Integration of transcriptomics and network analysis deciphers the mechanisms of baicalein in improving learning and memory impairment in senescence-accelerated mouse prone 8 (SAMP8)

European Journal of Pharmacology xxx (xxxx) xxxx

Contents lists available at ScienceDirect

European Journal of Pharmacology journal homepage: www.elsevier.com/locate/ejphar

Full length article

Integration of transcriptomics and network analysis deciphers the mechanisms of baicalein in improving learning and memory impairment in senescence-accelerated mouse prone 8 (SAMP8) Jiaqi Lia, Yuzhi Zhoua, Guanhua Dua,b, Xuemei Qina,∗∗, Li Gaoa,∗ a b

Modern Research Center for Traditional Chinese Medicine, Shanxi University, Taiyuan, 030006, PR China Institute of Materia Medica, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing, 100050, PR China

ARTICLE INFO

ABSTRACT

Keywords: Baicalein Learning and memory Transcriptomics Network analysis SAMP8

Our previous work suggested that baicalein could delay senescence and improve cognitive dysfunction in senescence-accelerated mouse prone 8 (SAMP8). Although baicalein has shown therapeutical benefits in improving learning and memory impairment, the exact molecular mechanisms have not been fully understood. In the present work, transcriptomics was integrated with gene network analysis for revealing the potential mechanisms of baicalein in improving learning and memory in SAMP8 mice. The results showed that baicalein regulated fifty hub differently expressed genes (DEGs) that were enriched in eight signaling pathways. Janus kinase-signal transducer and activator of transcription (JAK-STAT) signaling pathway was especially significant among them, therefore, the DEGs (Gh1, Il5, Il7, Il20rb, Prlr and Socs1) in JAK-STAT signaling pathway were verified by quantitative real-time PCR, and the key proteins (JAK2, p-JAK2, STAT1 and p-STAT1) and related proteins including β-amyloid peptide (Aβ1-42) and receptor for advanced glycation end products (RAGE) were assayed by Western blot and enzyme-linked immunosorbent assay. Our results suggest that baicalein prevented the activation of JAK2/STAT1 signaling pathway and decreased the levels of Aβ1-42 and RAGE in the cortex of SAMP8 mice. Taken together, our study unmasks the mechanism of baicalein on improving learning and memory impairment in SAMP8 mice, which is dependent upon the inhibition of Aβ1-42 and RAGE/JAK2/STAT1 cascade.

1. Introduction Learning and memory impairment is a key hallmark of brain aging, and has been linked to multiple age-onset physiological and pathological mechanisms. Neuroinflammation has recently attracted a substantial amount of attention due to a central role in neurodegeneration. In the aging brain, pathological changes associated with neuroinflammation including significant decreases in certain neuronal populations, post-synaptic densities, synapses and cortical volume, and increase in β-amyloid peptide (Aβ) deposition, leading to cognitive and behavioral disorders (Chinta et al., 2015). Learning and memory process is a complicated physiological activity that is regulated by multiple genes. Genome-wide RNA sequencing (RNA-seq) is a powerful technology to unveil differentially expressed genes (DEGs) among samples under different conditions. However, screening of key DEGs from a large number of DEGs has become increasingly challenging (Min et al., 2016). Traditionally, the

∗

screening of DEGs only considers the expression fold change and P value of a single gene, while the functional interrelationships between genes were ignored. Construction of a gene-gene network based on available data combining with network analysis may be conducive to mine functionally key genes. Therefore, transcriptomics combining with gene network analysis was utilized to reveal the potential mechanisms of baicalein. The Janus kinase-signal transducer and activator of transcription (JAK-STAT) signaling pathway, a highly conserved pathway, is uncovered to be related to neurogenesis, synaptic plasticity, gliogenesis, and microglial activation in brain (Nabavi et al., 2019). Receptor for advanced glycation end products (RAGE), a receptor for the immunoglobulin superfamily, acts as both an inflammatory intermediary and a critical inducer of oxidative stress. RAGE can be combined with different ligands, such as Aβ, which activates multiple signaling pathways including JAK-STAT signaling pathway thus inducing learning and memory impairment (Kierdorf et al., 2013; Zhang et al., 2018a).

Corresponding author. No.92 Wu Cheng Road, Taiyuan, 030006, China. Corresponding author. No.92 Wu Cheng Road, Taiyuan, 030006, China. E-mail addresses: [email protected] (X. Qin), [email protected] (L. Gao).

∗∗

https://doi.org/10.1016/j.ejphar.2019.172789 Received 18 June 2019; Received in revised form 4 November 2019; Accepted 7 November 2019 0014-2999/ © 2019 Elsevier B.V. All rights reserved.

Please cite this article as: Jiaqi Li, et al., European Journal of Pharmacology, https://doi.org/10.1016/j.ejphar.2019.172789

European Journal of Pharmacology xxx (xxxx) xxxx

J. Li, et al.

Baicalein is a flavonoid derived from the roots of Scutellariae baicalensis Georgi. Studies have shown that baicalein could extend longevity by attenuating oxidative stress in Drosophila melanogaster (Gao et al., 2016), and ameliorate memory deficits by reducing inflammation and metabolic dysfunction in D-galactose-induced aging rats (Duan et al., 2017) or by intensifying synaptic plasticity in APP/PS1 mice (Gu et al., 2016). In previous work, our results suggested that baicalein ameliorates senescence status and improves spatial learning and memory abilities, objective recognition memory and olfactory memory in senescence accelerated mouse P8 (SAMP8) that might be attributable to suppression of cortical proinflammatory cytokines ( Gao et al., 2018a). However, the precise mechanism of baicalein for improvement of aging-related cognitive deficits in SAMP8 mice needs further exploration. In this work, an integration of RNA-seq and network analysis was used to decipher the key genes and pathways in the brain cortex for the learning and memory improvement effect of baicalein. Furthermore, the key genes in JAK-STAT signaling pathway were validated by quantitative real-time PCR (qPCR) and the related proteins such as JAK2, STAT1, Aβ and RAGE were tested by Western blot and enzymelinked immunosorbent assay. Herein, we revealed the mechanism of baicalein for its learning and memory improvement effect is related to the inhibition of Aβ1-42 and RAGE/JAK2/STAT1 cascade.

