Goff

, Pertea

, Kim

, Kelley

, Pimentel

, Salzberg

, Rinn

, Pachter

(2012) Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks. Nat Protoc 7, 562–578.

12.

Subramanian

, Tamayo

, Mootha

, Mukherjee

, Ebert

, Gillette

, Paulovich

, Pomeroy

, Golub

, Lander

, Mesirov

(2005) Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 102, 15545–15550.

13.

Ashburner

, Ball

, Blake

, Botstein

, Butler

, Cherry

, Davis

, Dolinski

, Dwight

, Eppig

, Harris

, Hill

, Issel-Tarver

, Kasarskis

, Lewis

, Matese

, Richardson

, Ringwald

, Rubin

, Sherlock

(2000) Gene ontology: Tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 25, 25–29.

14.

Kanehisa

, Sato

, Kawashima

, Furumichi

, Tanabe

(2016) KEGG as a reference resource for gene and protein annotation. Nucleic Acids Res 44, D457–D462.

15.

Huang

da W

, Sherman

, Lempicki

(2009) Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 4, 44–57.

16.

Eppig

, Blake

, Bult

, Kadin

, Richardson

, Mouse Genome Database Group (2015) The Mouse Genome Database (MGD): Facilitating mouse as a model for human biology and disease. Nucleic Acids Res 43, D726–D736.

17.

Cheng

, Zhou

, Zhang

, Zhou

, Yang

, Chen

(2007) Differential gene expression profiles in the hippocampus of senescence-accelerated mouse. Neurobiol Aging 28, 497–506.

18.

Kumar

, Franko

, Farr

, Armbrecht

, Morley

(2000) Identification of age-dependent changes in expression of senescence-accelerated mouse (SAMP8) hippocampal proteins by expression array analysis. Biochem Biophys Res Commun 272, 657–661.

19.

Alvarez-Lopez

, Castro-Freire

, Cosin-Tomas

, Sanchez-Roige

, Lalanza

, del Valle

, Parrizas

, Camins

, Pallas

, Escorihuela

, Kaliman

(2013) Long-term exercise modulates hippocampal gene expression in senescent female mice. J Alzheimers Dis 33, 1177–1190.

20.

Galvin

, Ginsberg

(2005) Expression profiling in the aging brain: Aerspective. Ageing Res Rev 4, 529–547.

21.

Cheng

, Cui

, Zheng

, Zhang

, Li

, Huang

, Zhao

, Bo

, Wang

, Zhou

, Zhang

(2013) Nodes and biological processes identified on the basis of network analysis in the brain of the senescence accelerated mice as an Alzheimer’s disease animal model. Front Aging Neurosci 5, 19.

22.

van Meeteren

, Moolenaar

(2007) Regulation and biological activities of the autotaxin-LPA axis. Prog Lipid Res 46, 145–160.

23.

Mills

, Moolenaar

(2003) The emerging role of lysophosphatidic acid in cancer. Nat Rev Cancer 3, 582–591.

24.

Choi

, Herr

, Noguchi

, Yung

, Lee

, Mutoh

, Lin

, Teo

, Park

, Mosley

, Chun

(2010) LPA receptors: Subtypes and biological actions. Annu Rev Pharmacol Toxicol 50, 157–186.

25.

Geach

, Faas

, Devader

, Gonzalez-Cordero

, Tabler

, Brunsdon

, Isaacs

, Dale

(2014) An essential role for LPA signalling in telencephalon development. Development 141, 940–949.

26.

Sun

, Zhang

, Evans

, Floreani

, Zou

, Nishio

, Qi

, Leung

, Bowlus

, Gershwin

(2016) Autotaxin, pruritus and primary biliary cholangitis (PBC). Autoimmun Rev 15, 795–800.

27.

Okudaira

, Yukiura

, Aoki

(2010) Biological roles of lysophosphatidic acid signaling through its production by autotaxin. Biochimie 92, 698–706.

28.

Lasho

, Finke

, Zblewski

, Patnaik

, Ketterling

, Chen

, Hanson

, Tefferi

, Pardanani

(2015) Novel recurrent mutations in ethanolamine kinase 1 (ETNK1) gene in systemic mastocytosis with eosinophilia and chronic myelomonocytic leukemia. Blood Cancer J 5, 275.

29.