of 29.1%, 30.3% and 36.8% after 4, 6, and 8 week treatments, respectively. Furthermore, baicalein ameliorated cognitive functions including spatial learning and memory, object recognition memory, and olfactory memory after 8 week treatment, all of which have been reported by our previous study (Gao et al., 2018a), and the mice in this study are the same with those mice. After 8 weeks of administration, the brain cortex was collected and stored at −80 °C until processing. All experimental procedures complied with Institutional Animal Care and Use Committee (IACUC)-approved animal protocols that were based on the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. 2.3. Cortical RNA sequencing and sequence data analysis RNA-seq and sequence data analysis were performed by Shanghai Personal Biotechnology Co., Ltd. In brief, total RNA was isolated from brain cortex samples (three samples per group) using the Trizol Reagent, after which the concentration, quality and integrity were determined using a NanoDrop spectrophotometer. Sequencing libraries from 3 μg RNA were generated using the TruSeq RNA Sample Preparation Kit. Briefly, mRNA was purified from total RNA using polyT oligo-attached magnetic beads and then fragmented in an Illumina proprietary fragmentation buffer. First strand cDNA was synthesized using random oligonucleotides and SuperScript II, and then the first strand cDNA was used as a template to synthesize second strand cDNA synthesis. Remaining overhangs were converted into blunt ends via exonuclease/polymerase activities and the enzymes were removed. After adenylation of the 3’ ends of the DNA fragments, Illumina PE adapter oligonucleotides were ligated to prepare for hybridization. The library fragments were purified to select cDNA fragments of the preferred 200 bp in length. DNA fragments with ligated adaptor molecules on both ends were selectively enriched using Illumina PCR Primer Cocktail in a 15 cycle PCR reaction. Products were purified (AMPure XP system) and quantified on a Bioanalyzer 2100 system (Agilent). The sequencing library was then sequenced on a Hiseq platform (Illumina) by Shanghai Personal Biotechnology Co., Ltd. Firstly, raw data was filtered to remove sequence tags, 3′ end joints and average mass scores lower than Q20 using Cutadapt (Version 1.2.1). Then, we conducted quality control analysis on the raw data using FastQC, including single base quality, base content distribution, GC content distribution, and sequence base quality. The clean data was aligned by Tophat 2 to the reference genome of the species which was derived from the Ensembl database (version 86.38) (http://www. ensembl.org/) and annotated referring to the collected information from UniProtKB, Ensembl, GO, KEGG and eggNOG. Finally, based on the number of sequences aligned to the gene, the expression level was calculated using HTSeq 0.6.1p2, and the difference in expression between genes in different groups was further compared. In order to compare the gene expression levels between different groups, RPKM was applied to normalize the data. The genes for RPKM > 1 were considered to be expressed genes and could be analyzed further.

2. Materials and methods 2.1. Materials Baicalein (purity 98%) was purchased from Jingzhou Biotechnology Co., Ltd. (Nanjing, China). RNAiso Plus, SYBR® Premix Ex Taq™Ⅱ, PrimeScript™ RT Master Mix were purchased from TaKaRa Biotechnology Co., Ltd. (Dalian, China). Chloroform, isopropanol, and 75% ethyl alcohol were purchased Tianjin Damao Chemical Co., Ltd. (Tianjin, China). The Aβ1-42 enzyme-linked immunosorbent assay kit was obtained from Beijing Andy Huatai Technology Co., Ltd. (Beijing, China). Acetylcholinesterase (AchE) kit was purchased form Nanjing Jiancheng Bioengineering Co., Ltd. (Nanjing, China). Polyclonal antibodies against JAK1, STAT1, and β-actin were obtained from Proteintech Group Co., Ltd. (Chicago, Illinois), phospho-JAK2 (Try1007/1008) and phospho-STAT1 (Tyr701) were purchased from Cell Signaling Technology Inc. (Beverly, MA, USA), and RAGE was purchased from abcam Co., Ltd. (Cambridge, UK). RIPA lysis buffer and phenyl methyl sulphonyl fluoride were purchased from Beyotime Biotechnology Co., Ltd. (Jiangsu, China). SDS-PAGE Preparation kit and BCA protein assay kit were obtained from Sangon Biotech Co., Ltd. (Shanghai, China). 2.2. Animals, drug administration and behavioral tests Twenty-seven male 8-month-old SAMP8 mice and fifteen age-matched male senescence-accelerated mouse resistance 1 (SAMR1) mice were provided by the First Teaching Hospital of Tianjin University of Traditional Chinese Medicine (Tianjin, China). Mice were housed in single cages and allowed unrestricted access to food and water in a room at 22 ± 2 °C and 50 ± 10% humidity, with light cycles (12-hr light/dark). SAMP8 mice were divided into two experimental groups: SAMP8 + baicalein group (n = 14) and SAMP8 group (model group, n = 13). SAMR1 mice were used as a control group (n = 15). Mice in SAMP8 + baicalein group were intragastrically administered baicalein (200 mg/kg/day) dissolved in saline solution-0.9% NaCl by ultrasonic dissolving. Mice in SAMP8 group and SAMR1 group were intragastrically administered saline solution-0.9% NaCl. From the seventh week of administration, animals were assessed with the grading score of senescence, olfactory memory test, novel object recognition test, and Morris water maze test. The results revealed that baicalein could decrease the grading score of senescence in SAMP8 mice with percentages

2.4. Construction of differentially expressed genes interaction network and network analysis Genes that are differentially expressed > 3 × |median value of the fold change| or p-value < 0.05 between SAMP8 and SAMR1 groups or between “SAMP8 + baicalein” and SAMP8 groups were considered as DEGs used for further analysis. The inner interactions among DEGs were retrieved from STRING (Christian et al., 2003) (https://string-db. org/), and the gene-gene network was visualized in Cytoscape software (Smoot et al., 2011) (version 3.5.1, http://cytoscape.org/). NetworkAnalyzer (http://med.bioinf.mpi-inf.mpg.de/network analyzer/) (Yassen et al., 2008) was used to calculate topological parameters of genes in network, including average shortest path length (ASPL) and betweenness centrality (BC). The ASPL gives the expected 2

European Journal of Pharmacology xxx (xxxx) xxxx

J. Li, et al.

distance between two connected nodes and is the most commonly used measure of functional integration. The BC is defined as the fraction of all shortest paths in the network that pass through a given node, indicates that a node influences the communication between nodes in the network (Yoon et al., 2006). According to our previous method (Gao et al., 2018b), the R value was calculated by the following formula:

Xi Xi min R= × 50 % + Xi max Ximin

1 Xj

1 min Xj

1 max Xj

1 min Xj

2.7. Assay of Aβ1-42 and AchE The brain cortex was homogenized in a volume of 9 times ice-cold, 0.1 M phosphate buffer saline (pH 7.4). The supernatant was used to detect the level of Aβ1-42 and AchE activity using corresponding kits according to the manufacturer's instructions. 2.8. Statistical analysis

× 50 %

All data are expressed as the mean ± S.E.M. Significance was determined by one-way ANOVA with Dunnett's test using GraphPad Prism 5.0 (GraphPad Software, Inc., La Jolla, CA, USA). P < 0.05 was considered statistically significant.

where Xi is the ASPL, Xj is the BC. The smaller the R value, the more important the node is in the network. The top 40% DEGs in the network according the R value were selected as important genes, which then were taken intersection with 230 reversely regulated DEGs to give the DEGs as key genes regulated by baicalein. The selected DEGs were submitted to DAVID (version 6.8, https://david.ncifcrf.gov/). KEGG pathway was selected from DAVID settings panel. The gene-pathway network was visualized in Cytoscape plugin Cluepedia (Version 1.3.3).

3. Results 3.1. Transcriptional response in SAMP8 mice after baicalein treatment The RNA-seq results from the cortical samples showed that 1035 DEGs (584 up-regulated DEGs and 451 down-regulated DEGs) were found in 10-month-old SAMP8 mice versus the age-matched SAMR1 mice. Treatment of baicalein in SAMP8 mice altered 654 DEGs (349 upregulated DEGs and 305 down-regulated DEGs) including 230 reversely regulated DEGs, resulting in similar mRNA levels to SAMR1 mice. A total of 1448 DEGs were obtained (Supplementary Table 1), and the inner interactions were obtained from STRING. Finally, a gene-gene network containing 908 nodes and 3771 edges was constructed using Cytoscape software (Fig. 1). The topological parameters of ASPL and BC of genes in the network were analyzed using NetworkAnalyzer and the R value is calculated (Supplementary Table 2). According to the R values, 50 DEGs reversely regulated by baicalein were selected as hub genes for pathway analysis (Table 2).

2.5. qPCR analysis The total RNA of cortex was extracted and purified using the RNAiso Plus-chloroform-isopropanol-75% ethyl alcohol, followed by the reverse transcription to cDNA using the PrimeScriptTM RT Master Mix and SYBR® Premix Ex Taq TM Ⅱ, which were the same as those in our previous study (Li et al., 2018). The primers were synthesized by Sangon Biotech Co., Ltd., and the sequences are listed in Table 1. The crossing point (Cq) value for each target gene was normalized to that of housekeeping gene (β-Actin) and relative mRNA expressions were analyzed using 2 -△△Ct method as described previously (Livak and Schmittgen, 2001).

3.2. Pathway analysis for prediction of crucial pathways affected by baicalein in SAMP8 mice

2.6. Western blot analysis

In order to gain insights into the functional convergence of transcriptional changes triggered by baicalein, the 50 reversely regulated DEGs by baicalein were submitted to DAVID for pathway analysis. These genes were linked primarily to Cytokine-cytokine receptor interaction, JAK/STAT signaling pathway, PI3K/Akt signaling pathway, Neuroactive ligand-receptor interaction, Chemokine signaling pathway, Prolactin signaling pathway, Rheumatoid arthritis, and T cell receptor signaling pathway (Fig. 2). Among these signaling pathways, JAK-STAT signaling pathway was selected for further study due to its strong links to synaptic plasticity, neural function and neuroinflammation (Shariq et al., 2018), as well as the smaller P value in pathway analysis and a core position in DEG-pathway network (Fig. 2B).

Western blot was used to measure protein levels of RAGE, JAK2, pJAK2, STAT1 and p-STAT1 as previously described. Six samples of each group were analyzed. Briefly, brain cortex samples after fully grinding with liquid nitrogen were dissolved with RIPA lysis buffer containing 1% PMSF for 1 h on ice. Protein concentrations were determined by the BCA protein assay kit after centrifuging at 12,000 rpm for 10 min at 4 °C. Equal amounts of protein were separated by electrophoresis on 6% or 8% sodium dodecyl sulphate polyacrylamide gels and were transferred onto PVDF membranes. These membranes were soaked in 5% skimmed milk dissolved with TBST buffer (Tris Buffer Saline supplemented with 0.1% Tween-20) for 2 h to block nonspecific binding sites. The membranes were then incubated overnight at 4 °C with the primary antibodies. After washing with TBST, the membranes were incubated for 2 h at room temperature with fluorescent secondary antibodies. After rewashing with TBST, the membranes were scanned using a fluorescent scanner (Odyssey CLX, Gene Company Limited, USA). The optical density of the bands was normalized to β-actin levels, the loading control.

3.3. JAK2-STAT1 signaling pathway was suppressed by baicalein in SAMP8 mice Five DEGs (Gh1, Il5, Il7, Il20rb, and Prlr) enriched in JAK-STAT signaling pathway were validated by qPCR. Given suppressor of cytokine signaling (Socs) plays an important regulatory role in the JAKSTAT signaling pathway, the mRNA level of Socs1 was also detected. The expression levels of these genes obtained from the transcriptome analysis were shown in Supplementary Table 3. We observed that the mRNA levels of Il7, Il20rb, Prlr, Socs1 were significantly up-regulated and the mRNA level of Il5 was significantly down-regulated with the fold change of 3.26, 14.33, 2.71, 2.07, and −1.47 respectively in 10-month-old SAMP8 mice versus age-matched SAMR1 mice, while the mRNA levels of Gh1, Il7, Il-20rb, Prlr, Socs1 were significantly reduced and the mRNA level of Il5 had an increasing trend with the fold change of −1.96, −2.69, −7.02, −3.57, −1.85, and 1.28 respectively after treatment with baicalein in SAMP8 for 8