Gambacorti-Passerini

, Donadoni

, Parmiani

, Pirola

, Redaelli

, Signore

, Piazza

, Malcovati

, Fontana

, Spinelli

, Magistroni

, Gaipa

, Peronaci

, Morotti

, Panuzzo

, Saglio

, Usala

, Kim

, Rea

, Zervakis

, Viniou

, Symeonidis

, Becker

, Boultwood

, Campiotti

, Carrabba

, Elli

, Bignell

, Papaemmanuil

, Campbell

, Cazzola

, Piazza

(2015) Recurrent ETNK1 mutations in atypical chronic myeloid leukemia. Blood 125, 499–503.

30.

Lykidis

, Wang

, Karim

, Jackowski

(2001) Overexpression of a mammalian ethanolamine-specific kinase accelerates the CDP-ethanolamine pathway. J Biol Chem 276, 2174–2179.

31.

Kosmider

(2015) Mutations of ETNK1 in aCML and CMML. Blood 125, 422–423.

32.

Khor

, Do

, Jia

, Nakano

, George

, Abu-Amero

, Duvesh

, Chen

, Li

, Nongpiur

, Perera

, Qiao

, Wong

, Sakai

, de Melo

, Lee

, Chan

, Azhany

, Dao

TLH

, Ikeda

, Perez-Grossmann

, Zarnowski

, Day

, Jonas

, Tam

POS

, Tran

, Ayub

, Akhtar

, Micheal

, Chew

PTK

, Aljasim

, Dada

, Luu

, Awadalla

, Kitnarong

, Wanichwecharungruang

, Aung

, Mohamed-Noor

, Vijayan

, Sarangapani

, Husain

, Jap

, Baskaran

, Goh

, Su

, Wang

, Yong

, Yip

, Trinh

, Makornwattana

, Nguyen

, Leuenberger

, Park

, Wiyogo

, Kumar

, Tello

, Kurimoto

, Thapa

, Pathanapitoon

, Salmon

, Sohn

, Fea

, Ozaki

, Lai

JSM

, Tantisevi

, Khaing

, Mizoguchi

, Nakano

, Kim

, Tang

, Fan

, Wu

, Meng

, Nguyen

TTG

, Tran

, Ueno

, Martinez

, Ramli

, Aung

, Reyes

, Vernon

, Fang

, Xie

, Chen

, Foo

, Sim

, Wong

, Quek

, Venkatesh

, Kavitha

, Krishnadas

, Soumittra

, Shantha

, Lim

, Ogle

, de Vasconcellos

JPC

, Costa

, Abe

, de Souza

, Sng

, Aquino

, Kosior-Jarecka

, Fong

, Tamanaja

, Fujita

, Jiang

, Waseem

, Low

, Pham

, Al-Shahwan

, Craven

, Khan

, Dada

, Mohanty

, Faiq

, Hewitt

, Burdon

, Gan

, Prutthipongsit

, Patthanathamrongkasem

, Catacutan

MAT

, Felarca

, Liao

, Rusmayani

, Istiantoro

, Consolandi

, Pignata

, Lavia

, Rojanapongpun

, Mangkornkanokpong

, Chansangpetch

, Chan

JCH

, Choy

BNK

, Shum

JWH

, Than

, Oo

, Han

, Yong

, Ng

, Goh

, Chong

, Hibberd

, Seielstad

, Png

, Dunstan

, Chau

NVV

, Bei

, Zeng

, Karkey

, Basnyat

, Pasutto

, Paoli

, Frezzotti

, Wang

, Mitchell

, Fingert

, Allingham

, Hauser

, Lim

, Chew

, Ebstein

, Sakuntabhai

, Park

, Ahn

, Boland

, Snippe

, Stead

, Quino

, Zaw

, Lukasik

, Shetty

, Zahari

, Bae

, Oo

, Kubota

, Manassakorn

, Ho

, Dallorto

, Hwang

, Kiire

, Kuroda

, Djamal

, Peregrino

JIM

, Ghosh

, Jeoung

, Hoan

, Srisamran

, Sandragasu

, Set

, Doan

, Bhattacharya

, Ho

, Tan

, Sihota

, Loon

, Mori

, Kinoshita

, den Hollander

, Qamar

, Wang

, Teo

, Tai

, Hartleben-Matkin

, Lozano-Giral

, Saw

, Cheng

, Zenteno

, Pang

, Bui

HTT

, Hee

, Craig

, Edward

, Yonahara

, Neto

, Guevara-Fujita

, Xu

, Ritch

, Liza-Sharmini

, Wong

, Al-Obeidan

, Do

, Sundaresan

, Tham

, Foster

, Vijaya

, Tashiro

, Vithana

, Wang

, Aung

(2016) Genome-wide association study identifies five new susceptibility loci for primary angle closure glaucoma. Nat Genet 48, 556.