Table 1 Primers sequences used in quantitative real-time PCR experiments. Gene

Primer sequence forward (5′-3′)

Primer sequence reverse (5′-3′)

β-Actin Gh1 Il5 Il7 Il20rb Prlr Socs1

GTGGGAATGGGTCAGAAGGA CTGGCTGCTGACACCTACAA ATCAAACTGTCCGTGGGGGTACT GGAACTGATAGTAATTGCCCG GCCCCAGTTTGAGTTCCTTG AGCAACGTGGGAGTCTGAAG GTGGTTGTGGAGGGTGAGAT

CTTCTCCATGTCGTCCCAGT GACCAGCTTGTGTCTCGTCA TCTCTCCTCGCCACACTTCTCTTT TTCAACTTGCGAGCAGCACG CGCCTTCACACAGTACATGG CCACAGTGAAGGAGGAGCTG CCTGAGAGGTGGGATGAGG

3

European Journal of Pharmacology xxx (xxxx) xxxx

J. Li, et al.

Fig. 1. Gene-gene interaction network constructed using Cytoscape software. (A) graphical illustration of 908 DEGs, the darker nodes represent the 50 hub genes; (B) a sub-network visualization of the 50 hub genes. The nodes in red color represents upregulated genes; the nodes in green color represents down-regulated genes; the smaller the R value, the bigger the node size is in the network.

weeks (Fig. 3). The validation results revealed that the 6 genes involved in JAK-STAT signaling pathway followed the same trend as transcriptome analysis. We further validated the key proteins in this pathway by Western blot. As shown in Fig. 4, the phosphorylations of p-JAK2 and p-STAT1 in 10-month-old SAMP8 mice were significantly higher than that in SAMR1 mice, which were inhibited by treatment with baicalein for 8 weeks. These results showed that JAK2-STAT1 was activated in the cortex of 10-month-old SAMP8 mice, and baicalein could inhibit the activation. These results suggest that the integration of transcriptomics and network analysis has a definitely guiding significance for mechanism of action of drugs.

Table 2 50 hub DEGs reversely regulated by baicalein. DEGs

Mapk13 Actbl2 Il5 Il7 Ubl4b Tbx21 Fpr2 Ccl3 Cxcl5 Gm8225 Gm10036 Ctla4 Gcgr Hcrt Trh Th Ibsp Ppm1n Gh1 Cxcr6 Txndc2 Mst1 Scgn Cabp7 Gm5741 Tbxt Omp Tnni2 Igll1 Prlr Lhx3 Lrrn4 Lrrc55 Il20rb Eomes Prkag3 Dct Mc1r Myo7b Comp Dhrs9 Pitx1 Tfpi2 Krt15 Tnfrsf9 Mybpc3 Oxct2a Kctd19 Cyp2a5 Ttr

SAMP8 vs SAMR1

SAMP8+baicalein vs SAMP8

Fold Change

P Value

Fold Change

P Value

3.32 4.36 −5.88 2.40 −4.76 −100.2 5.40 3.63 2.21 9.20 −8.33 −3.34 1.36 −4.76 −5.56 −12.52 2.07 3.40 1.32 3.66 −4.55 −9.68 −33.33 3.66 −2.18 −5.56 −50.10 −5.03 4.19 6.93 10.84 5.60 −2.03 27.09 −14.29 −7.69 −3.57 −4.17 4.18 5.23 −14.29 6.29 −14.71 13.51 −49.86 8.64 −4.02 3.43 −6.67 38.61

0.039 P > 0.05 P > 0.05 0.030 P > 0.05 P > 0.05 P > 0.05 P > 0.05 0.049 P > 0.05 P > 0.05 P > 0.05 0.047 P > 0.05 P > 0.05 P > 0.05 0.043 P > 0.05 0.047 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 0.047 P > 0.05 P > 0.05 P > 0.05 P > 0.05 0.043 P > 0.05 0.049 0.026 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 0.000 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05

−3.70 −12.50 2.75 −4.05 6.85 47.54 −2.55 −1.67 −3.70 −6.67 6.69 3.19 −10.02 3.44 5.47 6.88 −4.76 −14.29 −1.20 −2.44 6.48 6.74 14.03 −3.29 4.50 3.43 37.08 5.39 −4.17 −1.87 −4.06 −5.26 1.76 −50.21 7.73 4.0 4.79 5.33 −3.33 −3.85 7.84 −6.13 6.11 −3.58 8.01 −10.89 4.99 −11.11 6.15 −7.46

0.021 P > 0.05 0.046 P > 0.05 P > 0.05 P > 0.05 0.042 0.048 P > 0.05 P > 0.05 P > 0.05 0.041 P > 0.05 P > 0.05 0.005 P > 0.05 P > 0.05 P > 0.05 0.046 0.046 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 0.042 P > 0.05 P > 0.05 0.009 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05 P > 0.05

R

0.138 0.146 0.159 0.161 0.163 0.163 0.163 0.164 0.169 0.171 0.171 0.173 0.173 0.176 0.176 0.177 0.179 0.179 0.180 0.180 0.102 0.184 0.185 0.185 0.185 0.189 0.189 0.190 0.190 0.191 0.191 0.192 0.192 0.192 0.193 0.194 0.194 0.197 0.198 0.198 0.198 0.199 0.202 0.203 0.206 0.206 0.208 0.210 0.214 0.216

3.4. Effect of baicalein on Aβ1-42 and AchE in the brain cortex of SAMP8 mice Aβ deposition is one of the pathological features of cognitive decline in SAMP8 (Zhang et al, 2018b). As shown in Fig. 5A, compared with SAMR1 mice, the level of Aβ1-42 in the cortex of SAMP8 mice was dramatically increased (46.7%); nevertheless, it was significantly decreased in baicalein-treated SAMP8 mice (18.2%). Given acetylcholinesterase (AchE) plays a key role in accelerating Aβ plaques deposition (Shen et al., 2006), we examined the effect of baicalein on the AchE activity in the cortex of SAMP8 mice. Compared with SAMR1 mice, the activity of AchE was significantly increased in the cortex of SAMP8 mice (55.92%), and there is a tendency to decrease after administration of baicalein (Fig. 5B), which seems to suggest that the inhibitory effect of baicalein on Aβ is not associated with AchE in the cortex of SAMP8 mice. We have previously demonstrated that baicalein could significantly inhibit the level of pro-inflammatory factors including IL-6, IL-1β and TNF-α (Gao et al., 2018a), which might contribute to its Aβ inhibition effect. 3.5. Baicalein reduced the expression of RAGE Studies have shown that RAGE could increase Aβ accumulation in brain via promoting translocation of Aβ from the extracellular to the intracellular space or transporting across the blood-brain barrier (Deane et al., 2003; Takuma et al., 2009). As noted above, we found that the protein level of RAGE was significantly higher in the SAMP8 mice than that in SAMR1 mice (108.6%), while it was significantly reduced after treatment with baicalein in SAMP8 mice (25.0%) (Fig. 6). 4. Discussion SAMP8 mice shows age-related deterioration in learning and memory ability with comprehensive brain pathological changes including Aβ deposition (Akiguchi et al., 2017), which is an ideal model for studying age-related cognitive impairment and Alzheimer's disease (AD). In previous work, our results have showed that baicalein

Note: The 3 × |median value of the fold change| is 3.33, −3.28, 3.32 and −3.34 in SAMP8 vs SAMR1_up, SAMP8 vs SAMR1_down, SAMP8+bai vs SAMP8_up, SAMP8+bai vs SAMP8_down, respectively.

4

European Journal of Pharmacology xxx (xxxx) xxxx

J. Li, et al.

Fig. 2. Pathway analysis for prediction of crucial pathways affected by baicalein in SAMP8 mice. (A) Pathway analysis of 50 reversely regulated DEGs by baicalein, P < 0.05. (B) DEG-pathway network for each functional group.

Fig. 3. Validation of genes in JAK-STAT signaling pathway by qPCR. n = 6 per group. Data are expressed as the mean ± S.E.M.; # P < 0.05, versus SAMR1 mice; **P < 0.01, *P < 0.05, versus SAMP8 mice.

###

P < 0.001,

##

P < 0.01,

Fig. 4. The effect of baicalein on the protein levels of phosphor-JAK2 (p-JAK2) (Try1007/1008), total JAK2, phosphor-STAT1 (p-STAT1) (Try 701) and total STAT1 in the brain cortex. The ratios of phosphorylated to total JAK2 (JAK2 P/T) and phosphorylated to total STAT1 (STAT1 P/T) are expressed as mean ± S.E.M.; n = 6 mice per group. ###P < 0.001, ##P < 0.01 versus SAMR1 mice; *P < 0.05, versus SAMP8 mice.

significantly ameliorates the senescence status as manifested by improving the reactivity and passivity, furglossiness and density, and attenuating skin ulcers and periophthalmic lesions, improves cognitive functions, and decreases the levels of cortical proinflammatory cytokines in 10-month-old SAMP8 mice (Gao et al., 2018a). Based on the preliminary work, we attempted to investigate the mechanism of action of baicalein from a holistic perspective. With the rapid advance in high-throughput sequencing, RNA-seq has been of the essence in transcriptome quantification, and a large

number of DEGs have been detected (Manuel et al., 2011). However, many distinct biological phenotypes are shown to be related to the regulation of gene network modules more than individual genes (Nam et al., 2008). Gene network-based topological analysis and enrichment analysis not only broadly consider all DEGs and gene interactions (Li et., 2018), but also help to ameliorate the multiple hypothesis testing problem inherent in RNA-seq data. The results obtained by integration of RNA-seq and gene network analysis may be more reliable than single-gene investigation (Le et al., 2018). Currently, gene network 5

European Journal of Pharmacology xxx (xxxx) xxxx

J. Li, et al.

Fig. 5. Effects of baicalein on the levels of Aβ1-42 and AchE in the brain cortex. (A) Aβ1-42, (B) AchE. n = 6 mice per group. Data are expressed as the mean ± S.E.M.; ### P < 0.001, #P < 0.05 versus SAMR1 mice; **P < 0.01 versus SAMP8 mice.

analysis based on RNA-seq data has revealed the pathogenesis of many diseases (Li et., 2018; Le et al., 2018) and the mechanism of drug action (Wang et al., 2017). In current study, RNA-seq combining with network analysis was used to reveal potential mechanisms that may mediate how baicalein improves learning and memory impairment in SAMP8 mice. Our results showed that the effects of baicalein on delaying senescence and improving cognitive functions might be mainly associated with the regulation of Cytokine-cytokine receptor interaction, JAK-STAT signaling pathway, PI3K-Akt signaling pathway, Neuroactive ligand-receptor interaction, Chemokine signaling pathway. Cellular aging often produces a senescence-associated secretory phenotype accompanied by the secretion of cytokines and chemokines, which can alter the microenvironment of surrounding tissue and affect the function of neighboring cells, ultimately leading to chronic lowgrade inflammation and aging-related diseases (Campisi, 2011). In the brain, senescent astrocytes produce neurotoxicity by secreting cytokines and chemokines, leading to chronic neuroinflammation (Turnquist et al., 2016) which is known to impair synaptic plasticity and cognitive ability by attenuating both NMDA receptor and L-type voltage gated Ca2+ channels-dependent LTP (Min et al., 2009). In addition, chronic neuroinflammation is closely related to the occurrence of AD, in turn, Aβ deposition promotes the production of cytokines and chemokines and accelerates cognitive impairment (Baker et al., 2018). These studies hint that cytokines and chemokines-induced chronic neuroinflammation may be the bridge between brain aging and cognitive impairment or AD. However, baicalein can suppress the levels of proinflammatory factors including IL-6, IL-1β and TNF-α in the cortex of SAMP8 mice (Gao et al., 2018a) and in the serum of D-galactoseinduced aging rats (Duan et al., 2017). The neuroactive ligand-receptor interaction pathway, which is a collection of neuroactive ligand and receptors, is visibly involved in the