33.

Gregorio-King

, McLeod

, Collier

, Bolton

, Van Der Meer

, Apostolopoulos

, Kirkland

(2002) MERP1: A mammalian ependymin-related protein gene differentially expressed in hematopoietic cells. Gene 286, 249–257.

34.

Frittoli

, de Oliveira Pelicari

, Bellini

, Marini

, Fernandes

, Appenzeller

(2016) Association between academic performance and cognitive dysfunction in patients with juvenile systemic lupus erythematosus. Rev Bras Reumatol (Rio J) 56, 252–257.

35.

Tay

, Mak

(2015) Anti-NR2A/B antibodies and other major molecular mechanisms in the pathogenesis of cognitive dysfunction in systemic lupus erythematosus. Int J Mol Sci 16, 10281–10300.

36.

Tay

, Ho

, Mak

(2015) 25-hydroxyvitamin D3 deficiency independently predicts cognitive impairment in patients with systemic lupus erythematosus. PLoS One 10, e0144149.

37.

Cesar

, Dwyer

, Shucard

, Polak

, Bergsland

, Benedict

, Weinstock-Guttman

, Shucard

, Zivadinov

(2015) Cognitive and white matter tract differences in ms and diffuse neuropsychiatric systemic lupus erythematosus. AJNR Am J Neuroradiol 36, 1874–1883.

38.

Hanly

(2014) Diagnosis and management of neuropsychiatric SLE. Nat Rev Rheumatol 10, 338–347.

39.

Jeltsch-David

, Muller

(2014) Neuropsychiatric systemic lupus erythematosus: Pathogenesis and biomarkers. Nat Rev Neurol 10, 579–596.

40.

Wahl

, Luhrmann

(2015) SnapShot: Spliceosome dynamics I. Cell 161, 1474–e1471.

41.

Papasaikas

, Valcarcel

(2016) The spliceosome: The ultimate RNA chaperone and sculptor. Trends Biochem Sci 41, 33–45.

42.

Schutze

, Ulrich

, Apelt

, Will

, Bartlick

, Seeger

, Weber

, Luhrmann

, Stelzl

, Wahl

(2016) Multiple protein-protein interactions converging on the Prp38 protein during activation of the human spliceosome. RNA 22, 265–277.

43.

Tian

, Liu

, Zhu

, Tian

, Zhong

, Chen

, Lu

, Zou

, Shen

, Qian

, Li

, Miao

, Wang

(2015) SF3A1 and pancreatic cancer: New evidence for the association of the spliceosome and cancer. Oncotarget 6, 37750–37757.

44.

Stark

, Dunn

, Grisdale

, Daniele

, Halstead

, Fast

, Rader

(2015) Dramatically reduced spliceosome in Cyanidioschyzon merolae. Proc Natl Acad Sci U S A 112, E1191–E1200.

45.

Tresini

, Warmerdam

, Kolovos

, Snijder

, Vrouwe

, Demmers

, van

IWF

, Grosveld

, Medema

, Hoeijmakers

, Mullenders

, Vermeulen

, Marteijn

(2015) The core spliceosome as target and effector of non-canonical ATM signalling. Nature 523, 53–58.

46.

Spalloni

, Longone

(2016) Cognitive impairment in amyotrophic lateral sclerosis, clues from the SOD1 mouse. Neurosci Biobehav Rev 60, 12–25.

47.

Consonni

, Catricala

, Dalla Bella

, Gessa

, Lauria

, Cappa

(2016) Beyond the consensus criteria: Multiple cognitive profiles in amyotrophic lateral sclerosis? Cortex 81, 162–167.

48.

, Ji

, Li

, He

, Liu

, Fan

(2016) The Edinburgh Cognitive and Behavioural ALS screen in a Chinese amyotrophic lateral sclerosis population. PLoS One 11, e0155496.