nerve conduction and is closely associated with the neuronal functions (Su et al., 2009). Several lines of studies have shown that PI3K/Akt signaling pathway is shown to play a distinct role on the neuroprotection. Mitochondrial dysfunction is an important feature of brain aging and AD, which could result in ROS-induced oxidative damages, destroying neuronal function. PI3K/Akt signaling pathway could inhibit ROS-induced oxidative damages by activating Nrf2 to enhance expression of the SODs (Matsuda et al., 2018). Furthermore, dysfunction of PI3K/Akt/mTOR pathway is linked to disrupted clearance of Aβ, synaptic loss, autophagy disruption and cognitive decline (David et al., 2014), highlighting the importance of PI3K/Akt pathway in regulation of learning and memory function. Numerous studies have shown that baicalein improved cognitive deficits in diabetic rats (Qi et al., 2015) and rat model of ischemia (Yang et al., 2019) via regulation of PI3K/ Akt pathway. The JAK/STAT signaling pathway, by its strong links to cell proliferation, survival, differentiation, synaptic plasticity, and inflammation, is considered as one of the most important signaling pathways involved in regulation of cognitive function. In addition, oxidative stress and some pro-inflammatory cytokines could activate both STAT1 and STAT3 by a JAK2-dependent mechanism (Planas et al., 2006). Activated STAT1 (phosphorylated STAT1), in the nucleus, upregulates BACE1 and subsequent Aβ production, in turn, Aβ could activate JAK2 to escalate phosphorylation of STAT1 (Zhang et al., 2018a). Furthermore, p-STAT1 could specifically induce Scos1 whose upregulated expression as a consequence of JAK2/STAT1 signaling pathway activation (Kershaw et al., 2013), which feeds back to negatively regulate JAK2/ STAT1 signaling (Baker et al., 2009). Hence, Scos1 may be as an indicator for the activation of JAK2/STAT1 signaling pathway. Xu et al. showed that baicalein can down-regulate the mRNA expression levels of STAT3 and STAT4 in JAK/STAT signaling pathway in T cells and exhibited therapeutic effects on autoimmune diseases (Xu et al., 2018).

Fig. 6. The effect of baicalein on the protein level of RAGE in the brain cortex. Results are expressed as a percentage of the control and the data are expressed as the means ± S.E.M.; n = 6 mice per group. ###P < 0.001, versus SAMR1 mice; **P < 0.01, versus SAMP8 mice. 6

European Journal of Pharmacology xxx (xxxx) xxxx

J. Li, et al.

According to the amyloid cascade hypothesis, Aβ1-42 triggers a series of amyloid cascade, and is a key factor to drive neuroinflammatory responses (Shen et al., 2006). Studies have shown that baicalein could significantly reduce Aβ levels in a variety of cognitive impairment models such as APP/PS1 mice (Gu et al., 2016) and Tg2576 mice models (Zhang et al., 2013). In recent years, considerable attention has been paid to the interaction between Aβ and RAGE. Indeed, studies have shown that RAGE could enhance Aβ cytotoxicity or accumulation in brain via translocation of Aβ from the extracellular to the intracellular space or transporting across the blood-brain barrier (Deane et al., 2003), and then mediate neuroinflammation and cognitive impairment. Inhibition of RAGE is expected to become a new target for the treatment of AD (Cai et al., 2016). In addition, several lines of evidence have suggested RAGE ligands could activate the JAK/STAT signaling pathway in human renal tubular cells (Jau-Shyang et al., 2015), dorsal root ganglia neurons from adult rats (Saleh et al., 2013), human articular chondrocytes (Yammani et al., 2009) and so on. In this work, given the close relationship between JAK/STAT signaling pathway and neuroinflammation or learning and memory impairment, the mRNA levels of five DEGs (Gh1, Il5, Il7, Il20rb, and Prlr) in this pathway and Socs1 (whose expression had a significantly reversal trend after baicalein intervention in the result from transcriptome) were validated by qPCR and the protein expression levels of JAK2, p-JAK2, STAT1, and p-STAT1 were tested by Western blot. The results from qPCR showed that the five DEGs involved in JAK-STAT signaling pathway were reversibly changed by baicalein and followed the same trend as transcriptome results, which demonstrates the reliability of RNA-seq results. The mRNA level of Socs1 in the cortex of 10-month-old SAMP8 mice was higher than that in SAMR1 mice, suggesting that JAK2/STAT1 signaling pathway was activated with aging, while the mRNA level of Socs1 was reduced after baicalein treatment for 8 weeks in SAMP8 mice, suggesting that baicalein might inhibit the activation of this pathway. Consistent with the previous results, the results from Western blot showed that baicalein reduced the protein levels of p-JAK2 and p-STAT1, indicating that baicalein prevented the activation of JAK2/STAT1 signaling pathway. In line with previous reports (Angela et al., 2013), the levels of Aβ142 and RAGE, and the activity of AchE in the cortex of 10-month-old SAMP8 mice were higher than those in SAMR1 mice, while the levels of Aβ1-42 and RAGE were reduced after baicalein treatment for 8 weeks in SAMP8 mice. But the activity of AchE was not decreased significantly by treatment of baicalein in the cortex of 10-month-old SAMP8 mice, which may be attributed to the weak effect of baicalein. In conclusion, by integration of transcriptomics and network analysis, our results revealed that baicalein improved learning and memory impairment possibly dependent upon the inhibition of Aβ1-42 and RAGE/JAK2/STAT1 cascade. This integrated approach provides new ideas for transcriptome-based drug mechanism research, and might have broad application prospects.

Declaration of competing interest The authors declare no conflict of interest. Acknowledgements This work was supported by the National Natural Science Foundation of China (81603319), Youth Science and Technology Research Fund of Shanxi Province (201801D221374), Postgraduate Education Innovation Project (2018SY008), and Key laboratory of Effective Substances Research and Utilization in TCM of Shanxi province (201705D111008-21). Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.ejphar.2019.172789. References Akiguchi, I., Pallàs, M., Budka, H., Akiyama, H., Ueno, M., Han, J., Yagi, H., Nishikawa, T., Chiba, Y., Sugiyama, H., 2017. SAMP8 mice as a neuropathological model of accelerated brain aging and dementia: Toshio Takeda's legacy and future directions. Neuropathology 37, 293–305. Angela, K., Mandy, H., Kai, S., Brigitte, V., Dietmar, Z., 2013. Senescence-dependent impact of anti-RAGE antibody on endotoxemic liver failure. Age 35, 2153–2163. Baker, B.J., Akhtar, L.N., Benveniste, E.N., 2009. SOCS1 and SOCS3 in the control of CNS immunity. Trends Immunol. 30, 392–400. Baker, S.K., Chen, Z.L., Norris, E.H., Revenko, A.S., 2018. Blood-derived plasminogen drives brain inflammation and plaque deposition in a mouse model of Alzheimer's disease. Proc. Natl. Acad. Sci. U.S.A. 115, E9687–E9696. Campisi, J., 2011. Cellular senescence: putting the paradoxes in perspective. Curr. Opin. Genet. Dev. 21, 107–112. Cai, Z., Liu, N., Wang, C., Qin, B., Zhou, Y., Xiao, M., Chang, L., Yan, L.J., Zhao, B., 2016. Role of RAGE in Alzheimer's disease. Cell. Mol. Neurobiol. 36, 1–13. Chinta, S.J., Woods, G., Rane, A., Demaria, M., Campisi, J., Andersen, J.K., 2015. Cellular senescence and the aging brain. Exp. Gerontol. 68, 3–7. Christian, v.M., Martijn, H., Daniel, J., Steffen, S., Peer, B., Berend, S., 2003. STRING: a database of predicted functional associations between proteins. Nucleic Acids Res. 31, 258–261. David, H.S., Pérez-Rojas, J.M., Jacqueline, H.D., José, P.C., 2014. The role of PI3K/AKT/ mTOR pathway in the modulation of autophagy and the clearance of protein aggregates in neurodegeneration. Cell. Signal. 26, 2694–2701. Deane, R., Du, Y.S., Submamaryan, R.K., Larue, B., Jovanovic, S., Hogg, E., Welch, D., Manness, L., Lin, C., Yu, J., 2003. RAGE mediates amyloid-beta peptide transport across the blood-brain barrier and accumulation in brain. Nat. Med. 9, 907–913. Duan, D.D., Wang, K.X., Zhou, Y.Z., Qin, X.M., Gao, L., Du, G.H., 2017. Baicalein exerts beneficial effects in D-galactose induced aging rats through attenuation of inflammation and metabolic dysfunction. Rejuvenation Res. 20, 506–516. Gao, L., Duan, D.D., Zhang, J.Q., Zhou, Y.Z., Qin, X.M., Du, G.H., 2016. A bioinformatic approach for the discovery of antiaging effects of baicalein from Scutellaria baicalensis Georgi. Rejuvenation Res. 19, 414–422. Gao, L., Li, J.Q., Zhou, Y.Z., Huang, X., Qin, X.M., Du, G.H., 2018a. Effects of baicalein on cortical proinflammatory cytokines and the intestinal microbiome in senescence accelerated mouse prone 8. ACS Chem. Neurosci. 9, 1714–1724. Gao, L., Wang, K., Zhou, Y., Fang, J., Qin, X., Du, G., 2018b. Uncovering the anticancer mechanism of Compound Kushen Injection against HCC by integrating quantitative analysis, network analysis and experimental validation. Sci. Rep. 8, 624. Gu, X.H., Xu, L.J., Liu, Z.Q., Wei, B., Yang, Y.J., Xu, G.G., Yin, X.P., Wang, W., 2016. The flavonoid baicalein rescues synaptic plasticity and memory deficits in a mouse model of Alzheimer's disease. Behav. Brain Res. 311, 309–321. Jau-Shyang, H., Ying-Ho, L., Lea-Yea, C., Jinn-Yuh, G., Jean-Yu, H., 2015. Cinnamaldehyde and nitric oxide attenuate advanced glycation end products-induced the Jak/STAT signaling in human renal tubular cells. J. Cell. Biochem. 116, 1028–1038. Kershaw, N.J., Murphy, J.M., Lucet, I.S., Nicola, N.A., Babon, J.J., 2013. Regulation of Janus kinases by SOCS proteins. Biochem. Soc. Trans. 41, 1042–1047. Kierdorf, K., Fritz, G., 2013. RAGE regulation and signaling in inflammation and beyond. J. Leukoc. Biol. 94, 55–68. Le, T.T., Savitz, J., Suzuki, H., Misaki, M., Teague, T.K., White, B.C., 2018. Identification and replication of RNA-Seq gene network modules associated with depression severity. Oncogene 34, 1608–1615. Li, J., Zhou, Y., Du, G., Qin, X., Gao, L., 2018. Bioinformatic analysis reveals key genes and pathways in aging brain of senescence-accelerated mouse P8 (SAMP8). CNS Neurol. Disord. - Drug Targets 17, 712–722. Livak, K.J., Schmittgen, T.D., 2001. Analysis of relative gene expression data using realtime quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25, 402–408. Manuel, G., Grabherr, M.G., Mitchell, G., Cole, T., 2011. Computational methods for transcriptome annotation and quantification using RNA-seq. Nat. Methods 8, 469–477.

Author contribution JL and LG contributed to conception and design. JL performed the experiments and wrote the first draft of the manuscript. LG revised the manuscript and final approval of the manuscript submitted. YZ contributed to technical or material support and study supervision. XQ and GD participated in designing of the study and writing the protocol. Ethical approval All procedures were in accordance with the institutional guidelines and ethics for laboratory animal care and use of Shanxi University and were approved by the Institutional Animal Care and Use Committee (NO: SXULL2019011). 7