49.

van der Hulst

, Bak

, Abrahams

(2015) Impaired affective and cognitive theory of mind and behavioural change in amyotrophic lateral sclerosis. J Neurol Neurosurg Psychiatry 86, 1208–1215.

50.

Chambers

, Androschuk

, Rosenfelt

, Langer

, Harding

, Bolduc

(2015) Insulin signaling is acutely required for long-term memory in Drosophila. Front Neural Circuits 9, 8.

51.

Takach

, Gill

, Silverman

(2015) Modulation of insulin signaling rescues BDNF transport defects independent of tau in amyloid-beta oligomer-treated hippocampal neurons. Neurobiol Aging 36, 1378–1382.

52.

Gontier

, George

, Chaker

, Holzenberger

, Aid

(2015) Blocking IGF signaling in adult neurons alleviates Alzheimer’s disease pathology through amyloid-beta clearance. J Neurosci 35, 11500–11513.

53.

van Exel

, Eikelenboom

, Comijs

, Deeg

, Stek

, Westendorp

(2014) Insulin-like growth factor-1 and risk of late-onset Alzheimer’s disease: Findings from a family study. Neurobiol Aging 35, 725e727–710.

54.

Talbot

, Wang

(2014) The nature, significance, and glucagon-like peptide-1 analog treatment of brain insulin resistance in Alzheimer’s disease. Alzheimers Dement 10, S12–S25.

55.

De Felice

, Vieira

, Bomfim

, Decker

, Velasco

, Lambert

, Viola

, Zhao

, Ferreira

, Klein

(2009) Protection of synapses against Alzheimer’s-linked toxins: Insulin signaling prevents the pathogenic binding of Abeta oligomers. Proc Natl Acad Sci U S A 106, 1971–1976.

56.

Craft

(2012) Alzheimer disease: Insulin resistance and AD–extending the translational path. Nat Rev Neurol 8, 360–362.

57.

Rozing

, Westendorp

, Frolich

, de Craen

, Beekman

, Heijmans

, Mooijaart

, Blauw

, Slagboom

, van Heemst

, Leiden Longevity Study Group (2009) Human insulin/IGF-1 and familial longevity at middle age. Aging (Albany NY) 1, 714–722.

58.

Tao

, Zhao

, Wang

, Yang

, Du

, Qiu

, Jiang

(2016) Biotransformation and metabolic profile of catalpol with human intestinal microflora by ultra-performance liquid chromatography coupled with quadrupole time-of-flight mass spectrometry. J Chromatogr B Analyt Technol Biomed Life Sci 1009-1010, 163–169.

59.

Xue

, Ma

, Zhang

, Li

, Zhu

, Liu

, Wu

, Wang

, Wu

(2015) Pharmacokinetics and tissue distribution of Aucubin, Ajugol and Catalpol in rats using a validated simultaneous LC-ESI-MS/MS assay. J Chromatogr B Analyt Technol Biomed Life Sci 1002, 245–253.

60.

Wang

, Lv

, Han

(2015) Determination and pharmacokinetic study of pachymic acid by LC-MS/MS. Biol Pharm Bull 38, 1337–1344.

61.

Sheng

, Zhi

, Yuan

, Zhang

, Jia

, Zhang

, Wang

(2014) Pharmacokinetic and excretion study of three secoiridoid glycosides and three flavonoid glycosides in rat by LC-MS/MS after oral administration of the Swertia pseudochinensis extract. J Chromatogr B Analyt Technol Biomed Life Sci 967, 75-83.

62.

Sheng

, Yuan

, Zhi

, Cui

, Zhang

, Jia

, Zhang

, Wang

(2014) Application of a liquid chromatography-tandem mass spectrometry method to the pharmacokinetics, tissue distribution and excretion studies of sweroside in rats. J Chromatogr B Analyt Technol Biomed Life Sci 969, 1–11.

63.

Lee

, Ho

, Wu

, Omar

, Huang

, Wang

, Hung

(2014) Modulation of cyclins, p53 and mitogen-activated protein kinases signaling in breast cancer cell lines by 4-(3,4,5-trimethoxyphenoxy)benzoic acid. Int J Mol Sci 15, 743–757.

64.

, Tsai

, Lin

, Tsai

(2007) Liquid chromatographic method with amperometric detection to determine acteoside in rat blood and brain microdialysates and its application to pharmacokinetic study. J Chromatogr B Analyt Technol Biomed Life Sci 853, 281–286.