European Journal of Pharmacology xxx (xxxx) xxxx

J. Li, et al. Matsuda, S., Nakagawa, Y., Tsuji, A., Kitagishi, Y., Nakanishi, A., Murai, T., 2018. Implications of PI3K/AKT/PTEN signaling on superoxide dismutases expression and in the pathogenesis of Alzheimer's disease. Diseases 6, E28. Min, S., Lee, B., Yoon, S., 2016. Deep learning in bioinformatics. Briefings Bioinf. 18, 851–869. Min, S.S., Quan, H.Y., Ma, J., Han, J.S., Jeon, B.H., Seol, G.H., 2009. Chronic brain inflammation impairs two forms of long-term potentiation in the rat hippocampal CA1 area. Neurosci. Lett. 456, 20–24. Nabavi, S.M., Ahmed, T., Nawaz, M., Devi, K.P., Balan, D.J., Pittalà, V., Argüelles-Castilla, S., Testai, L., Khan, H., Sureda, A., de Oliveira, M.R., Vacca, R.A., Xu, S., Yousefi, B., Curti, V., Daglia, M., Sobarzo-Sánchez, E., Filosa, R., Nabavi, S.F., Majidinia, M., Dehpour, A.R., Shirooie, S., 2019. Targeting STATs in neuroinflammation, the road less traveled!. Pharmacol. Res. 14, 173–184. Nam, D., Kim, S.Y., 2008. Gene-set approach for expression pattern analysis. Briefings Bioinf. 9, 189–197. Planas, A.M., Gorina, R., Chamorro, A., 2006. Signalling pathways mediating inflammatory responses in brain ischaemia. Biochem. Soc. Trans. 34, 1267–1270. Qi, Z., Xu, Y., Liang, Z., Li, S., Wang, J., Wei, Y., Dong, B., 2015. Baicalein alters PI3K/ Akt/GSK3β signaling pathway in rats with diabetes-associated cognitive deficits. Int. J. Clin. Exp. Med. 8, 1993–2000. Saleh, A., Smith, D.R., Tessler, L., Mateo, A.R., Martens, C., Schartner, E., Ploeg, R.V.D., Toth, C., Zochodne, D.W., Fernyhough, P., 2013. Receptor for advanced glycation end-products (RAGE) activates divergent signaling pathways to augment neurite outgrowth of adult sensory neurons. Exp. Neurol. 249, 149–159. Shariq, A.S., Brietzke, E., Rosenblat, J.D., Pan, Z., Rong, C., Ragguett, R.M., Park, C., McIntyre, R.S., 2018. Therapeutic potential of JAK/STAT pathway modulation in mood disorders. Rev. Neurosci. 30, 1–7. Shen, Q., Peng, Q., Shao, J., Liu, X., Huang, Z., Pu, X., Ma, L., Li, Y.M., Chan, A.S.C., Gu, L., 2006. Corrigendum to “Synthesis and biological evaluation of functionalized coumarins as acetylcholinesterase inhibitors”. Eur. J. Med. Chem. 41, 793. Smoot, M.E., Ono, K., Ruscheinski, J., Wang, P.L., Ideker, T., 2011. Cytoscape 2.8: new features for data integration and network visualization. Bioinformatics 27, 431–432. Su, S.Y., Hsieh, C.L., Wu, S.L., Cheng, W.Y., Li, C.C., Lo, H.Y., Ho, T.Y., Hsiang, C.Y., 2009. Transcriptomic analysis of EGb 761-regulated neuroactive receptor pathway in vivo. J. Ethnopharmacol. 123, 68–73. T Takuma, K., Fang, F., Zhang, W., Yan, S., Fukuzaki, E., Du, H., Sosunov, A., McKhann, G., Funatsu, Y., Nakamichi, N., Nagai, T., Mizoguchi, H., Ibi, D., Hori, O., Ogawa, S.,

Stern, D.M., Yamada, K., Yan, S.S., 2009. RAGE-mediated signaling contributes to intraneuronal transport of amyloid-beta and neuronal dysfunction. Proc. Natl. Acad. Sci. U.S.A. 106, 20021–20026. Turnquist, C., Horikawa, I., Foran, E., Major, E.O., Vojtesek, B., Lane, D.P., Lu, X., Harris, B.T., Harris, C.C., 2016. p53 isoforms regulate astrocyte-mediated neuroprotection and neurodegeneration. Cell Death Differ. 23, 1515–1526. Wang, J., Liu, Y., Cheng, X., Zhang, X., Liu, F., Liu, G., Qiao, S., Ni, M., Zhou, W., Zhang, Y., 2017. The effects of LW-AFC on the hippocampus transcriptome in senescenceaccelerated mouse prone 8 strain, a mouse model of Alzheimer's disease. J. Alzheimer's Dis. 57, 227–240. Xu, J., Liu, J., Yue, G., Sun, M., Li, J., Xiu, X., Gao, Z., 2018. Therapeutic effect of the natural compounds baicalein and baicalin on autoimmune diseas. Mol. Med. Rep. 18, 1149–1154. Yammani, R.R., David, L., Loeser, R.F., 2009. Interleukin-7 stimulates secretion of S100A4 by activating the JAK/STAT signaling pathway in human articular chondrocytes. Arthritis Rheum. 60, 792–800. Yang, S., Wang, H., Yang, Y., Wang, R., Wu, C., Du, G., 2019. Baicalein administered in the subacute phase ameliorates ischemia-reperfusion-induced brain injury by reducing neuroinflammation and neuronal damage. Biomed. Pharmacother. https://doi. org/10.1016/j.biopha.2019.109102. Yassen, A., Fidel, R., Sven-Eric, S., Thomas, L., Mario, A., 2008. Computing topological parameters of biological networks. Bioinformatics 24, 282–284. Yoon, J., Blumer, A., Lee, K., 2006. An algorithm for modularity analysis of directed and weighted biological networks based on edge-betweenness centrality. Bioinformatics 22, 3106–3108. Zhang, Z., Li, X.G., Wang, Z.H., Song, M., Yu, S.P., Kang, S.S., Liu, X., Zhang, Z., Xie, M., Liu, G.P., Wang, J.Z., Ye, K., 2018a. delta-Secretase-cleaved Tau stimulates Abeta production via upregulating STAT1-BACE1 signaling in Alzheimer's disease. Mol. Psychiatry. https://doi.org/10.1038/s41380-018-0286-z. Zhang, Z., Yang, J., Liu, C., Xie, J., Qiu, S., Yang, X., Wu, C., 2018b. PseudoginsenosideF11 alleviates cognitive deficits and Alzheimer's disease-type pathologies in SAMP8 mice. Pharmacol. Res. 139, 512–523. Zhang, S.Q., Obregon, D., Ehrhart, J., Deng, J., Tian, J., Hou, H., Giunta, B., Sawmiller, D., Tan, J., 2013. Baicalein reduces β-amyloid and promotes nonamyloidogenic amyloid precursor protein processing in an Alzheimer's disease transgenic mouse model. J. Neurosci. Res. 91, 1239–1246.

8